Kinetics and mechanism of the electrooxidation of formic acid - Spectroelectrochemical studies in a flow cell

Article abstract of DOI:10.1002/anie.200502172

(Figure Presented) Go with the flow: A novel spectroelectrochemical flow cell with well-defined mass transport allows time-resolved electrochemical and in situ FTIR spectroscopy measurements under continuous electrolyte flow (e.g. during electrolyte excha

Full text of DOI:10.1002/anie.200502172

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Angewandte
Chemie
Electrocatalysis
DOI: 10.1002/anie.200502172
Kinetics and Mechanism of the Electrooxidation
of Formic Acid—Spectroelectrochemical Studies
in a Flow Cell**
Yan Xia Chen,* Martin Heinen, Zenonas Jusys, and
Rolf Jürgen Behm*
The application of in situ infrared spectroscopy in electro-
chemistry is well established. It has been particularly useful in
electrocatalytic studies, in which the knowledge of the
adsorbed species can contribute significantly to a fundamen-
tal understanding of the reaction mechanism (see, for
example, reviews [1–3] and references therein). Most of
these studies were conducted by using IR reflection absorp-
tion spectroscopy (IRRAS) techniques in a thin-layer con-
figuration, where the thickness of the electrolyte layer
between the electrode and the IR window is limited to
about 10 mm to keep spectral contributions from the electro-
lyte at a tolerable level. Under these conditions mass trans-
port in the thin-layer gap is practically inhibited, which may
result in strongly misleading conclusions on the reaction
process owing to the depletion of reactants as well as
accumulation and possibly readsorption of by-products.[2]
These limitations can be reduced by performing infrared
absorption reflection spectroscopy measurements in an
attenuated total reflection configuration (ATR-FTIRS),[4,5]
in which a thin metal film deposited on an optical prism serves
as a working electrode and the IR beam is transmitted
internally from the back side of the prism to the film and
totally reflected at the prism/film interface. In this technique
the working electrode is freely accessible to the electrolyte.
Nevertheless, despite the obvious improvements over the
thin-layer configuration, diffusion of reactants to and prod-
ucts from the electrode is still ill-defined.
Herein we describe an extension of the present in situ
ATR-FTIRS technique by coupling it to a thin-layer electro-
chemical flow cell. We demonstrate the potential of this
technique for mechanistic and kinetic studies of electro-
[*] Dr. Y. X. Chen, M. Heinen, Dr. Z. Jusys, Prof. Dr. R. J. Behm
Dept. Surface Chemistry and Catalysis
University of Ulm
89069 Ulm (Germany)
Fax: (+49)731-502-5452
E-mail: yan-xia.chen@uni-ulm.de
juergen.behm@uni-ulm.de
[**] We gratefully acknowledge the help of M. Osawa and S. Ye
(University of Hokkaido, Japan) in introducing the ATR-FTIRS
technique as well as financial support by the Forschungsallianz
Brennstoffzellen (FABZ) Baden-Württemberg and by the Deutsche
Forschungsgemeinschaft (BE 1201/8-4, BE 1201/11-1). Y.X.C. is
grateful for a fellowship from the Alexander von Humboldt
Foundation.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 981 –985
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
981

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communications
catalytic reactions, with the electrooxidation of formic acid on
a Pt film electrode as an example. This reaction can be
considered as a model reaction for the electrooxidation of
small organic molecules. In addition, it is also of technical
interest, as formic acid is a reaction intermediate in direct
methanol fuel cells[6] and was also proposed as a fuel for low-
temperature fuel cells.[7] It is generally accepted that the
reaction proceeds through two different reaction pathways
(“dual-path mechanism”): an “indirect” pathway involving
formic acid dehydrogenation and subsequent electrooxida-
tion of the resulting adsorbed CO to CO2, and a “direct”
pathway, which proceeds directly, without COad formation, to
CO2.[3,8,9] Recently, adsorbed formate was suggested as a
stable intermediate.[10] Our data demonstrate that this picture
is incomplete and that the reaction involves at least three
different parallel reaction pathways. The relative contribu-
tions of the three pathways under current reaction conditions
are estimated from the electrochemical and IR spectroscopic
data.
the current electrolyte was changed back to pure base
electrolyte (adsorption/oxidation potential 0.4, 0.5, and 0.6).
The nature of the adsorbed species is evident from the
sequence of IR spectra in Figure 1, which were taken during
This study demonstrates the general potential of in situ
ATR-FTIRS studies under continuous flow conditions with
well-defined mass transport to/from the electrode, which is
particularly important for reaction studies involving dissolved
gases or other dilute reactants. Among others, this technique
allows:
Figure 1. Selected IR spectra recorded during adsorption/oxidation of
formic acid on a Pt thin film electrode at 0.5 V. Electrolyte exchange at
t=0 s from 0.5m H2SO4 to 0.1m HCOOH + 0.5m H2SO4 and back
after ꢀ4.5 min (270 s). Electrolyte flow rate: 50 mLsÀ1, cell volume:
10–20 mL.
1. transient, time-resolved spectroelectrochemical measure-
ments upon sudden exchange of the electrolyte under
potential control;
2. spectroelectrochemical adsorption/stripping measure-
ments of nonvolatile species;
these transient experiments at a constant potential of 0.5 V.
Characteristic features of these spectra are three positive
bands, one with a signal at n˜ ꢀ 1322 cmÀ1, which was
attributed
to
bridge-bonded
adsorbed
formate
3. modification of electrode surfaces by chemical or electro-
chemical adsorption/deposition in the flow cell prior to the
actual measurements, which opens the way for a direct
comparison of different surfaces.
(HCOOad),[10,13] and two peaks centered at n˜ ꢀ 2010 and
1790 cmÀ1, which are related to linearly bonded (COL) and
multiply bonded (COM) adsorbed CO (COad). In addition, the
spectra exhibit a two negative bands in the region n˜ = 1000–
1250 cmÀ1, which correspond to displaced bisulfate and
sulfate.[2] The presence of significant amounts of other
adsorbates can be excluded based on the IR data. Spectra
taken at 0.4 and 0.6 V exhibit similar characteristics as those
above.
More quantitative information on the build-up of the
adsorbate layer and on the relationship between adsorbate
coverage and the Faraday reaction current is derived from the
current densities in the chronoamperometric transients (Fig-
ure 2a) and from the integrated intensities of the IR signals as
a function of time at the different adsorption potentials
(Figures 2b–d). The current densities increase steeply when
switching from the supporting electrolyte to HCOOH-con-
taining solution (t = 0), reaching a maximum after approx-
imately 3 s, and then decrease slowly with time under a
continuous flow of HCOOH-containing electrolyte. When
changing back to pure supporting electrolyte after about
4.5 min, they instantaneously drop to 0.
Thus, this technique is of general relevance for kinetic and
mechanistic spectroelectrochemical studies, by far exceeding
applications in electrocatalysis.
The thin-layer spectroelectrochemical flow cell used in
this study is analogous to that of a dual thin-layer flow cell
design introduced recently,[11,12] which was modified for the
present measurements. It consists of a circular Kel-F plate
with openings for inlet and outlet capillaries, counter- and
reference electrodes, and a hemicylindrical silicon prism,
which is pressed against the Kel-F plate through a circular
gasket and a Cu foil current collector (Kel-F = poly(chloro-
trifluoroethylene)). A thin Pt film deposited on the flat
reflecting side of the Si prism serves as the working
electrode.[5] The electrolyte flow through the cell can be
switched between different electrolytes.
The interaction of formic acid with Pt electrodes at
potentials below oxide formation leads to the build-up of an
adsorbate layer, whose chemical nature/composition depends
critically on the electrode potential.[9] The accumulation of
the adsorbed species was monitored in potentiostatic, tran-
sient experiments under continuous electrolyte flow, changing
from pure base electrolyte (0.5m H2SO4) to formic acid
containing electrolyte (0.5m H2SO4 + 0.1m HCOOH, t = 0–
4.5 min); after reaching steady-state conditions (t = 4.5 min),
In contrast to the steep increase of the current densities,
the intensities of the signals for adsorbed COL (Figure 2b)
and COM (Figure 2c) develop slowly upon changing to 0.1m
HCOOH containing solution over the entire adsorption time
(4.5 min). Both the final COad coverage and the initial rate for
COad formation are highest for adsorption at 0.4 V, lower at
0.5 V, and very low at 0.6 V. When the electrolyte is switched
982
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 981 –985

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Angewandte
Chemie
bisulfate and sulfate intensity. Only at 0.4 V does the bisulfate
and sulfate intensity remain virtually constant, despite the
abrupt decay in formate intensity. This is tentatively
explained by a combination of two effects: We expect that
formate is randomly distributed in the CO adlayer and that at
the high COad coverage a relaxation of the CO adlayer is
energetically more feasible than adsorption of the large and
weakly adsorbed sulfate anions.
The above results clearly show that oxidation of formic
acid is accompanied by the formation of adsorbed formate, in
agreement with findings in a previous study,[10] and that under
the present reaction conditions the build-up of these formate
species is much faster than the formation of COad. Hence, the
rate of formic acid dehydrogenation/HCOOad formation (in
this term we also include contributions from adsorption of
formate anions from the electrolyte) is much higher than that
for formic acid dehydration (COad formation) under the
present reaction conditions.
The reaction rates for the dehydration of formic acid to
COad and subsequent oxidation of COad can be determined
quantitatively from the IR data by using the relationship
between the COad coverage and the IR intensities for the COM
species at low COad coverages and the COL-related intensity
in the medium COad-coverage regime, which were determined
in calibration measurements (see Supporting Information).
The COad oxidation rate under steady-state conditions is
determined from the initial slope of the COad intensity–time
curve, right after changing back from a HCOOH-containing
solution (1m) to HCOOH-free electrolyte (Figure 2b),
assuming that the COL/COM population ratios do not vary
significantly between 0.4 and 0.6 V.[14] For clarity, we magni-
fied the intensities of the COL species at 0.5 V (squares) by a
factor of 3 and at 0.6 V (circles) by a factor of 10. We used a
saturation coverage of 0.7 monolayers (ML) toe calculate
turnover frequencies (TOFs) of 6.4 ” 10À5 and 3.3 ”
10À4 molecules per Pt site per second for COad oxidation to
CO2 at 0.5 and 0.6 V, respectively, under the present steady-
state reaction conditions. At 0.4 V the rate for COad oxidation
is below our detection limit (Table 1). The increase in COad
oxidation rate with potential agrees with the common
explanation of enhanced OHad formation at higher potentials
(“electrochemical activation”).
Figure 2. Chronoamperometric transients (a) and evolution of the IR
peak intensities related to adsorbed COL (b) and COM species (c) as
well as to adsorbed formate (d) and bisulfate and sulphate (e) during
~
the interaction of formic acid with the Pt thin film electrode at 0.4 ( ),
*
0.5 (&) and 0.6 V ( ) (experimental conditions see Figure 1).
back to HCOOH-free solution, the intensities of the COad
signals (Figure 2b) decrease only slowly, as expected from the
slow kinetics of COad electrooxidation at potentials of 0.6 V
and below. The HCOOad-related intensities (Figure 2d) show
an abrupt increase when switching to HCOOH-containing
solution and reach an approximately constant value approx-
imately 3 s after the electrolyte exchange. For adsorption at
0.4 V we find a slight subsequent decay of the formate band
intensity with time, probably owing to formate replacement
by COad. Formate formation is accompanied by a decrease in
the bisulfate and sulfate band intensity (Figure 2e), which
approximately mirrors the temporal evolution of the formate
intensity (Figure 2d), implying that sulfate/hydrogen sulfate
is displaced by more strongly adsorbed formate species.
Similar to the current density, the formate intensity drops
instantaneously to 0 when changing back to pure supporting
electrolyte, accompanied by a similarly fast increase in the
In the same way we can determine the initial rates for
COad formation (HCOOH dehydration) on the initial, COad-
and HCOOad-free Pt surface from the initial slopes of the
COad intensity profiles, after switching from base electrolyte
Table 1: Turnover frequencies (TOF).[a]
Reaction
TOF [moleculessÀ1 siteÀ1
]
0.4 V
0.5 V
0.6 V
HCOOH dehydration[b]
COad oxidation[c]
2.0”10À2
1.8”10À3
6.4”10À5
1.4
3.2”10À4
3.3”10À4
2.5
ꢀ0
total HCOOH oxidation[d]
0.3
[a] TOF for [b] formic acid dehydration to COad on a clean Pt surface (in
the limit of very small COad and OHad coverages), [c] COad oxidation to
CO2 (steady state), [d] total oxidation of formic acid to CO2 under steady
state conditions, determined after 4.5 min under continuous flow of
0.1m HCOOH + 0.5m H2SO4 at different reaction potentials.
Angew. Chem. Int. Ed. 2006, 45, 981 –985
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
983

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communications
to HCOOH-containing solution. This yields values of 2.0 ”
10À2, 1.8 ” 10À3, and 3.2 ” 10À4 molecules per Pt site per second
for turnover frequencies of 0.4, 0.5, and 0.6 V, respectively
(Table 1).
In the later stages of the adsorption process the increase in
COad coverage represents the difference between COad
formation (formic acid dehydration) and COad oxidation;
under steady-state conditions the two rates are equal. At
0.6 V the rates for COad oxidation (on a partly COad-covered
Pt surface) and formic acid dehydration (on a bare Pt surface)
are of comparable magnitude, which explains the rather low
steady-state COad coverage detected at this potential. At
0.4 V the COad oxidation rate is negligible, resulting in the
observed pronounced accumulation of COad with time up to
saturation. The absolute values of the initial dehydration rates
as well as the fact that at 0.4 V the initial rate for formic acid
dehydration is about (more than) one order of magnitude
higher than at 0.5 V (0.6 V) agree well with results derived by
Sun from electrochemical transients.[3]
The relative contribution of the indirect reaction pathway
to the total rate of oxidation of formic acid to CO2 under the
present steady-state conditions can be calculated from the
ratio of the rate for oxidation of COad determined above and
the Faraday current under these conditions by using the
values obtained from the chronoamperometric transients
(Figure 2a) after oxidation of formic acid for 4.5 min
(Table 1). Comparing the TOFs of 0.28, 1.4, and 2.5 molecules
per Pt site per second for the oxidation of formic acid to CO2
at potentials of 0.4, 0.5, and 0.6 V, respectively, and the above
turnover frequencies for COad oxidation at these potentials,
we conclude that the indirect pathway, through dehydration
of a HCOOHad precursor to COad and its further oxidation to
CO2, contributes less than 0.1% to the total rate of the
oxidation of formic acid at 0.5 Vand less than 0.01% at 0.6 V.
Interestingly, the ratio of the formate band intensities at
the end of the HCOOH oxidation period at different
potentials (Figure 2d) differs significantly from that of the
Faraday currents (Figure 2a): 90% (0.5 V) and 50% (0.4 V)
of the formate intensity at 0.6 V result in Faradaic currents of
55% (0.5 V) and ꢀ 10% (0.4 V) of the value at 0.6 V,
respectively. Since the dehydration pathway does not con-
tribute significantly under these conditions, this observation
requires either a potential-dependent variation (electrochem-
ical activation) of both adsorbed formate formation and its
oxidation to CO2, or the existence of another, third reaction
pathway. For more information on this point we performed
similar transient oxidation measurements as in Figure 2 at the
same potential (0.6 V), using two different formic acid
concentrations.
Figure 3. Chronoamperometric transients (a, b) and IR spectra of
adsorbed formate (c, d) on a Pt thin film electrode at 0.6 V in a
solution containing 70 mm (a, c) and 0.7 mm HCOOH (b, d). The
spectra were acquired 100 s after changing to HCOOH-containing
solution.
contribution of the indirect pathway under these conditions
and assuming that the rate of formate decomposition is
related linearly to formate coverage, these data clearly
indicate that there must be a third reaction pathway.
Its contribution to the total reaction rate cannot be
calculated precisely from these data, but we can estimate a
higher limit for its contribution. Assuming that for 0.7 mm
HCOOH solution the Faraday current results completely
from the formate pathway, the fivefold-higher formate-band
intensity in 70 mm HCOOH solution should at most result in a
fivefold-higher current density. The experimentally observed
increase by a factor of 20 is only possible if the reaction
pathway via the adsorbed formate species detected by IR and
dominating the formate IR signal (“formate pathway”)
contributes less than 25% to the total anodic current.
Although our IR spectra do not give any indication of
another adsorbed species, this third pathway is dominant for
room-temperature oxidation of a 0.1m formic acid solution.
This can either be explained by an additional, thus far not
detected adsorbed reaction intermediate, whose concentra-
tion is very low (low lifetime under reaction conditions), or by
a direct reaction of weakly adsorbed HCOOHad species. (The
theoretical possibility of a flat-lying formate species with a
very low IR cross-section is ruled out as considerable
substrate–adsorbate bonding would be necessary for a
stable reaction intermediate.)
We favor the latter explanation and propose a triple path
mechanism; which starts with a weakly adsorbed HCOOHad
precursor that can subsequently either be directly oxidized to
CO2 (direct pathway), undergo dehydration to COad (“indi-
rect pathway”), or is dehydrogenated to stable bridge-bonded
adsorbed formates (formate pathway), as it is schematically
depicted in Figure 4. The stable, adsorbed intermediates
resulting in the last two pathways can then, in a third step, be
oxidized to CO2. In this scheme adsorbed formates are indeed
reaction intermediates, but, in contrast to recent concepts,[13]
are not in the dominant reaction pathway.
Representative results are shown in Figure 3. Comparing
the Faraday currents for the oxidation of 0.7 mm and 70 mm
HCOOH solution at 0.6 V, we find current densities of
0.045 mAcmÀ2 and 0.85 mAcmÀ2, respectively, after 100 s,
that is, roughly a factor of 20 between the two electrolytes.
The IR spectra acquired at t = 100 s, in contrast, show that the
respective formate intensities differ only by about a factor of
5. Hence, even at similar reaction potential the total steady-
state rates for HCOOH oxidation are not at all proportional
to the respective formate coverages. Based on the negligible
In summary, we have 1) significantly improved the
capability of in situ ATR-FTIRS measurements by combining
this technique with a thin-layer flow cell, which allows
mechanistic and quantitative spectroelectrochemical kinetic
studies of electrocatalytic reactions under well-defined mass-
984
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 981 –985