Participation of linear methoxy species bonded to Ti4+ sites in the methanol carbonylation catalyzed by TiO2-supported rhodium: An infrared investigation

Article abstract of DOI:10.1016/j.molcata.2012.03.020

TiO₂-supported rhodium samples synthesized from the Rh ₆(CO)₁₆ complex were catalytically active and selective for the gas-phase methanol carbonylation at 140°C and atmospheric pressure in the presence of methyl iodide as

Full text of DOI:10.1016/j.molcata.2012.03.020

Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Participation of linear methoxy species bonded to Ti4+ sites in the methanol
carbonylation catalyzed by TiO2-supported rhodium: An infrared investigation
Gerardo A. Flores-Escamilla, Juan C. Fierro-Gonzalez∗
Departamento de Ingeniería Química, Instituto Tecnológico de Celaya, Celaya, Guanajuato 38010, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
TiO2-supported rhodium samples synthesized from the Rh6(CO)16 complex were catalytically active and
selective for the gas-phase methanol carbonylation at 140 ◦C and atmospheric pressure in the presence
of methyl iodide as promoter. Infrared (IR) spectra recorded during catalysis allowed the identification
of molecularly adsorbed methanol, together with linear, doubly, and triply-bridged surface methoxy
species on Ti4+ sites of the support. IR bands characteristic of rhodium complexes that might be regarded
as reaction intermediates were also observed, in addition to bands assigned to surface acetate species
attributed to the formation of methyl acetate. Our results reveal that only linear methoxy species on Ti4+
sites react with flowing CO in the presence of CH3I to give methyl acetate, whereas bridged methoxy
species and molecularly adsorbed methanol are only spectators in methanol carbonylation.
© 2012 Elsevier B.V. All rights reserved.
Received 11 November 2011
Received in revised form 4 February 2012
Accepted 22 March 2012
Available online 30 March 2012
Keywords:
Methanol carbonylation
Supported rhodium
Infrared spectroscopy
1. Introduction
Rh complexes comes from X-ray crystallography, nuclear magnetic
resonance (NMR) and infrared (IR) spectroscopic studies [3,5,6].
The carbonylation of methanol is the most effective catalytic
route for the production of acetic acid and methyl acetate. BASF
and Monsanto were pioneers in the design of robust and prof-
itable industrial processes for methanol carbonylation with cobalt
and rhodium catalysts, respectively [1–3]. Both processes involve
the use of methyl iodide as cocatalyst. The BASF process uses a
cobalt carbonyl complex (i.e., [CoH(CO)4]) as catalyst and occurs
at relatively high pressure (680 bar) and temperature (250 ◦C),
whereas the Monsanto process uses a rhodium carbonyl complex
(i.e., [Rh(CO)2I2]−) as catalyst and the reaction occurs at moder-
ate pressure (30–60 bar) and temperature (150–200 ◦C), achieving
higher methanol conversion than that obtained by the BASF pro-
cess [3,4]. The advantages of the Monsanto process have motivated
increasing research on the mechanistic aspects of the methanol car-
bonylation catalyzed by rhodium [5,6]. An accepted catalytic cycle
representing the Monsanto process is shown in Fig. 1.
It is widely accepted that the rate determining step for the r−eac-
tion consists of the oxidative addition of CH3I to Rh(CO)2I2 to
give RhCH3(CO)2I3− (Fig. 1). Then, CO insertion into the Rh-methyl
bond occurs to give a Rh-acetyl complex, which after CO addition
and reductive elimination produces CH3COI and regenerates the
Rh(CO)2I2− complex (Fig. 1). Hydrolysis of CH3COI gives acetic acid
and HI, which reacts with methanol to produce water and regener-
ate the CH3I promoter (Fig. 1). Evidence of the various intermediate
Because recovery of the catalyst in the homogeneous process is
typically complex and expensive, there have been various attempts
to synthesize solid catalysts for methanol carbonylation by stabi-
lizing rhodium complexes on polymers, metal oxides and zeolites
[7–10]. Haynes et al. [8] anchored a [Rh(CO)2I2]− complex on a
polymer support and tested the resulting catalyst for methanol
carbonylation. They characterized the catalyst under reaction con-
ditions by IR spectroscopy and identified rhodium complexes
during catalysis that are analogous to the intermediates found in
the homogeneous process (Fig. 1). The authors concluded that the
methanol carbonylation reaction proceeds by the same sequence
of organometallic reactions for both the homogeneous and the het-
erogeneous processes.
For the homogeneous process, it is widely accepted that
methanol does not coordinate to rhodium atoms during the reac-
tion, but it is only involved in the cycle to regenerate CH3I (Fig. 1).
Nevertheless, the participation of methanol during the reaction
catalyzed by supported rhodium has not been described.
Because methanol is a simple molecule with vibration modes
that are sensitive to the nature of its interactions with solid sur-
faces, its adsorption on metal oxides has been widely investigated
by IR spectroscopy [11–25]. It has been observed that methanol can
be adsorbed both molecularly and dissociatively on metal oxides,
including TiO2 [11–16], ZrO2 [17,18], Al2O3 [19,20], and others
[21–25]. In some cases, it has been possible to determine which
of the methanol-derived surface species participate as intermedi-
ates in some chemical reactions [12,26]. For example, Wu et al.
[12] found that linear methoxy species bonded to TiO2 have a
∗
Corresponding author. Tel.: +52 4616117575; fax: +52 4616117744.
E-mail address: jcfierro@iqcelaya.itc.mx (J.C. Fierro-Gonzalez).
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.03.020

50
G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
Fig. 1. Catalytic cycle representing the Monsanto process for methanol carbonylation [3].
higher photo-oxidation rate than bridged methoxy species to give
CO2. Similarly, Rousseau et al. [26] used IR spectroscopy to investi-
gate the total oxidation of methanol catalyzed by CeO2-supported
gold catalysts. They found that only linear and doubly-bridged
methoxy species bonded to Ce4+ cations participated in the catal-
ysis, whereas doubly-bridged Ce4+ methoxy species having an
oxygen vacancy in the neighborhood were only spectators.
In this work, we report the methanol carbonylation to give
methyl acetate catalyzed by TiO2-supported rhodium samples
prepared from Rh6(CO)16. We used TiO2 as support because the
adsorption of methanol on its surface is well established in the lit-
erature [11–16], with IR spectra being sensitive to the identification
of various types of alcohol-derived surface species. Here we focused
on investigating the reactions of methanol-derived surface species
with CO by IR spectroscopy. Our results show the presence of var-
ious surface methoxy species bonded to Ti4+ sites during catalysis,
and demonstrate that only linear methoxy species react to give the
carbonylation product.
carbon film grid. Samples were analyzed in both bright field and
dark field imaging modes.
2.3. IR spectra characterizing TiO2-supported rhodium samples
during methanol carbonylation
Samples of the bare TiO2 support and TiO2-supported rhodium
were loaded into an IR cell that was closed and isolated with two
standard three-way valves. IR spectra were recorded in the dif-
fuse reflectance Fourier transform (DRIFT) mode with a Nicolet
FTIR 6700 spectrometer equipped with a SpectraTech CollectorTM
attachment fitted with a high-temperature and high pressure envi-
ronmental chamber. Spectra were recorded by coadding 128 scans
with a resolution of 4 cm−1. KBr powder was normally used as a
reference material. Difference IR spectra of the adsorbed species
were obtained by absorption subtraction of the cell and catalyst
background spectra using the installed software.
The samples were characterized by IR spectroscopy during
methanol carbonylation. In one type of experiment, IR spectra
were recorded as each sample (ca. 25 mg) was treated in a flowing
2. Experimental
mixture of methanol (PCH = 108.96 Torr), methyl iodide (PCH
=
I
I
3
3
108.96 Torr), CO (PCO = 52.43 Torr) and He (total flow rate 20 ml
(NTP) min−1 at 760 Torr) at 140 ◦C for 90 min. In another type of
experiment, the TiO2-supported rhodium sample was exposed to
flowing methanol at 140 ◦C for 5 min and then it was evacuated with
He for 1 h. IR spectra were then recorded as the sample was treated
in a continuous flow of a mixture of CO and He (PCO = 76 Torr, total
flow rate 20 ml (NTP) min−1 at 760 Torr) and as intermittent pulses
of CH3I (ca. 20 ␮L) were injected to the cell each 10 min at 140 ◦C.
2.1. Synthesis of TiO2-supported rhodium samples
The syntheses and sample transfers were performed in absence
of air and moisture with a N2-filled glove box (Labconco 5070).
Samples of TiO2-supported rhodium containing 2 wt% Rh were
prepared by slurrying Rh6(CO)16 (Sigma-Aldrich) in n-pentane
with TiO2 powder (Evonik, P25; approximately 30% rutile and 70%
anatase) that had been partially dehydroxylated under vacuum at
400 ◦C. The surface area of the TiO2 support was 50 m2 g−1. The
slurry was stirred for 24 h and the solvent removed by evacua-
tion (pressure < 10−3 Torr) for 12 h. The resultant TiO2-supported
rhodium sample was stored in the glove box.
2.4. Mass spectra characterizing the effluent gases from the flow
reactor/DRIFT cell
Mass spectra of the effluent gases from the flow reac-
tor/DRIFT cell during methanol carbonylation were measured
with an on-line Pfeiffer OmniStarTM mass spectrometer running
in multi-ion monitoring mode. In the experiments, the changes
in the signal intensities of the main fragments of methanol
(m/e = 31), CO (m/e = 28), methyl acetate (m/e = 43, 74), acetic acid
(m/e = 43, 45), dimethyl ether (m/e = 45, 46) and CO2 (m/e = 44)
were recorded as the bare TiO2 support and the TiO2-supported
2.2. Transmission electron microscopy characterizing
TiO2-supported rhodium samples
Samples were characterized by transmission electron
microscopy (TEM) on
a Jeol JEM-2100F HRTEM operated at
200 kV. Each powder sample was dispersed and placed on a holey

G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
51
Fig. 3. ꢀCO IR spectra characterizing (a) Rh6(CO)16 diluted in KBr, (b) TiO2-supported
rhodium and (c) TiO2 calcined at 400 ◦C. Inset: ꢀOH IR spectra characterizing (b) the
TiO2-supported Rh sample and (c) the bare support.
Fig. 2. Transmission electron micrograph characterizing a TiO2-supported Rh sam-
ple.
rhodium sample were treated in a flowing mixture of methanol
rhodium species represented as RhI(CO)2 and schematized in Fig. 4
[32–40]. Our results indicate that the rhodium complex might have
also reacted with adsorbed water, as evidenced by the decrease
in the intensity of the band at 1630 cm−1 upon adsorption of the
Rh6(CO)16 complex on TiO2 (Fig. 3). Our proposal is consistent with
that of Smith et al., [38] who reported that the octahedron frame
of Rh6(CO)16 was fragmented when this complex reacted with OH
groups or adsorbed water on the surfaces of alumina and magnesia
to give oxidized rhodium species. Others have also reported similar
results [39,41] including the possible reactions of Rh6(CO)16 with
hydroxyl groups on metal oxides.
(PCH
= 126.66 Torr), methyl iodide (PCH = 108.96 Torr), CO
OH
I
3
(PCO = 52.43 Torr) and He (total flow rate 230 ml (NTP) min−1 at
760 Torr) as the temperature was ramped from 25 to 400 ◦C. All
signals are reported relative to that of the He carrier gas at m/e = 4
to remove any effects of pressure fluctuations.
3. Results and discussion
3.1. Characterization of TiO2-supported rhodium samples
After impregnation of the rhodium precursor on the TiO2 pow-
der, the resultant solid was gray color. A transmission electron
micrograph characterizing a TiO2-supported rhodium sample is
shown in Fig. 2. The results indicate the presence of well-dispersed
small rhodium particles with an average diameter of 5.4 nm.
IR spectra in the C O stretching (ꢀCO) region characterizing the
pure Rh6(CO)16 precursor, the bare support and the TiO2-supported
rhodium sample are shown in Fig. 3. The IR spectrum characteriz-
ing Rh6(CO)16 includes three intense ꢀCO bands at 2074, 2023 and
1801 cm−1 (Fig. 3a). The bands at 2074 and 2023 cm−1 have been
assigned to the symmetric and asymmetric ꢀCO vibrations of termi-
nal carbonyls bonded to rhodium, respectively; whereas the band
at 1801 cm−1 is assigned to the asymmetric ꢀCO vibration of bridged
carbonyls attached to three rhodium atoms [27].
The IR spectrum characterizing the TiO2-supported rhodium
sample (Fig. 3b) includes bands at 2088 and 2019 cm−1, which can
be assigned to the symmetric and asymmetric ꢀCO vibrations of two
terminal carbonyls bonded to a RhI atom [28–30]. Because these
ꢀCO bands are quite different from those observed in the IR spec-
trum characterizing the pure precursor (Fig. 3a), we conclude that
Rh6(CO)16 reacted with the TiO2 surface during the synthesis of
the sample. IR spectra in the O H stretching (ꢀOH) region com-
paring the bare support and our TiO2-supported rhodium sample
(Fig. 3, inset) are in agreement with this conclusion, showing that
a band at 3686 cm−1, assigned to isolated hydroxyl groups on the
TiO2 surface [11,13,31], appeared with lower intensity in the spec-
tra characterizing the TiO2-supported rhodium sample than in that
characterizing the bare support. This observation indicates that the
Rh6(CO)16 complex reacted with surface hydroxyl groups of TiO2,
leading to the fragmentation of the precursor on the support to give
To test our hypothesis of the reaction of Rh6(CO)16 with surface
hydroxyl groups on TiO2, we performed titration experiments with
HNO3 of surface hydroxyl groups on both the bare support and the
supported rhodium sample [42]. The results (Supplementary Data)
indicate that the total amount of surface hydroxyl groups on the
supported rhodium sample was lower than that on the sample of
the bare support. These results confirm our proposal that some sur-
face hydroxyl groups reacted with the Rh6(CO)16 complex during
the synthesis of the TiO2-supported rhodium sample. Although the
titration experiments are in agreement with our IR data, we cau-
tion that it is not possible to make accurate quantitative statements
regarding the exact amount of Rh6(CO)16 that reacted with surface
hydroxyl groups because of the presence of water on the surface
of the samples, as evidenced by the band at 1630 cm−1 (Fig. 3). It
is possible that the Rh precursor also reacted with surface water,
as others have proposed [38] and as our own data suggest, as evi-
denced by the decrease in the intensity of the 1630-cm−1 band
upon adsorption of Rh6(CO)16 on the TiO2 support (Fig. 3).
Fig. 4. Schematic representation of a RhI(CO)2 complex on the TiO2 surface.

52
G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
Table 1
Observed frequencies/cm−1 and band assignments of main surface species formed
as samples of the bare support and TiO2-supported Rh were exposed to a flowing
mixture of CH3OH, CO, CH3I and He.a
Samples treated in reactive
flowing mixture
Assignment
Refs.
Rh/TiO2
TiO2
2974
2942
2922
2865
2838
2820
1510
1441
1437
1262c
1235c
1138
1119
1056
1035
1013
2063
1988
1731
2974
2940
2920
2872
2838
2819
ꢀa CH, CH3OH gas-phase
ꢀa CH, CH3OH(ads)
ꢀa CH, CH3O(ads)
[45]
[11–14]
[11–14]
[13]
[11–14]
[11–14]
[12]
2ıCH3, CH3O(ads)
ꢀs CH, CH3OH(ads)
ꢀs CH, CH3O(ads)
ꢀa COO, CH3COO(ads)
ꢀs COO, CH3COO(ads)
b
b
[12]
[13,16]
1437
1262c
1235c
1155
1125
1057
1036
1012
ıCH, CH3O(ads)
ꢁCH3, CH3O(ads) linear
ꢀ OC, CH3O(ads) linear
ꢀ OC, CH3OH(ads)
ꢀ OC, CH3O(ads) doubly-bridged
ꢀ OC, CH3O(ads) triply-bridged
[12,13]
[12,13]
[13]
[13]
[13]
[7,8,10]
[7,8,10]
[8,10]
Fig. 5. Changes in the intensity of the mass spectral signals of the effluent gases from
the flow reactor when the TiO2-supported Rh sample was treated in the flowing
mixture of CH3OH, CO, CH3I and He at increasing temperature. (•) Methyl acetate
(m/e = 74) and (ꢀ) dimethyl ether (m/e = 46).
ꢀ CO
ꢀ CO
ꢀ CO
,
,
d
3.2. Methanol carbonylation catalyzed by TiO2-supported Rh
Notation:
Mass spectra characterizing the effluent gases from the flow
reactor that contained a sample of TiO2-supported rhodium as it
was treated in a flowing mixture of methanol, CO, CH3I and He at
increasing temperature are shown in Fig. 5. The results show the
formation of methyl acetate starting at approximately 80 ◦C, as evi-
denced by the increase in the intensity of the signal of the fragment
at m/e = 74 (Fig. 5). The intensity of the mass fragment of methyl
acetate increased rapidly from 80 ◦C until it reached a maximum
value at approximately 140 ◦C. Then, it remained nearly constant
from 140 to 170 ◦C, and it finally decreased from that temperature
until the set point (400 ◦C) was reached. As the intensity of the mass
fragment of methyl acetate decreased above 170 ◦C, dimethyl ether
started to form, as evidenced by the rapid increase in the intensity
of the signal of the mass fragment at m/e = 46 (Fig. 5). The formation
of dimethyl ether increased with increasing temperature until the
set point was reached.
These results indicate that the TiO2-supported rhodium sam-
ple was active towards the carbonylation of methanol to give
methyl acetate. The catalyst was selective in the temperature range
between 100 and 170 ◦C under the reaction conditions used in
this work. At higher temperatures, the formation of dimethyl ether
occurred probably by the dehydration of methanol, as others have
proposed [43,44].
a
Reaction
conditions:
PCH
= 126.66 Torr,
PCH = 108.96 Torr,
I
3
OH
3
PCO = 52.43 Torr, catalyst weight = 25 mg, reaction temperature = 140 ◦C, reaction
time = 90 min.
b
Bands at 1510 and 1449 cm−1 were observed when a sample of the bare TiO2
support was exposed to flowing methyl acetate at 140 ◦C.
c
Bands at 1261 and 1238 cm−1 were observed when CH3I was adsorbed on KBr
powder.
d
ꢀCO vibration of a Rh-acetyl complex. (ads).- adsorbed.
oxides (e.g., TiO2 [11–16], ZrO2 [17,18], CeO2 [21]) has been widely
reported. Details regarding the assignment of the bands observed in
this work are discussed in the following paragraph and summarized
in Table 1.
The bands observed at 2942, 2838 and 1056 cm−1 (Fig. 6A and
B) can be assigned to the symmetric ꢀCH, asymmetric ꢀCH and ꢀCO
vibration modes of molecularly adsorbed methanol on TiO2, respec-
tively [12,13]. The bands at 2922 and 2820 cm−1 can be assigned to
the symmetric and asymmetric ꢀCH vibrations of methoxy species
bonded to Ti4+ sites [12,13], and the band at 2866 cm−1 can be
assigned to the overtone of the ıCH3 deformation mode in Fermi
resonance with ꢀCH3 vibrations of methoxy species (Fig. 6A) [13].
The results indicate the presence of three types of surface methoxy
species bonded to Ti4+ during the experiment, as evidenced by the
appearance of bands at 1138, 1119, 1035 and 1013 cm−1 (Fig. 6B).
The bands at 1138 and 1119 cm−1 can be assigned to the CH3
rocking and the ꢀOC modes of linear methoxy species bonded
to Ti4+ sites, respectively [12,13,23]. The band at 1437 cm−1 is
assigned to ıCH3 mode of methoxy species [13,16] and the bands
at 1035 and 1013 cm−1 can be assigned to the ꢀCO vibrations of
doubly and triply-bridged methoxy species on Ti4+ sites, respec-
tively [13]. In summary, these results show that methanol was
adsorbed both molecularly and dissociatively on the surface of
the TiO2-supported rhodium sample. Specifically, three types of
surface methoxy species on the support were observed: (a) lin-
ear, (b) doubly- and (c) triply-bridged. Bolstering the assignment
of these bands to methanol-derived species bonded to TiO2, IR
spectra characterizing the bare support as it was exposed to the
reactive mixture at 140 ◦C show the same bands as those observed
when the experiment was done with the TiO2-supported rhodium
sample (Supplementary Data). A schematic representation of the
various types of methoxy species bonded to the support is shown
in Fig. 7. The formation of methoxy species on various metal oxides,
including TiO2, ZrO2 and CeO2 after cleavage of the O H bond of
When the same experiment was done in the presence of the bare
TiO2 support, no formation of methyl acetate was observed and
only dimethyl ether was produced (Supplementary Data). These
results confirm that the presence of rhodium is necessary for the
catalyst to be active for the methanol carbonylation.
3.3. Surface species formed during methanol carbonylation
catalyzed by TiO2-supported rhodium at 140 ◦C
In order to identify surface species formed during catalysis, we
recorded IR spectra of a TiO2-supported rhodium sample as it was
treated in a flowing reactor/DRIFT cell with the flowing mixture of
CH3OH, CO, CH3I and He at 140 ◦C during 90 min.
The results (Fig. 6) show the appearance of a band at 2974 cm−1
,
assigned to the asymmetric C H stretching (ꢀCH) mode of gas-
phase methanol [45], and the evolution of various IR bands
characteristic of both, molecularly and dissociatively adsorbed
methanol on the surface of the catalyst. The formation of those
surface species upon adsorption of methanol on various metal

G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
53
Fig. 6. Evolution of IR spectra characterizing the TiO2-supported Rh sample as it functioned as catalyst for methanol carbonylation at 140 ◦C. (A) ꢀCH region, (B) 1600–950 cm−1
region and (C) ꢀCO region. Reaction conditions: PCH
= 126.66 Torr, PCH = 108.96 Torr, PCO = 52.43 Torr. Total flow rate 20 ml (NTP) min−1
.
OH
I
3
3
Fig. 7. Schematic representation of (a) linear, (b) doubly and (c) triply-bridged methoxy species bonded on Ti4+ sites.

54
G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
methanol has been widely reported [12,17,21]. It has been estab-
lished that the presence of doubly and triply-bonded methoxy
species is caused by bonding of the surface species at defect sites
with oxygen vacancies on metal oxides [12,23].
The IR bands observed at 1262 and 1235 cm−1 (Fig. 6B) can be
attributed to vibrations of CH3I, as they also appeared in IR spectra
characterizing a sample of KBr impregnated with CH3I (Table 1).
Fig. 6C shows ꢀCO IR spectra characterizing the TiO2-supported
rhodium sample during the same catalysis experiment at 140 ◦C.
The results show the initial appearance (after 2 min time on stream
(TOS)) of bands at 2088 and 2015 cm−1. These bands can be assigned
to the symmetric and asymmetric ꢀCO vibrations of two terminal
carbonyls bonded to a RhI atom [28–30]. These bands shifted to
lower frequencies with increasing TOS, until they were observed at
2063 and 1988 cm−1. Similar ꢀCO band-shifts have been observed
during methanol carbonylation catalyzed by zeolite-supported
rhodium [10] and by organorhodium complexes in solution [5].
The shift in the frequencies of the ꢀCO bands is consistent with
the reaction between the initially present rhodium dicarbonyls
and CH3I (most probably by oxidative addition), because the pres-
ence of iodine ligands bonded to the rhodium causes an additional
electron-donating effect, which lowers the frequencies of the ꢀCO
bands [10,46].
Fig. 8. IR subtraction spectra characterizing a sample of TiO2-supported rhodium
in flowing He at 140 ◦C after it had been exposed to methanol adsorption at 140 ◦
for 5 min.
C
methanol were first stabilized on the surface of a TiO2-supported
rhodium sample at 140 ◦C. Then, the reactions of those species were
monitored by IR spectroscopy in the presence of flowing CO and as
intermittent pulses of CH3I (ca. 20 ␮L) were admitted to the flow
reactor/DRIFTS cell.
It is generally agreed that the oxidative addition of CH3I to the
rhodium dicarbonyls is followed by the insertion of CO to the Rh-
methyl bond, leading to Rh-acetyl species [3,5,7]. Our results are
consistent with that proposal, showing the appearance of a broad
band centered at 1731 cm−1 (Fig. 6C), characteristic of the ꢀCO
stretching mode of Rh-acetyl species [8,10].
Fig. 8 shows IR spectra characterizing the TiO2-supported
rhodium sample in flowing He at 140 ◦C after methanol had
been adsorbed on its surface. The results show the presence of
molecularly adsorbed methanol, linear, doubly and triply-bridged
methoxy species, as evidenced by the appearance of bands at 1142,
1119, 1051, 1038 and 1008 cm−1. Bands characteristic of surface
formate species also appeared in the spectra at 1555, 1357 and
1376 cm−1 (Fig. 8). The bands at 1357 and 1555 cm−1 are assigned
to the symmetric and asymmetric ꢀCOO vibrations of adsorbed for-
mate species, respectively [23,47]; whereas the band at 1376 cm−1
is assigned to their ıCH vibrations [23,47]. Surface formate species
might have been formed by decomposition of methanol on the TiO2
support [12,14,26,47,48].
In addition to the bands described previously, bands at 1441 and
1510 cm−1 appeared almost instantly when the TiO2-supported
rhodium sample was treated with the reactive flowing mixture
(Fig. 6B). Because those bands were not observed when the experi-
ment was done with the bare support (Table 1 and Supplementary
Data), one could hypothesize that they correspond to vibration
modes of a product formed on the surface of the catalyst. To assign
those bands, IR spectra were measured after the bare TiO2 sup-
port had been exposed to methyl acetate at 140 ◦C, followed by
evacuation in flowing He for 60 min. The results (Table 1 and
Supplementary Data) show the appearance of bands at 1449 and
1510 cm−1. The frequencies of these bands are very similar to those
of the bands observed as the TiO2-supported rhodium sample was
used as catalyst for methanol carbonylation at 140 ◦C (Fig. 6B).
Therefore, we assign the bands at 1441 and 1510 cm−1 to the sym-
metric and asymmetric ꢀCOO vibrations of surface acetate species.
Indeed, Wu et al. [12] assigned bands at 1535 and 1453 cm−1 to
acetate species bonded to Ti4+ sites of TiO2. Therefore, our results
show evidence of methyl acetate adsorbed on the surface of the
catalyst. This carbonylation product might form by the reaction
between methanol and surface Rh-acetyl complexes. The question
that arises is which of the various methanol-derived surface species
reacted with the Rh-acetyl complex to give methyl acetate.
Fig. 9 shows the evolution of IR spectra measured at 140 ◦C as
the TiO2-supported Rh sample with adsorbed methanol was treated
in flowing CO and in the presence of intermittent pulses of CH3I.
The results indicate that the introduction of flowing CO did not
cause significant changes on the intensities of the various bands.
However, once the first CH3I pulse was admitted to the cell, bands
3.4. Evidence of the participation of linear methoxy species
bonded to Ti4+ in the carbonylation of methanol catalyzed by
TiO2-supported rhodium
Although the results shown in Fig. 6 allow to distinguish
between various methanol-derived surface species bonded to the
surface of the catalyst, it is not possible to determine from those
data which surface species participated in the catalysis, because
the TiO2-supported rhodium sample was continuously in contact
with the reactants in gas-phase and the intensities of all IR bands
characteristic of adsorbed methanol, linear and bridged methoxy
species remained nearly constant during the experiment. Thus, we
performed an experiment in which surface species derived from
Fig. 9. IR subtraction spectra as a function of TOS in the 1600–980 cm−1 region char-
acterizing the TiO2-supported rhodium sample that had been exposed to methanol
adsorption as it was treated in flowing CO (10% CO in He) and with intermittent
pulses of CH3I at 20, 30, 40 and 50 min TOS at 140 ◦C.

G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
55
and (c) V OCH3/Ti OH species. Their results suggest that the
OCH3/Ti OH pairs are the reactive species to give formaldehyde
V
at low temperatures in presence or in absence of O2 [50]. Given the
bifunctional character of the catalysts used in the aforementioned
examples and in our own work, it is suggested that supports capable
of stabilizing specific alcohol-derived surface species might confer
particular catalytic properties to supported metal catalysts.
There are at least two possible routes to explain the formation
of methyl acetate involving the participation of linear methoxy
species bonded on Ti4+ sites. One possibility is the nucleophilic
attack of surface linear methoxy species on the Rh-acetyl com-
plex. Another possibility is from the reaction between the CH3COI
(produced by the reductive elimination of the Rh-acetyl complex)
and surface linear methoxy species. Although our results demon-
strate that linear methoxy species bonded to Ti4+ participate in the
catalysis, we do not have enough information to determine the
exact manner by which they react with other surface intermedi-
ates. Therefore, more data are needed before proposing a detailed
reaction mechanism for the carbonylation of methanol catalyzed
by our samples.
Fig. 10. Correlation between the intensities of the IR bands characteristic of sur-
face acetate species (1447 cm−1, filled circles), triply-bridged methoxy (1008 cm−1
,
filled squares) and linear methoxy species (1142 cm−1, open circles) as a function
of TOS characterizing the TiO2-supported rhodium sample that had been exposed
to methanol adsorption as it was treated in flowing CO (10% CO in He) and with
intermittent pulses of CH3I.
4. Conclusions
In summary, we used IR spectroscopy to monitor surface species
formed on TiO2-supported rhodium catalysts during methanol car-
bonylation. Our results, showing the formation of Rh-acetyl species,
indicate that the reaction route with our samples is analogous to
that of the Monsanto process, with two separate sites for the activa-
tion of CO and methanol. While CO is activated on rhodium centers,
our data show that methanol is activated on the support, with evi-
dence of various methanol-derived species during catalysis. Our
results demonstrate that only linear methoxy species bonded to
Ti4+ sites participated in the methanol carbonylation to give methyl
acetate. These results have important implications on the design
of solid bifunctional catalysts for methanol carbonylation, as they
suggest that the ability of the support to stabilize linear methoxy
species might be crucial to determine the activity and selectivity of
supported rhodium catalysts.
at 1262 and 1236 cm−1, characteristic of CH3I appeared (Fig. 9),
and the bands at 1119 and 1142 cm−1, attributed to the ꢀCO and
the CH3 rocking modes of linear methoxy species bonded to Ti4+
decreased in intensity (Fig. 9). The results also show the appear-
ance of bands at 1447 and 1509 cm−1 (Fig. 9), which are assigned
to the symmetric and asymmetric ꢀCOO vibrations of surface acetate
species [12]. These results indicate that linear methoxy species
bonded to Ti4+ sites are involved in the formation of methyl acetate.
In contrast, the data show that the bands characteristic of
molecularly adsorbed methanol (1051 cm−1) and brigded methoxy
species bonded to Ti4+ (1038 and 1008 cm−1) did not change in
intensity when the CH3I pulses were admitted to the flow reac-
tor/DRIFT cell (Fig. 9). Therefore, we conclude that only linear
methoxy species bonded to Ti4+ sites participated in the forma-
tion of methyl acetate, whereas the other alcohol-derived surface
species were only spectators [49].
Acknowledgments
Fig. 10 shows the dependence between the intensities of the
IR bands at 1142, 1447 and 1008 cm−1 as a function of TOS dur-
ing the experiment. It is clearly observed that the decrease in the
intensity of the band at 1142 cm−1, characteristic of linear methoxy
species bonded to Ti4+, correlates well with the increase in the
intensity of the band at 1447 cm−1, characteristic of surface acetate
species. Thus, our conclusion that methyl acetate was formed by
the carbonylation of linear methoxy species bonded to Ti4+ sites is
We acknowledge Sergio A. Gomez, Gustavo A. Fuentes and Patri-
cia Castillo from Universidad Autónoma Metropolitana-Iztapalapa
for their support in the masurements of transmission electron
micrographs of our TiO2-supported Rh samples. This research was
supported by the Consejo Nacional de Ciencia y Tecnología (CB-
2006 61856).
bolstered. Again, because the intensity of the band at 1008 cm−1
,
characteristic of bridged methoxy species did not change during
the experiment, we conclude that those species did not participate
in the reaction.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.molcata.2012.03.020.
Our results, showing that only one specific methanol-derived
surface species bonded to the support participated in the methanol
carbonylation are analogous to those of Rousseau et al. [26] for the
aerobic oxidation of methanol catalyzed by CeO2-supported gold
particles. The authors found that methanol adsorption led to the
formation of three types of methoxy species which were identi-
fied by IR spectroscopy, and their results show that the linear and
doubly-bridged methoxy species bonded to Ce4+ are intermediates
in the formation of CO2, while doubly-bridged methoxy species
with an oxygen vacancy in the neighborhood are spectators. Sim-
ilarly, Bronkema et al. [50] investigated the selective methanol
oxidation over TiO2-supported vanadium to give formaldehyde.
They found three types of species derived from methanol adsorp-
tion: (a) methanol molecularly adsorbed, (b) Ti OCH3/Ti OH
References
[1] M.J. Howard, M.D. Jones, M.S. Roberts, S.A. Taylor, Catal. Today 18 (1993)
325–354.
[2] N. Yoneda, S. Kusano, M. Yasui, P. Pujado, S. Wilcher, Appl. Catal., A 221 (2001)
253–265.
[3] C.M. Thomas, G. Süss-Fink, Coord. Chem. Rev. 243 (2003) 125–142.
[4] A.S. Merenov, A. Nelson, M.A. Abraham, Catal. Today 55 (2000) 91–101.
[5] D. Forster, J. Am. Chem. Soc. 98 (1976) 846–848.
[6] A. Haynes, B.E. Mann, G.E. Morris, P.M. Maitlis, J. Am. Chem. Soc. 115 (1993)
4093–4100.
[7] R.S. Drago, E.D. Nyberg, A. El A’mma, A. Zombeck, Inorg. Chem. 20 (1981)
641–644.

56
G.A. Flores-Escamilla, J.C. Fierro-Gonzalez / Journal of Molecular Catalysis A: Chemical 359 (2012) 49–56
[8] A. Haynes, P.M. Maitlis, R. Quyoum, C. Pulling, H. Adams, S.E. Spey, R.W. Strange,
particles leads to oxidation of at least part of the surface rhodium atoms to RhI
[33–35]. The oxidation process of surface rhodium atoms by reaction with CO
is similar to that proposed for finely dispersed supported ruthenium clusters
[33,40].
J. Chem. Soc. Dalton Trans. (2002) 2565–2572.
[9] A. Krzywicki, M. Marczewski, J. Mol. Catal. 6 (1979) 431–440.
[10] H.E. Maneck, D. Gutschick, I. Burkhardt, B. Luecke, H. Miessner, U. Wolf, Catal.
Today 3 (1988) 421–429.
[11] E.A. Taylor, G.L. Griffin, J. Phys. Chem. 92 (1988) 477–481.
[12] W.C. Wu, C.C. Chuang, J.L. Lin, J. Phys. Chem. B 104 (2000) 8719–8724.
[13] D.A. Panayotov, S. Burrows, M. Mihaylov, K. Hadjiivanov, B.M. Tissue, J.R. Morris,
Langmuir 26 (2010) 8106–8112.
[33] K.I. Hadjiivanov, G.N. Vayssilov, Adv. Catal. 47 (2002) 307–511.
[34] P. Basu, D. Panayotov, J.T. Yates Jr., J. Phys. Chem. 91 (1987) 3133–3136.
[35] P. Basu, D. Panayotov, J.T. Yates Jr., J. Am. Chem. Soc. 110 (1988) 2074–2081.
[36] J.T. Yates Jr., T.M. Duncan, R.W. Vaughan, J. Chem. Phys. 71 (1979) 3908–
3915.
[14] C.C. Chuang, C.C. Chen, J.L. Lin, J. Phys. Chem. B 103 (1999) 2439–2444.
[15] G. Ramis, G. Busca, V. Lorenzelli, J. Chem. Soc. Faraday Trans. 1 (83) (1987)
1591–1599.
[16] Y. Suda, T. Morimoto, M. Nagao, Langmuir 3 (1987) 99–104.
[17] F. Ouyang, J.N. Kondo, K. Maruya, K. Domen, J. Phys. Chem. B 101 (1997)
4867–4869.
[37] R.R. Cavanagh, J.T. Yates Jr., J. Chem. Phys. 74 (1981) 4150–4151.
[38] A.K. Smith, F. Hugues, A. Theolier, J.M. Basset, R. Ugo, G.M. Zanderighi, J.L. Bilhou,
V. Bilhou-Bougnol, W.F. Graydon, Inorg. Chem. 18 (1979) 3104–3112.
[39] G. Srinivas, S.S.C. Chuang, J. Phys. Chem. 98 (1994) 3024–3031.
[40] S.Y. Chin, C.T. Williams, M.D. Amiridis, J. Phys. Chem. B 110 (2006) 871–882.
[41] O.S. Alexeev, G. Panjabi, B.L. Phillips, B.C. Gates, Langmuir 19 (2003) 9494–
9503.
[18] F. Ouyang, A. Nakayama, K. Tabada, E. Susuki, J. Phys. Chem. B 104 (2000)
2012–2018.
[42] B. Ohtani, Y. Okugawa, S. Nishimoto, T. Kagiya, J. Phys. Chem. 91 (1987)
3550–3555.
[19] G. Busca, P.F. Rossi, V. Lorenzelli, J. Phys. Chem. 89 (1985) 5433–5439.
[20] R.G. Greenler, J. Chem. Phys. 37 (1962) 2094–2100.
[21] A. Badri, C. Binet, J.C. Lavalley, J. Chem. Soc. Faraday Trans. 93 (1997) 1159–1168.
[22] C. Binet, M. Daturi, Catal. Today 70 (2001) 155–167.
[23] G. Busca, Catal. Today 27 (1996) 457–496.
[24] M. Daturi, C. Binet, J.C. Lavalley, A. Galtayries, R. Sporken, Phys. Chem. Chem.
Phys. 1 (1999) 5717–5724.
[25] E. Finocchio, M. Daturi, C. Binet, J.C. Lavalley, G. Blanchard, Catal. Today 52
(1999) 53–63.
[43] H. Knözinger, Angew. Chem. Int. Ed. 7 (1968) 791–805.
[44] R.S. Schiffino, R.P. Merrill, J. Phys. Chem. 97 (1993) 6425–6435.
[45] M. Falk, E. Whalley, J. Chem. Phys. 34 (1961) 1554–1568.
[46] P. Gelin, C. Naccache, Y.B. Taarit, Pure Appl. Chem. 60 (1988) 1315–1320.
[47] A. Gazsi, T. Bánsági, F. Solymosi, Catal. Lett. 131 (2009) 33–41.
[48] Bands attributed to formate species were not observed when the experiment
was done in the presence of flowing methanol, CO and CH3I (Figure 6). This
result could indicate that under those conditions the alcohol reacted prefer-
entially with CO to give the carbonylation product. In absence of CO and CH3I,
some methanol reacted to give formate species.
[26] S. Rousseau, O. Marie, P. Bazin, M. Daturi, S. Verdier, V. Harlé, J. Am. Chem. Soc.
132 (2010) 10832–10841.
[27] E.R. Corey, L.F. Dahl, W. Beck, J. Am. Chem. Soc. 85 (1963) 1202–1203.
[28] J.F. Goellner, B.C. Gates, J. Phys. Chem. B 105 (2001) 3269–3281.
[29] W.A. Weber, B.C. Gates, J. Phys. Chem. B 101 (1997) 10423–10434.
[30] B.E. Hanson, M.E. Davis, D. Taylor, E. Rode, Inorg. Chem. 23 (1984) 52–56.
[31] D.J.C. Yates, J. Phys. Chem. 65 (1961) 746–753.
[49] Because the bands characteristic of surface formate species (1555, 1376 and
1357 cm−1) overlap with those characteristic of the surface acetates (1509 and
1447 cm−1), it was not possible to monitor changes in the intensities of the
bands characteristic of formate species when the CH3I pulses were admitted
to the flow reactor/DRIFT cell.
[32] The existence of RhI(CO)2 species on supported rhodium samples is widely
reported [28,29,33–39]. It has been proposed that bonding of CO with rhodium
[50] J.L. Bronkema, D.C. Leo, A.T. Bell, J. Phys. Chem. C 111 (2007) 14530–14540.