University of Southern California Dissertations and Theses

Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential

(USC Thesis Other)

Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Studies on Direct Oxidation Formic Acid Fuel Cells:
Advantages, Limitations and Potential

By

Marc Thomas Iuliucci

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

August 2017

ii

Epigraph

“Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
To where it bent in the undergrowth;

Then took the other, as just as fair,
And having perhaps the better claim,
Because it was grassy and wanted wear;
Though as for that the passing there
Had worn them really about the same,

And both that morning equally lay
In leaves no step had trodden black.
Oh, I kept the first for another day!
Yet knowing how way leads on to way,
I doubted if I should ever come back.

I shall be telling this with a sigh
Somewhere ages and ages hence:
Two roads diverged in a wood, and I—
I took the one less traveled by,
And that has made all the difference.”
-The Road Not Taken Robert Frost

iii

Dedication

I dedicate this work to the individuals from all walks of life with whom I have crossed paths
during my time in California, who have each impacted my life in different ways, and brought a
unexpected richness to my graduate school experience.

iv

Acknowledgements

I would like to direct my initial thanks towards Professor G.K. Surya Prakash for his
continued guidance and support over the past 6 years. Pursuing a PhD in Chemistry under his
tutelage has been a rewarding experience, which has helped me grow professionally and as an
individual. I would also like to extend my gratitude to the late Professor George Olah, for the
many philosophical insights that he offered during most of my years in the Olah-Prakash group
meetings.
Special thanks goes out to the members of the Fuel Cell lab; Dr. Frederick “Charlie”
Krause, Dr. John Paul Jones, Dr. Akihisa Saitoh, Dr. Bo Yang, Dean Glass, and Vicente Galvan for
their support at various stages of the PhD program. Moreover, thank you to Dr. Fang Wang for
imparting a strong foundation in organic chemistry, with his unparalleled expertise in the
subject. Each of them provided invaluable input at various times throughout my journey at
USC.
Deserving of special mention is the incredible staff at the Loker Hydrocarbon Institute;
Dr. Robert Aniszfeld, Jesse May, Carol Phillips, David Hunter, and Ralph Pan for the unique roles
they each played in assuring that all aspects of the institute were running smoothly, and making
themselves available. An extra special thanks to all those who ensured there were always
coffee grounds in the Bistro.

v

Table of Contents
Epigraph…………………………………………………………………………………………………………………………………….ii
Dedication…………….…………………………………………………………………………………………………………………..iii
Acknowledgements…………………………………………………………………………………………………………………..iv
List of Tables…………………………………………………………………………………………………………………………….vii
List of Figures…………………………………………………………………………………………………………………………..viii
Chapter 1: Direct Formic Acid Fuel Cells Reported in the Literature………………………………………..…1
1.1 Introduction……………………………………………………………………………………………………………..1
1.2 Direct Formic Acid Fuel Cells…………………………………………………………………………………….4
1.2.1 Research Devices…………………………………………………………………………………....6
1.2.1.1 Pt-based electrocatalysts…………………………………………………………….6
1.2.1.2 Pd-based electrocatalysts ………………………………………………………..…9
1.2.1.3 Novel Catalyst Supports…………………………………………………………….13
1.3 Mechanistic Studies………………………………………………………………………………………………..21
1.3.1 Pd Anode Deactivation/Reactivation…...……………………………………………….21
1.3.2 MEA Structure……………………………………………………………………………………….23
1.4 Practical Devices…………………………………………………………………………………………………….31
1.5 Direct Formate Fuel Cells………………………………………………………………………………………..44
1.5.1 Initial reports………………………………………………………………………………………..45
1.5.2 AEM Fuel Cells……………………………………………………………………………………...47
1.5.3 Alternative-Electrolyte Fuel Cells…………………………………………………………..54
1.6 Conclusion and Outlook…..……………………………………………………………………………………..59
1.7 References……………………………………………………………………………………………………………..61

Chapter 2: Re-Assessing the Efficacy of Palladium Anodes in Direct Formic Acid Fuel Cells………67
2.1 Introduction …………………………………………………………………………………………………………..67
2.2 Experimental Methods……………………………………………………………………………………………69
2.2.1 Electrochemical Measurements…………………………………………………………….69
2.2.2 Membrane Electrode Assembly…………………………………………………………….70
2.2.3 Fuel Cell Measurements………………………………………………………………………..70
2.2.4 Trace Analysis of Leached Palladium……………………………………………………..71
2.3 Results & Discussion…………………………………………………………………………………………….…71

vi

2.3.1 Formic Acid Oxidation Characteristics.………………………………………………….71
2.3.2 Fuel Cell Polarization.……………………………………………………………………………74
2.3.3 Potentiostatic Discharge……………………………………………………………………….77
2.3.4 Palladium Anode Deactivation/Reactivation………………………………………….82
2.3.5 Palladium Leeching……………………………………………………………………………….83
2.3.6 Towards More Viable Formic Acid Fuel Cells…………………………………………84
2.4 Conclusion……………..………………………………………………………………………………………………86
2.5 References……………..………………………………………………………………………………………………87

Chapter 3: Improved Hydrogen Fuel Cell Performance Using CFx-Pt Blends…………………………....89
3.1 Introduction……………………………………………………………………………………………………………89
3.2 Experimental Methods……………………………………………………………………………………………92
3.2.1 Membrane Electrode Assembly…………………………………………………………….92
3.2.2 Fuel Cell measurements………………………………………………………………………..93
3.3 Results & Discussion…………………………………………………………………………………………….…93
3.3.1 Effect of Catalyst Application Procedure……………………………………………….93
3.3.2 Simple Mixing of Platinum with XC72 or CFx…………………………………………96
3.3.3 Balled-Milled Mass Ratio Formulations………………………………………………...97
3.3.4 Reduced O 2-Flow Rate Requirements with CFx……………………………………101
3.3.5 Platinum-Carbon Mixtures Compared with State-of-the-Art Pt/C….……103
3.4 Conclusion and Outlook………………………………………………………………………………………..107
3.5 References……………………………………………………………………………………………………………108

vii

List of Tables
Table 1-1: Pt-based DFAFCs……………………………………………………………………………………………………...7
Table 1-2: Pd-based DFAFCs……………………………………………………………………………………………………10
Table 1-3: Examples of Novel Anode Catalyst Supports……………………………………………………………16
Table 1-4: Overview of DFAFCs as Practical Power Sources…………………………………………..…………34
Table 1-5: Recently Studied DFFCs……………………………………………………………………………………………46
Table 2-1: DFAFC Anode Performance Characteristics……………………………………………………………..76
Table 2-2: Concentration of Leached Pd by ICP-OES…………………………………………………………………84

viii

List of Figures
Figure 1-1: Cyclic voltammograms of FAO on Pt (red) and Pd (blue) catalysts supported on
reduced graphene oxide (rGO)……………….……………………………………………………………….…………………9
Figure 1-2. Sections of the CO-stripping cyclic voltammograms on Pt and Pd, after 1 h of FAO at
0.3 V vs. RHE in 5 M formic acid…………………………………………………………………………………………….…11
Figure 1-3. Chronoamperometry of FAO (0.65 V vs. SHE) on Pd/GF and PtRu/GF, in 3 M HCOOH
+ 0.5 M H2SO4 at 25°C…………………………………………………………………………………………………………….14
Figure 1-4. Cyclic voltammograms of Pt/C and sub-monolayer Pt deposited on nanoporous gold
(NPG-Pt) in 1 M HCOOH + 0.5 M H2SO4…………………………………………………………………………………..17
Figure 1-5. Power density curves for Pd-Ni2P/C and related catalysts. [HCOOH] = 3 M, O2 at the
cathode, cell T = 30°C……………………………………………………………………………………………………………… 20
Figure 1-6. Cyclic voltammograms for Pd electrode with 10 wt.% (top) and 30 wt.%(bottom)
Nafion content, in 12 M HCOOH +0.1 M H2SO4………………………………………………………………….……26
Figure 1-7. Polarization curves for a DFAFC with membranes of different thicknesses……………28
Figure 1-8. SEM images of hot-pressed (top) and decaled (bottom) MEAs………………………………30
Figure 1-9. Voltage transients at a constant current of 20 mA as a function of time………………..32
Figure 1-10. Anode, cathode (both vs. RHE) and cell potentials for a passive DFAFC operating at
100 mA/cm2…………………………………………………………………………………………………………………………….33
Figure 1-11. Schematic of a microfluidic DFAFC (left) and the electrical circuit used for
measurements (right). V1 and V2 represent the potentials at the anode and cathode vs. a
reference electrode (RE), respectively. V3 is the cell voltage measured across an external
resistor, R…………………………………………………………………………………………………………………………………35
Figure 1-12. Schematic of the nanofluidic DFAFC, with actively supplied fuel and oxidant……….37
Figure 1-13. A four-cell DFAFC stack. In the two-cell stack, there is only one MEA on each side
of the
reservoir…………………………………………………………………………………………………………………………………..38
Figure 1-14. Two-cell and four-cell DFAFC stacks discharging at 20 mA, with 3.5 mL of 5 M
formic acid (left). Stack voltage and internal resistance as a function of discharge cycle number
(right)……………………………………………………………………………………………………………………………………….39

ix

Figure 1-15. 10-cell DFAFC discharging at 1A for 240 h (left). Polarization curves (right) prior to
lifetime test (squares), after the lifetime test (triangles), and after reactivating the anode by
washing with pure water for 1 h (circles)…………………………………………………………………………………41
Figure 1-16. A 15-cell DFAFC (a) and the hybrid system used to power a laptop (b)………………..42
Figure 1-17. One operation cycle of the 15-MEA DFAFC-hybrid system. The fuel tank was filled
with 280 mL of 11 M formic acid………………………………………………………………………………………………43
Figure 1-18. Cyclic voltammograms in formic acid (dotted line) and potassium formate (solid
line), normalized to mgPd. Potentials at working electrode (WE) are reported with respect to
mercury sulfate electrode (MSE)………………………………………………………………………………………………45
Figure 1-19. Possible AEM degradation mechanisms that may occur in a DFFC……………………….48
Figure 1-20. Polarization curves of AFCs with 1 M KCOOH/2 M ethanol + 2 M KOH supporting
electrolyte at the anode and oxygen or air at the cathode. Fuel and cell at 60°C (left).
Chronoamperomograms of KCOOH and ethanol oxidation at various potentials with a Pd black
working electrode. Electrolyte solutions were 1 M KCOOH/1 M ethanol + 1 M KOH, or 1 M
HCOOH + 1 M H 2SO 4 (right)………………………………………………………………………………………………………50
Figure 1-21. Polarization curves of a DFFC with an optimized anode structure operating under
oxygen (left) and air (right)………………………………………………………………………………………………………51
Figure 1-22. Voltage profiles of a hydroxide-free DFFC operating at 40°C, discharging at 100
mA/cm2 during a 24 h period…………………………………………………………………………………………………..52
Figure 1-23. Polarization curve for a DFFC with a home-made AEM and non-noble ORR
catalyst…………………………………………………………………………………………………………………………………….53
Figure 1-24. Schematic of a membraneless DFFC…………………………………………………………………….55
Figure 1-25. Polarization data for the membraneless DFFC………………………………………………………56
Figure 1-26. Schematic of an alkaline formate-acidic peroxide fuel cell……………………………………57
Figure 1-27. Polarization data for an alkaline formate-acidic peroxide fuel cell……………………….58
Figure 2-1. Forward (a) and reverse (b) cyclic voltammetry scans of FAO on Pt (black), Pd (blue),
and PtRu (red)………………………………………………………………………………………………………………………….72
Figure 2-2. Current-voltage (I-V) Transients of DFAFCs with common anode
materials……………………………………………………………………………………………………………………….………...74
Figure 2-3. Polarization curves generated of DFAFCs with common anode materials……………...75
Figure 2-4. DFAFCs discharged at potentials corresponding to their peak power densities………77

x

Figure 2-5. (a) DFAFCs discharged after the activation overpotential. (b) Expanded view of the
oscillatory region of Pd black………………………………………………………………………………………………...…79
Figure 2-6. DFAFCs operating near complete discharge……………………………………………………………81
Figure 2-7. (a) Polarization curves consecutively collected from a DFAFC with Pd black at the
anode. (b) Before and after potentionstatic discharge, then after washing the anode…………….82
Figure 3-1. Fuel Cell data at three different temperatures comparing the brush-painting method
(solid lines) to the spray painting method (dotted lines).Gases humidified at 75°C. Hydrogen and
oxygen flowed at 100 sccm……………………………….……………………………………………………………………..94
Figure 3-2. Preliminary evidence suggesting improvements with CFx at the cathode. Cell T = RT.
Gases humidified at 75°C. Hydrogen and oxygen flowed at 100 sccm……………………………………..97
Figure 3-3. Polarization curves with increasing CFx content at the cathode. Cell T = RT. Gases
humidified at 75°C…………………………………………………………………………………………………………………..98
Figure 3-4. Voltages transients with increasing CFx content at the cathode. Cell T = RT. Gases
humidified at 75°C…………………………………………………………………………………………………………………100
Figure 3-5. Hydrogen fuel cells at various cathode flow rates. Cell T = 30°C. Gases humidified at
75°C……………………………………………………………………………………………………………………………………….102
Figure 3-6. A simple balled-milled mixture of 20:80 Pt:XC72 in contrast with commercially
available Pt/C catalyst, represented in units per mg Pt…………………………………………………………….103
Figure 3-7. Hydrogen Fuel cell polarization curves with various fluorinated cathode catalysts.
Cell T=30°C, Hydrogen and oxygen flowed at 50 sccm……………………………………………………………105
Figure 3-8. Mass-normalized, per mgPt, polarization curves from Figure 3-7…………………………106

1
1
Direct Formic Acid Fuel Cells Reported in the
Literature
1.1 Introduction
Direct Formic Acid Fuel Cell (DFAFC) research has received much attention over the past
fifteen years. Some of the problems encountered with methanol fuel cells (i.e. expensive anode
electrocatalysts, fuel crossover) are mitigated by using formic acid as the fuel instead. There are
many reports of novel, efficient anode materials, and a few examples of practical devices. Several
authors have pursued fundamental research to understand the problem of Pd deactivation and
developed methods to regenerate its catalytic activity. The structure and preparation of the MEA
and its effect on the DFAFC performance is discussed. In recent years, the direct formate fuel cell
(DFFC) has been developed, which offers multiple potential solutions to the problems encountered
with its acidic counterpart. The aim of this chapter is to summarize the aforementioned results and
bring the reader up to date on the current state-of-affairs of direct formic acid and formate fuel cell
research.
2
As the need for alternative power sources has become increasingly evident, the development
of viable fuel cell technologies has been a hotly researched topic for the past 25 years. Polymer
electrolyte membrane fuel cells (PEMFCs) based on the Nafion® proton-exchange membrane
have received the most attention, particularly the H2-PEMFC.
1
However, realizing a hydrogen-
based infrastructure has proven difficult due to the complications of transporting and storing
hydrogen gas. In response to this, much work has been directed towards direct liquid fuel cells
(DLFCs), which operate by the same principles involved in H2-PEMFCs, such that fuel oxidation
produces protons that cross the Nafion® membrane and react with oxygen to produce water.
In this field, direct methanol fuel cells (DMFCs) have been the most studied system over the
past two decades.
2
Despite the success achieved with methanol, serious complications still exist.
The products of methanol oxidation on Pt are potential dependent, such that oxidation to CO and
CO2 predominate at low and high potentials, respectively.
3
This is problematic because CO binds
strongly to the metal surface, blocking catalytically active sites, thus “poisoning” then preventing
further oxidation. It was later found that the presence of Ru adjacent to active Pt sites could
mitigate this problem, via the formation of active oxygen species on Ru, by completing the
oxidation of CO to CO2.
4
Fuel crossover from the anode to cathode compartment is problematic
with Nafion, resulting in a mixed potential at the cathode and a concurrent decrease in cell
performance.
5
Moreover, the catalysts employed at the anode and cathode, in the case of DMFCs,
are generally PtRu and Pt black, respectively. Although they have demonstrated excellent
performance, the overall cost of such systems still make their application prohibitively expensive.
Second to methanol, formic acid is the most-studied liquid fuel for acidic DLFCs.
6,7
The
oxidation proceeds as either a dehydrogenation (direct pathway) or dehydration reaction (indirect
pathway):
3

Direct: 𝐻𝐶𝑂𝑂𝐻 → 𝐶 𝑂 2
+ 2𝐻 +
+ 2𝑒 −
(1)
Indirect: 𝐻𝐶𝑂𝑂𝐻 → 𝐶𝑂 + 𝐻 2
𝑂 → 𝐶 𝑂 2
+ 2𝐻 +
+ 2𝑒 −
(2)

It has been found that Pd-based anode catalysts are the most effective, as they are able to
oxidize formic acid predominately via the dehydrogenation pathway, minimizing the formation of
CO, which is known to bind strongly to the electrocatalyst surface, preventing further oxidation of
fuel molecules.
8
In spite of this, it is well observed in the literature, and in experiments carried out
in our lab, that continued oxidation of formic acid at some constant potential results in complete
surface deactivation after several hours.
9,10
This effect may arise from the gradual production of
CO with time, or through some other strongly binding oxidation intermediate. Furthermore, Pd is
about half the cost of Pt at the time of this writing. Although this brings down the overall cost of
the system, the oxygen reduction reaction (ORR) still requires Pt-based catalysts at the cathode,
making practical, widespread adoption difficult due to the scarcity and concurrent costliness of
necessary materials.
Another problematic feature of formic acid-based devices is anode catalyst stability under
acidic conditions, as it is known that Pd corrodes in acidic conditions at E > 0.8 V.
11
A common
approach to improve anode oxidation kinetics and limit cost is to alter the electronic structure of
Pd through alloying with other metals such as Fe,
12
Sn,
13
and Cu.
14
However, prolonged cell
operation will likely result in the leaching of the alloyed metals, and a loss of the beneficial
electronic effect.
A potential solution to the problems inherent to formic acid fuel cells could be to operate it
under an alkaline environment. Operating a fuel cell with aqueous solutions of formate would
4
require an anion exchange membrane (AEM), instead of Nafion®, and provide a non-corrosive
environment that may allow for the use of non-noble metals at both the anode and cathode.
Furthermore, higher operating temperatures are possible because aqueous formate solutions are
non-volatile above 100° C. It is also an ideal fuel for use in an anthropogenic carbon cycle, since
it may be obtained from the electrochemical reduction of CO2.
15
Although a direct formate fuel
cell (DFFC) was demonstrated as early as 1965,
16
these devices have received little attention
compared to the acid-based counterparts, but interest has grown in recent years. The purpose for
this review is to summarize the progress of DFAFC research, compare the results of various
authors, and highlight the interesting developments in the emerging field of DFFCs.
1.2 Direct Formic Acid Fuel Cells
Following the success of the Direct Methanol Fuel Cell (DMFC), formic acid was tested as a
substitute fuel, and was found to be more active than methanol on both Pt and PtRu electrodes.
17

However, the study was conducted at 170°C so that the gaseous products of the oxidation
reaction could be analyzed in real-time. Although these conditions are not representative of a
practical fuel cell, the results were significant, in that CO2 was the only observed product of the
formic acid oxidation (FAO), suggesting that formic acid could be an excellent alternative to
methanol, which readily poisons the anode catalyst surface due to incomplete oxidation to CO.
The authors noted that the energy density of formic acid is about a third that of methanol, which
arises from the former being a two-electron oxidation, and the latter, a six-electron oxidation.

𝐻𝐶𝑂𝑂𝐻 → 𝐶 𝑂 2
+ 2𝐻 +
+ 2𝑒 −
E
0
anode-DFAFC = 0.02 V (3)
𝐻 3
𝐶𝑂𝐻 + 𝐻 2
𝑂 → 𝐶 𝑂 2
+ 6𝐻 +
+ 6𝑒 −
E
0
anode-DMFC = -0.25 V (4)
1
2
𝑂 2
+ 2𝐻 +
+ 2𝑒 −
→ 𝐻 2
𝑂 E
0
cathode = 1.23 V (5)
5
E
0
cell = E
0
cathode – E
0
anode (6)
𝐻𝐶𝑂𝑂𝐻 +
1
2
𝑂 2
→ 𝐶 𝑂 2
+ 𝐻 2
𝑂 E
0
DFAFC = 1.48 V (7)
𝐻 3
𝐶𝑂𝐻 +
3
2
𝑂 2
→ 𝐶 𝑂 2
+ 2𝐻 2
𝑂 E
0
DMFC = 1.21 V (8)
Moreover, considering the cathode reaction, the theoretical open circuit voltage (OCV) is
higher for DFAFCs than DMFCs. Although these potentials are never achieved in a real fuel
cell, due to thermodynamic and resistive losses, the difference is often reflected in the actual
OCV of such devices, in that it is generally higher for DFAFCs. Another intrinsic advantage to
working with formic acid is that it is structurally resistant to fuel crossover, when compared to
methanol, because of electrostatic repulsions between its carboxylic acid moiety and the
sulfonate groups present on the Nafion polymer backbone. On the contrary, methanol may be
protonated in the aqueous fuel solution, and carried across the membrane via an osmotic drag.
We categorize literature reports on DFAFCs into the following four sub-groups: research
devices, mechanistic studies, commercial prototypes, and FAO electrocatalysts. The fourth sub-
group consists of reports on the preparation of novel anode catalyst materials. Although, the
electrocatalysts in such reports are not tested in a membrane electrode assembly (MEA), and are
only stated to be potentially viable candidates as DFAFC anode materials. Masel and co-workers
observed that even when significant improvements are achieved in single-cell experiments, the
results do not necessarily carry over to the fuel cell environment.
18,19
Since the main goal of this
review is to summarize the achievements in DFAFC research in terms of demonstrable cell
performance, and current understanding regarding chemical process that occur in the MEA,
developments outside the scope of this work will not be discussed.

6
1.2.1 Research Devices
In this type of work, a fuel cell is constructed purely for research purposes and subjected to a
general performance characterization, usually in the form of a current-voltage (I-V) measurements.
Also called a polarization curve, it is obtained by scanning between two current or potential limits,
and plotting voltage and power vs. current on separate axes. With these results, a researcher is
able to gauge the performance of the cell based on a current density and an operating potential that
may yield the maximum power density. Extended operation may also be demonstrated to
determine how long the cell can survive before deactivation. Formic acid is usually stored in an
external reservoir and circulated through the anode compartment, while pure oxygen or air is fed
to the cathode. Ultimately, these devices provide proof-of-principle that the MEA in question has
the capacity to function in a fuel cell, but more engineering is required before realization of a
practical device. The following sections detail the majority of work undertaken in this area.
1.2.1.1 Pt-based electrocatalysts
The first report of a DFAFC appeared in the literature in 2002, providing proof-of-principle that
such devices could work at reasonable temperatures.
20
At 60°C, a power density of 48.8 mW/cm
2
was achieved at 0.4 V with formic acid, versus 32.0 mW/cm
2
with methanol, under similar
operating conditions. Nonetheless it should be noted that to achieve those performance values it
was necessary to use 12 M formic acid, whereas the methanol fuel cell was operated using 1 M
MeOH. Although the anode catalyst is proprietary, it is stated to be “Pt-based, with some other
noble metal additives”, which may have been Ru, since it was already known to be more for active
for FAO than Pt alone.
17

7
A few years later, another group achieved higher power densities at even lower operating
temperatures, employing a Pt/C anode catalyst.
21
At temperatures of 25°C and 50°C, DFAFCs
demonstrated peak power densities of 76.5 mW/cm
2
and 116 mW/cm
2
, respectively. As can be
seen in Table 1-1, experimental conditions under which research-grade fuel cells are tested vary
from author to author. Hence, the difference in reported values may be attributed to a thinner
Nafion membrane (i.e. less resistance to proton transfer), and a higher catalyst loading (8 mg/cm
2

vs. 4 mg/cm
2
). Cell performances are highly system dependent, so comparisons between the
results of different authors must be evaluated in relation to their experimental setup.
Since the early 2000s, several other Pt-M anode catalysts have been demonstrated with varying
degrees of success. Rhee et al. found that PtPd (1:1), at an optimal loading of 5 mg/cm
2
, yielded
a power density of 64.7 mW/cm
2
at a current density of 200 mA/cm
2
.
22
Moreover, this
performance was demonstrated with air at the cathode, rather than pure O2, which is desirable
Table 1-1
Pt-based DFAFCs
Anode catalyst GDL Membrane
Anode
Loading
[HCOOH] Oxidant Cell T Max. p.d. Ref.
Proprietary Pt-
based catalyst
(UIUC-B)
carbon cloth Nafion 117 4 mg/cm
2
12 M O 2 60°C 48.8 mW/cm
2
20
Pt/C carbon paper Nafion 112 8 mg/cm
2
3 M O 2
25°C
50°C
76.5 mW/cm
2

116 mW/cm
2

21
PtPd carbon cloth Nafion 117 5 mg/cm
2
9 M HCOOH
+ 1 M HBF 4
air 25°C 64.7 mW/cm
2
22
PtAu/C carbon paper Nafion 115 2 mg PtAul/cm
2
2 M
6 M
O 2 60°C
45 mW/cm
2
64 mW/cm
2

23
Pt 0.6 Au 0.4
PtRu
carbon paper Nafion 115 3 mg/cm
2
9 M O 2 60°C
200 mW/cm
2
155 mW/cm
2

24
†
PtPb(3:1)/C
†
PtSb(3:1)/C
carbon paper Nafion 115 1.6 mg/cm
2
5 M O 2 RT
*137.5 mW/cm
2

*112 mW/cm
2
25
†
PtBi(15:1)/C carbon paper Nafion 115 1.6 mg/cm
2
5 M O 2 RT - 26
†
PtSb(4:1)/C carbon paper Nafion 115 1.6 mg/cm
2
5 M O 2 RT *148.5 mW/cm
2
27
‡
Bi@Pt/C
Pt/C
carbon paper Nafion 112 1 mg Pt/cm
2
4.75 M air RT
11mW/cm
2
4mW/cm
2
28
*These values were calculated from data reported in mA/mg, using the catalyst loadings provided in the paper
†Numbers in the parentheses represent the metal weight ratios
‡@ signifies that Bi was irreversibly adsorbed onto the Pt surface

8
because commercial devices would ideally be “air-breathing”. PtAu/C can achieve power
densities of 45 mW/cm
2
and 64 mW/cm
2
at 60°C with a low catalyst loading of 2mg-PtAu/cm
2
and
formic acid concentrations of 2 M and 6 M, respectively.
23
Lee et al. reported that Pt0.6Au0.4 is the
optimal atomic ratio for unsupported PtAu catalysts, and that it even outperforms PtRu, when
compared under similar conditions.
24
Even at a moderate metal loading of 3 mg/cm
2
, Pt0.6Au0.4
gives the highest power density (200 mW/cm
2
) of all Pt-based FAO electrocatalysts developed
thus far.
Pickup and Yu published a series of papers in which they demonstrated good room temperature
performances of PtPb (3:1)/C,
25
PtSb (3:1)/C,
25
PtSb (4:1)/C,
27
and PtBi (15:1)/C.
26
The catalysts
incorporating Pb and Sb were either reduced from their salts onto a commercial Pt/C catalysts
25
or
codeposited on the carbon support via simultaneous reduction of their precursor salts, as in the
case of PtSb (4:1)/C.
27
XPS analysis suggested that the Pb or Sb was present as a surface oxide,
with little to no alloying having occurred. This surface oxide may then act as a co-catalyst to
complete the oxidation of intermediates such as CO. Among a series of PtBi/C catalysts, a Pt:Bi
atomic ratio of 15:1 was found to be optimal for both Bi-modifed Pt/C and codeposited PtBi/C,
which behaved similarly in polarization experiments regardless of their synthetic route.
26
Similar
to the Pb and Sb modified catalysts, Bi existed primarily as a surface oxide in both decorated and
codposited PtBi/C.
Although the last entry in Table 1-1 exhibits the lowest power densities among the Pt-based
anode catalysts, the results are noteworthy in that the performance of commercial Pt/C may be
improved by a simple post-treatment of the MEA.
28
By passing a 70% Bi2O3 saturated solution
in 0.5 M H2SO4 through the anode compartment of the fuel cell for a set amount of time, Bi
could be irreversibly adsorbed onto the Pt surface, denoted as Bi@Pt/C. As shown in Table 1-1,
9
the Bi-treatment of a Pt/C anode catalyst increased the peak power density by nearly a factor of
three. This synthetic route is advantageous because it does not require special preparation of a
new catalyst prior to MEA fabrication, hence commercially available Pt catalysts, or even
commercially available MEAs, may be modified to improve the catalytic activity.
1.2.1.2 Pd-based electrocatalysts

Figure 1-1. Cyclic voltammograms of FAO on Pt (red) and Pd (blue) catalysts
supported on reduced graphene oxide (rGO).

Nowadays, Pd is generally considered a superior electrocatalyst for the FAO, chiefly because
FAO on Pd favors the direct pathway (1) which avoids or mitigates the generation of CO from (2).
This phenomenon can be observed experimentally using cyclic voltammetry, as shown from our
in-house prepared rGO-supported Pd and Pt catalysts in Figure 1-1. During the forward scan on
Pt-rGO, oxidation begins around 0.3 V but quickly levels off to a plateau at 0.6 V, corresponding
to (2), which inhibits further reaction. Around 0.8 V, CO is oxidized, thereby allowing FAO to
10
continue until the formation of Pt oxides again cover the surface at high potentials. In the reverse
scan, the reduction of Pt oxides results in an adsorbate-free surface, as

evidenced by the significantly larger formic acid oxidation current over the entire potential
range. On the contrary, FAO on Pd-rGO is strikingly different, such that the forward and reverse
scans are more or less symmetric, indicating that reaction (1) predominates. Direct oxidation to
CO2 occurs on Pd-rGO until about 1 V, at which time the surface is covered with Pd oxides. Much
like on Pt-rGO, current is not produced in the reverse direction until the surface oxides are reduced,
which occurs around 0.8 V.
One of the earliest examples of a DFAFC compared carbon supported Pt and Pd anode catalysts
in a DFAFC, and found that Pd/C carbon gave significantly higher power densities at both 25°C
(120 mW/cm
2
vs. 76.5 mW/cm
2
) and 50°C (166 mW/cm
2
vs. 116 mW/cm
2
).
21
As shown in Table
1-2, entry 5, Choi et al. achieved similar power densities with Pd/C using lower catalyst loading,
Table 1-2
Pd-based DFAFCs
Anode catalyst GDL Membrane
Anode
Loading
[HCOOH] Oxidant Cell T Max. p.d. Ref.
*Pd Ti mesh Nafion 117 2 mg/cm
2
1 M air 60°C 20.2 mW/cm
2
29
Pd black Carbon cloth Nafion 112 8 mg/cm
2
3 M air 22°C 248 mW/cm
2
30
Pd black Carbon cloth Nafion 112 2.4 mg/cm
2
5 M air 30°C 260 mW/cm
2
31
Pd/C Carbon paper Nafion 112 8 mg/cm
2
3 M O 2
25°C
50°C
120 mW/cm
2
166 mW/cm
2
21
Pd/C Carbon cloth Nafion 117 4 mg/cm
2
3 M O 2
25°C
60°C
129 mW/cm
2
163 mW/cm
2
32
PdPb
PdSb
PdSn
Pd black
Carbon paper Nafion 117 10 mg/cm
2
10 M air 30°C - 18
†
Sb@Pd
Pd black
Carbon paper Nafion 117 10 mg/cm
2
10 M air 30°C
247 mW/cm
2
225 mW/cm
2
19
*Pd
Pd/C
Carbon paper Nafion 117 0.5 mg/cm
2
1 M HCOOH
+ 0.5 H 2SO 4
O 2 25°C
120 mW/cm
2
50 mW/cm
2
33
*Electrodeposited directly onto the GDL
†@signifies that Sb was electrodeposited onto the Pd surface
11
a thicker Nafion membrane, and a carbon cloth diffusion layer.
32
Although, the optimal anode
catalyst is likely to be system-specific, considering that Pt outperformed Pd (32 mW/cm
2
vs. 20.2
mW/cm
2
) when thermally deposited on Ti mesh.
29

Figure 1-2. Sections of the CO-stripping cyclic voltammograms on Pt and Pd, after 1 h of
FAO at 0.3 V vs. RHE in 5 M formic acid.
31

Some of the highest power densities yet reported for a DFAFC were among the first to be
published. Employing high surface area Pd black at the anode at only 22°C, the Masel group
reported that a power density of 248 mW/cm
2
was achieved at 0.4 V and a current density of 600
mA/cm
2
.
30
These results are all the more impressive since they were demonstrated with air as the
oxidant. A few years later, the same group reported that at a slightly higher temperature of 30°C
and significantly less Pd-loading (2.4 mg/cm
2
), a peak power density of 260 mW/cm
2
could be
achieved at 750 mA/cm
2
and a similar operating potential.
31
As shown in Table 1-2, entries 2 and
12
3, the experimental conditions differ only by the catalyst loading, fuel concentration, and operating
temperature. As such, comparable performance with lower Pd-laoding may be attributed to a
combination of improved kinetics from increased reactant concentration and temperature, and
decreased mass-transfer resistance due to a thinner layer of catalyst through which the fuel must
diffuse in the MEA.
The aforementioned work also provided experimental evidence for the direct pathway (1) by
electrochemically removing surface-CO that accumulated during FAO.
31
Upon subjecting the Pt-
coated working electrode to a potential at which FAO occurs, for 1 h, and then scanning to positive
potentials results in a peak for CO-oxidation. As seen in Figure 1-2, the oxidation peak is absent
on the Pd electrode, indicating that a significant amount of CO was not formed during FAO.
However, several year later the Masel group also observed that after 3 h of FAO in 12 M formic
acid on Pd, significant amounts of CO had covered the catalyst surface.
18
Moreover, it is well
known that there is a so-called “CO-like” intermediate that deactivates the Pd-surface with
prolonged reaction time.
8
The nature of this intermediate, its removal from the surface, and
consequences for DFAFCs will be further discussed in Section 2.2.1.
At very low loading (0.5mg/cm
2
), and fuel concentration of 1 M HCOOH + 0.5 M H2SO4 as
supporting electrolyte, Pd/C at the anode yields a power density of 50 mW/cm
2
at 25°C.
33
In the
same work, this value was increased by a factor of 2.4 to 120 mW/cm
2
by electrodepositing Pd
directly onto the carbon paper gas diffusion layer (GDL). Traditionally, the catalyst (including
support) is painted directly on top of the GDL or PEM prior to being assembled into an MEA.
With this type of MEA architecture, there exists an optimal loading, below and above which there
may be kinetic and mass transfer limitations, respectively.
34,35
Electrodeposition directly on the
GDL may be advantageous, relative to the traditional MEA structure, because electrochemically
13
active sites are now present inside, rather than just on the surface, thereby increasing the utilization
of precious metal and decreasing the thickness of the diffusion medium.
Another attribute of Pd that aids in its attractiveness as a fuel cell catalyst, is that it is
comparatively cheaper than Pt. At the time of this writing, the price of Pd is about 795 USD/ounce
based on information available from four different sources (InvestimentMine, Kitco, Nasdaq, and
Monex). Although, if Pd-based DFAFCs are to be developed commercially, the catalyst loadings
will have to be moderated such that the cost-benefit is not outweighed by the excessive use of
noble metal. As shown in Table 1-2, Pd-loading ranges from 0.5 mg/cm
2
to as high as 10 mg/cm
2
.
While this is acceptable for a research device, the goal of which is usually to maximize power
density, careful optimization of catalyst loading and a judicious selection of high utilization Pd
catalysts will be necessary for DFAFC technology to be practical.
1.2.1.3 Novel Catalyst Supports
In an effort to better utilize the precious metal that is required at the anode, a several types of
novel catalyst supports have been investigated. Metal catalysts may be supported on a high surface
area carbon material to increase the number of electrochemically active sites, which has the added
benefit of a concurrent decrease in the overall metal loading. Most, if not all, carbon supported
catalysts found in the literature (notated as Metal/C) are actually referring to Vulcan® XC-72 (or
XC-72R), which may be considered the standard against which other catalyst supports are judged.
Vulcan® carbon, as it is also commonly called, has become an industry standard as both an anode
and cathode catalyst support, due to its high surface area, excellent conductivity, and chemical
stability.
14

Figure 1-3. Chronoamperometry of FAO (0.65 V vs. SHE) on Pd/GF and PtRu/GF, in 3 M
HCOOH + 0.5 M H2SO4 at 25°C.
36

Alternatives to Vulcan® carbon that have appeared in the literature include other carbon
materials,
36–40
high porosity gold,
41
or transition metals/metal oxides combined Vulcan®
carbon,
42–46
which are listed in Table 1-3. PtRu supported on hollow core/mesoporous shell carbon
(HSMSC) yielded power densities that were almost twice as high as PtRu/C, at both 30°C and
60°C.
37
Interestingly, concurrent with this improvement was a two-fold increase in
electrochemically active surface area (ECSA), which was measured to be 83 m
2
/g and 39 m
2
/g for
polycrystalline Pt in the HSMSC and Vulcan® carbon supported PtRu catalysts, respectively.
15
Graphite felt (GF) that had been pre-treated by Shipley-type solution to increase the
hydrophilicity of its surface was tested as a support for both Pd and PtRu.
36
Since the metals were
electrodeposited on GF, it acted as both a catalyst support and anode diffusion medium, thereby
allowing for improved mass-transfer properties. Pd/GF exhibited a peak power density of 85.2
mW/cm
2
compared to 39.2 mW/cm
2
with Pd black in a traditional MEA architecture. Moreover,
Pd activity on a mass-specific basis was reported to be 15 W/g when electrodeposited on GF versus
9.8 W/g with Pd black, indicating an improvement in precious metal utilization. However, it is
worth noting that the authors found PtRu to be more active and more stable than Pd in both the
traditional MEA and supported on GF. As shown in Figure 1-3, PtRu/GF retains a pseudo-steady
state current density after three hours of FAO, whereas a significant deactivation is observed with
Pd/GF, likely due to adsorption of reaction intermediates, which are more easily oxidized on PtRu.
Han et al. studied the effect of pore size of a carbon support for Pd catlaysts.
40
It was reported
that, for both 30 wt.% and 50 wt.% Pd, the catalyst support with 20 nm pores gave higher power
densities than with 50 nm pores, or Pd black alone. Moreover, they examined the effects of fuel
concentration (1-9 M) and anode loading (2-5 mg/cm
2
) on the performance of their fuel cells. At
a concentration of 1 M, high current densities could not be sustained due to a mass transport
limitation. However, 3 M formic acid was found to give optimal performance and the highest
OCV (0.844 V) of all the tested concentrations. Above this concentration, decreases in peak power
density and OCV were observed, which may be attributed to an increase in fuel crossover, more
rapid adsorption of strongly binding surface species, or dehydration of the Nafion membrane due
to the hygroscopic nature of formic acid. Increasing catalyst loading at the anode from 2 to 3 to 4
mg/cm
2
resulted in significant improvements in performance, but this trend was not followed upon
further increasing the loadings to 5 mg/cm
2
. It is thought that a thicker catalyst layer may introduce
16
mass-transport limitations by inhibiting diffusion of formic acid and/or H
+
to the
catalyst/membrane interface.

Somewhat surprisingly, substituting a traditional carbon support for graphene did not result in
substantial increases in power density.
38
As listed in Table 1-3, entry 3, Pt/C achieved a maximum
power density of 53 mW/cm
2
, whereas this value only increased to 70 mW/cm
2
with Pt/graphene.
A much more substantial increase to 136.5 mW/cm
2
was observed with PtAu/graphene, which the
authors attribute to a change in the electronic structure of Pt due to the addition of Au. More
specifically, an upshift in the Pt d-band center decreases the adsorption strength of reaction
intermediates, preventing them accumulating on the catalyst surface. Although the authors did not
Table 1-3
Examples of Novel Anode Catalyst Supports
Metal
catalyst
Catalyst support
Anode
diffusion
medium
Metal
loading
[HCOOH] Cell T Max. p.d.
Max. p.d.
(standard)
Ref.
PtRu HSMSC Carbon paper 3 mg/cm
2
3 M
30°C
60°C
89 mW/cm
2
196 mW/cm
2
46 mW/cm
2

99mW/cm
2
(PtRu/VC)
37
Pd Graphite Felt None -
1 M HCOOH
+ 0.5 M H 2SO 4
60°C 85.2 mWc/m
2

39.2 mW/cm
2
(Pd black)
36
PtAu
Pt
Graphene Carbon paper 4 mg/cm
2
3 M 30°C
136.5 mW/cm
2
70 mW/cm
2
53 mW/cm
2

(Pt/C)
38
Pd
Carbon (20 nm pores)
Carbon (50 nm pores)
Carbon cloth 4 mg/cm
2
3 M 25°C
75.8 mW/cm
2
47.2 mW/cm
2

27 mW/cm
2

(Pd black)
40
Pd C-Sb 2O 5·SnO 2 Carbon cloth 1 mg/cm
2
8 M 100°C 56 mW/cm
2
30 mW/cm
2
(Pt/C)
42
Pt Nanoporous Au Carbon paper 0.013 mg 3 M 40°C 61 mW/cm
2
40 mW/cm
2
(Pt/C)
41
Pd
MWCNT
XC-72
BP
ACB
Carbon cloth 2 mg/cm
2
3 M 60°C
163.6 mW/cm
2
142.1 mW/cm
2
120.1 mW/cm
2
106.2 mW/cm
2
92.8 mW/cm
2
(Pd black)
39
Pd SmOx-C Carbon cloth 2 mg/cm
2
3 M 60°C 167 mW/cm
2
119 mW/cm
2

(Pd/C)
43
Pd
WO 3 (20 wt%)-C
WO 3 (8.05 wt%)-C
Carbon cloth 8 mg/cm
2
3 M RT
7.6 mW/cm
2

6.0 mW/cm
2

4.2 mW/cm
2
(Pd black)
44
Pd
TiO 2-C (Furfural)
TiO 2-C (Chitosan)
TiO 2-C (Saccharose)
Carbon cloth 0.5 mg/cm
2
3 M RT
*80.0 mW/mg
*39.9 mW/mg
*38.4 mW/mg
*24 mW/mg
(Pd/C)
45
Pd Ni 2P/C Carbon cloth 6 mg/cm
2
3 M 30°C 550 mW/cm
2
157 mW/cm
2

(Pd/C)
46
*Note that these values are reported on the basis of metal utilization, rather the power density
17
test PtAu/C to determine the degree to which the addition of Au affected performance, it is likely
that an improvement versus Pt/C would have been observed, as was reported elsewhere.
23

Figure 1-4. Cyclic voltammograms of Pt/C and sub-monolayer Pt deposited on nanoporous
gold (NPG-Pt) in 1 M HCOOH + 0.5 M H2SO4.
41

Contrary to co-precipitating Pt and Au on some support, directly depositing sub-monolayer Pt
on nanoporous Au (NPG) also provides a substantial enhancement in performance and stability.
41

It is clear from the cyclic voltammogram shown in Figure 1-4 that FAO on the novel Pt/NPG
catalyst proceeds via a different pathway than on Pt/C, suggesting that a change in the electronic
structure of Pt has occurred. Moreover, Pt mass-specific activity on NPG is remarkably higher
than with the commercial Pt/C, so much so that the voltammogram of the latter had to be magnified
18
10 times to clearly distinguish its peaks. Catalyst stability was investigated by an accelerated
stability test, in which the cell is cycled between 0.5 and 1 V prior to being tested in a polarization
experiment. Even after 10,000 potential cycles, the maximum power density remains the same,
suggesting that this catalyst would have an excellent lifetime in a commercial device. Although
the conditions under which stability was tested were not necessarily representative of a real-world
situation, the results are important because data of this sort is often omitted from publications on
new DFAFC catalysts.
A variety of Pd/C catalysts, with different carbon materials, were investigated for FAO activity
and compared against Pd black.
39
Along with the commonly encountered XC-72, Pd was also
supported on materials such as acetylene black (ACB), black pearls (BP), and multi-walled carbon
nanotubes (MWCNT). All Pd/C catlaysts gave improved maximum power densities versus Pd
black, as shown in Table 1-3, entry 7. In fact, MWCNT were the only support to outperform XC-
72, which demonstrates the logic in the widespread adoption of Vulcan® carbon in the fuel cell
community. Interestingly, XPS analysis revealed that the Pd 3d binding energy peaks of all carbon
supported catalysts were shifted positively respective to Pd black, which is known to increase
catalytic activity towards FAO.
47
Hence, the observed performance enhancements are not just a
function of a higher surface area and improved Pd utilization, but also the result of an electronic
effect.
Several materials consisting of transition metal oxides combined with Vulcan® carbon have also
been investigated as catalyst supports. Pd supported on a physical mixture of 85% Vulcan XC-72
and 15% Sb2O5·SnO2 gave a higher peak power density than Pt/C (56 mW/cm
2
vs. 30 mW/cm
2
),
although no comparision was made against Pd/C or Pt/ C-Sb2O5·SnO2
42
. Moreover, the authors
also reported PdAu (90:10)/C-Sb2O5·SnO2 further increases the peak power density to 61
19
mW/cm
2
, suggesting that the presence of Au in Pd catalysts is beneficial to fuel cell performance,
much like with Pt catalysts.
23,24,38,41
Wang et al. supported Pd on synthetically prepared SmOx-C
and observed that maximum power density increased by a factor of 1.4 versus commercially
available Pd/C.
43
This improvement was attributed to enhanced ECSA, an electronic effect
between the support and catalyst that weakened the adsorption strength of reaction intermediates,
and the presence of oxygen containing species on the support that could improve FAO kinetics.
Phosphotungstic acid and sodium tungstate were used as precursors to prepare WO3 (20 wt%)-
C and WO3 (8.05 wt%)-C, repectively.
44
Applied to a DFAFC, both Pd/WO3-C catlaysts yielded
higher peak power densities than with Pd black, which increased from 4.2 mW/cm
2
to 6.0 and 7.2
mW/cm
2
with the 8.05 wt% and 20 wt% examples, respectively. Although these values are very
low compared to other reports on Pd-based DFAFCs, it is likely that careful optimization of the
cell parameters (i.e. catalyst loading, membrane thickness, fuel concentration, etc.) would improve
the observed power densities by an order of magnitude. Nonetheless, as the primary focus of this
type of work is to improve catalyst design, this work may be considered a success in that Pd on a
novel support was successfully demonstrated to outperform Pd black.
Matos et al. reported on hybrid TiO2-C supported Pd catalysts that significantly improved the
metal utilization, compared to Vulcan® carbon.
45
Starting from titanium (IV) isopropoxide mixed
with either furfural, chitosan, or saccharose as the carbon sources, a solvothermal synthesis
resulted in three distinct TiO2-C materials. Of the three supports, Pd on the furfural-sourced TiO2-
C gave the highest mass-specific power density of 80 mW/mg, which was a drastic improvement
over Pd/C, at only 24 mW/mg. Chitosan- and saccharose-sourced supports were also observed to
improve Pd utilization, exhibiting mass-specific power density maxima at 39.9 and 38.4 mW/mg,
respectively. The improvement observed in all cases may be due to increased hydrophilicity,
20
arising from the incorporation of TiO2, which could increase the rate of both formic acid diffusion
to catalytically active sites and removal of the oxidation product, CO2.

Figure 1-5. Power density curves for Pd-Ni2P/C and related catalysts. [HCOOH] = 3 M, O2 at
the cathode, cell T = 30°C.
46

removed, 20 wt% Pd supported on Ni2P/C, which yielded 550 mW/cm
2
at only 30°C, holds the
record for maximum power density in a DFAFC.
46
The authors investigated Pd on various support
constitutions ranging from 10-50 wt% Ni2P on Vulcan carbon, and found that 30 wt% gave the
highest mass activity in cyclic voltammograms and best catalytic stability in chronoamperometric
studies. To confirm that the observed enhancement was specifically the result of the novel catalyst
support interacting with Pd particles, several other catlaysts were prepared for comparison. As
shown in Figure 1-5, DFAFC power density curves for PdNi/C, PdP/C, and commercial Pd/C
21
(dubbed Pd/C-C) were also obtained, but the power density maxima were significantly lower than
with Pd-Ni2P/C. The authors hypothesize that since Ni2P is known to be a hydrogen evolution
catalyst, then the adsorbed hydrogen might accelerate the FAO.
48

1.3 Mechanistic Studies
The mentioned works employ materials whose utility in DFAFCs has been well established (i.e.
Pd or Pt anode, Pt cathode, Nafion PEM, etc.) and seek to answer fundamental question about
device operation. In this section, the question of anode catalyst deactivation is analyzed, and the
poisoning mechanisms and reactivation methods from various authors are discussed. Also, the
influence of the MEA structure on DFAFC performance is examined.
1.3.1 Pd Anode Deactivation/Reactivation
The predominance of the direct pathway (1) for FAO on Pd is touted as one of the hallmark
benefits of the DFAFC. However, experience has shown that some strongly binding species
covers the surface with prolonged reaction time. In addition to the direct pathway (1), Zhou et
al. hypothesized that the two-electron oxidation of formic acid could proceed via the processes
below
47
:

𝑃𝑑 − 𝐻𝐶𝑂𝑂 𝐻 𝑎𝑑
→ 𝑃𝑑 − 𝐶𝑂𝑂 𝐻 𝑎𝑑
+ 𝐻 +
+ 𝑒 −
(9)
𝑃𝑑 − 𝐶𝑂𝑂 𝐻 𝑎𝑑
→ 𝑃𝑑 + 𝐶 𝑂 2
+ 𝐻 +
+ 𝑒 −
(10)

As such, it may be the buildup of COOHad that is responsible for the deactivation that is
often observed in FAO chronoamperomograms, as seen with Pd/GF in Figure 1-3. However,
Zhang et al. claim that CO is clearly formed on a Pd surface based on results from attenuated
total reflection-infrared (ATR-IR) spectroscopy.
49
On the contrary, Baik et al. studied the
22
nature of the “deactivated” Pd surface with the same spectroscopic technique, and did not
observe any peaks corresponding to Pd-CO or Pd-COOH.
50
Rather, they observed a Pd-OH,
which they attributed as the cause of performance degradation due to the presence of water.

𝑃𝑑 + 𝐻 2
𝑂 → 𝑃𝑑 − 𝑂 𝐻 𝑎𝑑
+ 𝐻 +
+ 𝑒 −
(11)

Work done by Yu and Pickup suggests that reaction (6) is an unlikely source of poisoning
species.
51
They found that 2 h of polarization at cell potentials of 0.7 V or less led to rapid
deactivation, but no deactivation was observed after polarization at 0.8V. Since hydroxide
adsorption on Pd should occur at cell potentials greater than 0.7 V the authors state that some
other poisoning mechanism must be responsible for performance decay, although they admit that
their results do not provide enough evidence to propose an alternative pathway.
52

Although there is no general consensus in the literature regarding the nature of this
intermediate species, its existence is widely recognized and several methods for its removal have
been proposed. One such method employs a dynamic hydrogen electrode (DHE) at the cathode,
in which hydrogen gas is flowed instead of oxygen so the Ecathode ≈ 0 V. Hence, Ecell = Eanode
according to (6), which allows for the facile application of a specific voltage at the anode. In the
presence of formic acid, potentials > 0.9 V vs. DHE can clean the Pd surface and regenerate
performance on a time scale of seconds to minutes.
30,53
If formic acid is completely removed
from the solution by consumption, then polarization for 10 min at a potential as low as 0.7 V vs.
DHE is sufficient to remove strongly bound surface species.
54

The practicality of polarization-regeneration is up for debate because, according to the
Pourbaix diagram, Pd is expected to dissolve under conditions present in a DFAFC (i.e. pH 1), at
23
the aforementioned polarization potentials. Alternatively, Zhou et al. reported that simply
flowing pure water through the anode and cathode compartments at a cell temperature of 60°C
for 1 h was sufficient to reactivate the Pd catalyst.
55
Jung et al. proposed yet another
deactivation mechanism, based on their observation that aggregation of Pd particles occurred
after 11 h of continuous operation at 200 mA/cm
2
.
56
According to the authors, an increase in
mean particle size resulted in a reduction in ECSA, which contributed to deactivation along with
the accumulation of surface species. Hence, even after applying 1.1 V vs. DHE to the anode,
performance could not be fully regenerated due to the loss in ECSA.
In contrast to the work by Jung et al., another group claimed that by holding the potential
of the DFAFC at 0.2 V (Eanode ≈ 0.7 V vs. DHE) for 20 h, with formic acid and air at the anode
and cathode, respectively, the cell performance was not only regenerated, but also better than
initially recorded.
57
This improvement was attributed to the removal of strongly bound surface
species and a reduction in catalyst particle size due to Pd dissolution, which increased ECSA and
the total oxidation charge at the anode by (12).

𝑃𝑑 2+
+ 𝐻𝐶𝑂𝑂𝐻 → 𝑃𝑑 + 𝐶 𝑂 2
+ 2𝐻 +
(12)

Therefore, the formation of Pd
2+
may be beneficial to charge production because reaction (12)
happens faster than reaction (10), so formic acid could be consumed before the formation of
strongly adsorbed intermediates occurs.
1.3.2 MEA Structure
The MEA may be considered the heart of the fuel cell, because it is where the key processes
for device operation occur (i.e. reactant diffusion, fuel oxidation/oxidant reduction, proton
24
transport, electrical conduction, etc.). Typically, the MEA is prepared with a catalyst-coated
diffusion layer (CCDL) or a catalyst coated membrane (CCM). In either case, a slurry is prepared
from powdered catalyst suspended in a small amount of water, isopropanol, and Nafion ionomer,
which is then brush- or spray-painted onto the GDL or directly on either side of the PEM. In the
case of a CCDL, the painted electrodes are usually hot-pressed around the PEM to complete the
process.
21
However, hot-pressing is not necessary with a CCM because a GDL may simply be
placed on top of the anode and cathode catalyst layers.
30
As an alternative to these methods, the
use of graphite felt as a simultaneous catalyst support and diffusion medium has been reported.
36,58

It may be worth noting that the record power density of 550 mW/cm
2
with Pd-Ni2P/C at the anode,
at a temperature of only 30°C, was achieved using a CCM.
46

Han et al. utilized an electrospray method in an effort to improve the dispersion of catalyst
particles on the GDL.
59
Using a liquid aerosol created via electro-static charging between the
substrate and spray-nozzle, charged ink particles were attracted (i.e. electrosprayed) to the opposite
polarity GDL. In addition to promoting adhesion to the substrate, charging is thought to mitigate
agglomeration due to electrostatic repulsion between catalyst particles. Anode loadings of 1, 3,
and 7 mg/cm
2
Pd black were electrosprayed on the GDL, and it was found that 3 mg/cm
2
was the
optimal loading at a low flow rate (5 mL/min). Below this value, performance was worse because
of fewer catalytically active sites, but above, catalyst layer thickness introduced a mass-transfer
limitation. Although, this could overcome upon increasing the flow rate to 10 mL/min or higher,
thus allowing for the MEA with 7mg/cm
2
Pd black to exhibit the best performance.
The effect of Nafion content in the anode
60,61
and cathode
62
catalyst layers had also been
studied. Heat-treatment of the anode catalyst ink at 80°C results in improved dispersion of Nafion
ionomers, and a concurrent increase in maximum power density to 70mW/cm
2
from 53.8 mW/cm
2

25
with the ink pre-treated at 25°C.
60
The origin of this improvement may in part arise from an
increase in Pd utilization. According to CO-stripping experiments, Pd electrodes pre-treated at
25°C and 80°C exhibited ECSAs of 8.39 m
2
/g and 21.2 m
2
/g, respectively. Moreover,
electrochemical impedance spectra of the pre-treated electrodes revealed that increasing the
temperature decreased the charge transfer resistance from 26.6 Ω·cm
2
to 4.72 Ω·cm
2
, suggesting
better dispersion and electrical contact between Pd and Nafion.
Masel et al. studied the effect of Nafion loadings of 10, 30, and 50 wt.% at the anode.
61

Optimal performance was observed with 30wt.%, which may be because 10 wt.% is an inadequate
amount for proton conductivity and a polymer film is formed at 50wt.%, thereby reducing ECSA.
Interestingly, Nafion may also act as a stabilizing binder in the catalyst layer matrix, to mitigate
dissolution of Pd at oxidative potentials >0.8 V vs. DHE. As seen in Figure 1-6, after 80 potential
cycles, FAO oxidation current is significantly decreased in the 10 wt.% sample. On the contrary,
the electrode with 30wt.% Nafion retained most of its catalytic activity. The authors state that this
fact is particularly important if anodic stripping pulses are to be used for CO-removal (i.e. cell
regeneration), such that the loss in surface area due to Pd corrosion can be minimized. On the
topic of catalyst poisoning, it was also observed that CO-coverage increased with increasing
Nafion content. Nafion, which contains sulfonic acid groups (-SO3-) may be able to directly
dehydrate formic acid to CO, as has been observed with polystyrenesulfonic acid,
63
which
contribute to Pd deactivation.
26

Figure 1-6. Cyclic voltammograms for Pd electrode with 10 wt.% (top) and 30 wt.%(bottom)
Nafion content, in 12 M HCOOH +0.1 M H2SO4.
61

27
In addition to the effect of Nafion at the cathode, Kim et al. studied the effect of substituting
Pt/C (46.5 wt.% Pt) for Pt black as the ORR catalyst.
62
With Pt black, a Nafion ionomer/catalyst
(NI/C) volume ratio of 1.0 gave the highest current density at 0.6 V. At 0.3 V, current density
increased dramatically as NI/C value increased from 0.25 to 0.50. However, further increasing the
volume ratios to 1.0 and 2.0 resulted in only minor increases in current density. With Pt/C at the
cathode, current density at both 0.3 V and 0.6 V increased relatively linearly between NI/C values
of 0.2 and 0.8, where it reached its maximum, but exhibited the poorest performance at a volume
ratio of 1.1. Thickness of the MEA increased with Pt/C as the cathode catalyst. In cell tests, Pt
black performed better than Pt/C, even though the latter had significantly higher ECSA, possibly
due to mass-transport limitation of oxygen through the carbon support. Improvement of mass-
transport in the cathode catalyst layer could potentially be achieved with the addition of a pore-
former to the ink, as was done with a PtRu black anode in a DFAFC.
64
The highest improvement
in performance was observed with 17.5 wt.% Li2CO3 in the catalyst ink, which is subsequently
removed by an acid wash, leaving behind a porous framework. At 0.3 V, current density increased
27 % from 300 mA/cm
2
to 380 mA/cm
2
with the incorporation of a more porous structure, which
is thought to allow for improved diffusion of formic acid and CO2 towards and away from
catalytically active sites, respectively.

28

Figure 1-7. Polarization curves for a DFAFC with membranes of different thicknesses.
65

Of equal importance to cell performance and MEA optimization as the catalyst layers are
the properties of the PEM itself. Recently, the effect of decreasing membrane thickness in a
DFAFC was examined.
65
Under the same conditions, formic acid experiences less fuel crossover
than methanol, by a factor of six,
66
so DFAFCs should be able to tolerate a thinner membrane
without experiencing cathode poisoning from adsorbed formic acid and /or CO. Comparing
Nafion 117 (.007 in), Nafion 115 (.005 in) and NR 212 (0.002 in) in a Pd-anode/Pt-cathode MEA,
a maximum power density of 301 mW/cm
2
was observed with the thinnest membrane, as shown
in Figure 1-7. Upon increasing membrane thickness, maximum power density decreased, which
was attributed to higher ohmic resistance of the thicker membranes and an associated increase in
29
CO-formation from Nafion-catalyzed dehydration of formic acid due to the presence of more -
SO3- sites.
63

Xing et al. reported on the MEA fabrication process, and brings into question the
widespread practice of hot-pressing.
67
They compared a hot-pressed MEA with CCDLs against a
non-pressed control (in which the CCDLs were simply placed on either side of the membrane) and
introduce a third procedure, called the “decal” method. In this new process, catalyst inks are
sprayed-onto a high density PTFE film, before being pressed against the Nafion 117 membrane
and covered with GDLs. A slight performance improvement was observed with carbon cloth as
the GDL, rather than carbon paper. Of the three MEAs, the one that was hot-pressed performed
the poorest. The others exhibited the same polarization profiles until reaching a current density of
about 100 mA/cm
2
, after which the decaled-MEA exhibited higher power densities. SEM images
of the hot-pressed and decaled MEAs, shown in Figure 1-8, reveal that the anode and cathode
catalyst layers appear to have been crushed, which may have impeded contacts between the
catalysts, PEM and reactants at the triple phase boundary.
68

30

Figure 1-8. SEM images of hot-pressed (top) and decaled (bottom) MEAs.
67

31
1.4 Practical Devices
In terms of the need to answer our increasing energy demands, the demonstration of DFAFCs
that may be applied to real world situations will be the most valuable. Design of such devices
differs between authors, and depends on the stated energy production goal. But generally, fuel is
stored in a reservoir adjacent to the anode which reaches the catalyst by means of mass-transport
through some diffusion medium, and the cathode is air-breathing such that oxygen may be obtained
from ambient air. Cell performance is measured in either constant power or constant current, and
a cell lifetime (i.e. time required before refuelling) is usually reported, a key data point often
omitted in DFAFC research publications. The manner of supplying the fuel and oxidant can be
classified as either passive or active. The majority of works in this area report passive devices
incorporating an onboard supply of formic acid, facilitating easy refill, and use oxygen from the
air. The latter active devices rely on externally supplied fuel that is stored in a reservoir, and may
be equipped with an air compressor or fed with a gas tank.
The first passive DFAFC was reported by Masel et al., and used PtRu black at the anode.
69

Lifetime tests were performed with a CCM, but no GDL was used so as to minimize mass transfer
limitations of the fuel cell. With 8.8 M formic acid, the cell produced about 10-11 mW/cm
2
over
8 h. In tests that utilized a GDL, carbon cloth that had been modified with oxygen plasma was
found to improve cell performance versus plain or teflonized carbon cloth, due to better mass-
transport with higher hydrophilicity. Overall, the authors showed that the PtRu-based DFAFC
performed at a level comparable to DMFCs and that the difference in fuel energy density could be
better compensated for by using higher concentrations of formic acid.
32

Figure 1-9. Voltage transients at a constant current of 20 mA as a function of time.
70

Although Pd catalysts favor the direct pathway, Pd/C still exhibits lower stability than carbon
supported Pt or PdPt.
70
As shown in Figure 1-9, a current density of 20 mA could be sustained for
13 h with Pt/C and only 9.5 h with Pd/C, whereas PdPt/C exhibits a cell lifetime of 17 h. With
Pt/C and PdPt/C, the rapid decline in cell voltage after about 10 h and 14.5 h, respectively, was
attributed to fuel consumption. The notion of which catalyst is “better” depends on the point of
view, since Pd/C delivers the highest output voltage (t < 3 h), and Pt/C is capable of the highest
stable operating voltage (~10 h), but the PdPt/C combination can stably discharge at constant
current for the longest period of time. As such, the authors state that the anode catalyst should be
chosen based on the specific energy demands of an application.
33
Two years after their original publication on a passive DFAFC, Masel et al. demonstrated
another fuel cell that employed Pd black at the anode, which could sustain a current density of
100mA/cm
2
for 4.4 h before refill.
71
With refueling at 4.4 h and 7.3 h, the aforementioned current
density could be sustained for almost 10 h, but the anode overpotential increased more rapidly
after each refuel as seen in Figure 1-10.

Figure 1-10. Anode, cathode (both vs. RHE) and cell potentials for a passive DFAFC
operating at 100 mA/cm
2
.
71

This is may be due to the accumulation of reaction intermediates on the Pd surface, which will
necessitate refueling at ever-smaller time intervals in order to sustain a given current density. As
mentioned in Section 2.2.1, anodic stripping pulses or washing with pure water removes strongly
bound surface species, hence, the incorporation of such techniques may allow for longer cell
operation times. In any case, the authors point out that their passive DFAFC with 10 M formic
acid can produce 140 mA/cm
2
at 0.72 V, whereas a passive DMFC with 4 M methanol must be
34
discharged to 0.28 V to achieve a comparable current density.
72
Hong et al. developed a passive
DFAFC based on printed circuit board (PCB) technology, which consisted of Au-coated Cu
current collectors.
73
With very low loadings (≤ 2 mg/cm
2
) of Pd (20 wt.%)/C at the anode and Pt
(40 wt.%)/C at the cathode, a constant current density of 5 mA/cm
2
was sustained 3 h. Even though
cell voltage dropped only 20% (0.64 to 0.51 V) during this time, the experiment was terminated at
the 3 h mark, so the cell lifetime and effects of refueling were not properly evaluated.

Size is another consideration in the development of commercial DFAFCs, hence there have been
some efforts to make miniaturized devices. Masel et al. reported on a DFAFC “battery” with a
Table 1-4
Overview of DFAFCs as Practical Power Sources
Cell Type Anode Catalyst
Fuel-Oxidant
Supply
[HCOOH]
MEA
type
Power
Sustainable
Current
Lifetime

Ref.
Single MEA PtRu passive 8.8 M CCM ~11 mW/cm
2
- 8 h 69
Single MEA
Pt/C
Pd/C
PdPt/C
passive 5 M CCM - 20 mA/cm
2
13 h
9.5 h
17 h
70
Single MEA Pd passive 10 M CCM - 100 mA/cm
2
4.4 h 71
Single MEA Pd/C passive 5 M CCM - 5 mA/cm
2
- 73
Mini-fuel cell battery Pd passive 12 M CCM
0.12 mW/μL
‡
0.27 mW/μL
‡

0.39 mW/μL
‡

-
25 min
15 min
8 min
74
Micro-tubular cell Pd passive 5 M n.a.†

>3 mW/cm
2

~1 mW/cm
2

-
>1 min
>10 min
75
Micro-fluidic cell Pd/Pt passive neat n.a.† ~0.2 mW/cm
2
- 6 h 76
Nano-fluidic cell Pd/C active 3 M n.a.† *100 mW/cm
2
*~500 mA/cm
2
- 77
4 MEA-stack Pt/C passive 5 M CCM - 20 mA

7 h 78
2 MEA-stack Pt/C passive 5 M CCM - 20 mA 6 h 79
10 MEA-stack Pd/C+ Pt/C (mixed) active 10 M CCDL - 1 A 50 h 80
15 MEA-stack
w/BOP components
PtRu active 11 M CCDL 30 W - 2 h 81
*Power density and current density maxima from polarization curves. Constant current/voltage operation was not demonstrated.
†These works did not use the traditional MEA structure
‡Converted from average power densities reported in W/L
35
total volume of 11 mm
3
.
74
With a built in reservoir of only 6 μL of 12 M formic acid, the cell
could be discharged with average power densities of 0.39, 0.27, and 0.12 mW/μL for 8, 15, and 25
minutes, respectively. While this work employed what was essentially a miniaturized MEA, other
devices have been reported that completely do away with the traditional MEA design. Qiao et al.
developed a “micro-tubular” DFAFC using a catalyst-coated Flemion® tube.
75
Flemion®, another
fluorinated cation-exchange membnrane, acts as the PEM, akin to Nafion® in a standard MEA,
but as 3.5 cm tube that was coated with the Pd anode on the inside and the Pt cathode on the
outside. When discharged at 200-300 mV with 5 M formic acid, the micro-tubular device could
produce > 3mW/cm
2
for at least 1 min, after which the fuel was quickly exhausted. At 400 mV
about 1 mW/cm
2
could be sustained for at least 10 min, however, it seems like the experiment was
terminated after this time point, before complete fuel consumption. Moreover, this device was
successfully used to power a small propeller.

Figure 1-11. Schematic of a microfluidic DFAFC (left) and the electrical circuit used for
measurements (right). V1 and V2 represent the potentials at the anode and cathode vs. a
reference electrode (RE), respectively. V3 is the cell voltage measured across an external
resistor, R.
76

36
Eriskson et al. proposed another alternative design, a microfluidic DFAFC, shown in Figure 1-
11, that consisted of parallel glass supported thin-film electrodes (a and c) embedded in a PDMS
support, and connected via a network of micrometer scale channels (green lines), further covered
by a PDMS membrane.
76
A PDMS block (purple cylinder) saturated with neat formic acid and
0.5 M H2SO4 supporting electrolyte, which can be thought of as a fuel cartridge, was placed on
top of the membrane over the anode channel. Oxygen was obtained from the air via passive
diffusion through the PDMS-covered cathode channel. Due to fuel consumption and evaporation,
the formic acid-charge PDMS block had to be replaced about every 6 h. Nonetheless, the cell
could sustain a constant power of ~0.2 mW/cm
2
for about 18 h, if the fuel cartridge was replaced
accordingly. Although, the lack of a PEM separating the electrodes means that poisoning of the
cathode is very likely, since formic acid can diffuse through the micro channels and reach the
cathode. In fact, the authors note that by the third formic acid-PDMS cartridge, incomplete fuel
capture at the anode causes poisoning at the cathode, via diffusion of the excess formic acid. In
addition, it seems likely that the cathode poisoning would be exacerbated by the inevitable
accumulation of surface poisons at the anode, which would reduce the number of catalytically
active sites, leaving formic acid free to reach the opposite electrode.
Recently, Ortega et al. developed another air-breathing microfluidic DFAFC, but deemed it to
be nanofluidic, perhaps because of the use of carbon nanofoam as the catalyst supports.
77
Unlike
the previously discussed passive DFAFCs, the fuel and oxidant were actively supplied to the anode
and cathode at a rate of 100 and 200 μL/min, respectively. Much like in other works, H2SO4 was
mixed with formic acid to prepare the anolyte.
33,36
Interestingly, a liquid catholyte was used rather
than a gas stream, prepared by bubbling pure oxygen through 0.5 M H2SO4, which then combined
with oxygen from the air that was obtained from the open window at the cathode, as seen in
37
Figure 1-12. Under these conditions, with 3 M formic acid + 0.5 M H2SO4 and only 0.15 mg Pd
at the anode, a maximum power density of 100 mW/cm
2
was demonstrated, and the device could
be discharged to a current density of almost 500 mA/cm
2
.

Figure 1-12. Schematic of the nanofluidic DFAFC, with actively supplied fuel and oxidant.
77

Although this cell design is very simple and the polarization data is comparable with traditional
MEA-based fuel cells, no lifetime experiments were performed to determine the overall stability
of the system. Since there is no divider between outlet streams, it seems plausible that formic acid
could easily reach the cathode and rapidly deactivate the cell.
38

Figure 1-13. A four-cell DFAFC stack. In the two-cell stack, there is only one MEA on each
side of the reservoir.
78

Along with single MEA- and miniaturized-DFAFCs, there exists a third approach towards
practical devices, in which MEAs are combined in stack around a central fuel reservoir. Hong et
al. have investigated both two-cell
79
and four-cell stacks
78
(Figure 1-13), both of which were
operated under passive conditions. With the two-cell DFAFC, stack voltage remained at about 1.1
V during the first 5 h, after which there was a rapid decrease to 0.20 V over the next 1.5 h, which
was attributed to fuel consumption. Since formic acid concentration decreased from 5.0 M to 0.20
M throughout the course of the experiment, mass transport to the catalyst surface becomes a
limiting factor after several hours of operation.

39

Figure 1-14. Two-cell and four-cell DFAFC stacks discharging at 20 mA, with 3.5 mL of 5 M
formic acid (left). Stack voltage and internal resistance as a function of discharge cycle number
(right).
79

40
As shown in Figure 1-14, the two-cell stack exhibited excellent reproducibility, capable of
generating 20 mA of current for about 6 h after each refueling, but internal resistance increased
from 150 mΩ after the first cycle to 350 mΩ after the fourth. It is interesting to note that both
the two-cell and four-cell DFAFCs utilized Pt/C at the anode rather than a Pd catalyst. Hence,
the rapid buildup of CO on Pt may have been a major contributing factor to the increase in
internal resistance and decrease in cell performance. The authors also state that the observed
increase may arise from other causes, including MEA de-lamination or contamination, and the
accumulation of excess water (i.e. flooding) at the cathode. Similarly, the four-cell stack could
sustain 20 mA for 7 h, before the stack voltage began to decline rapidly. The experiment was
terminated after 10 h, so it is not known for how long the system could have potentially
discharged, and the effects of refueling were not investigated. Although, it is likely that the same
trends in discharge reproducibility and resistance increase would have been observed, since the
two-cell and four-cell DFAFCs were identical except for the number of MEAs.
Even larger DFAFC stacks have also been reported. Cai et al. developed a 10-MEA DFAFC
with an active supply of formic acid and pure oxygen.
80
A mixture of Pt/C +Pd/C at the anode
was found to yield superior performance compared with either catalyst alone. To evaluate
lifetime, 1.5 L of 10 M formic acid was actively circulated through the system while a current of
1 A was drawn from the device. As seen in the top panel of Figure 1-15, 1 A could be sustained
for 240 h by replacing the anolyte about once every 50 h. Moreover, the bottom panel of the
above Figure shows polarization data obtained before (squares) and after (triangles) the 240 h
lifetime test, demonstrating that performance had decreased, probably due to the accumulation of
surface poisons. In agreement with the work by Zhou et al.,
55
the cell could be reactivated
41
(circles) to yield a peak power of 32 W, a value 60% greater than initially recorded, simply by
washing the anode with pure water for 1 h.

Figure 1-15. 10-cell DFAFC discharging at 1A for 240 h (left). Polarization curves (right)
prior to lifetime test (squares), after the lifetime test (triangles), and after reactivating the anode
by washing with pure water for 1 h (circles).
80

42

Figure 1-16. A 15-cell DFAFC (a) and the hybrid system used to power a laptop (b).
81

Only one DFAFC has been applied to a real-world application, in which a 15-MEA stack was
used to power a laptop.
81
Therein, the authors describe a PtRu-based hybrid system consisting of
15-MEA DFAFC (Figure 1-16-a) along with a fuel tank, tubing, miniature liquid pump,
miniature air compressor, cooling fans, a small battery, and a power conditioning control board
(Figure 1-16-b). The battery, which was used as the power source to start the system, was then
recharged by the DFAFC, when it reached steady-state operation. In fact, after the initial start-
up, the fuel cell stack alone was able to supply enough energy to power the component parts of
the hybrid system, and power the laptop for about 2 h.
43

Figure 1-17. One operation cycle of the 15-MEA DFAFC-hybrid system. The fuel tank was
filled with 280 mL of 11 M formic acid.
81

Figure 1-17 shows the voltage profile of the DFAFC-hybrid system while it was used to power
a laptop. At 0 min, the laptop and power source were started together, and the battery drives the
boot-up process while the DFAFC reaches a stable OCV (~8 V). Then, 1.3.2 is a sharp drop in
stack potential to 6 V at about 5 min, which corresponds to the time when current starts to be
drawn from the fuel cell. Upon drawing current, a moderate screen brightness and computational
load was set to create an environment in which the computer exhibited above average power
requirements. Also at this time, the authors selected a “comfortable screen brightness, and
medium computation load…to provide an above average draw by the computer throughout the
test”. As such, the experiment mimicked actual working conditions that a laptop user might
experience. The fuel cell stack operated around 6 V for about 120 min, until it could no longer
44
supply sufficient power to the laptop, due to a decrease in fuel concentration (from 11 M to 2.1
M) and resulting mass-transfer limitation. For the remaining 30 min of the experiment, the on-
board battery powered the system.
1.5 Direct Formate Fuel Cells
Unlike PEMFCs, which are based on proton transport from anode to cathode, operation
of alkaline fuel cells (AFC) relies on the transport of hydroxide ions through an AEM in the
reverse direction. The half-cell reactions for a DFFC are as follows:

HCOO
-
+ 3 OH
-
 CO3
2-
+ H2O + 2 e
-
E
0
anode

= -1.05V (13)
½ O2 + H2O + 2 e
-
 2 OH
-
E
0
cathode

= 0.40 V (14)
HCOO
-
+ OH
-
+ ½ O2  CO3
2-
+ H2O E
0
DFFC = 1.45 V (15)

The theoretical open circuit voltage (OCV) for a DFFC is higher than that of both alkaline
methanol fuel cells (1.21 V) and alkaline ethanol fuel cells (1.14). Although it is possible to
operate DFFCs with formate salts exclusively
82
better cell performance is observed when
hydroxide is added to the fuel stream.
10,83
This may be due to inefficient anion transport across
the membrane, partly resulting from sluggish oxidation kinetics, hence the necessity to supply
added hydroxide at the anode.
One of the motives for the development of DFFCs is the possibility to use non-noble
metal catalysts at both the anode and cathode. It is well known that oxygen reduction kinetics
are more facile in an alkaline environment.
84
Our group has studied the oxidation of formate and
formic acid on a PdAu catalyst supported Vulcan XC-7 TaC, as shown in Figure 1-18.
Interestingly, the peaks for the forward and reverse scans of the cyclic voltammograms for
45
potassium formate are cathodically shifted about 100mV relative to the formic acid peaks,
suggesting that the oxidation of formate is more facile. Formate salts are not known to form any
poisoning species upon oxidation under alkaline conditions.
10
On the contrary, adsorbed formate
is considered to be a poisoning species in the formic acid oxidation on Pd, the accumulation of
which eventually leads to a decrease in catalytic activity.
8

Figure 1-18. Cyclic voltammograms in formic acid (dotted line) and potassium formate
(solid line), normalized to mgPd. Potentials at working electrode (WE) are reported with
respect to mercury sulfate electrode (MSE).

1.5.1 Initial Reports
To the best of our knowledge, the very first report of a DFFC came from the Research
Division of the Allis-Chalmers Manufacturing Company.
16
Therein the authors describe a
“sandwich” type fuel cell consisting of grooved, stainless steel current collector plates, and
46
porous nickel electrodes insulated by an asbestos separator. A variety of Pd and Pt mixtures
were tested at both the anode and cathode, and the authors conclude that, at 90° C, Pt and Pd:Pt
(1:1 by weight) are the best performing catalysts at the anode and cathode, respectively.
Although, a Pd anode was superior to a Pt anode at 30°C. In a fuel mixture of 4 M KCOOH + 4
M KOH, with O2 as the oxidant, a current of 200 A/ft
2
(~215 mA/cm
2
) could be drawn from the

cell at 0.82 V. This device could also be operated using 4 M KCOOH and 4 M K2CO3, but it
was necessary to discharge the cell to 0.5 V in order maintain the same current density as with 4
M KOH, although it was unclear why hydroxide was superior to carbonate. In conclusion, the
authors stated the “The output of formate ion-oxygen fuel cells approaches that of hydrogen and
hydrazine-oxygen fuel cells. This allows the formate ion-oxygen cells to be applied in special
applications”.

Table 1-5
Recently Studied DFFCs
Anode
catalyst
Cathode catalyst Membrane
Anode:Cathode
loadings
Fuel Mixture Oxidant Cell T Max. p.d. Ref.
Pd/C Ag/C
Polybenzimidazole
-based (FuMA-
Tech GmBH)
A: 4 mg Pd/cm
2

C: 8 mg Ag/cm
2

6 M HCOOK
+2 M KOH
O 2 120°C 160 mW/cm
2
85
Pd Pt Tokuyama A201 2 mg/cm
2
1 M HCOOK
+ 2 M KOH
O 2
air
60°C
144 mW/cm
2
125 mW/cm
2
10
Pd Pt Tokuyama A201
†
A: 4+4 mg/cm
2
C: not stated
1 M HCOOK
+ 2 M KOH
O 2
air
60°C
267 mW/cm
2
167 mW/cm
2
83
Pd Pt Tokuyama A201
†
A: 2+2 mg/cm
2
C: 4 mg/cm
2
1 M HCOOK
O 2
air
50°C
106 mW/cm
2
76 mW/cm
2
82
Pd/C
HYPERMEC K14 [Fe-
Co] (Acta)
‡
QAPSF- based
(homemade)
2 mg/cm
2
5 M HCOOK
+ 1 M KOH
O 2 80°C 250 mW/cm
2
86
Pd/C
Pt/C
MnNiCoO 4/N-MWNCT
None
A: 1 mg Pd/cm
2

C: 1 mg Pt/cm
2
2 M HCOOK
+ 2 M KOH
O 2
60°C
50°C
75 mW/cm
2
90 mW/cm
2

87
88
PdAu/Ni-foam PdAu/Ni-foam Nafion 115 3 mg/cm
2

1 M HCOONa
+ 2 M NaOH
1 M H 2O 2 +
1 M H 2SO 4
25°C
60°C
214 mW/cm
2

331 mW/cm
2

89
†Equal amounts of catalyst were spray-painted directly on the membrane and also brush-painted on the GDL
‡Quaternary ammonia polysulfone (QAPSF)
47
1.5.2 AEM Fuel Cells
There was only one other report on a formate fuel cell in the 20
th
century; a short
communication that highlighted the benefits of using potassium salts rather than sodium salts in
the anolyte.
90
It was not until 2012 that interest in the field was rekindled by Jiang and
Wieckowski’s publication on a prospective DFFC.
85
Therein the authors described a device
capable of producing 160 mW/cm
2
at current densities ranging from 250 to 500 mA/cm
2
. They
also exploited some of the inherent capabilities of working in alkaline media, such as a non-
noble, Ag/C cathode catalyst, and a very high operating temperature of 120°C. Although, it is
likely that high temperatures such as this would be detrimental to cell lifetime due to the
degradation of the AEM via hydroxide-induced beta elimination, demethylation, and/or
debenzylation.
91
In order for hydroxide migration to occur, there must be a counter-ionic species
grafted to the polymer back-bone, which is usually a quarternary ammonium species, as shown
in Figure 1-19. During operation, hydroxide attack on hydrogen atoms beta to the quaternary
group results in the formation of water through Hoffman elimination, thereby decreasing anion-
exchange capability of the AEM.
Between 2012 and 2013, the Haan group released three reports on DFFCs that
investigated various aspects of cell performance.
10,82,83
In their intial publication, a device
exhibiting a peak power density of 144 mW/cm
2
at 60°C was demonstrated, a value comparable
to Jiang and Wieckoski’s work, but at half the operating temperature.
10
This is possibly due to
the fact that more active anode and cathode catalysts were used (Pd black vs. Pd/C, and Pt black
vs. Ag/C). Furthermore, improved hydroxide transport characteristics and membrane stability
may have played a role, since the Tokuyama A201 membrane is generally considered to be the
48
most well-developed AEM.
91
It was also observed that fuel cells perform better with potassium
formate than sodium formate, as had been reported previously.
90

Figure 1-19. Possible AEM degradation mechanisms that may occur in a DFFC.
91

The polarization curves of the aforementioned DFFC are shown in the left panel of
Figure 1-20. Air could be supplied to the cathode instead of oxygen, but power density was
lower than with oxygen. Although, with the same MEA, significantly better performance was
observed with formate than with ethanol in both cases. The oxidation of formate, ethanol and
formic acid were investigated in a 15 h constant potential experiment with a rotating disk
electrode, as shown in the right panel of Figure 1-20. Regardless of potential, formate oxidation
remains relatively stable throughout the duration of the experiment after a brief activation period,
whereas current densities of formic acid and ethanol oxidation ultimately fall to negligible
values. As discussed previously in Section 2.2.1, FAO on Pd results in the formation of a
49
strongly binding “CO-like” intermediate that deactivates the catalyst surface. Similarly, the
incomplete oxidation of ethanol on Pd also results in deactivation.
92

Building on their previous work, the Haan group proposed an optimized anode
fabrication method, wherein half of the catalyst, by mass, is brush-painted on the GDL and the
other half is spray painted on the AEM.
83
In so doing they were able to further push the peak
power density of their devices, using 1 M HCOOK + 2 M KOH, to 267 and 167 mW/cm
2
operating under oxygen and air, respectively. As seen in Figure 1-21, excellent performance was
also demonstrated in the absence of added hydroxide with both oxygen and air at the cathode,
yielding peak power densities of 157 and 105 mW/cm
2
, respectively. Hydroxide-free devices
may be advantageous in that less carbonate formation may occur, which could precipitate in the
electrolyte, thereby blocking mass-transport channels and increasing cell resistance.
Furthermore, the authors observed that it was feasible to operate a DFFC at low temperature, but
with hydroxide added to the fuel stream. Upon decreasing the temperature from 60°C to 45°C to
35°C, peak power densities decreased from 267 to 185 to 141 mW/cm
2
, respectively. Although,
the cell could not sustain high current densities at the lowest temperature, which was likely due
to sluggish reaction kinetics at the electrodes, and hydroxide transport across the membrane.
Applying their findings regarding the optimized anode structure, the concept of low
temperature, hydroxide-free DFFCs was further explored.
82
Even at temperatures as low as 23-
50°C, reasonable current and power densities could still be achieved, irrespective of the oxidant.
In fact, a peak power density of 27 mW/cm
2
was still possible at room temperature with a
formate-air fuel cell. Although, higher current densities could only be sustained at 40°C and
above with oxygen, and at 50°C with air at the cathode, respectively. In both cases, peak power
density was found to be linearly proportional to cell temperature.
50

Figure 1-20. Polarization curves of AFCs with 1 M HCOOK/2 M ethanol + 2 M KOH
supporting electrolyte at the anode and oxygen or air at the cathode. Fuel and cell at 60°C (left).
Chronoamperomograms of HCOOK and ethanol oxidation at various potentials with a Pd black
working electrode. Electrolyte solutions were 1 MHCOOK/1 M ethanol + 1 M KOH, or 1 M
HCOOH + 1 M H2SO4 (right).
10

51

Figure 1-21. Polarization curves of a DFFC with an optimized anode structure operating
under oxygen (left) and air (right).
83

52

Figure 1-22. Voltage profiles of a hydroxide-free DFFC operating at 40°C, discharging at 100
mA/cm
2
during a 24 h period.
82

As shown in Figure 1-22, a hydroxide-free DFFC could be operated at 100mA/cm
2
with
either air or oxygen for 24 h, during the course of three 8 h discharges. Although formate does
not poison Pd in alkaline media, the authors suggest that the accumulation of excess formate is
responsible for the voltage decay. Hence, a polarization measurement was taken every 8 h,
which regenerated the cell voltage to its initial level by “cleaning” the catalyst surface.
Polarization curves obtained under oxygen at the 0, 8, 16, and 24 h time points revealed that peak
power densities remained stable at 89, 88, 87, and 90 mW/cm
2
, respectively. The same trend
was observed with air, but the average power density was 61 mW/cm
2
throughout the 24 h
period.
53

Figure 1-23. Polarization curve for a DFFC with a home-made AEM and non-noble ORR
catalyst.
86

In 2014, another DFFC was proposed that incorporated a homemade AEM, Pd/C at the
anode, and a non-precious Fe-Co catalyst at the cathode.
86
These authors returned to the practice
of high-temperature operation (80°C), which was responsible for sustaining current densities
greater than 800 mA/cm
2
, even in the absence of hydroxide. In contrast with work carried out by
the Haan group, the optimal fuel:electrolye ratio was found to be 5:1, which yielded a peak
power density of about 250 mW/cm
2
and could sustain current densities greater than 1200
mA/cm
2
with oxygen, as seen in Figure 1-23. Furthermore, the homemade AEM, prepared from
quarternary ammonium poly sulfone (QAPSF) ionomer solution, demonstrated excellent stability
at a current density of 100 mA/cm
2
, and could discharge for over 130 h without significant
degradation. The internal resistance of the membrane increased to about 0.24 Ω cm
2
after 20 h of
operation, but remained stable for the duration of the experiment. It is worth noting that the
54
MEA was prepared with a CCDL, and it would be interesting to see if the power densities
reported here could be further improved by adopting Haan’s optimized anode structure, with
equal amounts of catalyst loaded on the GDL and membrane itself, as well.
83

1.3.3 Alternative-Electrolyte Fuel Cells
Most recently, two novel designs for a DFFC have been proposed that break away from
the standard, AEM-based devices. Based on the observation that Pd is very active for the
formate oxidation, whereas Pt is not, Yu and Manthiram developed a membraneless DFFC in
which both the anode and cathode were exposed to the fuel stream, as shown in the schematic in
Figure 1-24.
87
Their device is self-described as a “game changer” since it does away with the as-
of-yet unviable AEM, which easily allows for devices to be scaled as small or as large as the
electrodes are able to be fabricated. Unlike traditional DFFCs, the anolyte passes through a
flow-field between the electrodes, which are separated from the solution by diffusion layers.
Oxygen from air, that is supplied to the cathode, is reduced to hydroxide, which diffuses into the
fuel stream, then to the anode to participate the in the oxidation of formate to carbonate. The
polarization data presented in Figure 1-25 is reported in specific power (mW/mg), but may be
read as power density (mW/cm
2
) since Pd was loaded at the anode to 1 mg/cm
2
. Naturally, this
also applies to the specific current. As such, a peak power density of about 75 mW/cm
2
was
observed at 0.6 V and 100 mA/cm
2
, when operating at 60°C. In fact, when Pt/C was substituted
for MnNiCoO4/N-MWNCT, a non-noble ORR catalyst that is even less active for the formate
oxidation the power and current densities at the same operating potential increased to about 90
mW/cm
2
and 120 mA/cm
2
, respectively.
88
Moreover, these values were obtained in a DFFC
operated at an even lower temperature of 50°C, so it is likely that performance could be even
further improved if the cell temperature were raised to 60°C.
55

Figure 1-24. Schematic of a membraneless DFFC.
87

56

Figure 1-25. Polarization data for the membraneless DFFC.
87

The second novel approach was proposed by Li et al., wherein a direct formate-peroxide
fuel cell (DFPFC) is described that operates with alkaline formate at the anode, and acidic
peroxide at the cathode.
89
The cell employs a Nafion 115 membrane, but is driven by the
transport of sodium ions rather than protons, as in the case of a traditional PEMFC. Hydrogen
peroxide is still reduced to water, but charge balance is achieved via the formation of Na2SO4, as
seen in the schematic in Figure 1-26. Such a design may be advantageous to the eventual
commercialization of a formate fuel cell, since the chemistry of Nafion has been much more
thoroughly investigated than counterpart AEMs. According to the cathodic half-cell reaction
below, there is an inherent energetic advantage to using peroxide rather than gaseous oxygen:
57

Figure 1-26. Schematic of an alkaline formate-acidic peroxide fuel cell.
89

𝐻 2
𝑂 2
+ 2𝐻 +
+ 2𝑒 −
→ 2𝐻 2
𝑂 E
0
cathode

= 1.78 V (16)
𝐶𝑂𝑂 𝐻 −
+ 𝑂 𝐻 −
+ 𝐻 2
𝑂 2
→ 𝐶 𝑂 3
2−
+ 2𝐻 2
𝑂 EDFPFC = 2.83 V (17)
Subtracting (16) from (13) according to (6), the theoretical OCV reaches 2.83 V,
significantly higher than any other DLFC currently being studied. Although, in practice the
authors found that their DFPFC yields an OCV of 1.50 V, which is still higher than that achieved
with other formic acid, formate, methanol, ethanol or hydrogen fuel cells. The authors attributed
the voltage loss to fuel crossover and the decomposition of some hydrogen peroxide to oxygen,
lowering the cathode half-cell potential to some value between 1.23 V (E
0
ORR) and 1.78 V.
58

Figure 1-27. Polarization data for an alkaline formate-acidic peroxide fuel cell.
89

The latter explanation seems more plausible, since formate crossover through Nafion should be
even more disfavored than formic acid due to electrostatic repulsions between the formal
“charge” on the oxygen atoms and the sulfonate moieties of the Nafion backbone. In any case, a
peak power density of 214 mW/cm
2
was achieved at only 25°C, and further increased to 331
mW/cm
2
at 60°C, as seen in the polarization data from Figure 1-27. The chemistry of this
system may not support higher temperature operation, since it could exacerbate decomposition of
hydrogen peroxide to water, further lowering the potential of the cathode, and ultimately the
OCV.
The DFPFC also utilized a novel electrode design, at both the anode and cathode, that
removed the need for carbon-based GDLs and the associated painting or hot-pressing. Ni foam
served as both the catalyst support and reactant diffusion medium, on which PdAu was
spontaneously deposited from solution to form spheres exhibiting a Au-rich core and Pd-rich
59
surface. While the financial attractiveness of a Pt-free fuel cell is evident, it would be interesting
to see if a Pt cathode could further improve the OCV, since Pt decomposes H2O2 less readily
than a bimetallic PtPd surface.
93
There may also other viable cathode materials, but selection
will be more stringent since the peroxide reduction is taking place in the acidic rather in the than
alkaline media, which requires more robust materials, capable of tolerating such harsh
conditions. Despite the promising results, the newly reported DFPFC is still in its infancy,
leaving much work to be done before the viability of such technologies can be assessed.
1.6 Conclusion and Outlook
Since the beginning of the 21
st
Century, much work has been dedicated to the development and
study of DFAFCs. Initial investigations employed a Pt-based anode catlaysts, but it was found to
favor the formation of CO during FAO, which deactivates the cell, but can be improved via alloyed
materials such as PtRu, PtAu, or PtPd. Pd anode catalysts have been found to oxidize formic acid
“directly” to CO2, mitigiating the formation of CO, but still suffers from a controversial, albeit
manageable, deactivation problem. Much like with Pt, alloying Pd with Au yields particularly
active catalysts for the FAO. Although Pd-based materials are initially more active than those that
are Pt-based, the latter generally exhibit better long-term stability.
Completely passive DFAFCs have been demonstrated to have potential in small portable power
applications. Devices have been prepared with a single MEA, or in 2-15 MEA stacks, which
exhibit a lifetime of about a few hours. This value can be increased to as high as 50 h in a device
with actively supplied reactants. There are some examples of miniaturized DFAFCs that can
operate with a only a few μL of fuel solution, exhibiting lifetimes ranging from a few minutes to
several hours, depending on the cell design. One of these, a novel micro-tubular DFAFC, was
even able to power a small propeller. Although, there is only one example, from 2006, in which a
60
hybrid DFAFC-battery system was designed with a complete balance of plant and used in a real-
world situation to power a laptop for two hours.
In recent years, interest has reawakened in the use of formate as a fuel, in part due to the inherent
benefits of working in the alkaline media, such as improved reaction kinetics and the possibility
of using non-noble metals. Since 2012, there have been several reports on DFFCs that achieve
peak power densities comparable with some of the best performing DFAFCs. Thus far, only Pd-
based materials have been investigated for the anode reaction, but there are examples of DFFCs
that employ Ag/C, a proprietary Fe-Co catalyst at the cathode, and MnNiCoO4/N-MWNCTs at the
cathode. Two novel approaches towards designing DFFCs have been proposed. The first of these
removes the necessity for an AEM by using a cathode catalyst that is inactive for formate
oxidation, and may be scalable to the upper limit of electrode size. On the other hand, the second
approach calls for a Nafion membrane to transport sodium ions to the cathode compartment that
is fed with a mixture of hydrogen peroxide and sulfuric acid, rather than oxygen or air. This
formate-peroxide fuel cell exhibits an OCV of 2.83 V, but only realizes 1.50 V, which is still the
highest for any DLFC thus far. Nonetheless, the extremely high OCV leaves much room for
improvement, and continued study in this area may result in enhancements that improve the
efficiency of this technology. Over the past decade, it has been shown that both formic acid and
formate fuel cells have their own benefits and limitations. Commercial production may ultimately
yield devices that are tailored to a specific application, depending the energy requirements. As the
need for alternative power sources grows, the scientific community will be drawing upon a
plethora of available technologies to meet that demand, and it is very probable that formic acid
and formate fuel cells will be amongst the tools utilized for this task in the 21
st
century and beyond.

61
1.7 References
(1) Tori, C.; Baleztena, M.; Peralta, C.; Calzada, R.; Jorge, E.; Barsellini, D.; Garaventta, G.;
Visintin, a; Triaca, W. Int. J. Hydrogen Energy 2008, 33, 3588.
(2) Kamaruddin, M. Z. F.; Kamarudin, S. K.; Daud, W. R. W.; Masdar, M. S. Renew. Sustain.
Energy Rev. 2013, 24, 557.
(3) Iwasita, T. Electrochim. Acta 2002, 47, 3663.
(4) Li, L.; Xing, Y. Energies 2009, 2, 789.
(5) Ahmed, M.; Dincer, I. 2011, 1213.
(6) Uhm, S.; Lee, H. J.; Lee, J. Phys. Chem. Chem. Phys. 2009, 11, 9326.
(7) Yu, X.; Pickup, P. G. J. Power Sources 2008, 182, 124.
(8) Pan, Y.; Zhang, R.; Blair, S. L. Electrochem. Solid-State Lett. 2009, 12, B23.
(9) Haan, J. L.; Masel, R. I. Electrochim. Acta 2009, 54, 4073.
(10) Bartrom, A. M.; Haan, J. L. J. Power Sources 2012, 214, 68.
(11) Pourbaix, B. M. J. N.; Muylder, J. Van; Zoubov, N. De. Platinum Metals Rev. 1959, 3, 100.
(12) Jin, Y.; Ma, C.; Shi, M.; Chu, Y.; Xu, Y.; Huang, T. Int. J. Electrochem. Sci. 2012, 7, 3399.
(13) Zhang, Z.; Ge, J.; Ma, L.; Liao, J.; Lu, T.; Xing, W. Fuel Cells 2009, 9, 114.
(14) Lu, L.; Shen, L.; Shi, Y.; Chen, T.; Jiang, G.; Ge, C.; Tang, Y.; Chen, Y.; Lu, T.
Electrochim. Acta 2012, 85, 187.
(15) Jones, J.-P.; Prakash, G. K. S.; Olah, G. a. Isr. J. Chem. 2014, 54, 1451.
(16) Grimes, P. G.; Spengler, H. H. Formate Ion-Oxygen Fuel Cells; Milwaukee, Wisconsin,
1965.
62
(17) Weber, M.; Wang, J.-T.; Wasmus, S.; Savinell, R.F. J. Electrochem. Soc. 1996, 143, 1994.
(18) Haan, J. L.; Stafford, K. M.; Masel, R. I. J. Phys. Chem. C 2010, 114, 11665.
(19) Haan, J. L.; Stafford, K. M.; Morgan, R. D.; Masel, R. I. Electrochim. Acta 2010, 55, 2477.
(20) Rice, C.; Ha, S.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.; Barnard, T. Journal of Power
Sources 2002, 111, 83.
(21) Liu, Z.; Hong, L.; Tham, M. P.; Lim, T. H.; Jiang, H. J. Power Sources 2006, 161, 831.
(22) Kim, K. H.; Yu, J.; Lee, H. S.; Choi, J. H.; Noh, S. Y.; Yoon, S. K.; Lee, C.; Hwang, T.;
Rhee, Y. W. Korean J. Chem. Eng. 2007, 24, 518.
(23) Xu, J. B.; Zhao, T. S.; Liang, Z. X. J. Power Sources 2008, 185, 857.
(24) Lee, J. K.; Lee, J.; Han, J.; Lim, T.-H.; Sung, Y.-E.; Tak, Y. Electrochim. Acta 2008, 53,
3474.
(25) Yu, X.; Pickup, P. G. Electrochim. Acta 2010, 55, 7354.
(26) Yu, X.; Pickup, P. G. Electrochim. Acta 2011, 56, 4037.
(27) Yu, X.; Pickup, P. G. J. Power Sources 2011, 196, 7951.
(28) Sáez, a.; Expósito, E.; Solla-Gullón, J.; Montiel, V.; Aldaz, a. Electrochim. Acta 2012, 63,
105.
(29) Chetty, R.; Scott, K. J. New Mater. Electrochem. Syst. 2007, 10, 135.
(30) Ha, S.; Larsen, R.; Zhu, Y.; Masel, R. I. Fuel Cells 2004, 4, 337.
(31) Larsen, R.; Ha, S.; Zakzeski, J.; Masel, R. I. J. Power Sources 2006, 157, 78.
(32) Choi, J. H.; Noh, S. Y.; Han, S. D.; Yoon, S. K.; Lee, C. Korean J. Chem. Eng. 2008, 25,
1026.
63
(33) Meng, H.; Xie, F.; Chen, J.; Shen, P. K. J. Mater. Chem. 2011, 21, 11352.
(34) Kim, J. S.; Yu, J. K.; Lee, H. S.; Kim, J. Y.; Kim, Y. C.; Han, J. H. Korean J. Chem. Eng.
2005, 22, 661.
(35) Nakagawa, N.; Xiu, Y. J. Power Sources 2003, 118, 248.
(36) Cheng, T. T.; Gyenge, E. L. J. Appl. Electrochem. 2009, 39, 1925.
(37) Fang, B.; Kim, M.; Yu, J.-S. Appl. Catal. B Environ. 2008, 84, 100.
(38) Rao, C. V.; Cabrera, C. R.; Ishikawa, Y. J. Phys. Chem. C 2011, 115, 21963.
(39) Chang, J.; Li, S.; Feng, L.; Qin, X.; Shao, G. J. Power Sources 2014, 266, 481.
(40) Han, S. D.; Choi, J. H.; Noh, S. Y.; Park, K.; Yoon, S. K.; Rhee, Y. W. Korean J. Chem.
Eng. 2010, 26, 1040.
(41) Wang, R.; Liu, J.; Liu, P.; Bi, X.; Yan, X.; Wang, W.; Ge, X.; Chen, M.; Ding, Y. Chem.
Sci. 2014, 5, 403.
(42) Nandenha, J.; De Souza, R. F. B.; Assumpção, M. H. M. T.; Spinacé, E. V.; Neto, a. O.
Ionics (Kiel). 2013, 19, 1207.
(43) Wang, K.; Wang, B.; Chang, J.; Feng, L.; Xing, W. Electrochim. Acta 2014, 150, 329.
(44) Feng, L.; Yan, L.; Cui, Z.; Liu, C.; Xing, W. J. Power Sources 2011, 196, 2469.
(45) Matos, J.; Borodzinski, A.; Zychora, A. M.; Kedzierzawski, P.; Mierzwa, B.; Juchniewicz,
K.; Mazurkiewicz, M.; Hernández-Garrido, J. C. Appl. Catal. B Environ. 2015, 163, 167.
(46) Chang, J.; Feng, L.; Liu, C.; Xing, W.; Hu, X. Angew. Chem. Int. Ed. 2014, 53, 122.
(47) Zhou, W. P.; Lewera, A.; Larsen, R.; Masel, R. I.; Bagus, P. S.; Wieckowski, A. J. Phys.
Chem. B 2006, 110, 13393.
64
(48) Wang, X.; Xiao, P.; Thia, L.; Alam Sk, M.; Lim, R. J.; Ge, X.; Wang, J.-Y.; Lim, K. H.
Energy Environ. Sci. 2014, 30.
(49) Zhang, H. X.; Wang, S. H.; Jiang, K.; André, T.; Cai, W. Bin. J. Power Sources 2012, 199,
165.
(50) Baik, S. M.; Han, J.; Kim, J.; Kwon, Y. Int. J. Hydrogen Energy 2011, 36, 14719.
(51) Yu, X.; Pickup, P. G. J. Power Sources 2009, 187, 493.
(52) Arenz, M.; Stamenkovic, V.; Schmidt, T. J.; Wandelt, K.; Ross, P. N.; Markovic, N. M.
Phys. Chem. Chem. Phys. 2003, 5, 4242.
(53) Zhu, Y.; Khan, Z.; Masel, R. I. J. Power Sources 2005, 139, 15.
(54) Pan, Y.; Zhang, R.; Blair, S. L. Electrochem. Solid-State Lett. 2009, 12, B23.
(55) Zhou, Y.; Liu, J.; Ye, J.; Zou, Z.; Ye, J.; Gu, J.; Yu, T.; Yang, A. Electrochim. Acta 2010,
55, 5024.
(56) Jung, W. S.; Han, J.; Yoon, S. P.; Nam, S. W.; Lim, T. H.; Hong, S. A. J. Power Sources
2011, 196, 4573.
(57) Kwon, Y.; Baik, S. M.; Han, J.; Kim, J. Bull. Korean Chem. Soc. 2012, 33, 2539.
(58) Cheng, T. T.; Gyenge, E. L. J. Electrochem. Soc. 2008, 155, B819.
(59) Kwon, Y.; Baek, S.; Kwon, B.; Kim, J.; Han, J. Korean J. Chem. Eng. 2010, 27, 836.
(60) Kang, Y.; Ren, M.; Zou, Z.; Huang, Q.; Li, Z.; Akins, D. L.; Yang, H. Electrochim. Acta
2010, 55, 5274.
(61) Morgan, R. D.; Haan, J. L.; Masel, R. I. J. Power Sources 2010, 195, 6405.
(62) Kim, S.; Han, J.; Kwon, Y.; Lee, K. S.; Lim, T. H.; Nam, S. W.; Jang, J. H. Electrochim.
65
Acta 2011, 56, 7984.
(63) Mtalysis, O. F. J. Catal. 1969, 15, 430.
(64) Bauskar, A. S.; Rice, C. a. Electrochim. Acta 2012, 62, 36.
(65) Tsujiguchi, T.; Iwakami, T.; Nakagawa, N. ECS Transactions 2013, 58, 1885.
(66) Jeong, K. J.; Miesse, C. M.; Choi, J. H.; Lee, J.; Han, J.; Yoon, S. P.; Nam, S. W.; Lim, T.
H.; Lee, T. G. J. Power Sources 2007, 168, 119.
(67) Cai, W.; Liang, L.; Zhang, Y.; Xing, W.; Liu, C. Int. J. Hydrogen Energy 2013, 38, 212.
(68) Litster, S.; McLean, G. J. Power Sources 2004, 130, 61.
(69) Ha, S.; Adams, B.; Masel, R. I. J. Power Sources 2004, 128, 119.
(70) Hong, P.; Luo, F.; Liao, S.; Zeng, J. Int. J. Hydrogen Energy 2011, 36, 8518.
(71) Ha, S.; Dunbar, Z.; Masel, R. I. J. Power Sources 2006, 158, 129.
(72) Kim, D.; Cho, E. A.; Hong, S.-A.; Oh, I.-H.; Ha, H. Y. J. Power Sources 2004, 130, 172.
(73) Hong, P.; Liao, S.; Zeng, J.; Huang, X. J. Power Sources 2010, 195, 7332.
(74) Kim, H. S.; Morgan, R. D.; Gurau, B.; Masel, R. I. J. Power Sources 2009, 188, 118.
(75) Qiao, H.; Shiroishi, H.; Okada, T. Electrochim. Acta 2007, 53, 59.
(76) Erickson, E. M.; Mitrovski, S. M.; Gewirth, A. a; Nuzzo, R. G. Electrophoresis 2011, 32,
947.
(77) Goulet, M.; Lee, J. W.; Arjona, N.; Kjeang, E.; Arriaga, L. G. Lab Chip 2014, 14, 4596.
(78) Hong, P.; Zhong, Y.; Liao, S.; Zeng, J.; Lu, X.; Chen, W. J. Power Sources 2011, 196, 5913.
(79) Hong, P.; Liao, S.-J.; Zeng, J.-H.; Zhong, Y.-L.; Liang, Z.-X. J. Power Sources 2011, 196,
66
1107.
(80) Cai, W.; Yan, L.; Li, C.; Liang, L.; Xing, W.; Liu, C. Int. J. Hydrogen Energy 2012, 37,
3425.
(81) Miesse, C. M.; Jung, W. S.; Jeong, K.-J.; Lee, J. K.; Lee, J.; Han, J.; Yoon, S. P.; Nam, S.
W.; Lim, T.-H.; Hong, S.-A. J. Power Sources 2006, 162, 532.
(82) Nguyen, T. Q.; Bartrom, a. M.; Tran, K.; Haan, J. L. Fuel Cells 2013, n/a.
(83) Bartrom, A. M.; Ta, J.; Nguyen, T. Q.; Her, J.; Donovan, A.; Haan, J. L. J. Power Sources
2013, 229, 234.
(84) Christensen, P. a; Hamnett, a; Linares-Moya, D. Phys. Chem. Chem. Phys. 2011, 13, 5206.
(85) Jiang, J.; Wieckowski, A. Electrochem. commun. 2012, 18, 41.
(86) Zeng, L.; Tang, Z. K.; Zhao, T. S. Appl. Energy 2014, 115, 405.
(87) Yu, X.; Manthiram, A. Appl. Catal. B Environ. 2015, 165, 63.
(88) Yu, X.; Manthiram, A. Catal. Sci. Technol. 2015, 5, 2072.
(89) Li, Y.; He, Y.; Yang, W. J. Power Sources 2015, 278, 569.
(90) Taberner, P.; Heitbaum, J.; Vielstich, W. 1975, 439.
(91) Pivovar, B. Proceedings from the Alkaline Membrane Fuel Cell Workshop (NREL);
Golden,Colorado, 2012.
(92) Liang, Z. X.; Zhao, T. S.; Xu, J. B.; Zhu, L. D. Electrochim. Acta 2009, 54, 2203.
(93) Hasnat, M. a.; Rahman, M. M.; Borhanuddin, S. M.; Siddiqua, A.; Bahadur, N. M.; Karim,
M. R. Catal. Commun. 2010, 12, 286.
67

2
Re-Assessing the Efficacy of Palladium Anodes in
Formic Acid Fuel Cells
2.1 Introduction
With the invention of the direct methanol fuel cell (DMFC) in the early 90’s, proton
exchange membrane (PEM) fuel cells that utilized a liquid fuel were demonstrated to be viable.
1

The advantages of such a system, compared with hydrogen gas, are two-fold; transportation and
storage of pressurized gas is avoided, and changes to current energy infrastructure would be
minimal. Even with such advantages, DMFCs are plagued by their own shortcomings, some of
which are shared by other PEM technologies, namely, the high cost of the electrocatalysts (i.e.
PtRu and Pt). Moreover, the structure of methanol makes it prone to membrane crossover, since
its protonated form resembles the hydronium ion, which is easily carried across the Nafion
®

membrane, resulting in a mixed potential at the cathode.
2

Hydrogen fuel cells, on which the DMFC drew its initial inspiration, operate using Pt
electrocatlaysts at both the anode and cathode.
3
However, during the course of the methanol
oxidation on Pt, CO builds up on the surface, eventually blocking catalytically active sites and
preventing further oxidation of the fuel.
4
A solution was found by utilizing PtRu at the anode,
68

which can complete the oxidation of CO, due the presence of hydroxyl moieties (-OH) on Ru
atoms adjacent to Pt active sites.
5

Formic acid initially showed great promise as an alternative liquid fuel, since it had
inherent advantages over methanol. Due to electrostatic repulsions between the sulfonate groups
of the Nafion
®
backbone and the oxygen atoms of formic acid, crossover to the cathode is several
orders of magnitude lower. While the energy density of formic acid is less than half that of
methanol (2104 vs. 4333 Wh/L), the inherent resistance to crossover allows for significantly higher
fuel concentrations, which compensates for the theoretical difference.
6
Coupled with the direct
oxidation to CO2 on Pd, this should present an ideal to solution to the challenges of working with
methanol. But although FAO is generally considered to proceed directly to CO 2 on Pd, the reality
is that the an intermediate species builds up that deactivates the Pd surface.
7
In spite of this, Pd
remains a more attractive anode catalyst in formic acid fuel cells due to its lower cost, which is
about half that of Pt at the time of this writing. As such, much work has been done to improve the
efficacy of Pd anodes in DFAFCs.
8

To demonstrate improvements in cell performance, authors compare different catalyst
systems using “polarization curves” which measure cell voltage as a function of current density
(often called current-voltage transients), which can be used to determine the power density at a
given cell potential. These polarization curves, when plotted as power density as a function of
current density, are generally compared by the shape and size of the parabola formed by them. A
larger curve demonstrates higher current capacity and a greater peak power density, thereby
deeming some fuel cells to be “better” than others.
9
From this perspective, Pd eletrocatalysts are
far superior to Pt. However, as will be shown in this work, a polarization curve does not take into
account the effect of constant operation, which is of utmost importance for a practical
69

application. Hence, the alleged superiority of Pd anode catalysts in DFAFCs will be examined to
assess their true utility and future in the field.
2.2 Experimental Methods
All chemicals were purchased from commercial suppliers, and used as is, without further
purification. Pd black, Pt (40wt.%)/C, Pd (20wt.%)/C, Nafion
®
117 (sodium salt) ion exchange
membrane, and formic acid (97%) were purchased from Alfa-Aesar. Pt black, and PtRu black
were purchased from Premetek. Liquion
™
Solution (LQ-1105 1100E5, 5wt.% Nafion in
alcohols) was purchased from Ion Power. 18.2 MΩ Millipore water was used throughout the
experiments. Ultra high purity O2 (99.994%) was fed to the cathode for fuel cell tests.
2.2.1 Electrochemical Measurements
Cyclic voltammetry measurements were carried out in 0.5 M H2SO4 + 0.5 M HCOOH
electrolyte solution, which was purged with Ar for 30 min prior to data collection. A 1 mg/mL
dispersion of the appropriate catalyst was prepared with a 90:10 H2O:iPrOH solution containing
10 mg of Liquion
™
solution to improve powder adhesion to the glassy carbon surface. After
sonicating for 480 s to maximize dispersion, a 200 µ L automatic pipette was used to deliver 20
µ L of the solution to the glassy carbon electrode, which was then dried in a 70°C oven for 30
min. Measurements were conducted using a Pine Rotating Disk Electrode Apparatus, and a
Solartron SI 1287 Electrochemical Interface. A Pt wire and mercury sulfate electrode (MSE)
with 0.5 M H2SO4 electrolyte, were used as the working and counter electrodes, respectively.

70

2.2.2 Membrane Electrode Assembly
Nafion
®
117 was converted to its protonated form in four successive steps. Nafion squares, cut
to about 2” x 2”, were boiled in 3% H2O2 for 1 h, then washed in boiling water for 1 h.
Membranes were acidified by boiling in 0.56 M H2SO4 for 1 h, then washed again for 1 h.
The membrane electrode assembly (MEA), with an active area of 5 cm
2
was constructed with
non-teflonized and teflonized Toray carbon paper at the anode and cathode, respectively. 20 mg
of the appropriate catalyst was mixed with approximately 120 mg of water, and 20 mg 5 wt.%
Nafion
®
solution, in a small vial, then sonicated for 480 s. Pt (40wt%)/C was chosen as the
cathode for all fuel cells. The sonicated mixtures were then hand painted onto the appropriate
Toray carbon paper, until no more ink could be drawn from the vial. Painted electrodes were
dried in a 120°C oven for 30 min, then their catalyst loadings were determined by the mass
difference in the pre- and post-painted carbon papers. A PHI hot press was used to adhere the
painted electrodes to the Nafion-H membrane to form the completed MEA. Throughout the
pressing procedure, 500 lbf was applied. The components were heated to 140°C over the course
of 25 min, held at 140°C for 5 min, then cooled to RT within 25 min.
2.2.3 Fuel Cell Measurements
The freshly pressed MEA was then placed directly in the fuel cell housing, and hydrated over
night at 60°C. The next morning, the water circulation was stopped, and replaced with the fuel
and oxidant. 2 M formic acid was fed to the anode at 1 mL/min, while O 2 was fed to the cathode
at 400 sccm. A Fuel Cell Technologies humidification system (Model LFHS-C) with a built-in
mass flow controller was used to humidify the oxygen flow at 75°C. Both the fuel cell, and the
formic acid reservoir were held at 60°C throughout all experiments.
71

Fuel cell performance was evaluated with a Scribner Associates, Inc. fuel cell test system.
Polarization curves were obtained by scanning the current at a rate of 0.1 A/10 s. Potentiostatic
measurements were performed for about 1.5 h at three different cell voltages: near the activation
overpotential (when the cell initially began producing current), at the peak power density, and
near complete discharge.
2.2.4 Trace Analysis of Leeched Palladium
Inductively coupled plasma-optical emission spectrometry (ICP-OES) was performed on a
ThermoFisher Scientific iCAP
™
7400 ICP-OES analyzer. A solution of 100 ppm Pd
2+
in 0.5 M
formic acid was prepared with ammonium tetrachloropalladate, and then diluted to 2, 1, 0.5, and
0.1 ppm solutions, to obtain a standard curve with ICP-OES. Formic acid fuel samples obtained
during operation were diluted by a factor of 4 to 0.5 M so that the analyte would be in a similar
matrix as the standard solutions.
2.3 Results & Discussion
2.3.1 Formic Acid Oxidation Characteristics
Reports on novel DFAFC anode catalysts almost invariably begin with cyclic
voltammetry measurements. Here is where the first stark difference between formic acid
oxidation (FAO) on Pd and Pt can be seen, in the shape of the curves produced by their
voltammetric oxidation currents.

72

Figure 2-1. Forward (a) and reverse (b) cyclic voltammetry scans of FAO on Pt (black), Pd
(blue), and PtRu (red).

0
0.0005
0.001
0.0015
0.002
0.0025
-0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 0.7
I (A)
E (V) vs MSE
(a)
-0.0005
0.0005
0.0015
0.0025
0.0035
0.0045
0.0055
-0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 0.7
I (A)
E (V) vs MSE
(b)
73

As shown in Figure 2-1(a), Pt black exhibits two distinct oxidation peaks in the forward
scan. The small increase in current at -0.35 V (all potentials vs. MSE in cyclic voltammetry
data) during the forward scan is the indirect oxidation to CO, which adsorbs to Pt and prevents
further reaction, thereby resulting in the plateau region between -0.4 V and 0.08 V. Upon
reaching 0.1 V, CO is electrochemically stripped from the surface, which allows for FAO (to
CO2) to occur until electrochemically active sites are again blocked by the formation of surface
oxides at high potentials. Also of note in the forward scans is the fact that the onset potential on
Pd occurs some 200mV cathodic to Pt and PtRu, indicating more facil oxidation kinetics. On the
reverse scan, Figure 2-1(b), Pt oxides are removed, and the combined effect of preferential
oxidation to CO2 at high potentials, and maximized electrochemically active surface area, much
higher current is achieved and the plateau effect is not observed. When considering FAO on Pd,
an apparently simpler mechanism is at work, evidenced by single oxidation peaks, of comparable
intensities, in both the positive and negative direction, both occurring ca. 0.27 V. This would
seem to suggest that direct oxidation to CO 2 is occurring preferentially. PtRu black’s resistance
to CO poisoning is evident here because oxidation continues over the same region (starting -0.35
V) while Pt black experiences surface coverage, in the forward scan. Moreover, the reverse
oxidation peak occurs in the same position for both Pt and PtRu black, suggesting the same
reaction products, as shown in Figure 2-1(b). The cyclic voltammograms of carbon supported Pt
and Pd were omitted form the graph for clarity purposes, since they exhibited the same trends as
their pure metal counterparts.

74

2.3.2 Fuel Cell Polarization

Figure 2-2. Current-voltage (I-V) Transients of DFAFCs with common anode materials.
There are three characteristics of DFAFCs with Pd anodes that set them apart from the Pt
systems. Namely, higher open circuit voltage (OCV), lower activation overpotential, and greater
current capacity. Although the theoretical OCV for a DFAFC is 1.48 V, in practice it is
significantly lower, due to thermodynamic losses (i.e. inefficient oxidation and reduction at the
appropriate electrodes). As demonstrated in Table 2-1, Pd electrocatalysts generally exhibit
OCVs between 0.8-0.9 V, where Pt electrocatalysts ranged from 0.64 – 0.77 V. Furthermore, the
activation overpotential, or energy required to start drawing current from the fuel cell was much
greater for the Pt anodes. This is evident from the sharp drop in cell voltage before beginning to
steadily discharge to the short-circuit current density (E = 0 V), as shown in Figure 2-2. Ideally,
the activation overpotential is minimal, which allows for a greater percentage of the cell voltage
to be available for drawing current. It is interesting, that Pt/C exhibits comparable current
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 100 200 300 400 500 600 700
E (V)
i (mA/cm2)
Pt Blk
Pt-40%/C
PtRu Blk
Pd Blk
Pd-20%/C
75

capacity to the Pd catalysts. This may be due to a greater number of electrochemically active
sites on Pt/C, which allows for higher current density to be sustained. Moreover, the presence of
oxygen functionalities on the carbon support may act to mitigate CO poisoning, thereby
prolonging operation.

Figure 2-3. Polarization curves generated of DFAFCs with common anode materials.
From the same data set in Figure 2-2, the Polarization curves in Figure 2-3 are generated
by plotting power density as a function of current density. Power density, p, is obtained by
taking the cross product of cell voltage (V) and current density (i), according to
𝑝 = 𝑉 𝑥 𝑖 (1)
Considering the peak power densities (mW/cm
2
), the apparent advantage of Pd-based DFAFCs is
more clearly seen, with Pd black outperforming Pt black by a factor of 4 (127 mW/cm
2
vs. 31
mW/cm
2
). Both PtRu and Pt/C yielded power densities that were twice that of Pt black, which is
likely due to better resistance to CO. The presence of oxygen species on the carbon support may
0
20
40
60
80
100
120
140
160
0 100 200 300 400 500 600 700
p (mW/cm2)
i (mA/cm2)
Pt Blk
Pt-40%/C
PtRu Blk
Pd Blk
Pd-20%/C
76

contribute to this CO-resistance on Pt/C, akin to the way oxo-ruthenium species contribute to CO
tolerance on PtRu. The exceptionally poor performance of Pt black makes sense considering the
lack of such oxygen containing sites, resulting in rapid surface contamination, with no
mechanism for CO-removal.
Table 2-1
DFAFC Anode Performance Characteristics
Anode
Catalyst
OCV (V)
Activation Overpotential
(V)
peak power density
(mW/cm
2
)
Anode Loading
(mg/cm
2
)
Cathode Loading (mg/cm
2
)
Pt Black 0.69 0.32 31 2.04 2.53
Pt-40%/C 0.77 0.28 64 2 1.96
PtRu Black 0.64 0.21 61 2.14 2.66
Pd Black 0.86 0.08 127 1.74 2.2
Pd-20%/C 0.87 0.07 135 1.82 2.5

Table 2-1 summarizes the performance characteristics of the five anode catalysts, and
gives the loadings for each fuel cell. It should be noted that fuel cell data was not normalized
because the loadings, of total catalyst material, were comparable. Moreover, it is clear that Pd
black and Pt black are less ideal catalysts than their carbon supported analogues because equal or
greater performance is possible with significantly less precious metal content. Ultimately, the
data in Table 1 suggests that catalysts containing Pd are superior those with Pt because the cells
exhibit higher OCV, lower activation overpotential, and are capable of producing higher power
densities. As will be discussed below, potentiostatic experiments with the five anode catalysts
77

call into question this apparent superiority, demonstrating that the expected improvements with a
Pd anode are not observed when a DFAFC is required to operate for an extended period of time.
2.3.3 Potentiostatic Discharge
In order to assess the long-term performance characteristic of the DFAFC anodes, full
cells were discharged at three different potentials for 6000 s, to gauge their stability. Of most
interest, was the potential at which the peak power density occurred, as it should produce the
greatest power output. The cells were also discharged immediately after the activation
overpotential, when they were first capable of drawing current, and at a potential near short-
circuit. These potentials were different for each anode catalyst, and are listed in the legend for
each experiment in Figure 2-4.

Figure 2-4. DFAFCs discharged at potentials corresponding to their peak power densities.
The data presented in Figure 2-4 shows the power densities sustained by DFAFCs when
discharged at the “peak power-potential”. Pt/C, PtRu black, and Pt black remain steady
throughout the experiment, and at values that are in agreement with the relative positions of their
0
20
40
60
80
100
120
140
160
0 1000 2000 3000 4000 5000 6000
p (mW/cm2)
t (s)
Pt Blk (0.15 V)
Pt-40%/C (0.2 V)
PtRu Blk (0.25 V)
Pd Blk (0.4 V)
Pd-20%/C (0.45 V)
78

peak power densities, observed in Figure 2-3 (i.e. Pt/C > PtRu > Pt). On the other hand, Pd/C
experiences much more gradual deactivation, reaching a pseudo steady-state of about 60
mW/cm
2
, the highest of the five materials, but still less than half the expected power output.
Surprisingly, Pd black turned out to be the poorest performing of all five anode materials,
exhibiting a power density of ~130 mW/cm
2
that rapidly falls to a value < 20mW/cm
2
. This data
suggests that CO2 cannot be the sole product of FAO on Pd, else electrochemically active sites
would remain available, allowing for steady-state operation. The mechanism of FAO on Pd is
thought to occur via direct dehydrogenation of the C-H bond in formate to form a radical -
COOH. What follows is an auto-inhibitive process governed by the reaction below
10
:

𝐻𝐶𝑂𝑂𝐻 → ∗ 𝐶𝑂𝑂𝐻 + 𝐻 +
+ 𝑒 −
(2)
𝐻𝐶𝑂𝑂𝐻 + ∗ 𝐶𝑂𝑂𝐻 → ∶ 𝐶 (𝑂𝐻 )
2
+ 𝐶 𝑂 2
+ 𝐻 +
+ 𝑒 −
(3)
∶ 𝐶 (𝑂𝐻 )
2
→ 𝐶 𝑂 2
+ 2𝐻 +
+ 2𝑒 −
(4)

Reactions (2) and (3) occur rapidly, resulting in the formation of a carbene that binds to
catalytically active sites, and sluggishly oxidized to CO2 (4). Therefore, the steady drop in
current density with a Pd black anode can be attributed to the rapid buildup of this intermediate,
whose sluggish oxidation kinetics decreases activity towards FAO over time. On the contrary
Pd/C exhibited more resistance to this phenomenon, due to the carbon support exerting an
electronic effect as a co-catalyst.
11
The uneven current response may be due to the formation of
CO2 bubbles, since this catalyst is expected to convert formic acid to CO 2 most effectively.
79

Figure 2-5. (a) DFAFCs discharged after the activation overpotential. (b) Expanded view of the
oscillatory region of Pd black.

Discharging the DFAFCs minimally (i.e. just after surpassing the activation
overpotential) yielded somewhat surprising results, especially with Pd black as the anode.
0
10
20
30
40
50
60
70
80
0 1000 2000 3000 4000 5000 6000
p (mW/cm2)
t (s)
Pt Blk (0.35 V)
Pt-40%/C (0.45 V)
PtRu Blk (0.45 V)
Pd Blk (0.65 V)
Pd-20%/C (0.65 V)
0
2
4
6
8
10
12
14
16
18
20
1900 1950 2000 2050 2100 2150 2200
p (mW/cm2)
t (s)
(b)
(a)
80

Notice that again in Figure 2-5(a), the Pt catalysts remain relatively stable throughout the
experiment, while Pd/C undergoes a steady loss in power output, albeit much more gradually
than Pd black. Although the Pt catalysts are supposed to be susceptible to CO-poisoning
(especially Pt black), they exhibit the most steady current response throughout all poteniostatic
experiments. Between 0-2000 s, FAO reaction intermediates build up on Pd black until reaching
almost complete surface coverage. At this point, an oscillatory behavior is observed that as
periodic spikes in power density. Shortly after 1950 s in Figure 2-5(b), a repeating cycle is
observed about every 100 s in which power density rapidly increases to a local maximum, before
falling to the apparent baseline between 2-4 mW/cm
2
. As the DFAFC continues to draw current,
the carbene :C(OH)2

from reaction (3), and possibly other poisoning species, accumulate in
significantly greater quantities until the active sites are completely “filled”. Near complete
surface coverage, small quantities of deactivating intermediates may be able to be oxidized, to a
point where a sufficient number of active sites are suddenly available for FAO. At this point
oxidation intermediates rapidly re-cover the electrocatalytic surface, and the cycle repeats after a
small quantity of said poisoning species are removed.

81

Figure 2-6. DFAFCs operating near complete discharge.
Lastly, the DFAFCs were discharged near short-circuit. All five anode catalysts
coalesced near 20 mW/cm
2
, with the Pd catalysts begin slightly higher than the Pt catalysts, as
shown in Figure 2-6. Unlike the previous two experiments testing catalytic stability, where Pd/C
experienced a slower deactivation than Pd black, the opposite is true in this case. Considering
the voltage transients in Figure 2, at 0.1 V Pd black is drawing more current that Pd/C, which is
reflected in the amount of power produced. Notice that between 0.75 V and 0.15 V, Pd/C draws
slightly more current than Pd black, which explains the higher power density of Pd/C in the
previous potentiostatic experiments (Figres 2-4 and 2-5). Since the DFAFCs yield roughly the
same power outputs as a function of time, irrespective of the anode catalyst, Pd electrocatalysts
do not meet the expectation set by the preliminary fuel cell data.

0
20
40
60
80
100
120
0 1000 2000 3000 4000 5000 6000
p (mW/cm2)
t (s)
Pt Blk (0.05 V)
Pt-40%/C (0.05 V)
PtRu Blk (0.05 V)
Pd Blk (0.1 V)
Pd-20%/C (0.1 V)
82

2.3.4 Pd Anode Deactivation/Reactivation

Figure 2-7. (a) Polarization curves consecutively collected from a DFAFC with Pd black at the
anode. (b) Before and after potentionstatic discharge, then after washing the anode.
0
10
20
30
40
50
60
70
80
90
0 100 200 300 400 500
p (mW/cm2)
i (mA/cm2)
Initial
2
3
4
5
Final
(a)
0
20
40
60
80
100
120
140
0 100 200 300 400 500 600 700
p (mW/cm2)
i (mA/cm2)
Pre-0.4 V
washed anode
Post-0.4 V
(b)
83

There is more to consider regarding the role of Pd in a DFAFC. The coverage of
electrochemically active sites from the buildup of surface species on Pd is so rapid, that there is a
noticeable decrease in the size of the polarization curve after only a few consecutive current
scans. Figure 2-7(a) shows this trend, wherein the peak power density falls, from 85 mW/cm
2
in
the first polarization curve, to 52 mW/cm
2
after only five consecutive measurements. Between
the initial and finals scans, peak power density, and curve shape, appear to decrease
proportionally in linear manner. This suggests that successive current scans would result in
polarization curves that continue to shrink in the same manner, until reaching a pseudo-steady
state level of response, as best exemplified in Figure 2-6. As seen in Figure 2-7(b), polarization
curves collected before and after discharge always showed a significant decrease in their sizes,
irrespective of the operating potentials. It was reported in the literature that washing the anode
with 60°C DI water was sufficient to regenerate the original performance characteristics of the
fuel cells.
12
In our experience, at least 3 h was necessary, and fuel cells were generally washed
overnight to ensure maximum removal of surface adsorbates. In fact, the lower power density
observed in the initial scan of Figure 2-7(a) is smaller than that of 2-7(b) because the anode was
washed for an insufficient length of time before testing.
2.3.5 Palladium Leaching
So while the less expensive Pd catalysts are more susceptible to destabilization than the
Pt family, their performance can be recovered by a simple washing procedure. In spite of this,
there is one other property of Pd that may complicate its application in a DFAFC. According to
the Pourbaix diagram for Pd, Pd
2+
will form under conditions found in DFAFCs.
13
Therefore,
continued operation may result in the dissolution of the precious metal catalyst. Naturally, the
concern presented here for Pd applies even more to the transition metals it is often “alloyed” with
84

to improve efficiency.
14–16
To date, no studies have been performed that quantify the amount of
Pd leaching, if any, that occurs during DFAFC operation.
Table 2-2
Concentration of Leached Pd by ICP-OES
(t @ 0.5A)
Pd Black
Avg. Conc. (ppm)
Pd/C
Avg. Conc. (ppm)
1h -0.049 -0.047
3h -0.041 0
15h 0.027 -0.011

As such, a simple leaching experiment was designed with Pd black and Pd/C as anode
catalysts in a DFAFC, to probe for dissolved Pd ions. A small aliquot of 2 M formic acid was
taken from the fuel reservoir after the fuel cells were discharged at 0.5 A for 1, 3, and 15 h.
Then, ICP-OES was utilized to measure the presence of Pd
2+
, if any, in the fuel sample, as seen
in Table 2-2. Interestingly, almost every sample gave a negative reading, or a zero, indicating
that any Pd
2+
present was below the detectable limit of the instrument, the exception being the
DFAFC with a Pd black anode, after 15 h of discharge. This is not surprising because the Pd
black catalyst contained 5 times the amount of precious metal as the 20 wt.% carbon supported
Pd, which should be reflected in the concentration of leached metal. Nonetheless, ICP-OES
results suggest that Pd leaching does not occur to a degree that makes it a major contributing
factor to the poor performance observed with Pd anodes in DFAFCs.
2.3.6 Towards More Viable Formic Acid Fuel Cells
The question arises as to what kind of future lay in store for DFAFC technologies.
Recently, direct formate fuel cells have shown promise, boasting long-term stability surpassing
traditional fuel cells with formic acid.
17,18
However the anion exchange membrane (AEM), with
85

the quaternary nitrogen backbone necessary for the transport of hydroxide ions, is
underdeveloped compared to Nafion®. AEMs suffer from Hoffman Elimination from hydroxide
attack on the quaternary center which yields an alkene, trimethylamine, and water, thereby
decreasing the anion exchange capacity.
19
Moreover, limited availability inhibits further
research and the identification of deeper problems that may lead to their optimization, and
potential point of commercialization. A membrane-less proof-of-concept has been proposed and
demonstrated, that functions via a single electrolyte between the electrodes, taking advantage of
Pt’s inactivity towards the formate oxidation in base.
20,21
Thus, formate only reacts at the Pd
anode leaving hydroxide ions to migrate through solution to reach the cathode. Pt cathode in the
same electrolyte solution, claiming formate oxidation on Pt is negligible, allowing for selectivity
amongst the two surfaces.
Another use for formic acid in PEMFCs is as a source of reformed hydrogen. An
organometallic catalyst converts formic acid to hydrogen gas, which feeds directly to the fuel
cell. An onboard hydrogen reformer could be imagined in the future transportation sector,
needing only a supply of formic acid, and eliminating the need to carry pressurized hydrogen
tanks. The drawback is that such a system requires greater engineering complexity, and still
operates with a full Pt MEA. The use of formic acid in a reformation process is advantageous
because most of its stored energy can be utilized through conversion to H2. Catalysts exist that
promote direct oxidation to CO2, with sufficient selectivity to avoid detrimental concentrations
of CO in the fuel stream.
22
Lastly, these reformed hydrogen fuel cells have high potential to
become more viable than DFAFCS, since the actual power is generated by a hydrogen fuel cell,
the most developed PEMFC.
23
Continued research on DFAFCs may find ways to minimize Pt
86

content, or improve the stability of Pd in FAO through the use of ad-atoms or novel high surface
supports, which could exert an electronic effect on Pd.
While Pd electrocatalyst are comparatively cheaper than Pt, their efficacy as DFAFC
anodes is less than stellar, since they perform no better, and sometimes worse, when compared to
Pt anodes. Solutions to the challenges presented by Pd may be found in a new support materials
or alloy preparations that exert electronic effects, and can mitigate the formation of strongly
binding surface species. Ultimately, advances will have to be made to the AEM to make
possible more alkaline fuel cell research, allowing for investigation into transition metal-
containing electrocatalysts that would otherwise not survive DFAFC operation. In recent years,
our own research group has shown formic acid to be a useful hydrogen reformation feedstock, by
demonstrating that a catalytic system derived from IrCl 3 and 1,3-bis(2′-pyridyl-imino)-
isoindoline catalyst generates hydrogen gas from formic acid, which was successfully used to
power a PEM fuel cell.
24
Technologies such as this offer new options for the application of
formic acid as an energy storage material.
2.4 Conclusion
The assertion that Pd is a superior anode catalyst than Pt for DFAFCs is challenged when
cell performance is compared by ptoentiostatic current density profiles. The much higher power
densities initially observed with Pd black or Pd/C at the anode are not retained throughout
potentiostatic discharge. The Pt family of catalysts (Pt black, Pt/C, PtRu black) exhibited a more
stable current response across all experiments. However, when the fuel cells were discharged
near 0 V, all current density profiles coalesced around a similar value, irrespective of the anode
catalyst. The results suggest that although Pd catalysts benefit from reduced cost and apparently
selective formic acid oxidation to CO2, their practical cell performance is impeded due to the
87

buildup of a reaction intermediate, likely :C(OH)2, on the Pd surface. Although, Pd/C is less
susceptible to surface deactivation due an electronic effect from the carbon support, which resists
strongly binding adsorbates. Pd leaching does not appear to be a major limiting factor, under the
conditions studied in this work.
In the future, it is possible that formic acid will be more valuable as a feedstock for
hydrogen reformers, than as a direct liquid fuel in fuel cells. Independent of the method
employed, formic acid will continue to be an important candidate as an energy storage medium
for the power sector of tomorrow. Despite the challenges facing traditional DFAFCs, regular
advances are being made to improve their practicality, and emerging technologies such as the
formate fuel cell and reformed hydrogen fuel cell establish formic acid as a valuable energy
storage material for future power applications.
2.5 References
(1) García, B.; Weidner, J. J. W.; Garcia, B. L. Mod. Asp. Electrochem. No. 40 2007, 229.
(2) Ahmed, M.; Dincer, I. International Journal of Energy Research. 2011, 1213.
(3) Tori, C.; Baleztena, M.; Peralta, C.; Calzada, R.; Jorge, E.; Barsellini, D.; Garaventta, G.;
Visintin, a; Triaca, W. Int. J. Hydrogen Energy 2008, 33, 3588.
(4) Iwasita, T. Electrochim. Acta 2002, 47, 3663.
(5) Li, L.; Xing, Y. Energies 2009, 2, 789.
(6) Rice, C.; Ha, S.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.; Barnard, T. Journal of
Power Sources. 2002, 111, 83.
(7) Pan, Y.; Zhang, R.; Blair, S. L. Electrochem. Solid-State Lett. 2009, 12, B23.
(8) Yu, X.; Pickup, P. G. J. Power Sources 2008, 182, 124.
(9) Chang, J.; Feng, L.; Liu, C.; Xing, W.; Hu, X. Angew. Chem. int. Ed. 2014, 53, 122-126.
88

(10) Zhang, Z.; Ge, J.; Ma, L.; Liao, J.; Lu, T.; Xing, W. Fuel Cells 2009, 9, 114.
(11) Chang, J.; Li, S.; Feng, L.; Qin, X.; Shao, G. J. Power Sources 2014, 266, 481.
(12) Zhou, Y.; Liu, J.; Ye, J.; Zou, Z.; Ye, J.; Gu, J.; Yu, T.; Yang, A. Electrochim. Acta 2010,
55, 5024.
(13) Pourbaix, B. M. J. N.; Muylder, J. Van; Zoubov, N. De. Platin. Met. Rev. 1959, 100.
(14) Haan, J. L.; Stafford, K. M.; Masel, R. I. J. Phys. Chem. C. 2010, 114, 11665.
(15) Jin, Y.; Ma, C.; Shi, M.; Chu, Y.; Xu, Y.; Huang, T. Int. J. Electrochem. Sci. 2012, 7,
3399.
(16) Lu, L.; Shen, L.; Shi, Y.; Chen, T.; Jiang, G.; Ge, C.; Tang, Y.; Chen, Y.; Lu, T.
Electrochim. Acta 2012, 85, 187.
(17) Bartrom, A. M.; Haan, J. L. J. Power Sources 2012, 214, 68.
(18) Zeng, L.; Tang, Z. K.; Zhao, T. S. Appl. Energy 2014, 115, 405.
(19) Christensen, P. a; Hamnett, a; Linares-Moya, D. Phys. Chem. Chem. Phys. 2011, 13,
5206.
(20) Yu, X.; Manthiram, A. Appl. Catal. B Environ. 2015, 165, 63.
(21) Yu, X.; Manthiram, A. Catal. Sci. Technol. 2015, 5, 2072.
(22) Singh, A. K.; Singh, S.; Kumar, A. Catal. Sci. Technol. 2015, 6, 12.
(23) Shin, J.-H.; Jung, N.; Yoo, S. J.; Cho, Y.-H.; Sung, Y.-E.; Park, T. H. Chem. Commun.
(Camb). 2011, 47, 3972.
(24) Czaun, M.; Kothandaraman, J.; Goeppert, A.; Yang, B.; Greenberg, S.; May, R. B.; Olah,
G. A.; Prakash, G. K. S. ACS Catal. 2016, 6, 7475.

89

3
Improved Hydrogen Fuel Cell Performance
Using CFx-Pt Blends
3.1 Introduction
Hydrogen fuel cells are the most-studied proton exchange membrane fuel cell (PEMFC)
to date.
1
Since there are no byproducts generated by the H2 oxidation to form poisoning species
at the anode or cathode, hydrogen fuel cells still exhibit far superior performance characteristics
to both direct methanol fuel cells (DMFC) and direct formic acid fuel cells (DFAFC), producing
considerably more power than competing technologies. Although problems related to the fuel
source, such as crossover and catalyst deactivation, are not present with H2, some challenges are
still present. Namely, water management to mitigate cathode compartmental flooding, sluggish
oxygen reduction reaction (ORR) kinetics in acidic media, and stability of the carbon support in
the harsh oxidation conditions at the cathode.
Water is necessary for hydrogen oxidation, as well as for the transport of protons across
the membrane. As mentioned previously, water management at the cathode is an important
consideration, because it is being introduced into the cathode compartment by three separate
mechanisms; crossover in the membrane via electro-osmotic drag, supplied in the humidified
90

oxygen stream, and generated as a product of ORR. Coupled together, their cumulative effects
may lead to compartmental flooding, which saturates the catalytically active sites and hinders
ORR. This is managed by utilizing teflonzied carbon paper, to facilitate the removal of water
through the cathode exhaust.
One of the inherent challenges regarding PEMFC cathodes is that ORR is a slow reaction
at any pH.
2
For effective fuel cell performance, relatively high loadings of Pt (~0.4 mg/cm
2
) are
required to compensate for the sluggish ORR kinetics, increasing the cost of the device.
Supporting Pt particles on high surface area Vulcan carbon (XC72) is an effective way to
increase catalyst utilization, and optimize the price-to-performance ratio, but extremely corrosive
conditions experienced during ORR damage the support, thereby decreasing electrochemically
active surface area (ECSA).
3

Fluorinated electrode materials have already been used successfully in lithium batteries.
4

In fact, Li-CFx primary batteries have the highest theoretical energy density (2180 Wh kg
-1
) of
all primary Li-batteries.
5
However, unlike the commercially successful, secondary Li-batteries,
Li-CFx batteries are not rechargeable, which severely limits their utility.
6
CFx participates in the
discharge reaction in the following manner:
7

𝐶𝐹𝑥 + 𝐿𝑖 → 𝐶 + 𝐿𝑖𝐹
However, the insulative nature of CFx limits applications to low-medium level current devices.
8

Future work may make further application of Li-CFx batteries possible, but there is much
potential for CFx as a cathode additive in PEMFCs. Since the fluoride present in CFx is not
directly involved in the electrochemistry of PEMFCs, and the major limiting factor of CFx in Li-
batteries is its non-rechargability, CFx is a useful component in PEMFC cathodes considering
91

that its physical presence may be enough to accelerate the rate of ORR by facilitating oxygen
transport.
Although to a much lesser extent than Li-batteries, some research incorporating
fluorinated materials at a PEMFC cathode has been reported. Pt-C aerogels can be fluorinated
by the thermal decomposition of XeF2, which was shown to improve mass transport of oxygen
relative to commercially available Pt/C.
9
In addition, the fluorinated catalysts exhibited much
greater stability, measured as ECSA (CO-stripping) after 5000 cycles between 0.6 V and 1.05 V
vs RHE. However, experiments were limited to single-cell investigations, and a full fuel cell
was not tested. Fluorinated ethylene, propylene, and single-walled carbon nanotubes were found
to exert a synergistic effect when added to the catalyst layer of the cathode gas diffusion layer
(GDL) at 20wt% and 50wt%, respectively.
10
Their synergy increased peak power density in the
fuel cell polarization curve to 134 mW/cm
2
, from 80 mW/cm
2
when only Vulcan carbon was the
catalyst support. Other authors added only FEP to the cathode catalyst layer between 5-30 wt%,
but no improvement in fuel cell performance was observed.
11
A microfluidic, passive DMFC
with fluorinated XC72 at the cathode showed increased peak power densities and discharge
stability relative to the fuel cell with only XC72 as the cathode catalyst support.
12
It is plausible
that fluorinated carbon black (CFx) would promote migration of oxygen molecules to Pt active
sites, due to its fluorophilic nature. The goal of this work was to improve the efficacy of ORR
by incorporating CFx into the catalyst matrix, thereby accelerating oxygen delivery to the
catalytically active sites, considering that oxygen exhibits excellent solubility in liquid
perfluorocarbons.
13
A cathode catalyst consisting of Pt nanoparticles on CFx demonstrated
improved performance in hydrogen fuel cells, and especially DMFCs, at low O2 and air flow
rates.
14

92

3.2 Experimental Methods
All chemicals were purchased from commercial suppliers, and used as is, without further
purification. Pt (40wt.%)/C was purchased from Alfa-Aesar. Pt black was purchased from
Premetek. Nafion
®
211 (sodium salt) ion exchange membrane and Liquion
™
Solution (LQ-1105
1100E5, 5wt.% Nafion in alcohols) were purchased from Ion Power. CFx was provided by
Advance Research Chemicals, Inc. 18.2 MΩ Millipore water was used throughout the
experiments. During fuel cell tests, ultra-high purity H2 (99.999%) and O2 (99.994%) were fed to
the anode and cathode, respectively.
3.2.1 Membrane Electrode Assembly
Nafion
®
211 was converted to its protonated form in four successive steps. Nafion squares, cut
to about 2” x 2”, were boiled in 3% H2O2 for 1 h, then washed in boiling water for 1 h.
Membranes were acidified by boiling in 0.56 M H2SO4 for 1 h, then washed again for 1 h.
The membrane electrode assembly (MEA), utilizing a 5 cm
2
active area, was constructed with
non-teflonized and teflonized Toray carbon paper as the gas diffusion layer (GDL) at the anode
and cathode, respectively. 20 mg of the appropriate catalyst was mixed with approximately 120
mg of water, and 20 mg 5 wt.% Nafion
®
solution, in a small vial, then sonicated for 480 s. Pt
(40wt%)/C was used as the anode for all fuel cells. The sonicated mixtures were then hand
painted onto the appropriate Toray carbon paper, until no more ink could be drawn from the vial.
Painted electrodes were dried in a 120°C oven for 30 min, then their catalyst loadings were
determined by the mass difference of the pre- and post-painted carbon papers. A PHI hot press
was used to adhere the painted electrodes to the Nafion-H membrane to form the completed
MEA. Throughout the pressing procedure, 500 lbf was applied. The MEA was heated to 140°C
over the course of 25 min, held at 140°C for 5 min, then cooled to RT within 25 min.
93

3.2.2 Fuel Cell Measurements
The freshly pressed MEA was then placed directly in the fuel cell housing, and tested without
pre-hydration. Both hydrogen and oxygen gas were supplied to that anode and cathode
compartments at rates of 200 sccm, respectively, unless otherwise specified. A Fuel Cell
Technologies humidification system (Model LFHS-C) with built-in mass flow controller was
used to humidify both gas streams at 75°C. Fuel Cell temperature was controlled by adhesive
heating pads powered by the humidification system. Fuel cell performance was evaluated with a
Scribner Associates, Inc. fuel cell test system. Polarization curves were obtained by scanning the
current at a rate of 0.1 A/10 s.
3.3 Results & Discussion
3.3.1 Effect of Catalyst Application Procedure
Multiple reports with other Liquid PEMFCs report using a spray-paint technique to apply
catalyst layers to the membrane or electrodes.
1,15–17
In fact, the hot-press technique used in this
work has been shown to damage proton channels, displaying poorer fuel cell performance when
compared to other methods.
18
Herein, Pt (40wt.%)/C was both spray- and brush-painted onto
non-teflonized, and wet proofed 6 milliinch Toray paper to prepare the anode and cathode,
respectively. Due to the ultrathin nature of the Nafion 211 membrane, overnight pre-hydration
was unnecessary to commence drawing current, as is required with methanol and formic acid
fuel cells that incorporate thicker membranes. Supplying humidified (75°C) hydrogen and
oxygen gas streams for a few minutes prior to testing was sufficient to establish hydration of the
proton channels and yield satisfactory polarization curves.
94

Figure 3-1. Fuel Cell data at three different temperatures comparing the brush-painting method
(solid lines) to the spray painting method (dotted lines).Gases humidified at 75°C. H2=O2 =100
sccm.

As mentioned previously, many research groups report using a spray-painting method to
prepare their MEAs, claiming more uniform catalyst coverage and more efficient ink utilization,
effectively producing better results at lower cost, by reducing the waste stream. It is worth
noting that typically only 40wt.% of the catalyst from a prepared ink solution is deposited on the
GDL by the brush-painting method applied here, which represents a significant expense added to
the start-up cost of the fuel cell. It may be possible to recycle the remaining catalyst caught in
the bristles of the brush, and adhered to the walls of the ink vial, but is outside the scope of this
work, in an academic research setting. However, utilizing nearly 100% percent of the catalyst
material, and having it in the form of an easily sprayable solution would be ideal for commercial
application.
0
100
200
300
400
500
600
700
800
0 500 1000 1500 2000 2500 3000
p (mA/cm2)
i (mA/cm2)
30 C
60 C
80 C
95

When hand-painting a GDL to prepare an electrode, brush strokes are made from left to
right until the surface is covered, and the process is continued after drying and rotating the
square GDL by 90°. Interestingly, even with its more uniform catalyst coverage, the spray-
painted electrode did not perform as well as the brush-painted one, even at temperatures as high
as 80°C, as seen in Figure 3-1. Looking at the purple solid line and the green dotted line, brush
painted MEAs achieve a greater peak power density at 30°C than spray-painted MEAs at 60°C.
The superiority of the brush painted method, under our conditions, is more starkly evident at
higher temperature, being able to outperform an 80°C brush-painted MEA at 60°C.
Approaching higher temperatures, polarizations curves for both MEAs exhibit
irregularities at 80°C, which is evident throughout the measurement for the spray-painted MEA,
whereas the brush-painted MEA begins to exhibit extreme fluctuations in power density around
1500 mA/cm
2
. These anomalous features are likely due to inconsistences in the bulk catalyst
morphology, wherein oxygen molecules can accumulate and become trapped, which results in a
spike in current density when they are liberated and pushed towards electroactive sites, by the
continuous supply of oxygen. For the spray-painted MEA, current could not be drawn as high as
with the brush-painted MEA (~1600 mA/cm
2
vs. 2700 mA/cm
2
), suggesting that sufficient
oxygen transport could not be sustained to achieve higher current densities, due to the non-ideal
microstructure of the catalyst layer(s).
Also of note between these two MEAs is the open circuit voltage (OCV) attained with
each. The average OCV of the spray-painted and brush-painted MEAs were 0.61 V and 0.86 V,
respectively. Of particular interest is why the spray-painted MEA performed poorer than
expected. A common characteristic spray-painted electrodes reported in the literature is a lack of
adequate experimental description, which would make the procedure easily reproducible. On the
96

contrary, vague explanations of how to perform spray-painting leaves much to the imagination,
and lacking direct knowledge of the technique, replicating reported work requires more trial and
error than is generally acceptable. As the data reported here is from our initial investigations into
catalyst spray-painting, improved performance can be expected in future work, perhaps even
beyond that of brush-painted electrodes. It is possible that our ink preparation method does not
sufficiently disperse catalyst particles, such that they agglomerate upon spray-deposition onto the
GDL. Returning to the OCV, it was likely lower in the spray-painted MEA due to poor particle
dispersion, from agglomeration in the ink, which resulted in a loss of ECSA, which limited the
abilities of both the anode and cathode reactions to contribute to the cell voltage.
3.3.2 Simple Mixing of Platinum with XC72 or CFx
Initial testing of the hypothesis, that using CFx as the support could improve oxygen
transport at the cathode, was conducted by mixing 20 mg of Pt black and 80mg of CFx or XC72
with a micro spatula in a 2 dram vial. Said mixtures were prepared into catalyst inks for the
cathodes, then tested in the Fuel Cell housing at RT.

97

Figure 3-2. Preliminary evidence suggesting improvements with CFx at the cathode. Cell T =
RT. Gases humidified at 75°C. H2=O2 =100 sccm.

By substituting XC72 for CFx in the 1:4 Pt:C mixture, peak power density increased
from 100 mW/cm
2
to 120 mW/cm
2
, and short circuit current density increased to 600 mA/cm
2

from 400 mA/cm
2
, as shown in figure 3-2. It is noteworthy that the activation overpotential
clearly decreases in the presence of fluorine moieties at the cathode, costing 0.1 V less with CFx.
This may also be indicative of an electronic effect exerted by the fluorinated support on Pt
particles, rendering ORR more facile. At high current density, the mass-transfer limited region
of the curve, the presence of fluorine in the carbon support accelerates the delivery of oxygen to
Pt active sites, allowing for further cell discharge.
3.3.3 Ball-Milled Mass Ratio Formulations
In an attempt to improve electro-catalytic activity and fuel cell performance, Pt and
carbon mixtures were ball milled for 30 min to obtain more uniformly dispersed catalyst
formulations. Five catalyst mixtures were prepared as 5-part 1:4 mass ratios of Pt:C (C = XC72
98

or CFx), with increasing amounts of XC72 being substituted with CFx. For later experiments,
Pt(40wt%)/C was ball-milled directly with CFx to obtain formulations consisting of 1:1, 1:2, and
2:1 Ptmg:CFxmg.

Figure 3-3. Polarization curves with increasing CFx content at the cathode. Cell T = RT. Gases
humidified at 75°C.

As seen in figure 3-3, the fuel cell with with no CFx in the cathode catalyst mixture, Pt:XC72
(green), reaches a peak mass specific power of only 100 mA/mg, and short circuits before
reaching a specific current of 400 mA/mg. The incorporation of only 20 wt% CFx (dark blue) in
the Pt:C catalyst yielded a fuel cell that exhibited nearly double the specific power and specific
current limits of the fuel cell without a fluorinated cathode, reaching almost 200 mW/mg and 700
mA/mg, respectively. Interestingly, 60 wt% CFx (purple) performs only slightly better than the
20 wt%, whereas the presence 40 wt% CFx (red), exhibits poorer performance, reaching only
140mW/mg and <600mA/mg. This result is surprising, since partially fluorinated carbon blacks
retain electrical conductivity on the same order of magnitude as XC72.
14
Not considering this
99

data point, specific power and specific current capacity appears to increase with greater amounts
of CFx in the mixture. It is possible that the red curve, corresponding to 60 wt% CFx, is an
anomalous result, since the fuel cell in which the cathode components were only Pt:CFx (light
blue), outperformed all previously tested materials. As such, the sufficient electrical
conductivity of CFx and improved oxygen transport characteristics aided in achieving a peak
specific power over 250 mW/mg, not including the high-current density shoulder, and a short
circuit specific current greater than 1000 mA/mg.
Irrespective of the amount of fluorinated material at the cathode, a shoulder appears in the mass-
transport limited region of the polarization curves, which remains present in all the CFx catalyst
mixtures. This type of response in polarization curves, in which the current beings to drop then
suddenly spikes, forming a shoulder, can be understood by considering the mechanism of data
collection by the fuel cell test system. Polarizations curves are collected by scanning current
from 0 A (Ecell = OCV) to short circuit (Ecell = 0 V). At higher current density, the shape of the
curve takes a downward turn, as the overpotential to sustain increasingly greater current density
causes Ecell to drop, which in turn is reflected in the power density, since it is measured by the
fuel cell test system as:
𝑃 (𝑊 ) = 𝐸 (𝑉 ) × 𝐼 (𝐴 )
At a certain point in the mass transport-limited regions in the polarization curves of CFx
containing fuel cells a sudden influx of oxygen to the Pt-active sites, caused by the presence of
fluorine atoms in the catalyst matrix, results in a surge in power density. Since Pt:CFx naturally
has the most fluorinated sites, thus the greatest capacity to accelerate migration of O 2 to Pt active
sites, this mixture shows the largest shoulder both in height and width, allowing the fuel cell to
continue drawing current further into the mass transport-limited region, despite containing the
100

greatest amount of electrically resistive material. However, this improvement only lasts a short
time before oxygen transport is insufficient to sustain even higher current densities.

Figure 3-4. Voltages transients with increasing CFx content at the cathode. Cell T = RT. Gases
humidified at 75°C.

The effects of CFx at the cathode on fuel cell performance are seen more clearly when
considering the voltage transients for the aforementioned fuel cells, as shown in Figure 3-4.
Taking Pt:XC72 (green) as the performance baseline, there are three characteristic regions
present in fuel cell voltage transients. Initially there is the activation overpotential, between ~0.9
V and 0.7 V, which is the energetic cost required to “start” the fuel cell, before current can be
drawn. Next is the linear region governed by Ohm’s law, which continues until reaching the
third region, which is mass-transport limited. The first interesting feature of these devices, is that
the presence of CFx in the catalyst mixture lowers the activation overpotential for the fuel cells.
However, catalysts with 3:1 (60:20) and 1:3 (20:60) CFx:XC72 exhibit similar activation
overpotential, which is likely due to the competing effects of improved oxygen delivery and
101

increased electrical resistance, as discussed previously. Although resistance from the presence of
fluorine is a limiting factor, and the overpotential for the Pt:CFx catalyst was expected to be the
largest, this mixture exhibited the smallest activation overpotential, because of the ability of CFx
to accelerate the migration of oxygen through the catalyst matrix. The size of the ohmic region
of each voltage transient is inversely proportional to the activation overpotential, because a lower
activation energy equates to greater current capacity before mass-transport becomes a limiting
factor. Of particular interest is the effect that the incorporation of CFx had on the mass-transport
limited regions of these fuel cell voltage transients. As mentioned in relation to the polarization-
shoulders observed in figure 3-3, this phenomenon presents itself as horizontal extensions of the
voltage transients in figure 3-4. Unlike a conventional voltage transient (i.e. Pt:XC72), in which
mass-transport limitation results in a rapid voltage drop, the presence of CFx temporarily
removes this constraint, showing that CFx accelerates the migration O 2, as is evident from the
horizontal artefacts present at high current density in the voltage transients.
3.3.4 Reduced O
2
-flow Rate Requirements with CFx
Fully substituting XC72 for CFx in the catalyst mixture showed the greatest improvement
in polarization performance, so Pt:XC72 and Pt:CFx were tested at various O2-flow rates to
determine if fluorinated cathode materials allow for lower oxygen requirements. The results of
these experiments are significant, because future commercial fuel devices will likely need to be
“air-breathing” (i.e. no active oxygen supply).
102

Figure 3-5. Hydrogen fuel cells at various cathode flow rates. Cell T = 30°C. Gases humidified
at 75°C.

Figure 3-5 demonstrates that with the Pt-XC72 (blue curves) at the cathode, a flow rate of
100 sccm yielded a peak power density of ~45 mW/cm
2
. Increasing the flow rate to 250 sccm
.further decreased fuel cell performance ~35 mW/cm
2
. Doubling the flow rate to 500 sccm again
resulted in a drop in peak power density. This is likely due to compartmental flooding at the
cathode, which is caused by the build-up of excess water carried by the humidified oxygen
stream, leading the blockage of the Pt-active sites. Although fuel cell performance decreases
with increasing O2-flow rate when the cathode consists of only Pt and XC72, the presence of
CFx (red curves) causes oxygen saturation at catalytically active sites at only 100 sccm. Further
increasing O2-flow rate yields negligible improvements, except for a small shoulder wherein the
huge excess of O2 molecules is temporarily sufficient to sustain slightly higher current densities,
before reaching short circuit.
103

3.3.5 Platinum-Carbon Mixtures Compared with State-of-the-Art Pt/C

Figure 3-6. A simple balled-milled mixture of 20:80 Pt:XC72 in contrast with commercially
available Pt/C catalyst, represented in units per mgPt.

An important consideration when designing a fuel cell catalyst is the utilization of
precious metals, which are responsible for electrocatalytic activity. A larger quantity of precious
metal does not necessarily correlated with superior performance characteristics. As seen in
Figure 3-6, even when normalizing for Pt content, commercially available Pt/C achieves
significantly higher power per unit mass than the in-house catalyst, likely due to differences in
the preparation methods of the two materials. Only Pt:XC72 and commercial Pt/C were shown,
in order to highlight the stark difference in performance characteristics between the two
materials. The in-house catalyst was only ball-milled with a carbonaceous material to create
regions of fluorocarbons that enhanced oxygen transport at the cathode, relative to a mixture
containing only XC72. On the contrary, the commercial catalyst was likely synthesized via an
104

adsorption-reduction route in which Pt salts were reduced to metal particles that nucleate and
grow on the carbon surface, leading to substantially higher electrochemically active surface area.
Although utilization is much higher in Pt-40%/C than the ball-milled mixture, combining
commercially available catalysts with CFx in the same manner may yield further improvements.
The reasoning behind such work is that if simple mixing of ready-made precious metal catalysts
with fluorinated materials can improve cell performance while decreasing Pt-requirements, then
this method could be immediately available for commercial applications. Fuel Cells were
operated at 30°C and low flow rates (50 sccm), to simulate the ideal conditions for a commercial
device, which would be “air-breathing” (i.e. no supplied gas stream) and operate at room
temperature. Although, pure oxygen was used instead of air in these experiments to obtain a
clearer understanding of how CFx affects fuel cell performance.
Instead of ball milling CFx, active carbon, and Pt black, as in the initial experiments, it
was substituted for Pt (40%/C). This was done since the utilizations of our Pt:XC72:CFx
catalysts were shown to be far behind that of commercially available Pt/C, as seen in figure 3-6.
By mixing an already-optimized catalyst with fluorinated carbon, improved delivery of oxygen
to the high-utilization Pt active sites may occur, which would result in a larger polarization
response.
105

Figure 3-7. Hydrogen Fuel cell polarization curves with various fluorinated cathode catalysts.
Cell T=30°C,H2=50 sccm, O2=50 sccm.

Mixing CFx with commercially available Pt (40wt%)/C yielded interesting results. As
observed in the initial ball-milling experiment, increasing amounts of CFx in the catalysts
mixtures lead to increased utilization of Pt, by mass. As shown in figure 3-7, commercial Pt/C
(green) achieves the highest current and power densities, at 1750 mA/cm
2
and 400 mw/cm
2
,
respectively. The two mixtures with the least amount of CFx performed rather poorly, reaching
power densities of barely 100 mW/cm
2
, and exhibiting short-circuit current densities
significantly less than 500 mA/cm
2
. Although, in keeping with the trend observed previously,
the mixture with the most amount of CFx, exhibited the largest polarization curve relative to
formulations with less CFx. However, in this case, the increase in the sizes of the power density
curves is not proportional with CFx content. Nonetheless it is promising that similar
0
50
100
150
200
250
300
350
400
450
0
0.2
0.4
0.6
0.8
1
1.2
0 500 1000 1500 2000
Power Density (mW/cm
2
)
Potential (V)
Current Density (mA/cm
2
)
Pt-CFx 1-1
Pt-CFx 2-1
Pt-CFx 1-2
Pt40%/C
106

performance characteristics to Pt(40wt%)/C can be achieved using only 22.2wt.% Pt, when
mixed in a 1:2 Pt:CFx weight ratio within the total catalyst mixture.

Figure 3-8. Mass-normalized, per mgPt, polarization curves from Figure 3-7.

Fascinatingly, the CFx mixture with the greatest weight percent Pt (Pt-CFx 2-1,
33.3wt%), exhibited the least Pt utilization, by mass, as shown in Figure 3-8. Even though total
Pt content was hardly 7wt% less than commercial Pt/C, its specific current limit was five times
smaller. Moreover, in the mixture with twice as much CFx to platinum (Pt-CFx 1-2, 22.2wt%)
peak specific power is 1000 mW/mg greater than Pt-CFx 2-1, and the specific current limit is an
order of magnitude greater, despite containing only 10wt% less Pt. It is worth noting from the
previous figure that although not all CFx mixtures performed equally as well, all three exhibited
0
200
400
600
800
1000
1200
1400
0 1000 2000 3000 4000 5000 6000 7000
Specific Power (mW/mg)
Specific Current (mA/mg)
Pt-CFx 1-1
Pt-CFx 2-1
Pt-CFx 1-2
Pt40%/C
107

OCVs larger than commercially available Pt/C, 0.9-1.0 V vs 0.8 V, respectively, demonstrating
the feasibility of CFx-containing PEMFCs.
3.4 Conclusion and Outlook
The presence of CFx at the cathode does result in improved fuel cell performance
characteristics, due to the more facile delivery of oxygen to Pt active sites, relative to a mixture
consisting of only Pt black and Vulcan XC72. Power density and current density reached their
maxima when XC72 was fully substituted with CFx in both crude, and precision ball-milled
mixtures. Regardless, of O2-flow rate, fuel cells with Pt:CFx at the cathode exhibited negligible
differences in their polarization curves. That is, a fuel cell operated with 100 sccm O2 at the
cathode achieved the same peak power density as 500 sccm O2, demonstrating that the presence
of CFx improves the efficacy of oxygen delivery such that oxidant saturation at the cathode
surface is achieved more rapidly. On the contrary, increasing oxygen flow rates in a fuel cell
containing only Pt:XC72 resulted in a steady decrease in measured power and current outputs.
Although, these Pt-Carbon mixtures exhibited significantly lower mass utilization than
commercially available Pt/C.
Ball-milling CFx with commercially available Pt/C into a 2:1 CFx:Pt weight ratio resulted in
a 100% increase in peak specific power, and a short circuit density equal to Pt(40wt%)/C, despite
containing nearly 20% less Pt, by mass. Further evidence for the enhancement in electrocatalytic
activity by CFx, is demonstrated by the increase in OCV relative to commercial Pt/C, in all CFx-
containing catalyst mixtures. Higher OCVs are the result of decreased activation overpotential at
the cathode, caused by the presence of CFx, whose fluorine moieties accelerate oxygen delivery
to Pt active sites. This acceleration compensates for kinetic and thermodynamic losses, which
contribute to the traditionally sluggish oxygen reduction reaction, thereby increasing the cathode
108

half-cell potential, and OCV concurrently. Further study into CFx-containing cathode materials
will likely result in greatly improved fuel cell technologies in the future.
3.5 References
(1) Litster, S.; McLean, G. J. Power Sources 2004, 130, 61.
(2) Sun, X.; Zhang, Y.; Song, P.; Pan, J.; Zhuang, L.; Xu, W.; Xing, W. ACS Catal. 2013, 3,
1726.
(3) Castanheira, L.; Silva, W. O.; Lima, F. H. B.; Crisci, A.; Dubau, L.; Maillard, F. ACS
Catal. 2015, 5, 2184.
(4) Cui, X.; Chen, J.; Wang, T.; Chen, W. Sci. Rep. 2014, 4, 5310.
(5) Fulvio, P. F.; Brown, S. S.; Adcock, J.; Mayes, R. T.; Guo, B.; Sun, X. G.; Mahurin, S.
M.; Veith, G. M.; Dai, S. Chem. Mater. 2011, 23, 4420.
(6) West, W. C.; Whitacre, J. F.; Leifer, N.; Greenbaum, S.; Smart, M.; Bugga, R.; Blanco,
M.; Narayanan, S. R. J. Electrochem. Soc. 2007, 154, A929.
(7) Zhang, Q.; Takeuchi, K. J.; Takeuchi, E. S.; Marschilok, A. C. MRS Adv. 2016, 1, 403.
(8) Baughman, R. H.; Zakhidov, A. a; de Heer, W. a. Science (80-. ). 2002, 297, 787.
(9) Berthon-Fabry, S.; Dubau, L.; Ahmad, Y.; Guerin, K.; Chatenet, M. Electrocatalysis
2015, 6, 521.
(10) Purwanto, W. W.; Slamet; Wargadalam, V. J.; Pranoto, B. Int. J. Electrochem. Sci. 2012,
7, 525.
(11) Avcioglu, G. S.; Ficicilar, B.; Eroglu, I. Int. J. Hydrogen Energy 2016, 41, 10010.
(12) Chen, M.; Wang, S.; Zou, Z.; Yuan, T.; Li, Z.; Akins, D. L.; Yang, H. J. Appl.
Electrochem. 2010, 40, 2117.
(13) Dias, A. M. A.; Freire, M.; Coutinho, J. A. P.; Marrucho, I. M. Fluid Phase Equilib. 2004,
109

222–223, 325.
(14) Viva, F. A.; Olah, G. A.; Prakash, G. K. S. Int. J. Hydrogen Energy 2017, 1.
(15) Nguyen, T. Q.; Bartrom, a. M.; Tran, K.; Haan, J. L. Fuel Cells 2013.
(16) Chang, J.; Feng, L.; Liu, C.; Xing, W.; Hu, X. Angew. Chem. Int. Ed 2014, 53, 122.
(17) Ha, S.; Larsen, R.; Zhu, Y.; Masel, R. I. Fuel Cells 2004, 4, 337.
(18) Cai, W.; Liang, L.; Zhang, Y.; Xing, W.; Liu, C. Int. J. Hydrogen Energy 2013, 38, 212.

Abstract (if available)

Abstract Direct oxidation formic acid fuel cells were developed in the 1990's and showed great promise as an emerging competitor with hydrogen and methanol fuel cells. This work examines all previously reported formic acid fuel cell technologies that have demonstrated proof-of-concept as power generation devices, and organizes the achievements of various research groups in a concise format. The evolution of formic acid fuel cells into their alkaline, formate-based counterparts in recent years is also addressed. Then, a closer examination of the performance characteristics of "state-of-the-art" anode materials calls into the question the utility of palladium in this role. Lastly, the effect of the incorporation of fluorinated active carbon into the cathode matrix is discussed, using hydrogen fuel cells as a model, and experimental results are presented that demonstrate its potential for improving the performance of any fuel cell that relies on the oxygen reduction reaction at the cathode.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Studies on direct methanol, formic acid and related fuel cells in conjunction with the electrochemical reduction of carbon dioxide

PDF

Design and modification of electrocatalysts for use in fuel cells and CO₂ reduction

PDF

The modification of catalysts and their supports for use in various fuel cells

PDF

Electrocatalysts for direct liquid-feed fuel cells and advanced electrolytes for lithium-ion batteries

PDF

Integrated capture and conversion of carbon dioxide from air into methanol and other C1 products

PDF

Selective fluoroalkylation methods and synthesis of water-soluble organic molecules for organic redox flow batteries

PDF

Understanding the mechanism of oxygen reduction and oxygen evolution on transition metal oxide electrocatalysts and applications in iron-air rechargeable battery

PDF

Design of durable and efficient catalysts for the electro-oxidation of methanol

PDF

Studies on iron-chloride redox flow battery for large scale energy storage

PDF

Development of polystyrene sulfonic acid-polyvinylidene fluoride (PSSA-PVDF) blends for direct methanol fuel cells (DMFCS)

PDF

Reforming of green-house gases: a step towards the sustainable methanol economy; and, One-step deoxygenative fluorination and trifluoromethylthiolation of carboxylic acids

PDF

PSSA-PVDF semi-IPN blends for direct methanol fuel cells

PDF

Integrated carbon dioxide capture and utilization: catalysis enabled carbon-neutral methanol synthesis and hydrogen generation

PDF

Hydrogen fluoride addition reactions and a fluorinated cathode catalyst support material for fuel cells

PDF

Dynamics of direct hydrocarbon polymer electrolyte membrane fuel cells

PDF

New bifunctional catalysts for ammonia-borane dehydrogenation, nitrile reduction, formic acid dehydrogenation, lactic acid synthesis, and carbon dioxide reduction

PDF

Electrochemical pathways for sustainable energy storage and energy conversion

PDF

Carbon-hydrogen bond activation: radical methane functionalization; unactivated alkene coupling; saccharide degradation; and carbon dioxide hydrogenation

PDF

Direct C−H arylation for the synthesis of conjugated polymers

PDF

Survival and evolution of Shewanella oneidensis MR-1: applications for microbial fuel cells

Asset Metadata

Creator Iuliucci, Marc Thomas (author)

Core Title Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 07/31/2017

Defense Date 05/26/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag CFx,DFAFC,formic acid oxidation,fuel cells,hydrogen,OAI-PMH Harvest,ORR,palladium,PEMFC,platinum

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Prakash, G.K. Surya (committee chair), Narayan, Sri (committee member), Shing, Katherine (committee member)

Creator Email iuliucci@usc.edu,marc.iuliucci@nawkaw.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-419750

Unique identifier UC11264477

Identifier etd-IuliucciMa-5661.pdf (filename),usctheses-c40-419750 (legacy record id)

Legacy Identifier etd-IuliucciMa-5661.pdf

Dmrecord 419750

Document Type Dissertation

Rights Iuliucci, Marc Thomas

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA