EPA-600/2-76-057
March 1976
Environmental Protection Technology Series
CARBON OXIDATION CATALYST
MECHANISM STUDY FOR FUEL CELLS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-057
March 1976
CARBON OXIDATION CATALYST
MECHANISM STUDY FOR FUEL CELLS
by
Yen-Chi Pan
Exxon Research and Engineering Company
Government Research Laboratory
Linden, New Jersey 07036
Contract No. 68-02-1831
ROAP No. 21BKB-006
Program Element No. 1AB013
EPA Project Officer: S. J. Bunas
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
- i -
TABLE OF CONTENTS
PAGE
LIST OF TABLES ii
LIST OF FIGURES ill
1. INTRODUCTION 1
2. SUMMARY OF THE RESULTS 2
3. CONCLUSION 13
4. GENERAL DISCUSSION 15
5. EXPERIMENTAL METHOD 19
6. RESULTS AND DISCUSSIONS 25
ABBREVIATIONS AND SYMBOLS. 58
REFERENCES 60
-------
- ii -
List of Tables
Page
Table 1 - Electrochemical Potentials Related to Oxygen Reduction
in Various Solutions 20
Table 2 - Parameters Involved in the Determination of the Oxygen
Reduction Reaction Activity in KOH Solutions 30
Table 3 - Limiting Current For 02 + H+ 2e ^ H202(H02~),
(mA/cm2) 34
Table 4 - Parameters Involved in the Determination of the Oxygen
Reduction Reaction Activity in Solutions Containing
Carbonate and Bicarbonate Ions 41
Table 5 - Parameters Involved in the Determination of the Oxygen
Reduction Reaction Activity in Buffer Solutions
Containing Phosphate 49
Table 6 - Parameter Involved in the Determination of the Oxygen
Reduction Reaction Activity in Potassium Acetate-Acetic
Acid Buffers and Sulfuric Acid 50
-------
- iii -
List of Figures
Page
Figure 1 - The First Wave of the Polarization Curves in 0.1M/0.1M
Solutions at 25°C 4
Figure 2 - The First Wave of the Polarization Curves in 0.1M/0.1M
Solution at 75°C 5
Figure 3 - The First Wave of the Polarization Curves in 1M/1M
Solution at 25°C 6
Figure 4 - The First Wave of the Polarization Curves in 1M/1M
Solution at 75°C 7
Figure 5 - Half Wave Potential vs pH Value at 25°C 8
Figure 6 - Half Wave Potential vs pH Value at 75°C 9
Figure 7 - Exchange Current Density vs pH Value at 25°C 10
Figure 8 - Exchange Current Density vs pH Value at 75°C 11
Figure 9 - End View of Rotating Ring-Disk Electrode 17
Figure 10 - The Flow Pattern of a Rotating Disk Showing Stream.-
lines . 18
Figure 11 - Circuit Diagram for Rotating Ring-Disk Electrochemical
Measurements 21
Figure 12 - Rotating Ring-Disk Electrode Assembly 23
Figure 13 - Reduction of Oxygen on Graphite in 0.1 N KOH at 25°C . 26
Figure 14 - Reduction of Oxygen on Graphite in 0.5N KOH at 75°C . 27
Figure 15 - Dependence of i
-------
- iv -
List of Figures (Cont'd)
Page
Figure 19 - Reduction of Oxygen on Graphite in 2N/1M KOH-K,C03
at 50°C 7 ... 39
Figure 20 - Reduction of Oxygen on Graphite in 0.5M/0.5M
K2C03-KHC03 at 25° C 40
Figure 21 - Dependence of. Limiting Current on w^ in Solutions
Containing Carbonate and Bicarbonate at 25°C .... 42
Figure 22 - Dependence of Limiting Current on w^ in Solutions
Containing Carbonate and Bicarbonate at 75°C .... 43
Figure 23 - Reduction of Oxygen on Graphite in Pre-electrolized
0.5M/0.5M K2C03-KHC03 at 25°C 45
Figure 24 - Reduction of Oxygen on Graphite in 1M/1M
KH2P04-K2Hp04 at 75°C 46
Figure 25 - Reduction of Oxygen on Graphite in 1M/1M KAcO-HAcO
at 75°C 47
Figure 26 - Reduction of Oxygen and Hydrogen Peroxide on Graphite
in 1M/1M KAc - HAc at 75°C 48
Figure 27 - Dependence of Limiting Current on w^ in Solutions
Containing Phosphate at 25° C 51
Figure 28 - Dependence of Limiting Current on w^ in Solutions
Containing Phosphate at 75° C 52
Figure 29 - Reduction of Oxygen on Graphite in 1M/1M HoPOA-KHoPO,
at 75°C 55
Figure 30 - Reduction of Oxygen on Graphite in 0.1 N Sulfuric Acid
at 75°C 56
Figure 31 - Dependence of Limiting Current on w^ in Potassium
Acetate-Acetic Acid Buffer and O.lN Sulfuric Acid . . 57
-------
1. INTRODUCTION
The objectives of this study were to develop information about
the electrocatalytic behavior of carbon and the reaction kinetics of oxygen
reduction, and to determine whether any combination of conditions, (temperature,
electrolyte concentration, and pH value) exist in which a carbon cathode
could operate sufficiently well to be used in a practical fuel cell. A
systematic examination of the electrocatalytic activity for oxygen reduction
in various electrolytes was conducted, using a rotating ring-disk electrode.
The oxygen or air cathode is undoubtedly the most desirable
cathode for a variety of types of fuel cells, simply because of its
relatively high open-circuit potential and the convenience of using air.
However, there are difficulties involved in the widespread application of
the air cathode. One of the major obstacles to its use is the cost and
the scarcity of the best catalysts--the noble metals in group VIII, Pt, Pd,
Os and Ir. Large quantities of these metals would be required for fuel cell ap-
plications and there simply are not enough of them; and the demand for this sort
of use would drive the cost up higher than the present level. Therefore, an in-
expensive and abundant material such as carbon is preferable for use as a
cathode material.
Certain carbons are of considerable technical importance as substrates
for the reduction of oxygen in alkaline electrolyte metal-air batteries and
fuel cells (1). In some applications, they also have the advantage of selec-
tivity; that is they are virtually inert in the oxidation of some soluble fuels.
In fuel cells using either organic fuels or reformed C02—containing fuels,
an electrolyte other than alkali may be required to prevent carbonate and
bicarbonate formation. However, among the sparse literature concerning oxy-
gen reduction on carbons, the emphasis is on pyrolytic graphite, and the
measurements are limited to alkaline solutions at room temperature (2-5).
Because of its untypically high crystalline order, pyrolytic graphite was not
considered the appropriate carbon for this study. Therefore, a more repre-
sentative carbon, commercial polycrystalline graphite, was chosen.
There have been numerous investigations of the cathodic oxygen
reduction reaction, and it is generally agreed that hydrogen peroxide
is the intermediate product. The reactions which may occur in the oxygen
reduction reaction are listed below:
09 + 2H90 + 4e > 40H" (la)
K2
02 + 2H+ + 2e v R —^ H202 (or H02 )
a - K1
H202a (or H02 ) + 2e •* >20H
2H009 4 > 0, + 2H,0 (Id)
K5 x
H_0oa L—" H202b (le)
K
a b
Here H202a and H202b designate the adsorbed and desorbed H202 respectively
-------
- 2 -
A general scheme of the overall reaction, disregarding the material balance,
can be expressed as
OH" (or H20)
K
K/
K,(+2e) \ _. K
'- H202a(H02
K.2(-2e)
OH"(or H20)
(1)
H202
The rate constant K^, which in general depends on the electrode potential,
pertains to the ith reaction indicated in the scheme, e.g. K^ represents
the catalytic decomposition of H202(H02-) yielding 02 and OH". The performance
of the electrode depends to a large extent on the pathway of the reaction
and its reaction rate.
In this study, a rotating ring-disk electrode technique was
used to study the oxygen reduction mechanism. This technique is the
most informative method of investigating the complex series of reactions
involved in the oxygen reduction. In the ring-disk electrode, the concentric
ring which is very slightly separated from the disk, can be set at a suitable
potential to monitor quantitatively the H20£ (or H02-) production from the
disk. The plots of the disk to ring current ratio, id/ir, obtained at
constant potential versus co% (co-angular velocity of the rotating electrode)
can serve as diagnostic criteria in distinguishing the pathways of reduction,
as well as means to obtain rate constant of various processes involved.
A rotating graphite disk with a gold ring electrode has been used for this
study.
-------
- 3 -
2. SUMMARY OF THE RESULTS
The polarization curves of oxygen reduction reaction on graphite
were systematically examined by using a rotating disk-ring electrode.
The reaction was studied in various electrolytes, over a pH range from
14.6 to 1 and a temperature range from 25° C to 75° C. The major findings
of the study are listed below:
(1) pH dependence of oxygen reduction activity on carbon
The activity of the oxygen reduction decreases with decreasing
pH values. The first wave of the polarization curves with respect
to reversible hydrogen potential are shown in Figures 1, 2, 3
and 4 for comparison. There is a linear relationship between
the pH value and the half wave potential, %i/2> the latter de-
creasing at a rate of approximately 40 mV per decade of H+ con-
centration. This relationship holds true at both 25°C and 75°C
(See Figures 5 and 6) . Decrease of exchange current density was also
observed as pH value decreased at both temperatures (See Figures 7 and 8)
(2) Carbonate formation in KOH solution does not effect the
oxygen reduction activity on carbon:
There was no direct significant effect on the activity as
seen from the polarization curves obtained in COj containing
KOH solution. Indirectly, however, the formation of carbonate
affected the pH value and hence the activity. Also as a secondary
effect, an increase in carbonate formation may result in a
decrease in the conductivity of the solution.
(3) Reaction sequence:
The plot of id/ir vs (to)""3* for KOH solutions indicated the reaction
sequence to be as follows: oxygen reduces to hydrogen peroxide,
which is then further reduced to water at higher cathodic
polarization.
This may be expressed as:
2e H0 — - ^ — 201T .
The analyses also indicate that no four-electron reduction of
oxygen on carbon occurs.
-------
1000
800-
<
a.
LU
or
cc
^
o
CO
0
600-
400-
200-
0
1.0
- 4 -
FIGURE 1
THE FIRST WAVE OF THE POLARIZATION
CURVES IN 0.1M/0.1M SOLUTIONS AT 25° C
0.8
0.6 0.4 0.2
DISK POTENTIAL vs R.H.E., VOLTS
- KHC0
a. 0.1 N KOH
b. 0.1M/0.1
c. 0.1M/0.1M
d. 0.1M/0.1M KAcO- HAcO
e. 0.1M/0.1M
f. 0.1 N H2S04
pH
pH
pH
pH
pH
pH
13
10.3
6.8
4.6
2.1
1
-0.2
-------
- 5 -
FIGURE 2
lOOOi
800
~ 600
400
200
0
1.0
THE FIRST WAVE OF THE POLARIZATION
CURVES IN 0.1M/0.1M SOLUTION AT 75°C
0.8
0.6
0.4
0.2
DISK POTENTIAL vs R.H.E., VOLTS
a. 0.1 N KOH
b. 0.1M/0.1M
- KHC0
c. 0.1M/0.1M K2HP04-KH2P04
d. 0.1M/0.1M KAcO = HAcO
e. 0.1M0.1M KH2P04
f. 0.1 N H2S04
pH
pH
pH
pH
pH
pH
0
13
10.3
6.8
4.6
2.1
1
-0.2
-------
500
400-
300-
LU
CC.
O
ifi
C/)
a
200-
100-
- 6 -
FIGURES
THE FIRST WAVE OF THE POLARIZATION
CURVES IN 1M/1M SOLUTION AT 25°C
0.8 0.6 0.4 0.2
DISK POTENTIAL vs R.H.E., VOLTS
a. IN KOH pH = 14
b. 1M/1M K2C03- KHCO
c. 1M/1M K2HP04- KH2
d. 1M/1M KAcO - HAcO
e. 1M/1M
-0.2
pH = 10.3
pH = 6.8
pH = 4.6
pH = 2.1
*- T -* T
*The current unit for (b) and (c) is half of the scale shown
-------
- 7 -
FIGURE 4
500
THE FIRST WAVE OF THE POLARIZATION
CURVES IN 1M/1M SOLUTION AT 75°C
0.8
a.
b.
c.
d.
e.
0.6 0.4 0.2
DISK POTENTIAL vs R.H.E., VOLTS
0
- KHC0
IN KOH
1M/1M
1M/1M
1M/1M KAcO - HAcO
1M/1M
*The current unit for (b) and (c) is half of the scale shown.
-------
15
14
13
12
11
10
9
£ 8
_i
>
for concentration of 0.1M/0.1M or 0.1M
D for concentration of 0.5M/0.5M or 0.5M
O for concentration of 1M/1M or 1
X
a.
7
6
5
4
3
2
0
- 8 -
FIGURE 5
HALF WAVE POTENTIAL vs pH VALUE AT 25°C
D/
v
w >
a o
V
^
I
/• 00
300
400 500 600 700
HALF WAVE POTENTIAL vs R.H.E., mV.
800
-------
LU
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1 -
0
300
- 9 -
FIGURE 6
HALF WAVE POTENTIAL vs pH VALUE AT 75°C
D for concentration of 0.1M/0.1M or 0.1 M
A for concentration of 0.5M/0.5M or 0.5M Q /'
O for concentration of 1M/1M or 1M /
/ A°
/ D A O
D £/O
D /O
w
D / O
1
400 500 600 700 800
HALF WAVE POTENTIAL vs R.H.E., mV
900
-------
UJ
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1-
0
10
-7
- 10 -
FIGURE 7
EXCHANGE CURRENT DENSITY vs pH VALUE AT 25°C
/
n for concentration of 0.1M/0.IMorO.lM ° /
V
A for concentration of 0.5M/0.5M or 0.5M D /
/
O for concentration of 1M/1M or 1M /
o n A
y
o AD
CD,
a,
10
-6
10
-5
10
-4
10
-3
10
-2
EXCHANGE CURRENT DENSITY (A/crn )
10
-1
-------
x 7
- 11 -
FIGURES
EXCHANGE CURRENT DENSITY vs pH VALUE AT 75°C
15
D for concentration of 0.1/0.1 or 0.1 M o /
A for concentration of 0. 5M/0.5M or 0.5M
13 \- a/
O for concentration of 1M/1M or 1M /
12 h /
11 h /
10
9
- =L /
3 sk /
/
5- /
DO/
2-
o
10"6 10'5 10'4 10'3 ID'2 lO"1
EXCHANGE CURRENT DENSITY (A/cm2)
-------
- 12 -
(4) Temperature effects:
When each concentration of each electrolyte was tested at
different temperatures, a higher level of activity was generally
associated with higher temperature.
(5) Concentration dependence:
In general, higher levels of activity were observed in
the more concentrated solutions.
(6) pH dependence of oxidation of hydrogen peroxide on gold.
The activity of oxidation of hydrogen peroxide decreases with
decreasing pH value.
(7) The residual current:
The oxidation residual current measured in nitrogen atmosphere
are generally small in the potential range 1.0 to 0.6 V (RHE).
In KOH solution the residual current is in the order of O.lyA
per total disk surface. In buffer solutions, the residual current
was in the order of lyA at 75°C and O.SyA at 25°C. The maximum
residual current of 4yA was measured in 0.1 M H2S04 at 75°C, which
seems to indicate the oxidation of carbon is higher in ^SO^ than
in buffer solutions and higher in the latter than in alkaline
media.
(8) Effects due to trace amounts of platinum:
A complete oxygen reduction at low cathodic polarization was
observed when the electrolyte solution was pre-electrolized by
using platinum electrodes, evidently as a result of a trace
amount of platinum introduced into the solution during that
process. This suggests that an effective technique may
be developed to prepare a well dispersed and low platinum
loading on carbon.
-------
- 13 -
3. CONCLUSION
The catalytic action of carbon is enhanced in an alkaline
medium and is progressively retarded as the electrolyte become more acidic.
Alkaline medium is therefore the most suitable electrolyte for using carbon
alone as the oxygen cathode. The polarization curves obtained in alkaline
medium show that 45yA current can be obtained at a polarization of .38V
below the theoretical oxygen reversible potential, 1.23V RHE. Double
layer capacitance measurements indicate that a 1 cm^ electrode containing
10 mg of high surface area carbon has 4,000 times the wetted surface area
of our rotating graphite disk electrode. By extrapolating the results
obtained by rotating disk electrode to the high surface area carbon, it
suggests that a completely utilized 10 mg/cm^ high surface area carbon
electrode is capable of delivering 180 mA/cm^ at 0.38V polarization;
e.g. without including other efficiency losses the hydrogen fuel cell
could have a cell voltage of .85V and a power density of 140 watts per
square foot. Fuel cell power plants for utility applications require
a cell voltage of 0.65 to 0.80 volt and a power density of 100 to 200
watts per square foot(6). Therefore, carbon is an acceptable cathode
material in alkaline solution,if coupled with a highly efficient hydrogen
anode. However, our results show that carbon is a poor catalyst for
H202 reduction (below 0.88V polarization from oxygen equilibrium
potential). The concentration of l^Oo produced from 02 reduction will
build up in the electrolyte, and the diffusion of 1^02 to anode will
reduce the efficiency of the H2 anode. In order to decrease the 1^02
concentration in the electrolyte, either a carbon electrode incorporating
an effective I^O^ reduction catalyst or an electrolyte contained soluble
1^2 decomposition catalyst is necessary. Since the carbon electrode can
only be used in alkaline electrolyte to achieve the power density needed,
the essential problem of carbonate and bicarbonate formation by using
organic or reformed -C02- containing fuel still remains to be overcome.
Though carbon by itself may not be useful as a catalyst, its use
as a support is unquestionably of great importance due to its high surface
area and electrical conductivity. Also, if there is bifunctional catalytic
action and favorable metal-carbon interaction, one could have increased overall
activity. The fact that this study found the stability of carbon to decrease
substantially with decreasing pH values, but overall is still excellent when
compared to other known possible supports, is very useful information.
The above points indicate the necessity of studying carbon as an electrode
material.
We reiterate the observation of substantial improvement in
catalytic action for oxygen reduction due to the presence of extremely
small quantities of platinum introduced into the system just by pre-electrolysis
of the electrolyte with platinum electrodes. This favorably underscores the
points made previously and indeed suggests that an effective technique can be
developed to prepare a well dispersed, low platinum loaded carbon cathode.
-------
- 14 -
The dependence of the catalytic activity of carbon on pH is
most probably be due to the concentration of H+. The half wave potential
of the oxygen reduction reaction vs pH value plot would have a slope of
59 mV/decade, if the reaction is not dependent on H* concentration; and it
would have a zero slope if the reaction has a first order dependence on H"1"
concentration. The slope obtained in this study was approximately 40 mV/decade,
and it suggests the rate dependence of a fractional order of H+ concentration.
One possible explanation is the occurrence of two slow irreversible steps;
one containing H such as
*0~ + H+ ^ *H02,
the other being * + 02 + 2e > *C>2 and the rest of the steps being
in quasi-equilibrium. Here, * indicates the surface. However, effects
such as the influence of solvent (electrolyte) on species absorbed on
carbon and hence the effect on activated complexes could be important
and should be given further consideration in more detailed future work.
Fundamentally, the study of solvent-solid catalyst interaction is im-
portant in order to be able to understand and hopefully predict catalytic
activity. The critical steps, from among the large number of elementary
steps which could take place on the surface, have been reported (7).
These steps can be elucidated with more confidence using isotope studies
(D+ and 0^) coupled with kinetic parameters such as those obtained
here and with which we have made some mechanistic prediction.
-------
- 15 -
4. GENERAL DISCUSSION
4.1 Fuel Cell
A fuel cell is a device for continuously converting chemical
energy into electrical energy through electrochemical reactions. A fuel
cell requires two electronic-conducting electrodes separated by an electrolyte.
Fuel is oxidized at the anode furnishing electrons to an external circuit,
while the oxidant accepts electrons from the cathode and is reduced.
Simultaneously, an ionic current in the electrolyte completes the circuit.
In a typical hydrogen-oxygen fuel cell the following reactions take place:
Anode: H2 + 20H" —) 2H20 + 2e
Cathode: % 02 + H20 + 2e > 20H"
Fuel cell: H2 +
The theoretical cell potential, E, can be predicted from the molar free
energy, AG, of the cell reaction:
AG - -nEF
where n is the number of electrons per molecule of fuel that is oxidized,
and F is the Faraday constant.
4.2 Polarization Curve
When current is drawn from an electrochemical cell, the terminal
voltage of the cell decreases from that which would be anticipated thermo-
dynamically. This phenomenon is called polarization. This decrease in
cell voltage reflects not only the potential drop cause by resistance
of the circuit and the fuel concentration gradient within the cell, but
also potential losses associated with irreversibility of various process
at the electrode-solution interfaces. These deviations in the potentials
of the electrodes from the values predicted thermodynamically are referred
to as electrode polarization. These current vs potential plots are called
polarization curves. The reversibility of a given electrode reaction,
and therefore the electrode polarization, depends on the electrode material
and operating conditions, such as temperature and electrolyte concentration.
By studying the behavior of the polarization curves, the nature of the
electrode reaction can be determined. Several parameters, such as the half
wave potential E^, the exchange current density, i^, and the Tafel slope,
have been used for the polarization behavior studies. The half wave
potential, which is defined as the potential where the current equals half
of the diffusion-limited current, has often been used for simple comparison
for the reversibility of the electrode reaction. Definitions of the other
parameters are discussed below.
-------
- 16 -
4.3 Butler-Volmer Equation and Tafel Line
The electrode polarization may generally be expressed by the
well-known Butler-Volmer equation as
—i
I
-1
1=io Le~ - - e ^'i/mj (2)
Here, i, the reaction current, is a measure of the reaction rate; io,
the exchange current density, is the current flow in any one direction at
equilibrium per unit electrode area; r) is the overpotential which is de-
fined as the potential difference between the electrode potential and
equilibrium potential; a and a are the transfer coefficients. When the
overpotential, n is large enough (approximate > 0.1V), one of the terms
in Eq. (2) becomes negligible compared to the other. The equation is,
therefore, approximately
or -i0 e (3)
By taking logarithms of this equation, the Eq. (3) can be reduced into
the commonly-used Tafel equation n = - —— In io + —- In i .
Otr CtF
When r\ values are obtained at various currents, it is possible to obtain
linear n vs. log i plots. Such n vs. log i plots are refered to as Tafel
lines. The slope of these Tafel lines provides a means of determing the
transfer coefficient a, the transfer coefficient consists of the parameter
of the fundamental reaction steps and the rate-determined steps. An anal-
ysis of the Tafel slope of a polarization curve will therefore be of
primary importance in mechanism determination. The exchange current
densities can be obtained by extrapolating the Tafel lines to n =0. The
exchange current densities reflect the kinetic properties of the particular
electrode interfacial systems concerned.
4.4 Rotating Ring-Disk Electrode
The rotating ring-disk electrode was used to study the oxygen
reduction mechanism. An end on view of such electrode is shown in Figure 9.
Here we have two active electrodes: the central disk electrode and the ring
electrode. These electrodes are separated by a narrow insulation. The
studied reaction is carried out on the disk electrode. Any intermediates
which are not firmly bound to the electrode may diffuse from the electrode
into the solution. If the electrode is rotating, the solution adjacent to
the electrode will acquire a convictive and radial motion (See Figure 10).
This will tend to carry the intermediates from the region of the disk past
the ring electrode. The intermediate can..then be detected at the ring. The
collecting efficiency, defined as the ratio of the intermediate produced at
the disk detected by the ring to the total intermediate produced at disk,
is independent of the rotation speed and the reaction. By examining the
simultaneously measured disk and ring currents, the reaction mechanism can
therefore be studied.
-------
- 17 -
FIGURE 9
END VIEW OF ROTATING RING-DISK ELECTRODE
-------
- 18 -
FIGURE 10
THE FLOW PATTERN OF A
ROTATING DISK SHOWING STREAMLINES
\
-------
- 19 -
5. EXPERIMENTAL METHOD
The following set of experiments were carried out for each
electrolyte concentration and at each temperature (25°C and 75°C, and
in some cases 50°C):
a. In Q£ saturated electrolyte, the dependence of the cathodic
current and the hydrogen peroxide oxidation current on disk
potential were simultaneously recorded, at three different
rotating speeds: 500 RPM, 1000 RPM and 2000 RPM.
b. Polarization curves were taken at 1000 RPM before and after
2 drops of 3% 1^02 are added to the electrolyte in order to
confirm the potential dependence of H202 reduction on carbon.
c. Residual current, which is the current measured with the
absence of oxygen, was determined in a nitrogen saturated
electrolyte.
The measurements were always taken after several sweeps between hydrogen
and oxygen evolution potentials. By doing so, better reproducibility was
obtained and little hysteresis was observed on the reversed scan, which
may be due to the slow change of the electrode surface state caused by
scanning potential.
Due to the pH change of the electrolyte and a corresponding change
in the reversible hydrogen potential, the recorded sweeps were adjusted to
the suitable range of potentials. The ring potentials were also adjusted in
order to obtain the maximum hydrogen peroxide oxidation current without the
interference of the oxygen evolution current. The starting potentials for
the recorded sweeps and the ring potential for different pH values are
listed in Table 1.
The experimental results reported in this work were obtained by
the slow sweep voltammetry technique used in conjunction with rotating ring
and disk electrodes. The block diagram of the potentiostatic ring and disk
circuit is given in Figure 11. Current-voltage polarization curves were ob-
tained by slow linear sweep voltammetry with a Princeton Applied Research
potentiostat (Model 173) and a universal programmer (Model 175) as a source
of scanning voltage. The ring potential was controlled by a home built
potentiostat. An X-Y-Y1 plotter (Esterline Angus 540) was used to record
the currents as a function of disk potential. All results reported here
are for a scan rate of 5 mv/sec. Slower scanning rates also have been
examined but no significant differences were found.
The disk electrode studied (see Figure 9) was made out of a
Carbone Corporation 5890 grade graphite (density 1.85 g/cnr). Instead
of platinum, 99.999% purity gold from Engelhard Industries was used for
the ring and the counter electrode construction in order to prevent
platinum contamination. The ring-disk electrodes were molded together
by Delta Cast 153 ETC epoxy from Wakefield Engineering Inc. The Delta
Cast 153 ETC has excellent chemical resistance for the range of electrolytes
we have studied.
-------
Table 1
Electrochemical Potentials
Solution
4N KOH
2N KOH
IN KOH
0.5 KOH
0.1 KOH
2N/2N KOH-K2C03
1:1 K2C03 - KHC03
1:1 K2HP04 - KH2P04
1:1 KAcO- HAcO
1:1 KH2P04 - H3P04
0.1N H^SO,
pH
14.6
14.3
14
13.. 7
13
14.57
10.3
6.8
4.6
2.1
1
RHE v&
SCE(mv)
-1105
-1087
-1068
-1049
-1009
-1100
-850
-643
-513
-365
-300
Related to Oxygen Rt
Recording Sweep
Starting Potential
(SCE)
0
0
0
0
0
0
+100
+200
+250
+400
+450
Equilibrium Potential
vs RHE Ring Potential used
f>2+2n+ ^ ^I^Q^ in Our Experiments
used In this report vs. SCE
25°C 50°C 75°C
769 758 748
760 749
751 740
742 731
721 711
769 758
682 657
682 657
682 657
682 657
682 657
739
730
721
701
748
633
633
633
633
633
0
0
0
0
0
0
+350 mV
+570 mV
+1.02 V
+1.35 V
+1.50 V
to
o
-------
- 21 -
FIGURE 11
CIRCUIT DIAGRAM FOR ROTATING
RING-DISK ELECTROCHEMICAL MEASUREMENTS
ROTATING
SPEED
CONTROLLER
CE
D
PI
P2
R
RE
T
UP
X-Y-Y1
TEMPERATURE
REGULATOR
Counter Electrode
Disk Electrode
Potentiostat for the Disk Electrode
Potentiostat for the Ring Electrode
Ring Electrode
Saturated Caromel Electrode, SCE.
Teflon Cell
Universal Programmer
X-Y-Y1 Recorder
-------
- 22 -
This epoxy was chosen mainly for its low thermal expansion coefficient
(1.8 x 10"^ cm/cm/°C) relative to other commonly used materials. The
molding technique was adapted to prevent possible leakage due to differ-
ential expansion for elevated temperature study. After the machining of
molded electrodes, the electrode surfaces were polished with fine grade
emery papers and finally with 6 micron Matadi diamond polishing compound
in situ. Finished electrodes (Figure 12) were rinsed thoroughly with
diluted sulphuric acid and leached in 45% KOH. The collecting efficiency
for the ring was measured in 0.001 M K3Fe(CN)6 solution with 1M:1M K2CC>3
and KHC03 as supporting electrolyte. The measured efficiency, N = 33%,
agreed well with the theoretical predicted value of 32% for our electrode
dimensions (8).
Other than electrodes, the remainder of the cell was constructed
of Teflon. The cell compartment contained a Teflon coated thermometer, a
reflux condenser and air outlet, a Teflon Luggin capillary, an air inlet
and electrodes. A saturated Calomel electrode, SCE, was used as reference,
which was separated from the Luggin by a Teflon stopcock.
The reversible hydrogen potential in various KOH and K2C03-KHC03
electrolytes at 75°C were measured by evolving hydrogen on a palladium wire.
The palladimized palladium wire was potentiostated to evolve hydrogen, and
the open-circuit potentials were measured with respect to saturated Calomel
electrode, SCE, which was used throughout the experiment as the reference
electrode. The process was repeated several times until steady potential
was measured. No significant differences were detected between the rever-
sible hydrogen potentials at 75°C and 25°C. The values were accurate within
a few mV to the theoretical calculated potential.
In order to characterize the effective surface area for various
electrode conditions, the double layer capacitances were measured by the
transient-charging-curves (9). In these measurements, the electrode was
potentiostated at open circuit potential, and then a 25 mV potential step
was applied by the universal programmer (Princeton Applied Research Model
175). Following the potential step, the transient charging curve was
recorded on an Techtronix 5103 N oscilloscope. The double layer capaci-
tance was then approximated from the transient equation:
I = I0e-t/ReC
here I0 is the peak current at the step, C is the double layer capacity,
and Re is the effective resistance of the circuit, which can be directly
obtained from Re = Ay , where Av is the applied potential step.
In
-------
- 23 -
FIGURE 12
ROTATING RING-DISK ELECTRODE ASSEMBLY
A. Graphite disk (O.D. .483 cm, Area .183 cm2)
B. Gold ring (I.D. .531 cm, O.D. .635 cm)
C. Epoxy holder and spacer
D. Shrinkable Teflon tubing
E. Stainless steel screw
F. Stainless steel rotating shaft
G. Disk electrical connection
H. Ring electrical connection
I. Epoxy seal
-------
- 24 -
Except where otherwise noted, the electrolytes were prepared
from following listed chemicals and distilled deionized water.
Chemical Manufacture Purity Grade
KOH Mallinckrodt Analytical Reagent
KHC03 Carlo Erba Chemical, Italy Analytical Reagent
K2C03 BDH Chemical Ltd. Analytical Reagent
K2HPC>4 Matheson Coleman and Bell A.C.S. Reagent
KH20P4 Matheson Coleman and Bell A.C.S. Reagent
H3P04 Fisher Scientific Co. A.C.S. Reagent
KAcO Research Organic/Inorganic Chemical Co. Analytical Reagent
HAcO Mallinckrodt Analytical Reagent
H2S04 Baker Chemical Co. Ultra High Purity
Commercial oxygen and nitrogen gas from Matheson Company were used
for the measurement.
-------
- 25 -
6. RESULTS AND DISCUSSION
The potential and current relationships for electrochemical
reduction of oxygen on carbon were obtained over a wide range of pH
values (from 4N KOH to 0.1N H2SO^), electrolyte concentrations and
temperatures. The results are grouped into the various chemical
compositions of the electrolyte and discussed in the following five
subsections. The correlations between all the measurements are summarized
and concluded at the end of this section.
6.1 Studies in Potassium Hydroxide
The electrochemical reduction of oxygen on carbons in KOH solution
are practically important, owing to the high conductivity of the OH" ion and
good oxygen reduction activity observed on high surface carbon. Five
different KOH concentrations, 0.1,0.5, 1.0, 2.0 and 4.0 N, were investigated.
The information obtained is discussed in the following separate subsections.
6.1.1 Polarization Gurves
The dependence of the cathodic currents and the hydrogen peroxide
oxidation currents on disk potential were simultaneously recorded for the
various KOH concentrations mentioned above. For each concentration, data
were obtained at 25° C, 50°C and 75°C. In all KOH measurements, the recorded
sweep was started at 0 mV SCE (the theoretical hydrogen reversible potential
in IN KOH is - 1.07 V SCE) and polarized to hydrogen evolution.
For each chosen concentration and temperature, measurements were
taken at three different rotating speeds, 500, 1000 and 2000 RPM. A
typical result obtained for 0.1N KOH at 25°C is presented in Figure 13.
The curves are designated by rotating speed and solid and dotted line
indicate the disk or the ring current, respectively. In general, the disk
current vs its potential plot present two distinct waves which are noted to be
of equal height. The first wave is expected to be the reduction of oxygen to
hydrogen peroxide; the second is the further reduction of hydrogen peroxide
to water or hydroxide ion. This expected reaction sequence is consistent
with the hydrogen peroxide oxidation current simultaneously measured on
the ring.
The polarization curves obtained in 0.5N KOH and some of the IN KOH
solutions, (see Figure 14, obtained in 0.5N KOH) show a maximum before
reaching the first plateau. These maximums were also observed in Morcos and
Yeager's study of pyrolytic graphite(4). The heights of the maximums varied
somewhat during separate measurements. Nevertheless, the increase in the height
of the disk current peak always corresponded with a proportional decrease in the
ring current level, it is clearly indicated that the additional current is
contributed by the further reduction of H202 formed in the first step of
oxygen reduction, instead of the explanation given by Morcos and Yeager as
that secondary effect associated with a decrease in electrocatalytic activity
of the electrode surface at more cathodic potentials after passing through
the maximum. We believe this additional current is due to the oxidation
state of the carbon surface or some trace impurity in the electrolyte pref-
erentially adsorbed on the electrode surface at that potential which
catalized the chemical decomposition or further reduction of H202« For
unknown reasons, the reactions are more sensitive at these medium con-
centrations of 0.5N and IN.
-------
CO
UJ
o:
CO
1.05
0.80
0.75
0.60
o
C3
1 °'45
Q
2
<
0.30
0.15
0.00
- 26 -
FIGURE 13
REDUCTION OF OXYGEN ON GRAPHITE IN 0.1 N KOH AT 25°C
disk current
ring current
2000 RPM
100 RPM
500 RPM
2000 RPM
1000 RPM
500 RPM
0
-300
-600
-900
-1200
-1500
DISK POTENTIAL vs SCE
-------
- 27 -
FIGURE 14
REDUCTION OF OXYGEN ON GRAPHITE IN 0.5N KOH AT 75°C
1.05
CO
5
.90
Qi
Di
ID
o
.75
CO
t .60
^ .45
Qi
O
.30
.15
0
0
disk current
ring current
2000 RPM
1000 RPM
500 RPM
2000 RPM
100 RPM
500 RPM
,
I
-300
-600
-900
-1200
-1500
DISK POTENTIAL vs SCE
-------
- 28 -
The polarized potentials were plotted vs log ia to approximate the
Tafel slopes. Here ia is the calculated activation current via the formula
!-!-+!_
i ia lL
where i is the observed current and i^ is the limiting current at the
current plateau or the maximum current at the peak. Good linear relation-
ships were plotted over two decades. The slopes are listed in Table 2.
The slopes are scattered around 79 mv/decade for the polarization
curves without the maximum peak. Isotopic studies(10) with 0^8 have indicated
that the reduction of 62 to H02~ as well as the reverse anodic process do
not involve the rupture of the 0-0 bond but rather just the modification
of the bond type. Having examined all the possible reaction paths
listed by Gnanamuther and Petrocelli (7) for the oxygen reduction
reaction, the following three steps were considered to be the most
probable rate determining steps:
* + 0
2e
2H20
H
*H02
*°2~2
2*H202
here, * represents surface,
bc (mv/decade)
59
30
59
bc' (mv/decade)
118 (5)
59 (6)
79 (7)
where -b_ and -b ' are calculated on the assumption that the electrons are
v* c
transferred from the substrate to the solution or to the adsorbed reactants
respectively. Eq. (6) has been included for consideration because the
measured slope may be higher than the actual Tafel slope value for a
porous electrode. The observed current at a certain potential may be larger
for a porous electrode than for an ideally non-porous carbon, due to the
contribution of the pore surface. This contribution of pore surface is
negligible near the limiting current region. The slopes for the curves with
a maximum peak are consistantly lower than those with no maximum peak. The
lower slopes seem to suggest that another slow reaction may be taking place.
The catalytic decomposition,
2 H202
2H20 + 02 (ads.)
(8)
which gives a Tafel slope of 30 mv, seems to be the most probable rate
determing step.
-------
- 29 -
Some of the additional electrochemical parameters derived from
these polarization curves, such as exchange current density, half wave
potential and the potential at which H2Q2 starts to undergo further re-
duction, are also listed in Table 2. These parameters characterized
the behavior of the oxygen reduction reaction under various conditions
and were used for the later comparison in section 3.5.
The exchange current densities are obtained by extrapolating
the approximated Tafel line to the reversible potential of electrochemical
reduction of Q£ and oxidation of ^C^- The reversible potentials are
affected by the dissociation of H202 ^^ H + HCL , which has a dissociation
constant of 10'11'6. Therefore, the equation, Eo (at 25°C) » 0.338 +
0.0295 pH (RHE), was used in alkaline region to obtain the reversible
potentials. To evaluate the current density at 50°C and 75°C, the
approximated equations E (at 50°C) = 0.327 + 0.0295 pH and E (at 75°C) "
0.317 + 0.029 pH (RHE) were used to obtain the reversible potential due to
the temperature correction of the free energy and the dissociation constant
of H202« The half wave potentials are dependant on the rotating speed and
the area of the rotating disk electrode. However, under the defined conditions,
the half wave potential is a good criteria for the electrochemical polarization
comparison. The average values of three rotating speeds, 500, 1000 and 2000
RPM are listed on Table 2.
6-1.2 The plot of id/ir vs (10)^ and reaction mechanism
The oxygen reduction reaction sequence can be examined by
plotting the disk to ring current ratio, id/irt versus U)H3, where w is
the angular velocity of the electrode. As we mentioned before, a
general scheme of oxygen reduction can be expressed as
OH" (or H20)
K
K/
OH"(or
(1)
The reaction rate constant for each step is designated by the subscribed
number, respectively.
-------
Table 2
Parameters Involved in the Determination of the Oxygen Reduction
Reaction Activity in KOH Solutions
Concentration Temperature
4N
PH=14.6
2N
pH=14.3
IN
pH=14
0.5N
pH=13.7
0.1N
pH=13
75° C
50° C
25° C
75° C
50° C
25° C
75° C
50° C
25° C
75° C
50° C
25° C
75° C
50° C
25° C
Eo
(a)
(b)
(c) io,
Zero current potential
half wave potential
the exchange current density
the potential where H202 starts to undergo further reduction
-------
- 31 -
The id/ir relation for this scheme has been derived by the
author's previous unpublished work with Dr. Halina Wroblowa at the University
of Pennsylvania, as:
Kfi
"3 + **> + A Z = 0.62 D'/3 v-1/6
where N is the collecting efficiency of the ring, D is the diffusion co-
efficient, and v the kinematic viscosity.
One typical example of this plot is shown on Figure 15 for
0.1 N KOH electrolyte at 25° C. The plot of i6+3, Fe(CN)6+^ redox couple.
By examining the Eq. (9), it is clearly indicated that the oxygen reduces
only to 1^02 (HOo-) by a two -electron process, no further reduction taking
place before -600 mv SCE. Further cathodical polarization causes the
slopes to increase, and further reduction of H202(H02~) is indicated. How-
ever, the interception still remains the same. By analyzing Eq. (9), it is
indicated that KS and K5 must be relatively large and (2K3 + K^) has a
finite value. The possibility of having four-electron reduction from 02 to
OH~ directly is small, in other words K^ is negligible compared to lOj. At
this stage, no clear distinction can be made between K3 and K^ by our results.
However, the potential dependence tends to reveal that the existence of K3
is more favorable than K^.
The id/ir analysis obtained for all the KOH electrolytes meet this
description, except in the maximum peak region for the polarization curves
obtained in 0.5N and some IN KOH concentrations. The i
-------
- 32 -
FIGURE 15
LLl
QL
o:
ID
O
C3
•21
OH
CC
Z)
O
0
-1 /?
-
DEPENDENCE OF id/ir ON o>
FOR OXYGEN REDUCTION IN 0.1N KOH AT 25°C
Up to -600 mV SCE
D
-------
- 33 -
6.1.3 Diffusion Limited
The diffusion limiting currents for the two electron process,
~
2H+ _ if _ ^ H0 (H0~) (10)
are reported in Table 3 for various KOH concentrations, temperatures and
rotating speeds.
The theoretical limiting current can be expressed as (11):
v,,,,, w% V% ( 0.6205 Sc-2/3 \
bulk "2 Vl + 0.298 Sc-1/3 + 0.14514 Sc-2/3 ) (11)
where iL is the diffusion limiting current in A/cm. F the Faraday constant
in coulomb, C^uik t^e bulk concentration in mole/cm , o> the angular velocity
in rad/sec, V the kinematic viscosity in cnr/sec, D the diffusion coefficient
in cnr/sec, and Sc the Schmidt number ( V/D). The limiting currents were
plotted versus 0)1$ and the necessary linear relationship was obtained in
each case. The plots are shown in Figures 16-18. The slope of these plots
contain information of the solubility and the diffusion coefficient of oxygen
in various electrolytes.
6.1.4 Double Layer Capacitance and Residual Current
The double layer capacitance of the graphite electrode was
measured, as described in the Experimental section, by the transient charging
curves. The capacitance measured in 0.1 N KOH increased slowly from
4.6 yF to 40 yF during a period of two weeks and then remained constant.
We believe that this is due to the electrolyte wetting the pores of the
graphite .
The residual currents measured in N2 were neglegible compared to
the current of the polarization curves over the entire potential range.
The residual currents in the operating range of air or oxygen cells, 1.0-0.6 V
(FHE), are generally in the order of a tenth of yA. The oxidation current
of the carbon is one of the contributors to these residual currents; based
on the magnitude of the observed residual currents, the level of oxidation
of carbon in KOH is not significant.
-------
- 34 -
Table 3
AN KOH
2N KOH
IN KOH
0.5N KOH
0.1N KOH
mt For
25° C
.21
.29
.43
0.57
0.82
1.15
.96
1.34
1.89
1.24
1.75
2.51
1.26
1.78
2.46
°2 + H+
50° C
.28
.41
.60
.65
.96
1.34
1.03
1.45
2.02
1.36
1.91
2.77
1.34
1.91
2.68
2e >
7
75° C
.27
.38
.50
.63
.87
1.23
.90
1.28
1.79
1;31
1.79
2.45
1.42
2.02
2.73
H202(H02")
w(rpm)
500
1000
2000
500
1000
2000
500
1000
2000
500
1000
2000
500
1000
2000
-------
- 35 -
FIGURE 16
E
o
LJ
O
3.5 -
1.0-
0.5-
DEPENDENCE OF LIMITING CURRENT
ON <}/2 IN KOH SOLUTION AT 25°C
for 0.1N
for 0.5N
for IN
for 2N
for 4N
0
(RPM)
1/2
-------
- 36 -
FIGURE 17
DEPENDENCE OF LIMITING CURRENT
ON
1/2
IN KOH SOLUTION AT 50°C
3.5
CM
LU
o:
Qi
^>
o
C3
3.0
2.5
2.0
1.5
1.0
0.5
0
0
O
D
A
T
for 0.1N
for 0.5N
for IN
for 2N
for 4N
10
8
20 30
,1/2(RPM)1/2
40
50
60
-------
- 37 -
FIGURE 18
DEPENDENCE OF LIMITING CURRENT
ON o)1/2 IN KOH SOLUTION AT 75°C
3.5
E
o
O
3.0
2.5
2.0
1.5
1.0
0.5-
0
O
D
A
T
for 0.1N
for 0.5N
for IN
for 2N
for 4N
0
10
20
30
40
50
60
(RPM)
1/2
-------
- 38 -
6.2 Studies in Solutions containing Carbonate and Bicarbonate Ions
As most fuel cells under investigation use either reformed
H2 or organic fuels, the presence of C02 is difficult to avoid.
Carbonate and bicarbonate will therefore be formed in KOH, and hence
the information on the electrochemical reduction in such solutions
is important to the development of a practical fuel cell using KOH
as the electrolyte. The electrolyte with 2N/IM KOH-K2C03 was studied
for this purpose. The 1:1 KoCC^-KHCOo buffer giving a pH value of
10.3 was studied in three different concentrationsj IN/IN, 0.5N/0.5N
and 0.1N/0.1N. The results are presented and discussed in the following
subsections.
6.2.1 Polarization Curves
The same types of experiments as described in Section 5 were
carried out in the various 0)3 and HCO^ containing electrolytes mentioned
above. Same prerecording potential scanning was undertaken to achieve
more reproducible surface conditions. Similar to the polarization curves
obtained in KOH solution, the disk current vs its potential plot presented
two waves with the same wave height. By comparing the polarization curves
to the ring current behavior, the same reaction sequence as we observed in
the KOH solutions was expected. For a 2N/1M KOH-K2C03 solution, the
recorded sweeps were started at 0 mV SCE, since the pH value was measured to
be 14.59 (the theoretical hydrogen reversible potential is -1.10 vs. SCE).
However, the K2C03-KHC03 buffer solution has a pH value of 10.3, which de-
creases the theoretical hydrogen reversible potential to -850 mV SCE. Thus,
the recorded sweeps were started at +100 mV. SCE. The ring potential was
also adjusted to +350 mV in order to detect the maximum hydrogen peroxide
current. Typical results obtained in 2N/1M KOH-K2C03 at 50°C and 1M/1M
K2C03-KHC03 at 75°C are presented in Figure 19 and Figure 20.
Some of the electrochemical parameters derived from these
polarization curves such as, half-wave potential, slope of potential vs
log ia plot, exchange current density and the potential where H202 started
to undergo further reduction, are listed in Table 4. For the evaluation
of the exchange current density, the reversible potential for
02 + 2H+ + 2e ^ > H202 at 75° C was corrected for the temperature
change of free energy of H202: at 75°C potential 633 mV RHE was used
instead of the 682 mV RHE used at 25°C. The 2N/1M KOH-^CO* results
indicate that presence of K2C03 in the KOH solution did not affect the
polarization of oxygen reduction on carbon. In the IfyCQ^ + KHCO-J solution,
with a pH value of 10.3, activity was significantly lessened, as shown by
the comparison of the half wave potentials and the exchange currents. It
appears then that changes in oxygen reduction activity correspond to
change in pH value, and such a change did not occur with the addition of
K2C03 into KOH solution. The limiting currents for the two electron oxygen
reduction process, for various solution and temperature were obtained, and
limiting currents in each electrolyte at 25°C and 75°C were plotted vs tu5
The linear relationships predicted by Eq. (11) were obtained in both cases,
and the plots are shown in Figures 21 and 22.
-------
- 39 -
FIGURE 19
REDUCTION OF OXYGEN ON GRAPHITE IN 2N/1M KOH-K2C03 AT 50°C
180
150
C/)
(—
-ZL
LU
C£
o:
ZD
o
o
z
Di
co
Q
120
70
60
30-
6\sk current
ring current
-300
-600
-900
-1200
-1500
DISK POTENTIAL vs SCE
-------
- 40 -
FIGURE 20
to
I-
z:
LU
CC
O
O
c/2
Q
REDUCTION OF OXYGEN ON GRAPHITE IN 0.5M/0.5M K2C03~KHC03 AT 25°C
0.7
0.6
0.5
0.4
0.3
0.2
0.1
disk current
ring current
1 J
0
+100 0
-200
-400
-600
-800 -1000
DISK POTENTIAL vs SCE
-------
Table 4
Parameters Involved in the Determination of the Oxygen Reduction Reaction Activity
in Solutions Containing Carbonate and Bicarbonate Ions
Concentration
and
Composition
2N/ 1M
KOH-K2COo
(pH-14.57)
1M/1M
K2C03 - KHC03
(pH=10.3)
0.5M/0.5M
(pH=10.3)
0.1M/0.1M
K2C03-KHC03
(pH-10.3)
Temperature
°C
75° C
50° C
25° C
75° C
25° C
75° C
25° C
75° C
25° C
Slope of Potential vs log ia
plot (mV/decade)
78
78
79
72
100
59
81
92
125 *
E%
(mV vs SCE) (mV
-205
-223
-216
-99
-134
-86
-124
-128
-173
vs
895
877
884
751
716
764
726
722
677
RHE)(a) i
4
3
2
3
1
2
4
1
2
(A/cm2)0
x 10"2
x 10~|T
x 10"2
x 10'2
x 10'3
x 10'2
x 10'3
x 10"2
x 10'3
EH2°2
(mV vs
-600
-750
-760
-400
-400
-380
-360
-380
-400
(0
(a) E, , half wave potential
(b) id, the exchange current density
(c) %90o^ the potential where H202 starts to undergo further reduction
-------
- 42 -
FIGURE 21
DEPENDENCE OF LIMITING CURRENT ON u/2 IN SOLUTIONS
CONTAINING CARBONATE AND BICARBONATE AT 25°C
O
o
3.5
3.0
2.5
2.0
= 1.5
1.0
0.5 -
0
0
for0.1M/0.1M
for 0.5M/0.5M
- KHC0
A
T
for 1M/1M K2C03 - KHC03
for 2N/1M KOH - K2C03
10
20
30
40
50
60
a,1/2 (RPM)1/2
-------
43 -
FIGURE 22
C\J
o
<
QJ
DC
£E
=3
O
O
3.5 -
3.0 -
2.5 -
I/?
DEPENDENCE OF LIMITING CURRENT ON 1/2
-------
- 44 -
6.2.2 The Effect of Impurities
The 003 and HCO^ containing electrolyte, the collecting efficiency
of the ring decreased with decreasing of pH values and also with increasing
concentration of 0)3 and HCO^. The reproducibility of the ring current also
decreased. The approximated collecting efficiencies for 2N/2N KOH-^003,
and for 0.1M/0.1M, 0.5M/0.5M and 1M/1M K2C03-KHC03 buffer are 27%, 22%, 19%
and 18% respectively. In order to ascertain that this decrease in the oxi-
dation activity of H202 on gold is due to an intrinsic chemical effect and
not to the impurity presented in the K2C03 and KHC03 reagents used for pre-
paring the solutions, two different approaches were used. These also pro-
vided a check on the impurity effect on the carbon electrode.
(a) 1M/1M K2C03-KHC03 solution was prepared from 4N KOH by bubbling
the solution with 2> which was passed through a dry ice trap before entering
the solution. During the converting process, samples from the solution were
regularly titrated by 0.1N HC1 until 1M/1M K2C03-KHC03 concentration was
reached. The polarization curves were then studied in the solution prepared
in this way; the disk activities and ring efficiencies obtained from K2C03
and KHC03 reagents. However the oxidation activity of H202 was still lower
than that in the KOH solution, and did not decrease with time. The full
efficiency was restored in KOH right after the K2C03-KHC03 measurements.
Therefore intrinsic chemical effects are accepted as the reason for the
decreasing of the ring activities, and electrode poisoning due to the impur-
ities is considerably small, or is a secondary effect.
(b) The 1M/1M and 0.5M/0.5M K2C03-KHC03 solutions were purified
by pre-electrolysis using two platinum electrodes. The polarization curves
obtained in these solutions have only one wave (Figure 23) and the limiting
current is comparable to the limiting current observed in the second wave
of normally used solutions (Figure 20); also the hydrogen evoluation over-
potential is significantly reduced for the pre-electrolized solution. Appar-
ently, the trace amount of platinum introduced from the pre-electrolysis
electrode deposited on the carbon was the reason for this activity. Another
solution pre-electrolized by using a gold anode and a platinum cathode was
studied on a repolished electrode. The observed polarization curves were
gradually changing from two waves to one when the measurement went on. It
further suggested that platinum dissolved from the pre-electrolysis platinum
electrode was the main cause of the activity. It indicated that even with
platinum as a cathode in the pre-electrolysis, the solution already contains
enough platinum to change the results significantly.
-------
- 45 -
FIGURE 23
<
Jl
C/}
Ul
Q;
o
o
0.1 -
REDUCTION OF OXYGEN ON GRAPHITE IN
PREELECTROLIZED 0.5M/0.5M K2C03-KHC03 AT 25°C
disk current
ring current
-200
-400
-600
-800 -1000
DISK POTENTIAL vs SCE
-------
- 46 -
FIGURE 24
350
300
250
200
o;
o
t/2
Q
150
100
REDUCTION OF OXYGEN ON GRAPHITE IN 1M/1M
KH2 P04- K2HP04AT 75°C
disk current
ring current
disk surface area: 0.183 cm
0
-200
-400
-600
-800
DISK POTENTIAL MS SCE
-------
- 47 -
FIGURE 25
700 -
600 -
500 -
<
a.
UJ
C£
QL
O
O
^
CO
Q
400 -
300 -
200 -
100 -
REDUCTION OF OXYGEN ON GRAPHITE IN 1M/1I
KAcO - HAcO AT 75°C
disk surface area: 0.183 cm
+200
-200
-400
-600
-800
DISK POTENTIAL vs SCE
-------
- 48 -
FIGURE 26
REDUCTION OF OXYGEN AND HYDROGEN PEROXIDE ON
GRAPHITE IN 1M/1M KAc - HAc AT 75°C
700-
600-
500-
400-
2 drops of 3% H20o added
300-
200-
100-
+200
-200
-400
-600
-800
DISK POTENTIAL vs SCE
-------
Table 5
Parameters Involved in the Determination of the Oxygen Reduction Reaction Activity
in Buffer Solutions Containing Phosphate
Concentration
and
Composition
1M/1M
K2HP04-KH2P04
(pH=6.8)
0.5M/0.5M
(pH=6.8)
0 . 1M/0 . 1M
K2HP04-KH2P04
(pH=6.8)
1M/1M
H3P04-KH2P04
(pH-2.1)
0.1M/0.1M
Temperature
°C
75° C
25° C
75° C
25° C
75° C
25° C
75° C
25° C
75° C
25° C
Slop of Potential vs log ia
plot (mV/decade)
108
99
105
136
100
128
140
140
154
159
E%
(mV vs SCE) (mV
-59
-87
-83
-135
-107
-140
+75
+10
+12
-69
vs RHE$a
584
556
560
508
536
503
440
375
377
296
N io 0 (b~
; (A/cm2)
1.6 x 10"4
2 x IO"5
2 x 10"4
4 x IO"5
2 x IO"4
6 x IO"5
5 x 10"5
7 x 10"6
3 x 10~5
8 x 10-6
^202
(mV vs i
-25)
-250
-240
-360
-230
-350
beyond
evo!
beyond
evo!
-300
beyond
VO
(pH-2.1)
evolution
(a) E!, half wave potential
(b) id, the exchange current density
^ ^20-2' the P°tential where H202 starts to undergo further reduction
-------
Table 6
Parameter Involved in the Determination
in Potassium Acetate-Acetic
Concentrat ion
and
Composition
1M/1M
KAcO -HAcO
(pH-4.6)
0.1M/0.1M
KAcO -HAcO
(pH-4.6)
IN H2S04
(pH-1)
Temperature
°C
75° C
25° C
75° C
25° C
75° C
25° C
Tafel Slope
(mV/decade)
125
125
135
175
115
115
of the Oxygen Reduction Reaction Activity
Acid Buffers and Sulfuric Acid
E%
(mV vs SCE)(mV
-11
-57
-57
-180
+115
+23
vs RHE)
502
456
456
333
415
323
io EH2<>2
^a) (A/cm2)(b) (mV vs SCE)(°)
1 x 10~4 -180
1.5 x 10-5 -250
6 x 10~5 -200
1.4 x 10~5 -350
1.8 x 10 "5 beyond hydroger
evolution
1.1 x 10"° beyond hydroger
On
O
(a) E^, half wave potential
(b) io, the exchange current density
(c) EH-O' the P°tential «here H2°2 starts to undergo further reduction
-------
E
o
LU
O
- 51 -
FIGURE 27
1/2
DEPENDENCE OF LIMITING CURRENT ON
-------
- 52 -
FIGURE 28
_E
h-
o
3.5
3.0
2.5
2.0
1.5
1.0
0.5
1/2
DEPENDENCE OF LIMITING CURRENT ON
IN SOLUTIONS CONTAINING PHOSPHATE AT 75°C
O for 0.1M/0.1M KH2P04 - H3P04
• for 0.1M/0.1M K2HP04 - KH2P04
A for 1M/1M KH2P04 - H
T for 0.5M/0.5M K2HP04
• for 1M/1M K2HP04 - KH2P04
- KH2P04
(RPM)
1/2
-------
- 53 -
6.4 Studies in the Potassium Phosphate - Phosphoric Acid Buffer
Solution and Sulfuric Acid
In order to investigate the pH dependence of the electrochemical
reduction of oxygen on carbon, three solutions were studied: 1M/1M and
0.1M/0.1M KH2P04-H3P04 with a pH value of 2.1, and 0.1N H2SO, with a pH
value of 1. The polarization curves were studied at 25°C and 75 C.
The polarization curves in these solutions were studied following
the procedure described in the experimental section. In 1:1 KH-PO.-H-PO,
buffers, the current was recorded starting at 4-400 mV SEC, and in the 0.1N
H~SO, solution it was recorded starting at +450 mV SEC, with ring potential
or +1.35 V and +1.50 V SCE respectively to detect the hydrogen peroxide
formed on the disk electrode. The rate of oxidation of H.,0,, on the ring
was very low in these solutions-the collection efficiency was approximated
to be 5%. It varied with the concentration, lower collection efficiency
being always associated with higher concentration.
The patterns of the polarization curves obtained from these
solutions were also similar. Only one wave was observed (Figures 29 and 30).
Examination of the shape of the ring current indicated that H»0? was not
being further reduced. It is assumed then that the reaction continued as a
two-electron reduction process until hydrogen evolution potential. The
hydrogen peroxide reduction process starts after the hydrogen evolution
potential is covered by the overwhelming current from the hydrogen evolution;
thus this hydrogen peroxide reduction current was not shown by a second wave,
as it was in previous experiments. This assumption was confirmed by the response
of the polarization curves when 2 drops of 3% H~0? were added to the solution.
The additional HO in the solution did not increase the current (change
the normally obtained curve) at 25 C: only a slight additional current was
derived from these polarization curves, such as half-wave potential, Tafel
slope and exchange current density are listed in Tables 5 and 6. The limiting
current are approximated and plotted vs. u£$ for each concentration at two
different temperatures are 'shown in Figure 31.
-------
- 54 -
6.3 Studies in the Potassium Phasphate Buffer and the Potassium
Acetate - Acetic Acid Buffer
The pH values of the 1:1 K2HPOA-KH2P04 and the 1:1 KAcO-HAcO
buffer solution are 6.8 and 4.6, respectively. In order to investigate
the pH dependence and concentration effects of the electrochemical reduction
of oxygen on carbon, five compositions were studied: 1M/1M, 0.5M/0.5M
and 0.1M/0.1M of K2HP04-KH2PO^ solution, and 1M/1M and 0.1M/0.1M of KAcO-HAcO
solution. The polarization curves were studied at 25°C and 75°C.
The polarization curves for all the solutions described here
were obtained by the procedures described in Section 5. In 1:1 K2HP04~
KH2PO^ buffers, the current was recorded from a disk potential of
+200 MV SCE to the hydrogen evolution potential, with the ring potential
set at +570 mV SCE. In 1:1 KAcO-HAcO buffers, recording was begun at a
disk potential of +250 mV SCE and the ring potential was reset at +1.02 V
SCE. The patterns of the polarization curves obtained for both buffers are
similar, and are presented and discussed together below. In these solutions
the waves representing the two separate oxygen reduction steps were more
difficult to distinguish. Before the oxygen-reduction-to-hydrogen-peroxide
wave reached the plateau, the hydrogen peroxide had already started to
undergo further reduction, thus overlapping the first step of the reaction
as shown in Figures 24 and 25. However, examination of the maximum at the
ring current plot still enabled us to estimate with reasonable accuracy the
limiting current of two-electron oxygen reduction and the potential where
H202 reduction started taking place. After taking the normal polarization
curves, two drops of 3% H202 were added to the solution to obtain another
polarization curve. The results of the curves obtained (a) without and (b)
with the addition of H202 were in agreement; ie at that potential where the
ring current in (a) started to fall, the disk current in polarization curves
(b) rapidly increased beyond values normally obtained without addition of
H202 as shown in Figure 26.
Some of the electrochemical parameters derived from these polarization
curves related to the oxygen reduction activities such as, half wave potential
(E^), Tafel slope, exchange current density and potential where H202 starts
to undergo further reduction, are listed in Tables 5 and 6. To approximate
the exchange current densitieis, the Tafel lines were extrapolated to the
reversible potential of the o2 + 2H+ 2e "» H202 couple; the potential at 25°C
is +682 mV RHE was used at 75°C. The limiting current plots vs o% for each
electrolyte at two different temperatures are shown in Figures 27 and 28.
-------
- 55 -
FIGURE 29
REDUCTION OF OXYGEN ON GRAPHITE IN 1M/1M H3P04 - KH2P04 AT 75°C
700k
600
500 h
400h
cc
3D
O
M
c/2
Q
6\sk current
ring current
disk surface area: 0.183 cm
300h
200h
lOOh
+200
0
-200
-400
-600
-800
DISK POTENTIAL vs SCE
-------
- 56 -
FIGURE 30
REDUCTION OF OXYGEN ON GRAPHITE IN 0.1 N SULFURIC ACID AT 75°C
700
600
500
400
LU
Cm
cm
Z3
O
^
(S)
Q
300
200
100
0
disk current
ring current
2
disk surface area: 0.183 cm
2000 RPM
+400
+200 0 -200 -400
DISK POTENTIAL vs SCE
-100
-------
I-
Z
LLl
CC.
O
O
- 57 -
FIGURE 31
1/2
7
DEPENDENCE OF LIMITING CURRENT ON
-------
- 58 -
ABBREVIATIONS AND SYMBOLS
b calculated Tafel slope based on the assumption that the electrons
are transferred from the substrate through the double layer to
the reactants.
b'c calculated Tafel slope based on the assumption that the electrons
are transferred directly from the substrate to the adsorbed
reactants.
C double layer capacity.
C, , bulk concentration of the reactant.
D diffusion coefficient.
E equilibrium potential of the reaction 0 +2H 0+2e -—-^ 2H 0 .
Ej half wave potential, potential where the current equals half of
the diffusion limited current of a polarization curve.
EU Q , potential where H^O^ starts to undergo further reduction.
F Faraday constant, 96500 coulombs.
I measured current for transcient charging curves.
I current measured at the potential step occurs for transcient
charging curves.
i calculated activation limiting current.
cl
i, current measured on disk electrode.
d
iT diffusion limiting current.
i-i
± exchange current densities of the reaction 09+2H90+2e v 2H-0,,.
i current measured on ring electrode.
K. reaction rate constant.
i
N collecting efficiency for the ring electrode
Re effective resistance
-------
- 59 -
ABBREVIATIONS AND SYMBOLS (cont'd)
RHE potential with respect to the reversible hydrogen potential
in the same solution.
RPM revolutions per minute
Sc Schmidt number (v/D).
SCE potential with respect to the potential of saturated Caromel
electrode.
Z Abbreviation for 0.62 D2/3 v"1/6.
* surface of the electrode
H overpotential, the potential difference between the electrode
potential and equilibrium potential.
AV potential step applied by the potentiostat.
GJ angular velocity.
V kinematic viscosity.
-------
- 60 -
REFERENCES
(1) K. V. Kordesch in "Fuel Cells" Chemical Technology, Vol. 1,
I. W. Mitchell, Jr. Ed. Academic Press, New York (1963).
(2) M. R. Tarasevich, F. Z. Sabirov, A. P. Mertsalova and R.K. Burstein,
Electrokhimiya 4_, 432 (1968).
(3) E. Yeager, P. Krouse and K. V. Rac, Electrochim. Acta 9_, 1057 (1964).
(4) I. Morcos and E. Yeager, Electrochim. Acta L5, 953 (1970).
(5) A. J. Appleby and J. Marie, International Society of Electrochemistry,
25th Meeting Brighton - September 22-27, 1974.
(6) EPRI SR-13, Special Report, August 1975.
(7) D. S. Gnanamuthu and J. V. Petrocelli, J. Electrochem. Soc. 114,
1036 (1967).
(8) W. J. Albery and S. Bruckenstein, Trans. Faraday Soc. 62 (7),
1920 (1966).
(9) W. Tiedemann and J. Newman, J. Electrochem. Soc. 122, 70 (1975).
(10) M. Davies, M. Clark, E. Yeager and F. Hovorka, J. Electrochem. Soc.
106, 56 (1959).
(11) R. E. Davis, G. L. Horvath and C. W. Tobias, Electrochim. Acta 12,
287 (1967).
-------
- 61 -
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-057
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Carbon Oxidation Catalyst Mechanism Study for
Fuel Cells
5. REPORT DATE
March 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Yen-Chi Pan
8. PERFORMING ORGANIZATION REPORT NO.
GRU. 2DYBA. 75/715520
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Exxon Research and Engineering Company
Government Research Laboratory
Linden, New Jersey 07036
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21BKB-006
11. CONTRACT/GRANT NO.
68-02-1831
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
Final: 10/74-6/75
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES project officer S.J.Bunas is no longer with EPA; for details,
contact G.L.Johnson, Mail Drop 64, Ext 2815.
is. ABSTRACT Tne repOr^ gjves results of 2i systematic examination of the polarization
curves of oxygen reduction reaction on graphite, using a rotating disk-ring electrode.
The reaction was studied in various electrolytes over a pH range of 14. 6-1 and a
temperature range of 25-75 C. The activity of this reaction was found to increase
with decreasing H(+) concentration and increasing temperature. A linear relationship
was found between the half-wave potential and pH value, with a slope of 40 mV per
decade. The plot of ring current ratio versus the square root of angular speed in KOH
solution clearly indicated a reaction sequence of: O2 + 2H(+) yields 2e + H2O2
which yields 2e + 2OH(-). No 4-electron reduction of oxygen on carbon occurred.
The level of carbon oxidation was higher in more acidic medium.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Air Pollution
Fuel Cells
atalysis
Electrostatics
Graphite
Electrodes
Current Density
Diffusion
Activation
Polarization
Cathode
Air Pollution Control
Oxygen Reduction
Rotating Disk-Ring
Electrode
Tafel Slope
Residual Current
13B
frOB
07D
20C
08G
09A
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
67
?O SFCURITY CLASS ITIii* ncpp)
Unclassified
2?.. PRICE
EPA Form 2220-1 (9-73)
------- |