(141a) Influence of Carbon Dioxide and Carbonate on the Electrode Reactions in Alkaline Direct Methanol Fuel Cells

Introduction

Direct methanol fuel cells (DMFC) have been among the
first fuel cells to be commercialised. The high energy storage capacity and
easy handling of the liquid fuels makes the use of DMFC advantageous for
applications where a long running time with medium to low power demands is
required. Furthermore, the simple and thus robust system design and the typical
set-up which keeps the membrane wet for most of the time even after shut-down
leads to a high cycle stability. Therefore, DMFC are in particular suitable for
applications were random start-up and shut-downs are to be expected like
portable power supply, on-board vehicle battery charging or back-up power.
Whereas the market for portable fuel cell power systems never really took off
outside the military world, battery charging and in particular back-up power
are increasing in volume. However, for these markets larger systems in the
range of kilowatts are often required. Here the use of today’s proton exchange
membrane technology based DMFC experiences economical limits due to the need to
use platinum based catalyst at rather high loadings. Alkaline anion exchange
membrane based DMFC could be an alternative to overcome this issue. However,
classic alkaline fuel cells (AFC) are known to be sensitive against carbon
dioxide from air. As was shown by Inaba et al.
typical anion exchange membranes also suffer from a reduction of conductivity
if exposed to carbon dioxide¹.
In contrast to the situation in AFC the effect was, however, found to be
reversible. As the current density at which DMFC typically operate are much
lower than those of fuel cells operating on hydrogen, the effect of increased
ohmic resistance can also be expected to be less relevant.

Much less is known about the effect of CO₂
and carbonate ions on the electrochemical reaction. Vega et al. reported on the
effect of CO₂ on the ORR at Pt electrode in KOH electrolyte². A slight reduction of performance was
found. For the alcohol oxidation and the oxygen reduction at non PGM catalyst
even less is known.

In this study we will report on systematic tests of
the methanol oxidation at carbon supported platinum catalyst in KOH and KHCO₃
electrolytes as well as the oxygen reduction at commercial PGM-free cathode catalyst
K4020 by Acta.. The results will be compared to studies on single cell
level. Here tests were performed using also KOH or KHCO₃ as
supporting electrolyte. Additionally, the effect of dosing CO₂ to
the synthetic air feed of the cathode was investigated.

Experimental set-up and procedures

Catalyst tests were performed using a standard
rotating disk electrode set-up from PINE instruments with a glassy carbon disk
electrode. The disk was coated with an ink of catalyst power dispersed in a
mixture of DI water and 2-propanol with a very small amount of added Nafion^®. A platinum foil was used as counter
electrode and a HydroFlex^® reversible
hydrogen electrode by Gaskatel as reference
electrode. The used cell was equipped with a double jacked so that heating of
the cell was possible. Measurements there conducted between 30 °C and 60 °C.

Single cell tests were performed at membrane electrode
assemblies consisting of a Fumatech FumaPEM FAA3 electrolyte membrane, an anode GDE made from a
Sigracet 10AA GDL coated with Johson
& Matthey HiSpec3000 Pt/C catalyst at a platinum loading of 2 mg cm^-2
and a cathode GDE made from a Freudenberg H32C2 GDL coated with Acta K4020 catalyst at a loading of 5 mg cm^-2.
Test were performed in a qFC 100-25 cell fixture by balticFuelCells using an Autolab potentiostat as electric load. On the cathode side gas was
supplied using mass flow controllers from Bronkhorst
for synthetic air and CO₂.

Results

The first observation comparing the methanol oxidation
at Pt/C catalyst in 0.25 M KOH and 0.25 M KHCO₃ solution is that the
achievable peek mass activity in the KOH solution is about a factor of 5 higher
than in KHCO₃ solution. At the same time the peak potential in the
KHCO₃ electrolyte is shifted towards lower potentials, whereas the
onset potentials are comparable. In both cases a pronounced temperature
dependence was observed with the mass activity increasing by about a factor 5
if temperature is raised from 30 °C to 60 °C. A further significant difference
between the CV in KOH and KHCO₃ electrolyte is the presence of a
second oxidation wave in the latter. The CV in KHCO₃ resembles thus
a CV in acidic electrolyte.

Figure 1: CV of the methanol oxidation in 0.1 M
methanol solution at HiSPEC 3000 Pt/C covered electrode
as function of temperature left using 0.25 M KOH right 0.25 M KHCO₃
as electrolyte.

Chronoamperometric measurement of the methanol oxidation in KOH at different temperatures
showed that the initial current is also increasing with temperature. Up to 40
°C the increased temperature does not significantly change the poisoning rate
of the electrode. For temperatures of 50 °C and 60 °C an accelerated electrode
poisoning of the electrode is observed. In consequence the mass activity at 40
°C exceeds the mass activity at higher temperatures after 10 minutes.

In KHCO₃ electrolyte the CA experiments
revealed a significant thermal activation of the mass activity only for
temperatures of 50 °C and 60 °C. The accelerated degradation here is only
observed at 60 °C.

Figure 2: CA measurement  of the methanol oxidationm at HiSPEC 3000 Pt/C
covered electrodes in 0.1 M MeOH solution at 0.7 V vs. RHE as function of
temperature; left in 0.25 M KOH; right in 0.25 M KHCO₃.

For the oxygen reduction reaction at Acta K4020 catalyst in KOH electrolyte almost no
temperature effect was observed. In KHCO₃ at 30 °C the onset
potential and half wave potentials are shifted towards lower values by about
200 mV. However, the ORR in KHCO₃ exhibited a strong thermal
activation so that 60 °C the shift in the onset potential is reduced to about
50 mV.

Figure 3: LSV of the oxygen reduction at Actaka K4020
cobered electrodes, left comparison for 0.25 M KOH and 0.25 M KHCO₃
electrolyte at 30 °C; right temperature dependence in 0.25 M KHCO₃.

Single cell tests were first performed using 4M
methanol, 1 M KOH electrolyte solution as fuel on the anode side. The cathode
was supplied with 500 ml min¹ (STP) synthetic air to which CO₂
could be added to yield CO₂ concentrations between 0 ppm and 1600
ppm.

IV curves recorded at 60 °C with 0 ppm and 400 ppm CO₂
addition on the cathode side showed only a small effect of the CO₂
addition The Peak power was reduced by about 2 mW cm^-2.

Under galvanostatic
conditions of 30 mA cm² a stable voltage after an initial voltage
decay was observed for all cathodic CO₂
concentrations from 0 ppm to 1600 ppm. With increasing CO₂
concentration on the air side the stable voltage was however slightly reduced.

Figure 4: Single cell tests at 60 °C with 4 ml min^-1
4M MeOH, 1M KOH anode feed and 500 ml min^-1 (STP) synthetic air
cathode feed with variable addition of CO₂; left iV curves, right
galavnistaic measuremt at 30 mA cm^-2

If the electrolyte is changed to 4 M methanol, 2 M
KHCO₃ solution the cell performance is drastically reduced to values
in the 1 mW cm^-2 range. However, here the
addition of 400 ppm to the cathode air supply during iV
measurements yielded a raise in the peak performance by about 0.2 mW cm^-2. This positive influence of cathodic CO₂ addition on the cell performance
with anodic KHCO₃ electrolyte could be verified in galvanostatic measurements.

Figure 5: Single cell tests at 60 °C with 4 ml min^-1
4M MeOH, M KHCO₃ anode feed and 500 ml min^-1 (STP)
synthetic air cathode feed with variable addition of CO₂; left iV
curves, right galavnistaic measuremt at 30 mA cm^-2

Discussion

Full saturation of the electrolyte with CO₂
leading to KHCO₃ electrolyte solution change the electrochemical
behaviour of the catalyst significantly. In all cases the obtainable mass
activity was reduced. For the methanol oxidation the CV also exhibits typical
features of the methanol oxidation in acidic environment. Therefore, the change
in pH will play a strong roll. For the methanol oxidation no significant
differences in the thermal activation were observed. In contrast to this the
ORR in KHCO₃ is strongly thermally activated so that at high
temperature the differences to KOH become small.

In single cell tests the change from KOH to KHCO₃
electrolyte on the anode side had a much stronger effect than the addition of
CO₂ on the cathode side. This also indicates the strong influence of
the electrolyte on the methanol oxidation kinetics.  

1.            Inaba
M, Matsui Y, Saito M, et al. Effects of carbon dioxide on the performance of
anion-exchange membrane fuel cells. Electrochemistry.
2011;79(5):322-325.

2.            Vega
JA, Mustain WE. Effect of CO(2), HCO(3)(-) and CO(3)(-2) on oxygen reduction in
anion exchange membrane fuel cells. Electrochimica
Acta. Feb 1 2010;55(5):1638-1644.