Journal of Electroanalytical Chemistry 650 (2011) 233–240

Contents lists available at ScienceDirect

Journal of Electroanalytical Chemistry

journal homepage: www.elsevier.com/locate/jelechem

A method for kinetic study of methanol oxidation at Pt electrodes

by electrochemical in situ infrared spectroscopy
Ling Wen Liao a, Shao Xiong Liu a, Qian Tao a, Bin Geng a, Pu Zhang a,
Cheng Ming Wang a, Yan Xia Chen a,⇑, Shen Ye b
a
Hefei National Laboratory for Physical Sciences at Microscale, Department of Chemical Physics, University of Science and Technology of China, Hefei 230026, China
b
Catalysis Research Center, Hokkaido University, Sapporo 001-0021, Japan

a r t i c l e i n f o a b s t r a c t

Article history: In this contribution, we describe a method to estimate the kinetics of the indirect pathway in methanol
Received 25 February 2010 oxidation reaction (MOR) at Pt electrode by using electrochemical in situ infrared spectroscopy combined
Received in revised form 3 September 2010 with a thin-layer flow cell. Based on a quantitative relationship between the coverage of adsorbed carbon
Accepted 9 September 2010
monoxide (COad) and its infrared band intensities determined experimentally, COad coverages during
Available online 17 September 2010
MOR are estimated from the simultaneously recorded time-resolved IR measurements. By assuming that
COad oxidation goes through Langmuir–Hinshelwood mechanism, its rate constant is estimated in pure
Keywords:
supporting electrolyte under otherwise identical condition. Based on the result, the rates of COad forma-
Methanol oxidation
Platinum electrode
tion and COad oxidation during MOR are deduced from the COad coverage–time curve. The methodology is
Flow-cell exemplified with MOR on Pt electrode at +0.6 V (versus RHE) in 0.1 M HClO4 with 2 M methanol. It is
Infrared spectroscopy found that under forced-flow condition: (i) the maximum reaction rate for COad oxidation is ca. 0.004
Attenuated total reflection configuration molecule site1 s1 which is 100 times smaller than the maximum rate for COad formation from methanol
Kinetic study dehydrogenation; (ii) with increase in COad coverage from zero to 0.5 ML, the current efficiency of the
indirect pathway for MOR increases and reaches ca. 17% under steady state. The general applicability
of such a method is shortly discussed.
Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction amount of HCHO produced from MOR can also be estimated

Extensive studies on the electro-oxidation of methanol have
been carried out in the past decades, since methanol is believed
to be one of the promising fuels for low temperature fuel cells
[1–19]. Pt is found to be the best monometallic catalyst for meth-
anol oxidation reaction (MOR) [1,3,20,21], and has been the most
frequently used model electrocatalyst for the mechanistic and
kinetic studies on MOR. At Pt electrode, methanol can be either
completely oxidized to CO2 or be oxidized incompletely to side-
products such as HCHO, HCOOH and COad, which are shown in a
simplified reaction scheme in Fig. 1 [5–7,14,15,17]. For the com-
plete oxidation of methanol to CO2, a dual path mechanism with
the direct pathway going through non-COad adsorbates, while the
indirect pathway proceeding via the formation of COad and its sub-
sequent oxidation, is proposed [5–7]. By monitoring the mass sig-
nal of CO2 and methylformate using differential electrochemical
mass spectrometry (DEMS), the production rate of CO2 and HCOOH
can be evaluated [14,15,22]. By subtracting the equivalent currents
for CO2 and HCOOH production from the total MOR current, the
⇑ Corresponding author. Tel./fax: +86 551 3600035.
E-mail address: yachen@ustc.edu.cn (Y.X. Chen).
(Fig. 1). For the complete oxidation of methanol to CO2, however,
no quantitative information on the contribution of the direct and
indirect pathway is available so far.
In a recent study by Chen et al. on HCOOH oxidation at Pt sur-
face using the electrochemical in situ infrared spectroscopy under
attenuated total reflection configuration (EC-ATR-FTIRS) coupled
with a thin-layer flow cell, the rate of COad formation under condi-
tions where COad oxidation rate is negligibly small (i.e., hCO  0)
and the rates of both COad formation and COad oxidation under
steady state where both rates equal each other are deduced [23],
based on the time-resolved IR spectra recorded during potentio-
static HCOOH oxidation reaction and a linear relationship between
COad surface coverage and its IR band intensity [23–26]. The cou-
pling of the EC-ATR-FTIRS with the flow cell largely eliminates
the influence from the ill-defined mass transport as commonly
encountered in stationary cells and from the re-adsorption and
the subsequent reaction of side products (such as HCHO and
HCOOH in MOR), this enables the kinetic analysis for complicated
reactions such as MOR. However, procedure on how to carry out
such quantitative analysis, especially under conditions when
both COad formation and COad oxidation are active but with
different reaction rates, has not been provided. As a consequence,

1572-6657/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jelechem.2010.09.012
234 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
Fig. 1. A simplified reaction scheme for methanol oxidation at Pt electrode: (1) the
pathway for COad formation from methanol dehydrogenation; (2) the pathway for
COad oxidation to CO2 through the indirect pathway; (3) the pathway leads to
soluble HCHO production, (4) the pathway leads to soluble HCOOH production and
(5) the formate pathway.
quantitative information about the effects of COad coverage on the
kinetics of the formation and oxidation of COad during the electro-
catalytic oxidation of C1 molecules is not available so far.
In this paper, we describe in detail the procedure on how to
derive the instantaneous rates of COad formation and COad oxida-
tion as a function of reaction time (with any COad coverage) during
MOR by using EC-ATR-FTIRS, from which the current efficiency of
the indirect pathway in MOR to CO2 can be estimated. The paper
is organized as following: firstly, the control experiments and the
procedure for deriving the relationship of the COad surface cover-
age with the IR band intensities of COad species are described.
Then, time-resolved infrared spectra and current transient
recorded during MOR on Pt film electrode at +0.6 V (versus RHE)
are presented, this potential was chosen since all possible path-
ways involved in MOR operate. Afterwards, the process on how
to carry out kinetic analysis on MOR is exemplified by taking the
time-resolved IR spectral data and chronoamperometric transients
simultaneously recorded during MOR on Pt at +0.6 V. The mecha-
nism/kinetics of MOR and general applicability of such a method
are briefly discussed.
2. Experimental
The configuration of the thin-layer spectro-electrochemical
flow-cell used in this study has been described in detail in Ref.
[23]. The cell volume is ca. 15 ll. In order to minimize the effect
of mass transport on the kinetic analysis, in the present study
2 M methanol was used and the flow rate of electrolyte was ca.
250 ll/s. This means that the electrolyte in the cell will be renewed
every 60 ms, one can almost ignore the re-adsorption and oxida-
tion of HCHO and HCOOH products [23,27,28]. A Pt thin film
(thickness of 50 nm) chemically deposited on the flat face of a
hemi-cylindrical Si prism [29] was used as working electrode
(WE). The active surface area of the WE is ca. 4.8 cm2, as deter-
mined by the charge associated with oxidation of a monolayer of
under-potential deposited H(upd-H) atoms (calculated from the
integrated charge in the CV from 0.05 to 0.45 V) [30,31]. A revers-
ible hydrogen electrode (RHE) was used as reference electrode (all
the potentials in this paper are quoted against RHE) and the elec-
trode potential was controlled by a potentiostat (PAR 273 A, Ame-
tek, USA). Millipore Milli-Q water, ultrapure perchloric acid
(Suprapure, Sigma–Aldrich), methanol (for analysis, Fluka) were
used to prepare the solutions. Supporting electrolyte used in all
measurements in this study was 0.1 M HClO4 and constantly
purged by N2 (N2, 99.999%, Nanjing Special Gas Corp.). Before the
measurements, continuous potential cycles in the potential region
from +0.05 to +1.3 V at a scan rate of 0.05 V/s in 0.1 M HClO4 were
carried out to clean Pt thin film electrode until its cyclic voltammo-
gram (CV) showed the characteristic feature of polycrystalline Pt
electrode (Fig. 2, dotted line). All the experiments were carried
out at room temperature (ca. 18 ± 2 °C).
Fig. 2. Cyclic voltammograms of CO stripping at Pt electrode— in 0.1 M HClO4 with
various hCO (as denoted in the figure) formed either from the decomposition of
0.05 M CH3OH in 0.1 M HClO4 at 0.4 V or from direct adsorption of CO from CO
saturated 0.1 M HClO4 solution for different time duration, scan rate: 20 mV/s. The
base CV is also shown as dotted line.
The quantitative relationship between the surface coverage and
the IR band intensity of COad in the low coverage regime (hCO < 0.5
ML) is derived according to the following procedure. Firstly the
electrode potential was held at +0.4 V in 0.1 M HClO4, then 0.1 M
HClO4 + 0.05 M CH3OH was flowed through the cell and the IR
spectra were recorded simultaneously. The reasons for using
0.05 M CH3OH solution instead of CO saturated solution are: (i) di-
rect adsorption of CO from bulk solution saturated with CO is too
fast to accurately control of its surface coverage; (ii) and it is well
confirmed that at E < 0.65 V, CO is the only stable adsorbate formed
from methanol dehydrogenation at Pt surface [13,32], thus 0.05 M
MeOH instead of CO was used to make sure that the rate of COad
formation from methanol dehydrogenation is relatively slow, so
that COad coverage can be easily controlled by switching the valve
between the two electrolyte supplying bottles with and without
methanol. After COad with a desired coverage is formed from meth-
anol dehydrogenation for a certain holding time at the potential
(+0.4 V), the electrode potential was stepped to 0.05 V and the
electrolyte was immediately switched back to 0.1 M HClO4. And
the COad coverage was determined by the subsequent cyclic vol-
tammetric stripping after completely washing away the residual
methanol in the cell by continuously flushing the cell with 0.1 M
HClO4 at +0.05 V. The further uptake of COad at +0.05 V during
the above potential step process is negligible, since methanol
adsorption/dehydrogenation is largely hindered by the upd-H
atoms [33] and the previously formed COad at +0.4 V [13]. The
corresponding IR spectra of COad at different CO coverages are
recorded during the methanol adsorption and subsequent CO
stripping measurements. Since maximum COad surface coverage
achievable from methanol dehydrogenation is below 0.5 ML, the
data points in the hCO–ICOL plot with hCO > 0.5 ML are derived from
direct adsorption of CO from CO saturated 0.1 M HClO4 solution for
various time duration.
Potentiostatic MOR was carried out as following. Firstly, the
electrode potential was held at +0.6 V in 0.1 M HClO4, and a back-
ground infrared spectrum was recorded. After that the electrolyte
supplying bottle was switched to 0.1 M HClO4 + 2 M CH3OH solu-
tion (pre-deaerated by N2 for at least 10 min). After ca. 5 min of
MOR at +0.6 V, the electrolyte supplying bottle was switched from
methanol-containing solution back to the supporting electrolyte
again. Meanwhile the cell is immediately washed with 0.1 M HClO4
solution while the electrode potential was kept constant there
for another 5 min to follow the COad oxidation kinetics. The
 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
time-resolved infrared spectra and chronoamperometric transient
of methanol oxidation were simultaneously recorded throughout
the above mentioned processes. The in situ IR measurements were
carried out with a Varian FTS 7000e spectrometer equipped with a
mercury cadmium telluride (MCT) detector cooled by liquid nitro-
gen. All the spectra were obtained with a resolution of 4 cm1. The
spectrum integration time for the calibration measurements was
1 s and in the potentiostatic MOR measurements was 0.2 s in order
to follow the fast spectral changes in short time interval. All spec-
tra were shown in absorbance, i.e., log(R0/R), where R0 is the reflec-
tance of the background spectrum recorded at the same potential
on clean Pt electrode in the supporting electrolyte without
methanol.
3. Results and discussion
3.1. Deriving the calibration curve of COad band intensity versus COad
surface coverage
Some selected CVs for the stripping of COad formed from di-
rectly adsorption of CO from solution or after exposing to 0.05 M
CH3OH for various specified time duration when holding the elec-
trode potential at +0.4 V are shown in Fig. 2. The COad coverage is
derived according to the following equation [34–36]:
Q CO oxidation
¼ Q CO stripping
 Q Pt ð1Þ
net pzc to Eh pzc to Eh
where Q CO stripping
is the charge from pzc to the higher potential limit
pzc to Eh
(Eh) where CO oxidation current is still appreciable, Q Pt pzc to Eh is the
charge recorded in the same potential region in the 2nd scan after
stripping off the CO adlayer. From the CO displacement experiment,
the pzc of the Pt film electrode in 0.1 M HClO4 is found to be ca.
0.3 V [34]. From Fig. 2 it is found that with increase in COad cover-
age, the currents for the oxidative removal of upd-H atoms
decrease. Furthermore, the peak potential for CO stripping shifts
positively accompanying with an increase in COad oxidation current,
as in agreement with previous results for COad stripping on Pt single
crystal surfaces and carbon-supported Pt catalysts [16,32]. The cor-
responding IR spectra recorded at 0.4 V from the as formed COad
Fig. 3. IR spectra of COad on Pt electrode at 0.4 V with various coverages of COad as
formed according to the conditions given in Fig. 2.
235
adlayer are shown in Fig. 3. The bands at around 2020–2082 cm1
and 1795–1853 cm1 are assigned to COL and COM on Pt surface,
respectively. Inspection from Fig. 3, it is found that the peak fre-
quencies of both COL (2016 cm1) and COM (1795 cm1) at
hCO = 0.07 ML are quite low and the intensity of COL is smaller than
that of COM. This indicates that at such low COad coverages CO is
preferentially adsorbed in multiply-bonded configuration. Toward
higher COad coverage, the intensity of COL increases, which is
accompanied with a blue-shift in its peak frequency (e.g., from
2016 cm1 at h = 0.07 to 2082 cm1 at h = 0.7 ML). On the other
hand, the spectral changes of the COM are rather complicated: with
the increase in COad coverage, the peak height of COM band does not
show obvious change, while its band widths first increases with
coverage up to ca. 0.4 ML, and at higher COad coverage it decreases
again. The increase in the COM band width is associated with the
development of a new component at ca 1850 cm1. In contrast,
with further increase in COad coverage from 0.4 ML up to saturation,
the components below 1800 cm1 disappears, which leads to the
subsequent decrease in the band width. All these phenomena imply
that with increase in COad surface coverage, there is a clear shift of
CO adsorbate from multiply bonded sites to bridge or atop sites.
Despite of such complexities, from the data of COad coverage de-
rived from the CO stripping charge shown in Fig. 2 and IR spectra
shown in Fig. 3, an approximately linear relationship between
the COad surface coverage and IR band intensity of COL at COad cov-
erages below 0.6 ML are derived (Fig. 4), i.e.,
hCO =hCO;sat ¼ kCOL  ICOL ð2Þ
with kCOL ¼ 0:78  0:05, as was well agreement with previous cali-
bration curve using the same EC-ATR-FTIRS technique and similar
optical setup [23,26], although in the previous study CO is formed
from the decomposition HCOOH rather than from methanol used
in the present study. From our systematic EC-ATR-FTIRS study on
the adsorption of CO with submonolayer coverage from various of
C1 molecules, [37] it is found that at the same coverage the struc-
ture of COad layer is not affected by the history for COad formation:
(i) where the sites for COad formation from different precursors may
be different and (ii) whether COad with a certain coverage is formed
by directly dosing CO to clean Pt surface or by partial oxidation of a
CO adlayer with initial high coverage. These facts indicate that un-
der present conditions COad molecules diffuse rapidly at Pt surface.
In previous FTIRS studies in external reflection configuration,
similar linear relationship between the COad surface coverage
and IR band intensity of COL at COad coverages below 0.6 ML
1.0
0.8
θ CO / θ CO,sat
0.6
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
COL band intensity / a.u.
Fig. 4. Plot of IR band intensity of COL as a function of COad surface coverage at
0.4 V, the corresponding data points are obtained from the raw data partly shown in
Figs. 2 and 3.
236 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
have also been confirmed for CO adsorption at Pt electrode by
Kunimatsu [25] and Weaver and co-workers [24]. Furthermore,
such a linear relationship is also confirmed at 0.6 V as demon-
strated by a recent study using combined ATR-FTIRS and DEMS,
where the coverage is determined from DEMS measurement
during the oxidative removal of COad (Fig. 4 in Ref. [26]).
Furthermore, it is found that with nearly fixed COad surface cov-
erage, the change of COL band intensity at potentials between
+0.35 and +0.6 V is negligibly small [23,26,38], as can be seen
clearly from two representative sets of potential dependent IR
spectra with COad coverage of 0.08 and 0.16 ML Fig. 5. [39]. This
is well explained by the fact that the adsorption of the anions

and water molecules from the supporting electrolyte (e.g., ClO4
2
or SO4 ) at Pt surface are much weaker than that of CO. As a con-
sequence, the slopes of hCO–ICOL plot do not show obvious change
in the potential region from +0.35 to +0.6 V. Thus, by exploiting
the linear relationship between COad band intensity and COad sur-
face coverage measured at +0.4 V, hCO(t) at other potentials in the
potential region from +0.35 to +0.6 V can be deduced from COL
band intensity–time curve. It should be noticed that such an cali-
bration curve is not applicable at E 6 +0.3 V, since CO islands is
formed at low COad coverages when co-adsorption of CO and H
takes place [37]. In the following section, we will describe the
time-resolved spectral-electrochemical data recorded during
MOR at +0.6 V.
3.2. Electrochemical in situ infrared study on MOR at +0.6 V
Selected IR spectra recorded during MOR on Pt electrode at
+0.6 V in the first 50 s after switching from 0.1 M HClO4 to 0.1 M
HClO4 + 2 M CH3OH (from t = 0) and within first 50 s after switch-
ing back to 0.1 M HClO4 (from t = 300) are given in Fig. 6a and b,
respectively. The corresponding COL band intensity and the current
transient are plotted as a function of time (Fig. 7). From Figs. 6 and
7, it is seen that immediately after exposing the Pt film electrode to
methanol-containing solution, two bands from CO adsorbed at
atop (COL), bridge and hollow sites (the latter two are denoted as
a b
1812
2019
θ CO=0.08 ML
0.003
0.3 V
0.35 V
0.4 V
0.45 V
0.5 V
2050
0.55 V
1826
0.6 V
2100 2000 1900 1800
Wavenumber/cm-1
Fig. 5. IR spectra of COad on Pt electrode at various electrode potentials recorded during the votammetric stripping of the COad layer in 0.5 M H2SO4 at a scan rate of 10 mV/s:
(a) 0.08 ML and (b) 0.16 ML.
COM due to the superimposition of these two modes) appear. The
band intensities increase sharply in the first 3 s after the solution
exchange, which is followed with a slower increase from 3 to
10 s until the maximum intensity is reached. Then the COL band
intensity decays slowly with time. When switching back to meth-
anol free solution after MOR for ca. 300 s, a gradual decay of COL
band intensity with time is seen. Furthermore, it is found that
the COM band gives similar time profiles, except that the band
intensity is much smaller and the time for reaching maximum
band intensity is slightly shorter than that of COL.
On the other hand, from the simultaneously recorded chrono-
amperometric transient (Fig. 7a) it is seen that the anodic MOR
current appears as soon as the methanol-containing solution is
introduced. And it increases with time and reaches maximum at
ca. 0.5 s after the solution switch. Then it drops sharply within next
8 s, which is followed by a much slower decay at longer reaction
time. The initial steep increase in the current after the solution
switch is preliminary attributed to the increase in methanol con-
centration at the electrode surface (it should be noticed that
although it takes less than 0.1 s to fill the cell with methanol-
containing solution, however the electrolyte composition at the
electrode surface is still controlled by diffusion as well as the
change in MOR mechanism the latter is discussed in detail in
[15] through the diffusion layer). The abrupt drop in the anodic
current correlates with the sharp increase in COL band intensity,
indicating that the MOR activity at the Pt surface is strongly
reduced due to the ‘‘poisoning” effects from COad formed at this
potential. From 10 to 300 s, the Faradic current displays a steady
slow decay which is accompanied with a small decrease in COL
band intensity. This is explained by the redistribution of COad
occurring simultaneously with COad formation and COad oxidation
processes, which leads to more uniformly domain structure of CO
adsorbates. As a consequence, the dipole–dipole coupling interac-
tion among the neighboring CO oscillators decreases slightly and
the intensity transfer from COM to COL drops slightly. On the other
hand, the extended distribution of COad at the Pt surface also
reduces the numbers of free ensembles necessary for methanol
2036 θ CO=0.16 ML
0.004 1845 1815
0.3 V
0.35 V
2060
0.4 V
0.45 V
0.5 V
0.55 V
1827
0.6 V
1700 2100 2000 1900 1800 1700
Wavenumber/cm-1
 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
Fig. 6. The time-resolved IR bands of COad produced during methanol oxidation on
Pt film electrode at +0.6 V: (a) in the first 50 s after switching from 0.1 M HClO4 to
0.1 M HClO4 + 2 M MeOH and (b) in the first 50 s after switching back to 0.1 M
HClO4.
Fig. 7. (a) Chronoamperometric transient, (b) the time-resolved IR band intensities
of COL (circle) and time-dependent COad coverages (square) during methanol
oxidation on Pt film electrode at +0.6 V. The electrolyte was switched from 0.1 M
HClO4 to 2 M MeOH at 0 s and switch back to 0.1 M HClO4 at ca. 300 s.
adsorption. Immediately after switching back to methanol free
0.1 M HClO4 solution at the same potential (+0.6 V), COad band
intensities displays a slightly faster decay with time (Fig. 6b), this
indicates that at +0.6 V COad is oxidized at a certain rate in
237
methanol free solution. From this, it is inferred that at the interface
of Pt electrode/0.1 M HClO4 + 2 M MeOH, sequential COad produc-
tion and COad oxidation take place at +0.6 V, although so far it is
not clear whether the competitive adsorption/decomposition of
methanol affect the kinetics of COad oxidation or not.
The increase in COad coverage within first 10 s after solution
exchange suggests that COad formation rate is higher than that of
COad oxidation at clean Pt surface or at Pt surface with low COad
surface coverage, and the ratio between these two rates decreases
with further uptake of CO at the surface until the balance is
reached. The steep drop of the faradaic current after switching back
to CO free solution is explained by the depletion of methanol in the
cell, where the pathways of COad, HCHO and HCOOH formation
from methanol dehydrogenation are switched off. And the slower
decay after the steep current drop of the anodic current with time
(after ca. 308 s, Fig. 6a) is attributed to the slow oxidation of the
remained COad species at the Pt electrode surface, as confirmed
by the slow decay in the IR band intensities of COad species (see be-
low for detailed kinetic discussion).
3.3. Kinetic analysis of methanol oxidation on Pt film electrode
at +0.6 V
According to the simplified reaction scheme as shown in Fig. 1,
the total current density for methanol oxidation can be expressed
as below:
j ¼ jd þ jdec þ jox ð3Þ
where jd is defined as the sum of the current density for all path-
ways other than the indirect pathway, which includes the current
density from complete oxidation of methanol to CO2 in the direct
pathway (through non HCHO and HCOOH intermediates), those
from the incomplete oxidation of methanol to HCHO and HCOOH
and the CO2 production from the readsorbed HCHO and HCOOH.
One can imagine that re-adsorption and oxidation of HCHO and
HCOOH is largely eliminated under present condition with the
thin-layer flow cell (volume: 15 ll) and the fast flow rate
(250 ll/s). jdec and jox are the equivalent current densities for the
dehydrogenation of methanol to COad and the oxidation reaction
of COad to CO2, respectively.
The rate of methanol dehydrogenation to COad (denoted as
rdec(t)) can be expressed by [2,4]
 
bnFE
rdec ðtÞ ¼ kdec  exp ½chCO ðtÞ½CH3 OHd  exp ð4Þ
RT
where kdec is the rate constant for methanol dehydrogenation, c is a
coefficient which describes how the surface coverage of COad affects
methanol decomposition and d is the reaction order. And
exp ðbnFE=RTÞ describes the potential dependence of the rates
according to the Butler–Volmer equation. In principle, with the
information of the effects of COad coverage, methanol concentration
and electrode potential on the rates of methanol dehydrogenation,
the rate (rdec(t)) and the rate constant (kdec) can be derived. In the
present work, although without comprehensive information as
listed above, the rates of methanol dehydrogenation to COad at
any time during potentiostatic MOR can still be directly evaluated
from the time-resolved IR results according to the following
equation:
dhCO ðtÞ
¼ r dec ðtÞ  r ox ðtÞ ð5Þ
dt
where dhCOdt
ðtÞ
is the derivative of COad coverage–time plot at any time
(Fig. 7b, square), and rox(t) is the rate for COad oxidation. The equa-
tion describes that the net rate for COad build-up at Pt surface
equals the difference in the rate for COad production and that for
COad oxidation during MOR. Since the net build-up of COad can be
238 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240

derived from the IR band intensity–time plot, if rox(t) can be deter- 0.4 0.4
r dec

-1
mined experimentally, rdec(t) can be easily calculated. a

TOF / molecule site s

-1
In order to derive rox(t) during MOR, we will first calculate the r ox
0.3
rate constant for COad oxidation (kox) from the IR data. Assuming
that oxidation reaction of COad follows Langmuir–Hinshelwood 0.2
mechanism as proposed previously: [40] 0.2 X5

H2 O ! OHad þ Hþ þ e ð6Þ 0.1

0.0
þ 
COad þ OHad ! CO2 þ H þ e ð7Þ 0 3 6 9
0.0
The oxidation rate of COad at any fixed potential can be ex- 0.25
j
0.25
pressed as: tot
b 0.20
0.20 j
1 0.15
r oxðtÞ ¼ kox  hCO  ð1  hCO Þðmolecules site s1 Þ ð8Þ dec

-2
j / mA cm
0.15 0.10 j ind
It should be noticed that under present condition for MOR, (1  hCO) 0.05
0.10
can still be roughly taken as the free sites for OHad adsorption, since
0.00
the life time for other reactive intermediates formed during MOR is 0.05 0 3 6 9
short (under dynamic transfer) and whose surface coverage is small.
In the simplest case where the methanol solution is completely 0.00
replaced by pure electrolyte solution where COad formation from c

methanol decomposition is switched off, rox(t) can also be simplified
as:

0.10
dhCO d ICOL  kCOL
r oxðtÞ ¼  ¼
dt dt
where ICOL and kCOL are IR intensities for COad and calibration coef-
ficient given in Eq. (2), respectively. With the COad coverage–time
curve derived from time-resolved IR spectra recorded right after
changing back from a methanol-containing solution to the support-
ing electrolyte and Eqs. (8) and (9), kox for COad oxidation at +0.6 V
on Pt electrode is estimated to be ca. 0.015 s1. With kox, the rate of
COad oxidation at any COad coverage during MOR can be quantified
using Eq. (8). On the other hand, the rates for methanol dehydroge-
nation to COad (rdec(t)) can be derived by substituting rox(t) and
dhCO(t)/dt into Eq. (5). Both rates are given in Fig. 8a as a function
of reaction time [41].
From Fig. 8a it is seen that at the Pt film electrode used in the
present study, rdec(t) increases with reaction time up to ca. 0.5 s
after the solution switch then it decreases exponentially with time
until a steady-state value is reached at ca. 8 s after solution switch.
The initial increase in rdec(t) is probably due to the increase in
methanol concentration at Pt surface during the initial solution
switch as similar to the case with the total faradic current
(Fig. 7a). The following decay of rdec(t) is attributed to the poisoning
of Pt surface by COad species. On the other hand, the rox(t) increases
with COad coverage until the maximum COad coverage is reached.
The initial increase in rox(t) with COad coverage indicates that the
supplying of COad to the neighboring sites where OHad is adsorbed
may be the rate limiting step for COad oxidation process under
present reaction conditions. Furthermore, the maximum rate for
methanol dehydrogenation to COad at +0.6 V is ca. 0.4 molecule
site1 s1 at Pt surface with very low CO coverage, while the max-
imum rate for COad oxidation is observed at steady state where CO
coverage reaches Ca 0.5 ML with value of ca. 0.004 molecule site1
s1. These numbers demonstrated that at clean Pt surface rox(t) is
more than. 100 times smaller than rdec(t). The measured small COad
oxidation rate can be explained by its slow oxidation kinetics
(most probably the combination of COad and OHad is slow). At long-
er reaction time, COad formation from methanol dehydrogenation
is greatly slowed down by the accumulated COad at the surface.
As a consequence, the difference between rates for COad oxidation
and COad formation from methanol dehydrogenation becomes
smaller with time, until finally the same rates for both processes
and the steady state are reached.
Current Efficiency
0.15
r ind
ð9Þ
0.05
0.00
0 50 100 150 200 250 300
Time / s
Fig. 8. (a) rates for methanol dehydrogenation (rdec(t), square) and COad oxidation
(rox(t), triangle), (b) the current densities for complete oxidation of methanol to CO2
through the indirect pathway (jind, triangle), for methanol dehydrogenation to COad
(jdec, square) and the total MOR reaction (jtot, diamond) and (c) the current
efficiency for the indirect pathway as a function of time during methanol oxidation
at +0.6 V in 2 M CH3OH + 0.1 M HClO4. The insets show the magnification of the
data within first 10 s during MOR.
From rox(t), the equivalent current densities for COad oxidation at
any time (jox(t)) can be calculated according to the following equa-
tion (Fig. 8b):
jox ðtÞ ¼ 2  eNPt kox hCO ðtÞ½1  hCO ðtÞ ð10Þ
where e is the charge for elementary electron, 2e represents that
two electrons are involved in the oxidation of one COad molecule
to CO2, and NPt is number of surface Pt atoms per cm2, for present
calculation it is assumed that 1 cm2 of Pt film (real surface area)
contains 1.35  1015 Pt atoms. Since six electrons are involved in
complete oxidation of methanol to CO2 which is three times more
than that for COad oxidation, the equivalent faradic current densities
involved in the indirect pathway, which leads to the complete oxi-
dation of methanol to CO2, can be expressed as:
jindirect ¼ 3  jox ð11Þ
The equivalent current densities for the indirect pathway (Eq. (11))
derived from the in situ IR measurements as well as the measured
total faradic current density are displayed Fig. 8b, from which it is
clearly seen that the equivalent current density for the indirect
pathway increases with time (or COad surface coverage) within first
10 s until it reaches a steady-state value of ca. 17% of the total fara-
dic current density of MOR at Pt electrode (Fig. 8c).
On the other hand, taking that four electrons are involved in
methanol dehydrogenation to COad:
CH3 OH ! COad þ 4Hþ þ 4e ð12Þ
 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
From the rates, the equivalent current densities for methanol dehy-
drogenation to COad (jdec) can also be calculated according to the
following equation [42]:
jdec ðtÞ ¼ 4eNPt hmax r dec ðtÞ ð13Þ
where hmax is the saturation coverage of COad achievable at the Pt
film electrode surface, which is found to be ca. 0.7 ML from the
CO stripping charge [34]. Here one needs to multiply hmax, since
the COad formed at Pt surface may occupy one, two or three Pt sites
depends on its bonding configuration, thus in the course of forma-
tion of one COad from the decomposition one methanol molecule
the average charge flowed through each Pt atom is smaller than
4e[42]. The equivalent current densities for methanol dehydroge-
nation (Eq. (13)) derived from the above analysis based on the
in situ IR measurements are also displayed Fig. 8b. From Fig. 8b, it
is clearly seen that at ca. 1 s after solution switch, jdec(t) amounts
to ca. 80% of the observed jtot, and its absolute values and its frac-
tional ratio to the total faradic current decreases with increases in
COad coverage. It drops sharply in the next 2 s, which is followed
with a much slower decay and reaches a plateau at ca. 10 s after
the solution switch. This reveals that other pathways for HCHO
and HCOOH production take place in addition to methanol
dehydrogenation to COad (Fig. 1) under present condition, COad
formation process are inhibited and more side products are formed
with the increase in the extent of Pt surface poisoning by COad.
The above analysis demonstrates that if the surface coverage
and the population of reaction intermediates can be deduced from
the corresponding infrared band intensities, electrochemical in situ
infrared spectroscopy can be used as a powerful tool for unraveling
the mechanism and kinetics of complicated reactions such as MOR.
However, it should be noticed that when relating the adsorbate
coverage with its IR band intensities, one should be very careful
since there is not always linear relationship between these two
parameters due to various reasons. Taken CO adsorbate as an
example, the coefficient kCOL in Eq. (2) changes sensitively with
the structure of electrode substrate and the actual optical system
used for recording the IR spectra. Thus when applying such a cali-
bration curve, one should be very careful to ensure that other con-
ditions are unchanged. In addition, due to the electronic screening
effect from the dipole–dipole coupling interactions among the
neighboring COad oscillators, at higher coverages (hCO > 0.6 ML)
the COL band intensity-COad coverage curve deviates from the lin-
ear plot for (hCO < 0.6 ML) with a much slower increase in the band
intensity versus the COad coverage. Under such circumstance, alter-
native techniques to derive the information on the coverage of
reaction intermediates are necessary. Furthermore, when deriving
the kinetic information of COad formation from methanol in the po-
tential region from 0.05 to 0.3 V, great care need to be taken since
the COad coverage–band intensity relationship changes greatly due
to the formation of COad islands induced by co-adsorbed H atoms
[37].
4. Conclusion
In this paper, a way to unravel the mechanism and kinetics of
complicated reactions such as methanol oxidation reaction is
presented based on the electrochemical in situ infrared spectro-
scopic study combined with a thin-layer flow cell. Using such an
approach, quantitative information on the reaction kinetics can
be derived from the changes in the coverage of reaction intermedi-
ates. The procedures on how to derive the calibration curve relat-
ing the coverage of reaction intermediates with its IR band
intensities and how to carry out the kinetic analysis are described
in detail. The methodology is illustrated by analyzing the kinetics
of the indirect pathway involved in methanol oxidation reaction
on Pt electrode surface under potentiostatic conditions.
239
By exploiting a calibration curve relating the COad surface
coverage with the IR band intensity of CO stretching vibration,
COad surface coverage at any time can be derived. From the
changes in COad coverage with time in methanol free solution,
the rate constant for COad oxidation is estimated by assuming a
Langmuir–Hinshelwood mechanism. Then the rates for methanol
dehydrogenation and that for COad oxidation during MOR under
potentiostatic conditions are derived. Present analysis reveals that
the maximum rate for COad oxidation rate at +0.6 V observed under
steady state is only ca. 0.004 molecule site1 s1, which is ca. two
orders of magnitude smaller than that for COad formation from
methanol dehydrogenation at clean Pt surface. And at +0.6 V the
CO pathway is found to only contribute to ca. 17% of the total
methanol oxidation currents on Pt film electrode. By exploiting
such a strategy, we have carried out systematic studies on the
effect of electrode potential and methanol concentration on the
kinetics of CO pathway and its contribution into the total methanol
oxidation, the results will be reported elsewhere.
Acknowledgements
The authors thank Prof. W. Vielstich for invaluable discussion.
This work was supported by 100 Talents’ Program of the Chinese
Academy of Science, National Natural Science Foundation of China
(NSFC) (Project’s No. 20773116), 973 Program from the ministry of
science and technology of China (Project’s No. 2010CB923302).
References
[1] M.W. Breiter, Electrochim. Acta 8 (1962) 973–983.
[2] V.S. Bagotzky, Y.B. Vassilyev, Electrochim. Acta 12 (1967) 1323–1343.
[3] K. Shimazu, H. Kita, Denki Kagaku 53 (1985) 652–662.
[4] R. Inada, K. Shimazu, H. Kita, J. Electroanal. Chem. 277 (1990) 315–326.
[5] T.D. Jarvi, E.M. Stuve, in: J. Lipkowski, P.N. Ross (Eds.), Electrocatalysis, John
Wiley and Sons Ltd., NewYork, 1998 (Chapter 3).
[6] W.F. Lin, M.S. Zei, M. Eiswirth, G. Ertl, T. Iwasita, W. Vielstich, J. Phys. Chem. B
103 (1999) 6968–6977.
[7] A. Hamnett, in: A. Wieckowski (Ed.), Interfacial Electrochemistry: Theory,
Experiment, and Applications, Marcel Dekker Inc., New York, 1999 (Chapter
47).
[8] D. Kardash, J. Huang, C. Korzeniewski, Langmuir 16 (1999) 2019–2023.
[9] C.L. Green, A. Kucernak, J. Phys. Chem. B 106 (2002) 11446–11456.
[10] N.M. Markovic, P.N. Ross, Surf. Sci. Rep. 45 (2002) 117–229.
[11] W. Viestich, Electrochemistry, second ed., John Wiley and Sons Ltd., New York,
2007.
[12] A.A. El-Shafei, R. Hoyer, L.A. Kibler, D.M. Kolb, J. Electrochem. Soc. 151 (2004)
F141–F145.
[13] Y.X. Chen, A. Miki, S. Ye, H. Sakai, M. Osawa, J. Am. Chem. Soc. 125 (2003)
3680–3681.
[14] Z. Jusys, J. Kaiser, R.J. Behm, Langmuir 19 (2003) 6759–6769.
[15] T.H.M. Housmans, A.H. Wonders, M.T.M. Koper, J. Phys. Chem. B 110 (2006)
10021–10031.
[16] J.S. Spendelow, P.K. Babu, A. Wieckowski, Curr. Opin. Solid State Mater. Sci. 9
(2005) 37–48.
[17] H. Wang, H. Baltruschat, J. Phys. Chem. C 111 (2007) 7038–7048.
[18] P. Ferrin, A.U. Nilekar, J. Greeley, M. Mavrikakis, J. Rossmeisl, Surf. Sci. 602
(2008) 3424–3431.
[19] P. Ferrin, M. Mavrikakis, J. Am. Chem. Soc. 131 (2009) 14381–14389.
[20] J. Greeley, M. Mavrikakis, J. Am. Chem. Soc. 124 (2002) 7193–7201.
[21] D. Cao, G.Q. Lu, A. Wieckowski, S.A. Wasileski, M. Neurock, J. Phys. Chem. B 109
(2005) 11622–11633.
[22] H. Wang, T. Loffler, H. Baltruschat, J. Appl. Electrochem. 31 (2001) 759–765.
[23] Y.X. Chen, M. Heinen, Z. Jusys, R.B. Behm, Angew. Chem. Int. Ed. 45 (2006) 981–
985.
[24] L.W.H. Leung, A. Wieckowski, M.J. Weaver, J. Phys. Chem. 92 (1988) 6985–
6990.
[25] K. Kunimatsu, Ber. Bunsen Phys. Chem. 94 (1990) 1025–1029.
[26] M. Heinen, Y.X. Chen, Z. Jusys, R.J. Behm, Electrochim. Acta 52 (2007) 5634–
5643.
[27] Y.X. Chen, M. Heinen, Z. Jusys, R.J. Behm, Langmuir 22 (2006) 10399–10408.
[28] Y.X. Chen, M. Heinen, Z. Jusys, R.J. Behm, ChemPhysChem 8 (2007) 380–385.
[29] A. Miki, S. Ye, M. Osawa, Chem. Commun. (2002) 1500–1501.
[30] B.E. Conway, H. Angerstein-Kozlowska, Acc. Chem. Res. 14 (1981) 49–56.
[31] F.C. Nart, W. Vielstich, in: W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.),
Handbook of Fuel Cells – Fundamentals, Technology and Applications:
Electrocatalysis, John Wiley and Sons Ltd., England, 2003 (Chapter 21).
[32] T.H.M. Housmans, M.T.M. Koper, J. Phys. Chem. B 107 (2003) 8557–8567.
240 L.W. Liao et al. / Journal of Electroanalytical Chemistry 650 (2011) 233–240
[33] Z. Jusys, R.J. Behm, J. Phys. Chem. B 105 (2001) 10874–10883.
[34] Q.S. Chen, J. Solla Gullón, S.G. Sun, J.M. Feliu, Electrochim. Acta. 55 (2010)
7982–7994.
[35] A. Cuesta, A. Couto, A. Rincon, M.C. Perez, A. Lopez-Cudero, C. Gutierrez, J.
Electroanal. Chem. 586 (2006) 184–195.
[36] R. Gomez, J.M. Feliu, A. Aldaz, M.J. Weaver, Surf. Sci. 410 (1998) 48–61.
[37] Y.X. Chen, M. Heinen, Z. Jusys, R.J. Behm, J. Phys. Chem. C 111 (2007) 435–438.
[38] S.C. Chang, M.J. Weaver, J. Chem. Phys. 92 (1990) 4582–4594.
[39] Data given in Fig. 5 is measured in 0.5 M H2SO4, where the amount of COad at
E = 0.6 V is ca. 90% of its total initial COad coverage due to partial oxidation at
E < 0.6 V, while for the data set measured in 0.1 M HClO4, at E > 0.4 V due to
more facile COad oxidation, less COad remained on the surface.
[40] N.P. Lebedeva, M.T.M. Koper, J.M. Feliu, R.A. van Santen, J. Electroanal. Chem.
524–525 (2002) 242–251.
[41] It should be noticed that since the data points for the COL band intensity–time
curve are limited (0.2 s per points) and due to the noise associated in collecting
IR spectra, directly differentiating the coverage–time curve derived from the
band intensity–time plot gives rather noisy rdec(t)–time curve. To overcome
this drawback, polynomial simulation has been firstly used to fit the coverage–
time curve (Fig. 3b), before carrying out the differentiation procedure.
[42] K. Franaszczuk, E. Herrero, P. Zelenay, A. Wieckowski, J. Wang, R.I. Masel, J.
Phys. Chem. 96 (1992) 8509–8516.