Necessity of Oxygenated Surface Species for the Electrooxidation of Methanol on Iridium

Article abstract of DOI:10.1021/jp953185i

The electrooxidation of methanol on an iridium disk in the pH range 0-14 was studied by Fourier transform infrared (FTIR) spectroscopy.The following solution were used: 1 M H2SO4, 0.1 M HClO4, 1 M H3PO4, and 1 M Na2SO4 whose pH had been adjusted to 2 with H2SO4, 0.5 M phosphate buffer of pH 7, 0.1 M borax buffer, and 1 M NaOH.Linear potential sweep (LPS) FTIR spectroscopy showed that the products of 0.1 - 1.0 M methanol electrooxidation on smooth Ir were CO2 and formic acid in acidic media and bicarbonate/carbonate and formate ions in alkaline media.Only at pH 7 and 9.2 neither formic acid nor formate ion was detected.The activity of Ir for methanol electrooxidation was very low.At all pH values dissociative chemisorption of methanol yielded chemisorbed CO, always in the linear form.At pH 14 this chemisorbed CO was completely electrooxidized by 0.6 V vs RHE, a value 0.4 V less positive than in perchloric acid and phosphoric acids of pH 1.This behavior is parallel to that of the appearance of formate/formic: at pH 14 formate is detected already at 0.4 V, while in perchloric and phosphoric acids of pH 1 formic acid is first detected at 0.9 V.This dependence on pH of the electrocatalytic activity of Ir both for the electrooxidation of chemisorbed CO and for that of methanol to formate/formic is in perfect agreement with the fact that the peak of the main Ir surface oxidation process occurs at 0.6 V vs RHE at pH 14, but at a rather more positive potential, 0.9 - 1.0 V vs RHE, in acidic media (super-Nernstian pH dependence).This agreement lends support to the hypothesis that the electrooxidation of organic compounds requires the presence of oxide, hydroxide, and/or oxyhydroxide groups on the Ir surface, which provide the O atoms necessary for the formation of CO2 and formic/formate.However, in sulfuric acid, both at pH 0 and 2, formic acid was already observed at 0.7 V, possibly because in sulfuric acid a prepeak of Ir oxidation at 0.6 V that precedes the main peak at 1.0 V is better defined.The stretching frequency of linear CO increased by up to 60 cm^-1 with increasing coverage and at the same coverage was higher in HClO4 than in H3PO4 solutions.The Stark shift varied over the range 39 - 45 cm^-1 V^-1 with pH, CO coverage, and nature of the anion in the case of acids.

Full text of DOI:10.1021/jp953185i

J. Phys. Chem. 1996, 100, 8389-8396
8389
Necessity of Oxygenated Surface Species for the Electrooxidation of Methanol on Iridium
R. Ortiz,† O. P. Ma´rquez,‡ J. Ma´rquez,‡ and C. Gutie´rrez*,†
Instituto de Qu´ımica F´ısica “Rocasolano”, C.S.I.C., C. Serrano, 119, 28006-Madrid, Spain, and
Laboratorio de Electroqu´ımica, Depto. de Qu´ımica, Facultad de Ciencias, UniVersidad de los Andes,
Me´rida 5101, Venezuela
ReceiVed: October 30, 1995; In Final Form: February 28, 1996X
The electrooxidation of methanol on an iridium disk in the pH range 0-14 was studied by Fourier transform
infrared (FTIR) spectroscopy. The following solutions were used: 1 M H2SO4, 0.1 M HClO4, 1 M H3PO4,
and 1 M Na2SO4 whose pH had been adjusted to 2 with H2SO4, 0.5 M phosphate buffer of pH 7, 0.1 M borax
buffer, and 1 M NaOH. Linear potential sweep (LPS) FTIR spectroscopy showed that the products of 0.1-
1.0 M methanol electrooxidation on smooth Ir were CO2 and formic acid in acidic media and bicarbonate/
carbonate and formate ions in alkaline media. Only at pH 7 and 9.2 neither formic acid nor formate ion was
detected. The activity of Ir for methanol electrooxidation was very low. At all pH values dissociative
chemisorption of methanol yielded chemisorbed CO, always in the linear form. At pH 14 this chemisorbed
CO was completely electrooxidized by 0.6 V vs RHE, a value 0.4 V less positive than in perchloric and
phosphoric acids of pH 1. This behavior is parallel to that of the appearance of formate/formic: at pH 14
formate is detected already at 0.4 V, while in perchloric and phosphoric acids of pH 1 formic acid is first
detected at 0.9 V. This dependence on pH of the electrocatalytic activity of Ir both for the electrooxidation
of chemisorbed CO and for that of methanol to formate/formic is in perfect agreement with the fact that the
peak of the main Ir surface oxidation process occurs at 0.6 V vs RHE at pH 14, but at a rather more positive
potential, 0.9-1.0 V vs RHE, in acidic media (super-Nernstian pH dependence). This agreement lends support
to the hypothesis that the electrooxidation of organic compounds requires the presence of oxide, hydroxide,
and/or oxyhydroxide groups on the Ir surface, which provide the O atoms necessary for the formation of CO2
and formic/formate. However, in sulfuric acid, both at pH 0 and 2, formic acid was already observed at 0.7
V, possibly because in sulfuric acid a prepeak of Ir oxidation at 0.6 V that precedes the main peak at 1.0 V
is better defined. The stretching frequency of linear CO increased by up to 60 cm-1 with increasing coverage
and at the same coverage was higher in HClO4 than in H3PO4 solutions. The Stark shift varied over the
range 39-45 cm-1 V-1 with pH, CO coverage, and nature of the anion in the case of acids.
Introduction
tions can be reached if, as is usually the case, only CVs at sweep
speeds higher than say 10 mV/s are carried out.
Already in 1963 Breiter1 concluded that Ir was less active
than Rh, Pd, and Pt for the electrooxidation of methanol.
Bagotzky et al.2 found that the plot of the coverage of smooth
Ir by residues of the dissociative chemisorption of methanol vs
potential had the typical flat maximum, with a coverage of about
0.65. They also found that the pH greatly influenced the
coverage by methanol residues, which showed a pronounced
minimum of only 0.1 in the pH range 3-5. The steady-state
electrocatalytic activity of Ir for methanol electrooxidation also
showed a pronounced minimum, but at pH 7.
Aramata et al.5 studied by CV and IR spectroscopy the
electrooxidation of methanol on Ir in acid medium and observed
linearly chemisorbed CO. For a positive limit of 0.8 V an
anodic peak appeared in the return negative sweep, obviously
due to a more active surface produced by the reduction of a
surface oxide. This anodic peak was much smaller for a positive
limit of 1.2 V and disappeared completely for a 1.4 V limit,
due to increasing irreducibility of the oxide, as is frequently
the case.
Podlovchenko et al.3 found in stationary measurements that
the activity of their “surface skeleton” Ir (with a roughness
factor, determined assuming a hydrogen monolayer charge of
220 µC cm-2, of up to 5300) for methanol electrooxidation in
0.5 M H2SO4 was decreased by UPD Bi and Ag. This is in
agreement with results of Ureta-Zan˜artu et al.,4 who found that
eight UPD metals decreased the methanol electroactivity of an
Ir wire in 1 M HClO4 at sweep speeds lower than a few mV/s,
i.e., those more related to industrial practice. It should be
remarked that they found that the same UPD metals increased
the Ir activity at higher sweep speeds, which shows how wrong
conclusions on the electrocatalytic effect of electrode modifica-
As is well-known, the CV of Ir shows a reversible oxidation,
whose peak lies at 1.0 V vs RHE in acid. Upon changing the
pH the position of this peak does not remain the same with
respect to the RHE. On the contrary, it shows a super-Nernstian
pH dependence of about 30 mV vs RHE/pH unit (i.e., about 90
mV/pH unit when using a fixed reference electrode).6 A thick
electrochromic oxide layer can be grown on Ir by potential
cycling at pH <3.5, in spite of which the hydrogen area does
not decrease, probably due to a very hydrated, permeable nature
of the oxide.7
It is usually accepted in electrocatalysis that electrooxidation
of organic compounds on metallic electrodes occurs first at
potentials positive enough for electrooxidation of the metal to
begin, the metal oxycompounds providing the required O atoms.
This hypothesis has recently been confirmed by XANES8 (for
CO electrooxidation on Pt50Ru50 nanoparticles) and by potential-
† C.S.I.C.
‡ Universidad de los Andes.
* Author for correspondence.
X Abstract published in AdVance ACS Abstracts, May 1, 1996.
S0022-3654(95)03185-6 CCC: $12.00 © 1996 American Chemical Society

8390 J. Phys. Chem., Vol. 100, No. 20, 1996
Ortiz et al.
modulated reflectance spectroscopy9 (for ethanol electrooxida-
tion on polycrystalline Ni and Fe). Further confirmation has
been provided by a XANES and EXAFS study of a carbon-
supported Pt-Ru electrocatalyst, whose high activity for
methanol electrooxidation was attributed to the formation of
RuOH at low potentials.10 However, it was concluded that
electronic effects should also be taken into account.
The super-Nernstian dependence of the electrooxidation of
the Ir surface yields a good opportunity to check that elec-
trooxidation of organics coincides with the start of iridium oxide
formation. For this purpose we have studied by means of FTIR
spectroscopy the electrooxidation of methanol on an Ir disc in
acidic, neutral, and alkaline aqueous solutions.
Experimental Section
The Ir electrode was a 12 mm disc cut from Ir sheet, 99.9%
pure and 0.25 mm thick, from Alfa Ventron. It was polished
with successively finer grades of alumina down to 0.05 µm,
copiously rinsed with ultrapure water, and cleaned for 5 min in
an ultrasonic bath just prior to use. Methanol (Riedel, 99.8%),
with a nominal water content <0.01% confirmed by Karl Fischer
test, and an ethanol content of <0.1% were used as received.
The other reagents were of analytical grade.
The Ir electrode was glued with Araldit to the end of a 0.5
cm long, 12 mm o.d. Pyrex tube, and electric contact to a copper
cable was made with Ag epoxy. This assembly was joined with
heat-shrinkable polyolefin tubing with an inner adhesive lining
to another Pyrex tube of 6 mm diameter, the flexible intercon-
necting tubing allowing to obtain uniform films of electrolyte
solution by pressing the electrode against the fluorite window.
A conventional glass cell and a Teflon cell with a capacity
of about 12 cm3 were employed for voltammetric and FTIR
studies, respectively. The Teflon cell was positioned in a
Spectra-Tech variable angle specular reflectance accessory so
as to maximize the detected IR signal. The p-polarized light
reached the CaF2 window at an incidence angle of about 67°.
A P.A.R Model 362 scanning potentiostat was used. The
reference electrode was SCE for the experiments in a conven-
tional electrochemical cell and, due to space limitations, Ag/
AgCl in saturated KCl for the FTIR experiments. However,
the potentials are given versus the reversible hydrogen electrode
(RHE).
Figure 1. Dashed curves are the cyclic voltammograms (CVs) at 50
mV s-1 of an Ir disk in 1 M H2SO4 (A), 1 M H3PO4 (B), 0.5 M, pH 7
phosphate buffer (C), and 1 M NaOH (D), recorded after potential
cycling at 500 mV s-1 until the CV was reproducible. Potentials are
given versus the reversible hydrogen electrode (RHE), with which the
super-Nernstian dependence of the main Ir oxidation process becomes
clearly apparent. The CVs in the presence of 1 M methanol are shown
as solid curves.
Results and Discussion
Cyclic Voltammetry. The dashed curves in Figure 1 are
cyclic voltammograms at 50 mV s-1 of Ir in 1 M H2SO4 (A),
1 M H3PO4 (B), pH 7 phosphate buffer (C), and 1 M NaOH
(D) obtained after potential cycling at 500 mV s-1 until a
reproducible CV was obtained. In the RHE scale the main Ir
oxidation process appears at a 0.3-0.4 V more negative
potential at pH 14 than in acidic and neutral media, showing
the super-Nernstian dependence mentioned in the Introduction.
In the presence of 1 M methanol, the CVs (solid lines) show
a very slight increase in the anodic current over the potential
range in which Ir electrooxidation occurs and a steep increase
of the anodic current at about 1.4 V vs RHE, a potential only
about 0.1 V less positive than in base electrolyte, showing the
poor electrocatalytic activity of Ir for methanol oxidation. (The
equilibrium potential for the electrooxidation of methanol to
CO2 and water is 0.01 V vs RHE11). The hydrogen area
decreased slightly, due to dissociative chemisorption of metha-
nol, yielding linearly chemisorbed CO, as will be seen below.
Infrared reflectance spectra in the region 1000-3000 cm-1
were collected with a Perkin-Elmer Model 1725-X FTIR
spectrometer connected to a 486DX2P.C. A liquid nitrogen-
cooled narrow-band MCT detector was used. The sample
compartment was purged with dry, CO2-free air from a Peak
purge gas generator at a flow rate of 20 L min-1. The spectral
resolution was 8 cm-1
.
Differential FTIR spectra were obtained from interferometric
scans continuously carried out during a linear potential sweep
(LPS) at 1 mV/s, each 80 successive interferograms (taken in
about 35 s) being separately collected (LPS-FTIRS). The
reference spectrum was always taken at 0.1 V vs RHE, unless
otherwise indicated.
Normalized differential reflectance is defined here as the
difference between the reflectance at the sample potential, Es,
and that at the reference potential, Er, divided by the reference
reflectance, i.e., ∆R/R ) R(Es)/R(Er) - 1. With this definition
upward and downward bands correspond to loss and gain of
compounds, respectively, in the thin electrolyte layer and/or
electrode surface. The bands originated by a Stark shift of a
chemisorbed species are the so-called bipolar bands, with a
positive and a negative lobe.
The lack of electrocatalytic activity of Ir required the use of
high methanol concentrations for its oxidation products to be
detectable in FTIR experiments. Actually, the methanol con-
centration was of the order of, or even higher than, that of the
buffers. Consequently, the electrooxidation of methanol pro-
duced large pH changes and these, in turn, large artifacts in the
FTIR spectra, as will be seen below. Another consequence is
that especially in neutral and alkaline media the potentials vs
RHE shown against FTIR spectra should be viewed as maxi-
mum values.
FTIRS Results in Acidic Medium. In the differential
spectrum obtained by positive polarization of Ir in 1 M H2SO4,

Electrooxidation of Methanol on Iridium
J. Phys. Chem., Vol. 100, No. 20, 1996 8391
Figure 3. Differential LPS-FTIR spectra obtained during a positive
linear potential sweep (LPS) at 1 mV s-1 of an iridium disk in 0.1 M
HClO4, 1 M in methanol.
Figure 2. Differential LPS-FTIR spectra obtained during a positive
linear potential sweep (LPS) at 1 mV s-1 of an iridium disk in 1 M
H2SO4, 1 M in methanol. During the LPS interferometric scans were
continuously recorded, each 80 successive interferograms being
separately collected in about 35 s. The reference spectrum was taken
at the initial potential of 0.1 V vs RHE.
little tenable, since only a fraction of the Ir surface is covered
by chemisorbed CO.
Although formic acid has two peaks at 1732 and 1216 cm-1
,
1 M in methanol linearly chemisorbed CO is evidenced by a
bipolar band (due to the well-documented Stark shift) of linear
(on top) CO at about 2045 cm-1 (Figure 2). This bipolar band
indicates that methanol already dissociates on Ir, yielding linear
CO, at the potential of the reference spectrum, 0.1 V. The
bipolar band becomes a monopolar upward band at 2045 cm-1
at 0.7 V, indicating that at this potential the chemisorbed CO is
completely electrooxidized, since the band intensity does not
increase with further potential increase.
The maximum normalized reflectance change, ∆R/R, of
chemisorbed CO was 0.7%. The value for a (maximum) CO
coverage of 0.60-0.65 on Ir(111) can be estimated as 1.9%
from Figure 1 in ref 12. Therefore, a sizable fraction of the Ir
surface is covered with chemisorbed CO, which is in apparent
contrast with the CVs shown in Figure 1, in which the hydrogen
area is about the same in the absence and presence of methanol.
The reason for this apparent contradiction is that the CVs in
Figure 1 are stabilized repetitive CVs at 50 mV/s, in which little
methanol chemisorption can take place during the negative
sweep, while the FTIR experimens were run at 1 mV/s from
the negative limit, which allowed ample time for dissociative
chemisorption of methanol to proceed.
as will be shown below, only the peak at 1720 cm-1 appears in
Figure 2, while the formic acid peak at 1216 cm-1 is completely
swamped by an artifact at 1208 cm-1. As a matter of fact, both
the downward peaks at 1208 and 1055 cm-1 in Figure 2 are
-
artifacts due to an increase of the HSO4 concentration in the
2-
thin layer, produced at the expense of SO4 ions, whose loss
originates an upward peak at 1100 cm-1. These artifacts are
originated by the acidification of the thin layer consequent to
the electrooxidation of methanol, either to formic acid, which
produces four H+ ions per methanol molecule, according to
CH3OH + H2O w HCOOH + 4H+ + 4e-
or to CO2, which frees six H+ ions per methanol molecule,
(1)
CH3OH + H2O w CO2 + 6H+ + 6e-
(2)
Even in base electrolyte, i.e., in the absence of organic
compounds, the processes of hydrogen adsorption and desorption
and of surface oxide formation and reduction can change the
pH in the thin layer.13,14 Furthermore, even in the double layer
region, i.e., in the absence of faradaic reactions, the change in
the state of charge of the electrode can lead to the migration of
ions into or out of the thin layer, a rapid process which can
also originate artifacts.14
Water is consumed in the electrooxidation of methanol in
acidic and neutral media, as seen in reactions 1 and 2. This
loss of water originates an upward peak at about 1650 cm-1 in
all other acidic and neutral media (Figures 3-5 and 7).
However, in 1 M H2SO4 it appears shifted to 1590 cm-1 (Figure
2), probably because of the higher acidity of the medium, since
this water band is very sensitive to pH, hydrogen bonding, and
ionic strength.
Neither bridge-bonded nor multiply bonded CO was detected
on the Ir surface in the presence of methanol over the pH range
0-14, in agreement with the results of Aramata et al.5 for
methanol electrooxidation on polycrystalline Ir in acidic media
and with those of Jiang et al.11 for adsorption of CO on Ir(111)
in 0.1 M HClO4.
Dissolved CO2 (downward peak at 2345 cm-1) is first
observed at 0.6 V. Formic acid (downward peak at 1720 cm-1
)
first appears at 0.7 V, the same potential at which chemisorbed
CO is completely electrooxidized, which indicates that both
processes require the same surface iridium oxide or, alterna-
tively, that chemisorbed CO completely poisons the electrooxi-
dation of methanol to formic acid. This latter hypothesis is
The results in 0.1 M HClO4 (Figure 3) and in 1 M H3PO4
(Figure 4) are essentially the same as those in 1 M H2SO4

8392 J. Phys. Chem., Vol. 100, No. 20, 1996
Ortiz et al.
Figure 5. Differential LPS-FTIR spectra obtained during a positive
linear potential sweep (LPS) at 1 mV s-1 of an iridium disk in a 1 M
Na2SO4 solution whose pH had been adjusted to 2 with H2SO4, 0.5 M
in methanol. In the inset a deconvoluted spectrum is shown in order to
reveal the band at 1730 cm-1 of formic acid which was swamped by
the upward band of water loss at 1652 cm-1. The band of formic acid
at 1214 cm-1 is completely hidden by the large downward band at
1204 cm-1 produced by the increase of the amount of HSO4- in the
thin layer due to the acidification inherent to the electrooxidation of
methanol.
Figure 4. Differential LPS-FTIR spectra obtained during a positive
linear potential sweep (LPS) at 1 mV s-1 of an iridium disk in 1 M
H3PO4, 1 M in methanol.
(Figure 2). Dissolved CO2 first appears at 0.4 and 0.7 V,
respectively, and formic acid is first observed at a higher
potential, 0.9 V in both cases. Chemisorbed CO is completely
electrooxidized at a slightly higher potential, 1.0 V in both cases.
It must be stressed that with these two acids the two bands of
formic acid, at 1730 and 1214 cm-1, appear simultaneously,
which rules out the presence of formaldehyde, since aldehydes
also show the carbonyl band at 1730 cm-1, but not the C-O
stretch at 1214 cm-1. This is in contrast with the behavior of
ethanol in 0.5 M HClO4, in which case acetaldehyde is formed
first, and acetic acid at higher potentials.15 No conclusions on
the presence of formaldehyde can be drawn in the case of
sulfuric acid, due to the large HSO4- peak at 1208 cm-1, but it
is reasonable to assume that also in this case no formaldehyde
is produced.
at 2 mV/s. The electrochemical reaction is simply
2HCOOH + 2e- w 2HCOO- + H2
(3)
The reference potential was 0.1 V vs SCE, the first spectrum
was recorded at a potential of -0.2 V and the last one at -0.4
V. Two upward peaks at 1216 and 1732 cm-1 corresponding
to the disappearance of HCOOH grow simultaneously with two
downward peaks at 1352 and 1580 cm-1 and a shoulder at 1382
cm-1, due to the formation of HCOO- ion at the expense of
HCOOH (Figure 6). A small peak of water loss at 1657 cm-1
is also present in the spectrum.
The downward band at 1108 cm-1 in perchloric acid is due
-
to migration of ClO4 ions into the thin layer to compensate
for the increase in H+ concentration produced by methanol
electrooxidation.
Neutral and Weakly Alkaline Media. In the electrooxi-
dation of 0.5 M methanol on Ir in a pH 7 phosphate buffer (0.5
M K2HPO4 + 0.5 M KH2PO4), CO2 (downward band at 2345
Also in 1 M Na2SO4 acidified with H2SO4 down to pH 2.0
(about the value of the second acidity constant of H2SO4, pK2
) 1.9616) the band of formic acid at 1730 cm-1 appeared in
the IR spectrum (Figure 5), but only after deconvolution (see
inset in Figure 5) from the upward band at 1652 cm-1 produced
by the loss of water in the thin layer. Formic acid appeared at
0.7 V, a slightly lower potential than that of 0.9 V at which
chemisorbed CO was completely electrooxidized.
Again, as in Figure 2 the downward bands at 1204 and 1050
cm-1 are due to an increase in the HSO4- concentration in the
thin layer, produced at the expense of SO42- ions (upward band
at 1115 cm-1) by the acidification inherent to the electrooxi-
dation of methanol.
Spectra of Formic Acid and Formate Ion. In order to
obtain the IR spectrum of both formic acid and formate ion,
we carried out the electroreductive deprotonation of 0.1 M
formic acid in 0.1 M NaClO4 aqueous solution, by means of an
LPS-FTIRS experiment in which interferometric scans were
continuously recorded during a negative linear potential sweep
cm-1) and bicarbonate ion (downward band at 1364 cm-1
)
appear at 0.6 V, and chemisorbed linear CO (upward band at
2003 cm-1) is completely electrooxidized at 0.7 V (Figure 7a).
Contrary to the behavior in acidic medium neither HCOOH nor
HCOO- (actually at pH 7 formic acid, if present, should be
ionized by more than 99.9%, since its acidity constant is pKa
) 3.7516) was detected. Deconvolution of the upward peak of
water loss at 1647 cm-1 failed to show a downward peak of
formate ion at 1580 cm-1
.
The presence in the spectrum of both CO2 and bicarbonate
ion was to be expected, since the first ionization constant of
carbonic acid is 6.35.16 However, there is an acidification of
the thin electrolyte layer, as evidenced by the large downward
band at 1165 cm-1 of H2PO4- formed at the expense of HPO4
,
2-
whose disappearance originates a large upward band at 1095
cm-1 (the second ionization constant of phosphoric acid is
6.8616). As a matter of fact, the acidification was so strong

Electrooxidation of Methanol on Iridium
J. Phys. Chem., Vol. 100, No. 20, 1996 8393
Figure 6. Differential LPS-FTIR spectra obtained in the electrore-
ductive deprotonation of 0.1 M formic acid in 0.1 M NaClO4 aqueous
solution during a negative linear potential sweep (LPS) at 2 mV/s. The
electrochemical reaction is 2HCOOH + 2e- w 2HCOO- + H2. The
two downward peaks at 1352 and 1580 cm-1 and a shoulder at 1382
cm-1 are due to the formation of HCOO- ion at the expense of
HCOOH, whose loss originates two upward peaks at 1216 and 1732
cm-1
.
that even H3PO4 (first ionization constant 2.1416) was formed
in the thin layer, as revealed by the downward band at 1027
cm-1 17
.
When the methanol concentration was increased from 0.5 to
1 M, CO2 in solution, but no HCO3-, appeared in the FTIR
spectra. (Linearly chemisorbed CO is also formed, although it
is barely seen at the scale of Figure 7b.) Obviously, the absence
of bicarbonate is due to exhaustion of the buffering power of
the 0.5 M buffer by the electrooxidation of 1 M methanol. There
is very little difference between two consecutive spectra at 1.4
V, showing that even at this potential the reaction rate is not
high.
Borax. In the electrooxidation of 0.1 M methanol on Ir in
0.1 M borax, pH 9.2 chemisorbed linear CO (upward band at
1993 cm-1) is completely electrooxidized at 0.5 V, and CO2
(downward band at 2345 cm-1) is first observed at a slightly
higher potential, 0.7 V (Figure 8). As at pH 7, no bands
assignable to formate ion or formic acid appear in the spectrum.
The acidification of the thin layer inherent to the electrooxidation
of methanol is stronger due to the use of a low buffer
concentration, 0.1 M, mandated by the low solubility of borax,
and consequently only CO2, and no bicarbonate, is seen in the
spectra. This acidification is also responsible for the downward
peaks at 1413 and 1156 cm-1, originated by an increase in the
concentration of boric acid, H3BO3.17 There is, correspondingly,
an upward band of H2BO3- ions at 1350 cm-1. The downward
peak at 1048 cm-1 is probably an artifact, since there is a peak
at this position in the energy spectrum with only air in the beam.
The upward and downward peaks that appear at 1694 and
1633 cm-1, respectively, straddling the position of the water
absorption at 1650 cm-1, have not been assigned.
Sodium Hydroxide. The electrooxidation of 0.25 M metha-
nol on Ir in 1 M NaOH produces chemisorbed linear CO
(upward band at 1996 cm-1), formate ion (downward bands at
1353 and 1582 cm-1), and carbonate ion (downward band at
1384 cm-1, which swamps the formate band at the same
frequency) (Figure 9a). (The two small upward peaks of loss
Figure 7. Differential LPS-FTIR spectra obtained in a positive linear
potential sweep (LPS) at 1 mV s-1 of an iridium disk in a pH 7
phosphate buffer (0.5 M K2HPO4 + 0.5 M KH2PO4) in the presence
of methanol at a concentration of 0.5 M (a, top) and 1 M (b, bottom).
It can be seen that, in contrast to the results in acidic and strongly
alkaline media, neither the peaks of formic acid at 1732, and 1216
cm-1 nor those of formate ion at 1580, 1382 and 1352 cm-1 appear in
the LPS-FTIR spectra. The downward peaks at 2341 and 1364 cm-1
-
in (a) indicate the formation of CO2 and HCO3 ion, respectively. At
a twice higher methanol concentration the acidification of the electrolyte
inherent to the electrooxidation of methanol is so high that only CO2
(peak at 2341 cm-1) and no HCO3-, is observed (b).
of gaseous CO2 at about 2340 cm-1 are due to an incomplete
purging of the sample chamber before beginning the experi-
ment.) It should be stressed that at this pH 14 an electrooxi-
dation product of methanol, formate ion, could be detected
already at 0.4 V, as compared with 0.8-0.9 V at pH 1, and
that electrooxidation of chemisorbed CO was already completed
at 0.6 V, as compared with 1.0 V at pH 1.

8394 J. Phys. Chem., Vol. 100, No. 20, 1996
Ortiz et al.
Figure 8. Differential LPS-FTIR spectra obtained in a positive linear
potential sweep (LPS) at 1 mV s-1 of an iridium disk in 0.1 M borax
in the presence of 0.1 M methanol. The acidification inherent to the
electrooxidation of methanol produces two downward peaks at 1413
and 1156 cm-1 due to the increase of the concentration of boric acid
at the expense of borate ions, whose loss originates an upward peak at
1350 cm-1
.
If the methanol concentration is increased to 2 M, the
acidification brought about by its electrooxidation decreases the
pH so much that CO2 appears instead of carbonate or bicarbon-
ate, which means that the pH must be lower than the pK1 of
carbonic acid, 6.3516 (Figure 9b). In this case there is no
interference by the carbonate ion, and therefore the three formate
bands clearly appear in the spectrum, which indicates that the
pH must be higher than the pKa of formic acid, 3.75.16
Consequently, the pH in the thin layer should be about 5.
Discussion
Products of Methanol Electrooxidation. We have found
that besides CO2 (or, obviously, bicarbonate or carbonate ions
in neutral and alkaline media) the only product of methanol
electrooxidation on Ir is formic acid and formate ions in acidic
and alkaline media, respectively. It is interesting that, within
the sensitivity of our FTIRS measurements, neither formic acid
nor formate ion was produced in a pH 7 phosphate buffer or in
a pH 9.2 borax buffer. This could perhaps be related to the
above-mentioned pronounced minimum at pH 7 for the steady-
state electrooxidation of methanol on smooth Ir.2
Necessity of Oxygenated Metal Surface Species for the
Electrooxidation of Methanol on Iridium, As Deduced from
Its pH Dependence. The FTIR spectra over the pH range 0-14
clearly show that in 1 M NaOH both complete electrooxidation
of the linearly chemisorbed CO produced by the dissociative
chemisorption of methanol and the electrooxidation of methanol
to formic/formate occur already at 0.4-0.6 V vs RHE.
Similarly, in sulfuric acid, both at pH 0 and 2, formic acid is
already observed at 0.7 V, but in both perchloric and phosphoric
acids at pH 1 a rather more positive potential, 0.9-1.0 V, is
required.
Figure 9. Differential LPS-FTIR spectra obtained in a positive linear
potential sweep (LPS) at 1 mV s-1 of an iridium disk in 1 M NaOH in
the presence of methanol at a concentration of 0.25 M (a, top) and 2
M (b, bottom). At this high pH chemisorbed CO is completely
electrooxidized at 0.6 V, at which potential the bipolar character of
the CO peak is lost, only an upward peak of CO loss remaining. Also,
formate begins to appear already at 0.4 V.
It is usually accepted that the electrooxidation of organic
compounds on metals starts at those potentials at which
oxidation of the metals begins, the metal oxycompounds
providing the oxygen atoms necessary for the electrooxidation
of the organic compounds.8,9 As mentioned above, the main
surface process in the CV of Ir, involving the formation of a
semimetallic Ir(IV) oxide and its reduction, probably to an
iridium(III) oxide,18 has a super-Nernstian dependence on pH
of about 30 mV vs RHE; i.e., at pH 14 the main process appears

Electrooxidation of Methanol on Iridium
J. Phys. Chem., Vol. 100, No. 20, 1996 8395
at a potential 0.4 V more negative (in the RHE scale, which
corresponds to 1.2 V more negative if a fixed reference electrode
is used) than at pH 0. This can be seen in Figure 1, where the
main surface process appears at 0.6 V at pH 14, and at 0.9-
1.0 V in acidic media.
The negative shift (in the RHE scale) with increasing pH of
the potential of the main Ir surface oxidation process is exactly
the same shift with pH observed in 1 M NaOH, in 0.1 M HClO4,
and in 1 M H3PO4 for the potential of electrooxidation of
chemisorbed CO and of methanol to formic/formate on Ir. In
other words, both Ir oxidation and complete electrooxidation
of chemisorbed CO and start of methanol electrooxidation to
formic/formate occur at 0.4-0.6 V vs RHE at pH 14 and at
0.9-1.0 V at pH 1 (in perchloric and phosphoric acids). This
coincidence lends full support to the above assumption that
electrooxidation of chemisorbed CO and organic compounds
necessitates metal surface oxycompounds.
It can be objected that the results in sulfuric acid, at both pH
0 and pH 2, do not fall in line with the above model, since in
both cases formic acid is already observed at 0.7 V, a potential
0.3 V lower than that of the peak of the main Ir oxidation
process. However, this can be due to the fact that in sulfuric
acid a prepeak at 0.6 V precedes the main oxidation peak of Ir,
the oxide produced in this prepeak providing the oxygen atoms
necessary for electrooxidation of the organics.
Over the pH range 0-2 formic acid is first observed at
potentials only 0-0.2 lower than those at which chemisorbed
CO is completely electrooxidized, probably because in both
processes the same surface Ir oxycompound is involved. On
the contrary, dissolved CO2 appears already at 0.4-0.7 V,
potentials about 0.5 V lower than those at which CO is
completely electrooxidized. This small peak of CO2 that appears
at low potentials proceeds from the electrooxidation of chemi-
sorbed CO, and it is only after electrooxidation of chemisorbed
CO has been completed that a large peak of CO2 is produced
by the electrooxidation of methanol.
As far as we know, there is only one well documented case
of electrooxidation of CO or organics on an oxide-free metal,
and this is the electrooxidation of dissolved CO on oxide-free
Pt (i.e., at potentials in the double layer region of Pt) first
reported by Kita et al.19 for polycrystalline Pt, later confirmed,20
and finally extended to single crystal Pt(111).21 Diffusion-
limited electrooxidation of dissolved CO occurs on only 10-
15% of the Pt atoms, namely, those which at the potentials
concerned are free of chemisorbed CO, these free Pt atoms,
either alone or in islands, probably acting as microelectrodes
whose diffusion fields overlap, which would justify the experi-
mental result that the stationary current is the same calculated
for an electrode without blocked areas.20,21
Dependence of the Frequency of the CO Stretching on
the CO Coverage, the pH, and the Anion. In Figure 10 we
show the dependence of the frequency of the stretching vibration
of chemisorbed CO on the normalized differential reflectance
change, for eight experiments at different pH values in the range
0-14, at the same pH 1 but with different acids (0.1 M HClO4
and 1 M H3PO4), or at the same pH 1 and the same acid (1 M
H3PO4) but at different coverages. The spectra were measured
at 0.1 V vs RHE, using as reference the spectrum at 1.0 V, at
which chemisorbed CO has been completely eliminated from
the Ir surface by electrooxidation to CO2.
Figure 10. Plot of the frequency of the CO stretching vibration versus
the differential reflectance (assumed to be approximately proportional
to the coverage) for eight experiments of methanol electrooxidation
on an iridium disk, conducted at different pH values, or at the same
pH value obtained with different acids, or at the same acid and pH
value but at different coverages. (a) 0.1 M borax, pH 9.2; (b) 1 M
H3PO4, pH 1, low CO coverage; (c) 1 M Na2SO4 solution whose pH
had been adjusted to 2 with H2SO4; (d) pH 7 phosphate buffer (0.5 M
K2HPO4 + 0.5 M KH2PO4); (e) 1 M NaOH; (f) 0.1 M HClO4, pH 1;
(g) 1 M H2SO4; (h) 1 M H3PO4, pH 1, high CO coverage.
observed only up to a coverage of 4.5 × 1014 CO molecules
cm-2, at which it reaches saturation.22 Also for CO chemisorbed
on Ir(110) from the gas phase the integrated peak intensity
increases linearly with CO coverage only up to 0.4 monolayer,
after which it decreases slightly with increasing coverage.23
Since the CO coverages reached in this work are only a fraction
of a monolayer, it can be safely assumed that the CO band
intensity is proportional to the coverage.
In Figure 10 it can be seen that the frequency of CO stretching
in 1 M H3PO4, pH 1, increases by 60 cm-1 when the differential
reflectance (and therefore the coverage) increases 6-fold, from
0.1% (point b) to 0.6% (point h). This is in agreement with
the reported increase in the frequency of CO chemisorbed on
Ir(111) in 0.1 M HClO4 by up to ca. 100 cm-1 with increasing
coverage.12
It can also be seen in Figure 10 that the pH would seem to
influence the frequency of CO stretching, since at about the
same differential reflectance, 0.1%, the CO frequency is about
30 cm-1 higher in a pH 9.2 borax buffer (point a) than in 1 M
H3PO4, pH 1 (point b). However, probably it is the nature of
the anion rather than the pH that is the origin of this change in
CO frequency, since for example at the same pH value of 1 the
CO frequency is about the same for different values of the
differential reflectance, but in different acids, 0.35% in 0.1 M
HClO4 (point f) and 0.61% in 1 M H3PO4 (point h), which
means that the frequency increase produced by the higher CO
coverage is compensated by the lower frequency (at the same
coverage) in H3PO4 as compared with HClO4. This influence
of the anion on the CO frequency is not new, since it has been
reported that at saturation coverage the frequency of CO
stretching of CO chemisorbed on Pt was 9-15 cm-1 higher in
1 M HClO4 than in 1 M H2SO4. (The difference increased
linearly with electrode potential because the Stark shift was
higher in the former than in the latter acid.24)
From published IR spectra of CO chemisorbed on Ir(111) in
12
0.1 M HClO4 it can be estimated that the height of the CO
band is roughly proportional to the CO coverage up to saturation
at 9.4 × 1014 CO molecules cm-2. On the contrary, with
polycrystalline Pt linearity of the integrated band intensity is

8396 J. Phys. Chem., Vol. 100, No. 20, 1996
Ortiz et al.
on pH). This coincidence of potentials lends support to the
hypothesis that oxidation of the Ir surface is necessary for the
electrooxidation of both chemisorbed CO and methanol to
formic/formate. However, in sulfuric acid, at both pH 0 and 2,
the electrooxidation of methanol to formic acid first occurred
at 0.7 V, a potential lower than the main peak of Ir oxidation,
perhaps because at this pH the main peak is preceded by a small
peak at 0.6 V.
The frequency of CO stretching of CO chemisorbed on Ir
increased by up to 60 cm-1 with increasing coverage and at
the same coverage was higher in HClO4 than in H3PO4 solutions.
The Stark tuning rate of CO chemisorbed on Ir varied from 39
to 45 cm-1 V-1 in the pH range 1-14, for a CO differential
reflectance range of 0.08-0.54% and for different acids.
Acknowledgment. This work was carried out with the help
of DGICYT, Ministerio de Educacio´n y Ciencia (Spain), under
Project PB93-0146. R.O. gratefully acknowledges a Fellowship
from CONICIT (Venezuela) under the Program of New
Technologies.
Figure 11. Dependence on the electrode potential of the frequency of
the CO stretching vibration at four different pH values, shown against
each curve. pH 0: 1 M H2SO4, ∆R/R ) 0.54%, Stark shift 35 cm-1
References and Notes
V-1. pH 9.2: 0.1 M borax, ∆R/R ) 0.08%, Stark shift 43 cm-1 V-1
.
(1) Breiter, M. W. Electrochim. Acta 1963, 8, 973.
(2) Bagotzky, V. S.; Vassiliev, Yu. B.; Khazova, A.; Sedova, S. S.
Electrochim. Acta 1971, 16, 913.
(3) Podlovchenko, B.; Shterev, G.; Semkova, R.; Kolyadko, E. Elec-
trochim. Acta 1990, 35, 191.
pH 14: 1 M NaOH, ∆R/R ) 0.26%, Stark shift 44 cm-1 V-1. pH 1:
1 M H3PO4, ∆R/R ) 0.10%, Stark shift 40 cm-1 V-1
.
(4) Ureta-Zan˜artu, M. S.; Bravo, P.; Zagal, J. H. J. Electroanal. Chem.
1992, 337, 241.
(5) Aramata, A.; Yamazaki, T.; Kunimatsu, K.; Enyo, M. J. Phys.
Chem. 1987, 91, 2309.
(6) Burke, L. D.; Whelan, D. P. J. Electroanal. Chem. 1984, 162, 121.
(7) Burke, L. D.; O’Sullivan, E. J. M. J. Electroanal. Chem. 1981,
117, 155.
(8) Aberdam, D.; Durand, R.; Faure, R.; Gloaguen, F.; Hazemann, J.
L.; Herrero, E.; Kabbabi, A.; Ulrich, O. J. Electroanal. Chem. 1995, 398,
43.
(9) Kowal, A.; Gutie´rrez, C. J. Electroanal. Chem. 1995, 395, 243.
(10) McBreen, J.; Mukerjee, S. J. Electrochem. Soc. 1995, 142, 3399.
(11) Bockris, J. O’M.; Reddy, A. K. N. Modern Electrochemistry;
Plenum: New York, 1976; p 1363.
(12) Jiang, X.; Chang, S.-C.; Weaver, M. J. J. Phys. Chem. 1991, 95,
7453.
(13) Iwasita, T.; Nart, F. C. J. Electroanal. Chem. 1990, 295, 215.
(14) Iwasita, T.; Nart, F. C.; Polligkeit, H. Ber. Bunsen-Ges. Phys. Chem.
1991, 95, 638.
Stark Shift. The value of the Stark shift of the frequency
of the CO stretching of chemisorbed CO varied with pH and/
or coverage, as can be seen in Figure 11. It was 43.6 ( 2.3
and 44.6 ( 4.5 cm-1 V-1 at pH 9.2 and 14, with a differential
reflectance of 0.08% and 0.26%, respectively. The Stark shift
was 38.96 ( 0.03 cm-1 V-1 at pH 0 and for a differential
reflectance of 0.54% and 40.0 ( 0.1 cm-1 V-1 in 1 M H3PO4
and for a differential reflectance of only 0.10%. Although these
Stark shifts are lower than at alkaline pH, the observed
difference is only two and one standard deviations lower than
the Stark shift at pH 9.2 and 14, respectively. Assuming that
this decrease of the Stark shift is true, it could be due either to
a decrease of the Stark shift with increasing coverage12 or to
the dependence of the Stark shift on the nature of the anion.24
(15) de Tacconi, N. R.; Lezna, R. O.; Beden, B.; Hahn, F.; Lamy, C. J.
Electroanal. Chem. 1994, 379, 329.
Conclusions
(16) Albert, A.; Serjeant, E. P. The Determination of Ionization
Constants. A Laboratory Manual,; Chapman and Hall: London, 1984.
(17) Cuesta, A.; Gutie´rrez, C. To be published.
(18) Ko¨tz, E. R.; Neff, H. J. Electroanal. Chem. 1985, 160, 517.
(19) Kita, H.; Shimazu, K.; Kunimatsu, K. J. Electroanal. Chem. 1988,
241, 163.
(20) Caram, J. A.; Gutie´rrez, C. J. Electroanal. Chem. 1991, 305, 259.
(21) Wieckowski, A.; Rubel, M.; Gutie´rrez, C. J. Electroanal. Chem.
1995, 382, 97.
(22) Tornquist, W.; Guillaume, F.; Griffin, G. L. Langmuir 1987, 3,
477.
(23) Lyons, K. J.; Xie, J.; Mitchell, W. J.; Weinberg, W. H. Surf. Sci.
1995, 325, 85.
Formic acid and CO2 are formed in the electrooxidation of
methanol on iridium in acid medium and formate ion and
carbonate or bicarbonate ion in basic medium. Neither formic
acid nor formate ion was detected at pH 7 or 9.2, at which only
bicarbonate and CO2 were produced.
On Ir both the electrooxidation of methanol to formic/formate
and the complete electrooxidation of chemisorbed CO at pH
14 occur at 0.4-0.6 V, which is also the potential of the main
surface process of Ir oxidation. On the contrary, in both
perchloric and phosphoric acid at pH 1 both the electrooxidation
of methanol and of chemisorbed CO occur at a more positive
potential, 0.9-1.0 V, which is also the potential of the main
surface process in acidic media (super-Nernstian dependence
(24) Russell, J. W.; Severson, M.; Scanlon, K.; Overend, J.; Bewick,
A. J. Phys. Chem. 1983, 87, 293.
JP953185I