Decomposition of Methyl Formate over Supported Pd Catalysts

Article abstract of DOI:10.1007/s10562-017-2011-y

Abstract: Methyl formate (MF) is used as alternative carbonylating agent in many organic transformations. The current study investigates nature of decomposition of MF over Pd catalysts on different support of varying physicochemical properties. Palladium catalysts with 1 and 5?wt% supported on Mg, and 5?wt% on γ-Al₂O₃ and α-Al₂O₃ were synthesized by wetness impregnation and their catalytic activities for MF decomposition were examined. MgO supported Pd with 1?wt% showed maximum MF conversion of about 40?mol% and produced predominantly carbon monoxide (CO) and methanol (MeOH) in the molar ratio of 1:1.2 at 150 °C, 1?atm. However, 5?wt% Pd/MgO catalysts showed decrease in the MF conversion to 33?mol% producing CO/MeOH in the ratio of 1:1. Catalysts with 5?wt% Pd supported on γ-Al₂O₃ and α-Al₂O₃ showed lower MF conversion compared to the Pd supported on MgO. Dimethyl ether (DME) was produced as the main by-product on Pd/γ-Al₂O₃ apart from CO and MeOH and it is attributed to dehydration of methanol on the acidic sites of the support. Higher MF conversion on Pd/MgO catalysts is mainly attributed to the support-metal interactions which are significant in case of Pd–MgO in addition to higher Pd dispersion with 1?wt% Pd on MgO. These results suggest that the presence of Pd over basic supports can be used to produce CO at lower temperatures; hence they may be used as a source of CO for carbonylation reactions at low pressures. Graphical Abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s10562-017-2011-y

Catal Lett (2017) 147:1285–1293
DOI 10.1007/s10562-017-2011-y
Decomposition of Methyl Formate over Supported Pd Catalysts
1
1
1
1
Satyanarayana Reddy Akuri · Chaitanya Dhoke · K. Rakesh · Shweta Hegde ·
1
1
1
Sreejit A. Nair · Raj Deshpande · P. Manikandan
Received: 18 October 2016 / Accepted: 20 February 2017 / Published online: 29 March 2017
Springer Science+Business Media New York 2017
©
Abstract Methyl formate (MF) is used as alternative
carbonylating agent in many organic transformations. The
current study investigates nature of decomposition of MF
over Pd catalysts on different support of varying physico-
chemical properties. Palladium catalysts with 1 and 5 wt%
supported on Mg, and 5 wt% on γ-Al O and α-Al O were
ether (DME) was produced as the main by-product on
Pd/γ-Al O apart from CO and MeOH and it is attributed
2
3
to dehydration of methanol on the acidic sites of the sup-
port. Higher MF conversion on Pd/MgO catalysts is mainly
attributed to the support-metal interactions which are sig-
nificant in case of Pd–MgO in addition to higher Pd disper-
sion with 1 wt% Pd on MgO. These results suggest that the
presence of Pd over basic supports can be used to produce
CO at lower temperatures; hence they may be used as a
source of CO for carbonylation reactions at low pressures.
2
3
2
3
synthesized by wetness impregnation and their catalytic
activities for MF decomposition were examined. MgO sup-
ported Pd with 1 wt% showed maximum MF conversion of
about 40 mol% and produced predominantly carbon mon-
oxide (CO) and methanol (MeOH) in the molar ratio of
Graphical Abstract
1
:1.2 at 150°C, 1 atm. However, 5 wt% Pd/MgO catalysts
showed decrease in the MF conversion to 33 mol% produc-
ing CO/MeOH in the ratio of 1:1. Catalysts with 5 wt% Pd
supported on γ-Al O and α-Al O showed lower MF con-
2
3
2
3
version compared to the Pd supported on MgO. Dimethyl
Keywords Methyl formate · Catalytic decomposition · Pd
catalyst · In-situ CO · Carbonylating agent
*
Satyanarayana Reddy Akuri
Satyanarayana.Akuri@SABIC.com
1
Introduction
Chaitanya Dhoke
Chaitanya.Dhoke@SABIC.com
Carbon monoxide (CO) is one of the key C1 building
blocks for carbonylation reactions such as aminocarbonyla-
tion, alkoxycarbonylation, double carbonylation and oxida-
tive carbonylation [1, 2] where high CO pressure is typi-
cally used [3, 4]. Thus alternate carbonylating agents that
would require lower pressure operation are sought after.
Methyl formate (MF), the smallest member of the ester
family, is considered as one of the alternative carbonylating
agents.
K. Rakesh
Rakesh.K@sabic.com
Shweta Hegde
Shweta.Hegde@SABIC.com
Sreejit A. Nair
Sreejit.Nair@SABIC.com
Raj Deshpande
raj.deshpande@sabic.com
P. Manikandan
Methyl formate is decarbonylated catalytically to pro-
duce CO and methanol (MeOH) and recently, it is widely
explored as a carbonylating agent. For example, carbonyla-
tion of formaldehyde with MF as carbonylating agent to
P.Manikandan@SABIC.com
1
SABIC Research and Technology Pvt. Ltd,
Chikkadunnasandra, Anekal Taluk, Off. Sarjapura - Attibele
State Highway, Bangalore, Karnataka 562125, India
Vol.:(0
1
123456789)
3

1
286
S. R. Akuri et al.
produce methyl glycolate (MG) and methyl methoxy ace-
tate (MMAc) [5]. In fact MF has been found to be a more
efficient carbonylating agent compared to the conventional
CO and can act as a source for in-situ CO formation [6, 7].
For instance, carbonylation of formaldehyde was reported
by Hendriksen et al. [8] using MF as a CO source. Further,
the use of MF has resulted in much higher carbonylated
product, methyl phenyl carbonate, as compared to CO and
MeOH. Additionally, Oxidative carbonylation of alkenes
by MF at low temperatures over Pd–Cu bimetallic catalysts
has also been reported [9]. Overall, carbonylation reactions
demand, a safer and sustainable carbonylating agent like
MF with low pressure requirement over gaseous CO [6].
Methyl formate is decomposed in-situ into CO and
MeOH over solid catalysts by either dissociating at car-
of varying physicochemical properties has been investi-
gated especially to understand the role of these supports,
Pd loadings and metal oxidation states on the nature of MF
decomposition.
2 Experimental
2.1 Catalyst Preparation
2.1.1 Preparation of MgO
Magnesium oxide was prepared by precipitation of a
dilute solution of Mg(NO ) ·6H O with NaOH using an
3
2
2
automated Syrris synthesis reactor. In a typical synthe-
bonyl–hydrogen or decarbonylation to CO and CH OH [2]
sis, 154.09 g of Mg(NO ) ·6H O (601 mmol) was dis-
3
3 2
2
as shown in Scheme 1.
solved in 600 ml of de-ionized water and 32.19 g of NaOH
(25 mmol) in 800 ml of deionized water were prepared
at room temperature. Both alkali and salt solutions were
added simultaneously into the reactor under vigorous stir-
ring at 60°C while maintaining the pH of the resultant mix-
ture at 9± 0.1 at 60°C. After completion of precipitation,
the pH of the gel was adjusted to ten and aged at 60°C for
1 h followed by aging the resultant solution for 16 h at room
temperature. The precipitate thus obtained was filtered and
washed thoroughly with deionized water until the pH of
the filtrate was ∼7.5. The sample was dried at 120°C for
about 12 h and calcined at 500°C in air for 5 h. The MgO
obtained in the above step was used for Pd impregnation.
The other supports, γ-Al O extrudes supplied by Sasol
Various efforts have been made on the catalyst front
and many catalysts have been explored for MF decomposi-
tion to CO and MeOH. For example, supported K CO on
2
3
activated charcoal indicated decomposition of MF at tem-
perature of 260°C and the pressure of 60 bar [11]. Sodium
hydroxide doped MgO showed substantial conversion of
MF at 210°C [12]. Na/MgO catalysts showed higher activ-
ity at 200°C and at high pressure of 50 bar [13]. Base cata-
lyzed zeolites have also been investigated for MF decom-
position, which indicated MF decomposition at 250°C and
1
bar [14].
On one hand, all these catalysts such as NaOCH , MgO
3
and zeolites studied for MF decomposition require either
2
3
2
higher temperatures greater than 200°C or high pressures
(surface area of 211 m /g and N pore volume of 0.8 cc/g)
2
2
[
13]. Apart from this, handling such catalysts for vapor
and α-Al O supplied by SudChemie India (7 m /g and N
2
3
2
phase reactions is a challenge [5]. Palladium based cata-
pore volume of 0.02 cc/g) were used.
lysts supported on common supports such as γ-Al O and
2
3
α-Al O , MgO are known for many important industrial
2.1.2 Impregnation of Pd on MgO, γ‑Al O and α‑Al O
3
2
3
2
3
2
catalytic reactions. These catalysts could be potential for
MF based carboxylation reaction especially at low temper-
ature and pressure. However, there are no reports on MF
decomposition on these well-known systems to the best of
our knowledge. In the present study, MF decomposition on
catalysts with different Pd loading and on different supports
Metal solution was prepared by dissolving requisite amount
of 10 wt% solution [Pd(NH ) ] (NO ) purchased from
3
4
3 2
Aldrich, in deionized water that is equivalent to the water
pore volume of the supports (WPV of MgO, γ-Al O and
2
3
α-Al O are 1.5, 0.7 and 0.5 cc/g, respectively). Pre-dried
2
3
supports were added to the metal solution and dried gently
under nitrogen flow. The impregnated samples were dried
at 120°C for 16 h in hot air oven followed by calcination
in a tubular furnace at 500°C for 5 h. Using this procedure,
1
and 5 wt% Pd/MgO and 5 wt% Pd/γ-Al O and α-Al O
2 3 2 3
were prepared for this study.
2
.2 Catalyst Characterization
The BET surface area of the samples was measured
by nitrogen sorption at −196 °C using a Micromerit-
ics Tristar Surface Area and Porosity Analyzer. The
Scheme 1 MF decomposition mechanism on catalyst surfaces [10]
1
3

Decomposition of Methyl Formate over Supported Pd Catalysts
1287
catalyst samples were evacuated at 300 °C for 3 h before
2
2.3 Catalytic Decomposition Experiments
N adsorption. Powder X-ray diffraction (XRD) data were
obtained on a Rigaku–Miniflex 600 X-ray diffractometer
equipped with a Ni-filtered Cu Kα (λ = 0.1541 nm, 40 kV
A, 30 mA) radiation and graphite crystal monochromator.
The data was collected in the 2θ range of 10°–80° with a
step size of 0.02° and a scan rate of 2°/min.
Catalysts were tested in a vertical down flow, fixed bed
Inconel tubular reactor placed in a two-zone split furnace
(ATS, USA). The reaction data were collected at atmos-
pheric pressure at the temperature of 150°C, using 1 g of
catalyst with a particle size of 0.8–1 mm. The temperature
of the catalyst bed was continuously monitored by a ther-
mocouple. Before the reaction, the catalysts were activated
Temperature-programmed reduction (TPR) profiles
were obtained using a Micromeritics Autochem II Chem-
isorption analyzer. The samples were pretreated in high-
purity (99.98%) argon (50 ml/min) at 500 °C for 3 h.
After cooling to ambient, argon was replaced with a 5%
in N at 400°C for 1 h, followed by reduction at 300°C in
2
presence of 5% H /N . Methyl formate was fed to the reac-
2
2
tor using a high-precision syringe pump (ISCO Model
H in argon mixture, and the catalyst was heated up to
500D), at a flow rate of 0.08 ml/min. Methyl formate was
2
5
50 °C at a heating rate of 5 °C/min. The flow rate of H /
further mixed with the controlled flow of N from MFC to
2
2
−
1
Ar mixture used for this purpose was 40 ml/min. The
water produced during the reduction step was condensed
and collected in a cold trap immersed in the slurry of
ice–liquid nitrogen mixture.
maintain the GHSV of 3100 h . The reactant was evapo-
rated and preheated to the reaction temperature in the upper
part of the reactor before it entered the catalyst bed. The
product mixture was analyzed through online gas chroma-
tograph (Agilent GC7890B) equipped a FID and two TCD
DRIFT-IR analysis was performed using thin wafer
of approximately 20 mg catalyst and diameter of 1 cm
is made using a pellet die. The pellet is carefully loaded
into the pellet ring holder of the high temperature high
pressure FTIR cell. The windows are closed and sealed.
The cell is evacuated and temperature is raised to the
desired reaction temperature i.e. 250 °C stepwise in steps
of 50 °C. An initial background of the empty cell was
taken before loading the catalyst pellet. The spectrum
of the catalyst pellet at 250 °C in vacuum is taken after
about 1 h of evacuation. All spectra were recorded on a
Perkin Elmer Spectrum one FTIR spectrometer equipped
with a Specac high temperature high pressure cell.
detectors using N as the internal standard. The analysis
2
was done over molecular sieve and Carbowax columns. All
the products were identified using pure authentic calibra-
tion product mixture. Methyl formate conversion (decom-
position) was calculated based on below equation;
Methyl formate conversion (%)
Normalized moles of methyl formate reacted
=
× 100.
Normalized moles of methyl formate feed
3
Results and Discussion
X-ray photoemission spectroscopic data was recorded
for the reduced palladium catalysts using SPECS spec-
trometer using non-monochromotized AlKα (1486.6 eV)
X-ray radiation from the source operated at 150 W
3
.1 Catalyst Characterization
It is essential to have physiochemical properties of cata-
lysts and supports of under study in order to establish
structure-activity correlations. BET surface area, pore vol-
ume and pore diameter details of the pure supports and
supported Pd catalysts are summarized in Table 1. As it is
seen in the table, there are no significant changes in the tex-
tural properties of 1 wt% Pd on MgO compared to that of
the neat MgO support. These results can be explained as
(
12 kV and 12.5 mA). Selected spectra were also
recorded with Mg Kα X-ray (hν = 1253.6 eV) source
to eliminate the overlap between different Auger and
core levels. All the survey spectra were obtained with
pass energy of 70 eV, whereas individual spectra were
recorded with pass energy of 70 eV and step increment of
0
.1 eV. Binding energies were calculated with reference
1
wt% Pd loading on MgO did not block the pores. How-
to C1s at 284.6 eV. The samples were made into 8 mm
diameter pellets, mounted on the sample holder and kept
into load-lock chamber with ultrahigh vacuum (UHV) at
ever, in case of 5 wt% Pd/MgO sample, the surface area
2
is decreased from 30 to 24 m /g compared to MgO. Simi-
larly, pore volume and pore diameter also decreased from
−
8
5
× 10 Torr for 5 h in order to desorb any volatile spe-
0
5
.21 to 0.11 cc/g and 275 to 236 Å, respectively. In case of
cies present on the sample surface. The sample was then
placed into analyzer chamber and spectra was recorded
wt% Pd loaded on γ-Al O and α-Al O , the surface area
2
3
2
3
2
2
reduced from 211 to 183 m /g and 7 to 2.1 m /g, respec-
−
10
at 5 × 10
Torr. Relative surface concentrations of the
tively. Similarly, pore volume was reduced from 0.64 to
elements present in the samples were calculated using
respective atomic sensitivity after subtracting Shirley
background employing CasaXPS program.
0
.55 cc/g and 0.02 to 0.005 cc/g during 5 wt% Pd loading
for γ-Al O and α-Al O respectively. Decrease in surface
2
3
2
3
area and pore volume upon metal loading is not surprising
1
3

1
288
S. R. Akuri et al.
Table 1 Textural properties
of MgO, γ-Al O and α-Al O
3
2
Catalyst composition
SA (m /g)
PV (cc)
PD (Å)
H consumption
TPR tem-
perature
2
2
3
2
(
μmol/g)
supported Pd catalysts
(
°C )
MgO
30
211
7
31
24
183
2.1
0.21
0.64
0.02
0.18
0.11
0.55
0.005
275
93
126
239
236
97
27
420
416
452
120
125
69
γ-Al O
0.25
0.23
419
422
77
2
3
α-Al O
2
3
1
5
5
5
wt% Pd/MgO
wt% Pd/MgO
wt% Pd/γ-Al O
2
3
wt% Pd/α-Al O
113
187
80
2
3
not detectable by XRD. This can also be attributed to the
high surface area of the support that possibly caused better
dispersion of the PdO species [17, 18]. Presence of smaller
crystallites size of PdO or Pd which are not big enough to
be detected by XRD cannot be ruled out. In contrast, clear
peaks of PdO are observed for 5 wt% Pd/α-Al O along
2
3
with the peaks corresponding to α-Al O at 2θ~25.88,
2
3
3
8
5.13, 37.78, 43.36, 52.55, 57.51 (JCPDS File No.
3-2080). As α-Al O exhibits low surface area and pore
2
3
volume, the PdO particles formed are of a considerable size
to appear in XRD peaks [19]. This was also confirmed by
TEM analysis reported in further sections.
TEM images of MgO supported Pd catalysts are shown
in Fig. 2. The morphology of MgO is non-uniform with
truncated spherical in shape with agglomeration of par-
ticles. In case of 1 wt% Pd loading on MgO, metal parti-
cles are uniformly distributed with average size of ~20 nm,
however, in case of 5 wt% loading the Pd particle sizes are
non-uniform with average size of ~40 nm.
Fig. 1 XRD patterns of MgO and supported Pd catalysts calcined at
00 °C in air
5
as these observations can be attributed to surface covering
and pore blockage of support by the Pd cluster formation.
Similar observations of reduction in textural properties are
reported elsewhere [15, 16] upon addition of metals/addi-
tives on the supports.
TEM images of 5 wt% Pd on various supports are shown
in Fig. 3. It is clear that for 5 wt% Pd loading, metal parti-
cles are uniformly distributed on MgO and γ-Al O , while
2
3
on α-Al O aggregates of Pd particles are seen in different
2
3
sizes (up to 200 nm). Uniform distribution of metal parti-
XRD patterns of Pd supported on MgO, γ-Al O and
cles of Pd on MgO and γ-Al O could be due to the rela-
2
3
2
3
α-Al O are shown in Fig. 1. The XRD peaks observed
tively strong metal support interaction or due to high sur-
2
3
at 2θ~37.16, 43.12, 62.40, 74.65 and 78.70 match well
with those reported for MgO (JCPDS File No. 78-0430).
Upon Pd loading on MgO, the new peaks are observed
at 2θ~3.87 and 54.79, which can be attributed to PdO
face area compared to α-Al O .
2
3
TPR profiles of Pd supported on MgO, α-Al O and
2
3
γ-Al O samples are given in Fig. 4. The catalyst, 1 wt% Pd/
2
3
MgO, showed a broad reduction peak with T at ~120°C.
max
(
JCPDS file no. 41-1107) and the intensity of these peaks
However, the 5 wt% Pd/MgO showed two distinct reduction
increased for 5 wt% Pd loading [17]. However, no peaks
corresponding to metallic Pd are observed in any of these
samples. These results indicate that Pd exists in the oxide
form and the intensity of the peaks corresponding to PdO
is also increased with metal loading. In case of 5 wt%
Pd/γ-Al O , only broad peaks of γ-Al O (JCPDS File No.
peaks with temperature maxima (T ) at 108 and 125°C,
max
0
these peaks could be ascribed to the reduction PdO to Pd
of smaller and larger Pd particles, respectively. An increase
in the low temperature peak area with increasing Pd load-
ing on MgO can be attributed to the higher hydrogen con-
sumption by 5 wt% Pd particles of larger size. Hydrogen
consumption is increased from 419 to 934 μmol/g when Pd
loading is increased from 1 to 5 wt% on MgO. It is inter-
esting to note that the amount of hydrogen consumed for
2
3
2
3
1
0-0425) are observed and no peaks of PdO or Pd are evi-
dent. The absence of XRD peaks of Pd or PdO loaded on
γ-Al O may be due to smaller crystallites size, which are
2
3
1
3

Decomposition of Methyl Formate over Supported Pd Catalysts
1289
Fig. 2 TEM images of a MgO, b 1% Pd/MgO and c 5% Pd/MgO
Fig. 3 TEM images of 5% Pd on a MgO, b α-Al O and c γ-Al O
3
2
3
2
2
+
reduction of Pd [20]). As pure MgO did not showed any
hydrogen consumption at <400°C (data not given) but
showed 27 mmol/g hydrogen consumption at 467°C. The
higher hydrogen consumption i.e. 419 mmol/g for 1 wt%
Pd/MgO could be due to the hydrogen spillover from
metal to support [21, 22]. This data shows that smaller and
highly dispersed Pd particles reduced at higher tempera-
ture i.e. 120°C, than larger particles. This could be due to
high electronegative surface oxygen species of basic metal
oxide i.e. MgO, which surrounded Pd particle [23]. This
metal-support interactions or highly coordinated Pd spe-
cies caused more resistance to reduce to their metallic state
compared to 5 wt% Pd/MgO. This behavior is predominant
with smaller Pd particles. Pd supported on γ- and α-Al O
2
3
showed only negative reduction peaks and these peaks are
due to decomposition of Pd-hydride (β-PdHx) species at 69
and 80°C, respectively [24, 25]. The negative TPR peaks
on γ-Al O and α-Al O are well reported for these samples
Fig. 4 TPR profiles of supported Pd catalysts
2
3
2
3
[
18, 26]. At ambient temperature in hydrogen atmosphere,
reduction of PdO is much higher than the stoichiometric
amount needed for the reduction (H/Pd=2). Theoretically,
the hydrogen uptake for 1 wt% Pd based catalyst is around
species like PdO gets reduced to metallic Pd, which further
interacts with hydrogen to form β-PdHx [27] temperature
program. The intensity of this peak is more on 5 wt% Pd/α-
Al O , which may be attributed to lower Pd dispersion
9
4 mol per gram of catalyst (attributed mainly due to the
2
3
1
3

1
290
S. R. Akuri et al.
−
1
in this catalyst [18]. The TPR patterns of Pd/MgO fresh
catalysts are found to be complex unlike the patterns of
Pd/Al O catalysts. No negative peak corresponding to
band seen at 870 and 1100 cm could be assigned to
Pd–C–O linkages. When the catalyst is calcined and there-
after used in the reaction the Pd atoms remain associated
with the support. This metal-support interaction has an
important role to play in the MF decomposition. On the
other hand, on the Pd/γ-Al O surface, the surface hydroxyl
2
3
β-PdHx decomposition is observed over Pd/MgO. The rea-
son for this could be high dispersion of Pd or decrease in
Pd crystallite size as reported on Pd/SiO which showed
2
2
3
decrease in the intensity of β-PdHx decomposition peak
groups were found to decrease significantly with respect
[
25]. Palladium loaded on alumina’s released much less
to the pure γ-Al O . Figure 3 has also the FTIR spectra for
2
3
hydrogen from PdHx than expected theoretically i.e. 470
μmols for 5 wt% Pd. These behaviors can be attributed to
both metal particle size and interactions with the supports.
Palladium on α-Al O is expected to have weak interaction
pure γ-Al O and Pd/γ-Al O samples, where no additional
2 3 2 3
peaks were found for Pd/γ-Al O unlike Pd/MgO.
2
3
Figure 6 shows Pd core level X-ray photoemission spec-
tra of 5- wt% Pd on MgO, γ-Al O and α-Al O catalysts.
2
3
2
3
2
3
and formed large particles which is capable to store more
hydrogen [28, 29], this is also supported by TPR results
with higher quantity of hydrogen released from PdHx i.e.
The Pd 3d signal was probed with Al Kα radiation on
γ-Al O and α-Al O while Mg Kα radiation was used for
2
3
2
3,
MgO supported Pd due to superimposition of Pd 3d signal
1
87 μmol/g. While Pd on γ-Al O with smaller particles
with the Mg KLL Auger peaks between 300 and 350 eV
2 3
0
size compared to α-Al O could store less hydrogen i.e.
[31]. Palladium on γ-Al O exhibited 55% Pd (Pd 3d
2
3
2
3
5/2
2
+
7
7 μmol/g [30] as shown in Fig. 4.
BE: 335.1 eV) and 45% Pd (Pd 3d BE: 337.1 eV) spe-
5/2,
0
Figure 5 shows the infrared spectra of pure MgO and
wt% Pd/MgO taken at 200°C in vacuum. At this tempera-
cies. Similarly, Pd on α-Al O also exhibited 19% in Pd
2 3
2
+
5
(Pd 3d BE: 335.3 eV) and 81% in Pd (Pd 3d BE:
5
/2
5/2
ture, it is expected that the physisorbed moisture is removed
and ideally the medium and high strength surface hydroxyl
groups are present. On MgO at 200°C surface carbonates
336.9 eV). Interestingly, unlike in case of alumina sup-
ports, very weak signal is observed for Pd on MgO, despite
observing XRD peaks corresponding to PdO (Fig. 1). The
difference in the nature of Pd species on MgO and alu-
mina supports can be explained based on the support tex-
tural properties, Pd particle size and metal dispersion and
absorbed from the atmospheric CO and binding to the sur-
2
−
1
face are observed in the region of 1300–1500 cm . The
carbonates typically desorb at temperatures above 350°C.
These carbonates and hydroxyl groups can bind to the Pd
ions through complete or partial ligand exchange. Thus Pd
ions anchored onto the MgO surface lead to the formation
of Pd–O–H or Pd–C–O type of linkages. Pd–O bonds are
typically observed in the lower wavenumber regions and
were not detected in the Pd/MgO sample. The additional
−
e donor sites of the support. The absence of XPS signal
for Pd could be due to nonexistence of Pd species on MgO
surface as it exhibited larger pores (pore dia ~240 Å) unlike
in case of alumina supports where Pd is located on the sur-
face due to smaller pore size (pore dia ~100 Å,,Table 1).
Although Pd core level peaks on MgO were not detected
by XPS, oxidation–reduction behavior of Pd, the unique
metal-support interactions well known in the literature
[
32–34] that electron rich Pd particles are formed due to
−
e donor sites of MgO resulting formation of electron rich
Pd particles formation [31].
3
.2 Methyl Formate Decomposition
The MF conversion data on pure MgO and Pd loaded
MgO catalysts at 150°C and 1 bar is represented in Fig. 7.
These results show that pure MgO does not decompose
MF at temperatures <200°C, however, on incorporating
1
wt% Pd on MgO, about 40% MF conversion to CO and
MeOH is seen even at 150°C. Hence the MF decomposi-
tion on Pd/MgO can be attributed due to the presence of
palladium [13]. However, further increase in Pd loading to
5
wt% showed a reduced MF conversion of 33% under the
same experimental condition. The higher activity of 1 wt%
Pd/MgO compared to 5 wt% Pd loading can be attributed
to PdO crystallites of smaller size as observed and TEM
results (Fig. 2). Furthermore, higher activity of 1 wt% Pd/
Fig. 5 FTIR spectra of MgO, Pd/MgO, γ-Al O and Pd/α-Al O cat-
2
3
2
3
alysts in vacuum at 250 °C
1
3

Decomposition of Methyl Formate over Supported Pd Catalysts
1291
Fig. 6 X-ray photoemission spectra of 5-wt% Pd on MgO, γ-Al O and α-Al O catalysts
2
3
2
3
Fig. 8 Methyl formate (MF) decomposition over 5 wt% Pd cata-
Fig. 7 Effect of Pd loading over MgO on methyl formate (MF)
lysts over different supports (reaction conditions: 1 g of catalyst, 150;
decomposition (reaction conditions: 1 g of catalyst, 150 °C; 1 atm;
−1
−
1
1
atm; GHSV=3100 h ; feed composition of MF=47.2%, metha-
GHSV=3100
h ; feed composition of MF=47.2%, metha-
nol=1.29%, balance N2)
nol=1.29%, balance N2)
However, to understand the role of the support on Pd
activity, supports with different partial charge of oxide
MgO could also be explained by the interaction of Pd with
MgO, which is basic in nature due to high partial charge of
surface oxide ions [21, 22]. In addition, TEM showed that
average particle size of 1 wt% Pd is 20 nm, while 5 wt% Pd
formed the size of 40 nm on MgO (Fig. 2). Furthermore
the role of support and metal interaction cannot be ruled
out towards higher activity for 1 wt% Pd/MgO. This inter-
action is also evident from the TPR results where reduc-
tion peak of PdO is broad from 1 wt% Pd/MgO. It can be
explained by the smaller Pd particles well dispersed on
MgO. Whereas in case of 5 wt% Pd/MgO lower activity for
MF decomposition is due to larger metal particle size.
ions like γ-Al O and α-Al O have been investigated with
2
3
2
3
5
wt% loading. Figure 8 shows the effect of the support on
MF decomposition over Pd catalysts. Palladium with 5 wt%
on γ-Al O and α-Al O showed MF conversions i.e. 16
2
3
2
3
and <1%, respectively. These results indicate that the sup-
ports are playing significant role on catalytic activity. The
role of the support on the catalytic activity can be explained
based on the fact that surface hydroxyl groups and surface
charges influence the Pd active sites. The TPR profiles of
Pd supported on MgO, γ-Al O and α-Al O revealed that
2
3
2
3
1
3

1
292
S. R. Akuri et al.
these interactions are different based on acid–base strengths
of the support [35, 36]. The support–metal interactions can
well reflect in the TPR profiles. The TPR (Fig. 2) results
show that the Pd with 5 wt% forms strong interaction with
MgO as is evident from the higher temperature required
for reduction compared to Pd supported on γ-Al O and
2
3
α-Al O . Based on the TPR data the order of metal-support
2
3
interactions is Pd/MgO>Pd/γ-Al O >Pd/α-Al O . The
2
3
2
3
catalytic activity of these catalysts for MF decomposition
also follows the same trend as shown in Fig. 8.
Figure 9 shows MF decomposition over Pd/MgO cata-
lyst, which showed formation of CO and MeOH predomi-
nantly, while Pd/γ-Al O produced dimethyl ether in addi-
2
3
tion to MeOH and CO. These results suggest that the MF
decomposition follows simple decarbonylation of MF to
form CO and MeOH on both 5 wt% Pd on MgO or γ-Al O
Fig. 9 Effect of different support over MeOH/CO ratio in methyl
formate (MF) decomposition (reaction conditions: 1 g of catalyst,
2
3
−1
1
50 °C; 1 atm; GHSV=3100 h ; feed composition of MF=47.2%,
[
2, 10]. The MeOH formed undergoes further dehydra-
methanol=1.29%, remaining N2)
tion on acidic support i.e. γ-Al O to form DME. The MF
2
3
decomposition products, MeOH and CO are formed in the
ratio of 1.1 over Pd/MgO catalyst. Since, the DME forma-
tion is a consecutive reaction from MeOH, the nature of
support is playing an important role than the Pd. Hence,
the mechanism of MF decomposition on these catalysts is
similar to the one observed on base catalysts [2]. It can be
inferred that the metal function catalyzes MF decomposi-
tion at lower temperatures i.e. 150°C compared to metal
oxide catalyst which otherwise requires higher activation
temperature. In addition, the nature of the support influ-
ences the activity of Pd active sites. However, the details of
Pd–support interactions, palladium size, the support–metal
interaction and nature of the support are critical for chemi-
cal transformations [37, 38]. XPS data (Fig. 6) revealed that
between Pd- and α-Al O that resulted in high concentra-
2
3
tion of PdO on the surface Fig. 9. However, compared to
α-Al O support, Pd/γ-Al O showed higher MF decompo-
2
3
2
3
sition up to 16 mol% due to higher concentration of surface
Pd metallic species due to support metal interactions.
4
Conclusions
Catalytic decomposition of MF to produce CO and MeOH
has been observed over supported Pd catalysts at 150°C
compared to pure MgO catalysts which decompose MF at
>200°C. Higher MF decomposition upto 40 mol% could be
achieved with 1 wt% Pd loading on MgO, which is attrib-
uted to high metal dispersion and strong metal-support
0
2+
palladium exists in the form of Pd and Pd on γ-Al O
2
3
0
and α-Al O and the surface concentration of Pd is higher
2
3
on γ-Al O than α-Al O . The MF decomposition on the
2
3
2
3
−
interaction mainly with e donor sites on the MgO sup-
supported Pd catalysts can be correlated with the oxidation
states. These observations corroborate that palladium in
metallic state is more active than oxide for decomposition
of MF, as 5-wt% Pd/γ-Al O showed higher activity than
port resulting in more metallic Pd species which could be
responsible for High MF decomposition. Palladium loading
and the surface properties of oxide support plays significant
role on MF decomposition. Acid–base property of the sup-
port determines the product distribution and acidic support
formed the by-product, DME by dehydration of MeOH.
Both Pd/MgO or Pd/γ-Al O occurs through decomposition
2
3
5
-wt% Pd/α-Al O due to higher concentration of Pd in
2 3
metallic form in case of 5-wt% Pd/γ-Al O . In case of Pd/
2
3
MgO, although Pd oxidation state for could not be obtained
by XPS due to absence of surface Pd species, several theo-
retical and experimental studies [39, 40] dedicated to the
interaction of Pd and MgO showed that Pd exists more in
the metallic form due to dominant e− donor sites present
on MgO. Metallic Pd is formed due to charges transfer from
the support to the Pd atom. Hence, the basic sites on MgO
support and e− donor sites enable Pd in metallic form in
higher concentration compared to alumina supports, result-
ing in higher activity for MF decomposition.
2
3
of MF to CO and MeOH, while Pd on α-Al O shows pure
2
3
activity. Formation of β-PdHx is prominent when the inter-
actions of metal and the support is weak, which probably
may not form active Pd for MF decomposition. Hence, sup-
ported Pd catalysts with about 1 wt% loading on basic sup-
ports can be potential catalysts for in-situ CO generation for
carbonylation reactions at low pressure.
Acknowledgements The authors thank Mr. Aurn Sikder for co-
coordinating XPS characterization of the catalysts for this study and
Dr. G Srinivasarao for his input in interpretation of the XPS data.
Catalytic decomposition of MF is negligible on 5 wt%
Pd/α-Al O , this could be mainly due to weak interactions
2
3
1
3

Decomposition of Methyl Formate over Supported Pd Catalysts
1293
2
2
2
2. Annalisa D, Elio G, Cesare P (1999) J Phys Chem B
References
1
03:3872–3876
3. Largentiere PC, Canon MG, Figoli NS (1995) Appl Surf Sci
1
2
3
4
5
.
.
.
.
.
Lee JS, Kim JC, Kim YG (1990) Appl Catal A 57:1–30
Gerard J (1995) Appl Catal A 57:25–44
8
9:63–68
4. Sai Prasad PS, Lingaiah N, Chandrashekar S, Rama Rao KS,
Kanta Rao P, Raghavan KV, Berry FJ, Smart LE (2000) Catal
Lett 66:201–204
Dake SB, Raghunath VC (1985) J Chem Eng Data 30:400–403
Veleckis E, Hacker DS (1984) J Chem Eng Data 29:36–39
Mohammad SY, Giulio L, Aurel W, Leslaw M, Thomas EM,
Walter L (2013) Green Chem 15:1146
2
5. Fagherazzi G, Beneditti A, Polizzi S, Di Mario A, Pinna F,
Signoretto M, Pernicone N (1995) Catal Lett 32:293–303
6. Victor HS, Carlos EG (1996) Appl Catal A 148:81–96
7. Stobbe-Kreemers AW, Soede M, Makkee M, Scholten JJF (1995)
Appl Catal A 131:367–381
6
.
Tsumoru M, Kiyomi K (2004) Angew Chem Int Ed Engl
2
2
4
3:5580–5588
7
8
.
.
Fuat EC, Henry L, Alexis TB (2008) J Mol Catal A 288:87–96
Hendriksen DE (1983) Prepr Pap-Am Chem Soc Div Fuel Chem
2
2
8. Bhogeswararao S, Srinivas D (2015) J Catal 327:65–77
9. Seshu Babu N, Lingaiah N, Sai Prasad PS (2012) Appl Catal B
2
8:176–190
9
.
Patrick P, Marc F, Yves C, Andre M, Francis P (1996) Appl
Catal A 135:329–339
1
11–112:309–316
3
3
0. Fernando C-L, Mark AK (2014) Appl Catal A 473:41–50
1
1
1
1
1
0. Joseph SF (2003) J Am Chem Soc 125:10475–10480
1. Claus P, Berndt H, Mohr C, Radnik J, Shin EJ, Keane MA
1. Lee SJ, Lee KH, Lee JS (1992) Appl Catal A: Gen 83:165–178
2. Fu-Qin M, Da-Shun L, Zhi-Ying G (1992) J Catal 134:644–653
3. Takashi U, Hideshi H, Kozo T (1984) Chem Lett 13:649–652
4. Fu-Qin M, Da-Shun L, Zhi-Ying G (1993) J Mol Catal
(
2000) J Catal 192:88–97
3
3
2. Ketteler G, Ogletree DF, Bluhm H, Liu H, Hebenstreit EL, Salm-
eron M (2005) J Am Chem Soc 127:18269–18273
3. Sterrer M, Risse T, Pozzoni UM, Giordano L, Heyde M, Rust
HP, Freund HJ (2007) Phys Rev Lett 98:096107
7
8:309–325
1
1
1
1
1
2
2
5. Damyanova S, Pawelec B, Arishtirova K, Fierro JLG, Sener C,
Dogu T (2009) Appl Catal B 92:250–261
3
3
3
3
4. Kim SM, Lee H, Park JY (2015) Catal Lett 145:299–308
5. Lercher JA (1982) React Kinet Catal Lett 20:409–413
6. Scirè S, Minicò S, Crisafulli C (2002) Appl Catal A 235:21–31
7. Flank AM, Delaunay R, Lagarde P, Pompa M, Jupille J (1996) J
Phys Rev B 53:R1737–R1739
6. Marina MVMS, Donato AGA, Martin S (2001) J Catal
2
04:498–511
7. Peter C, Heinz B, Christian M, Jorg R, Eun-Jae S, Mark AK
(2000) J Catal 192:88–97
3
3
4
8. Matveev AV, Neyman KM, Yudanov IV, Rösch N (1996) Surf
Sci 426:123–139
8. Hari Padmasri A, Venugopal A, Krishnamurthy J, Kishan G,
Niemantsverdriet JW (2002) J Phys Chem B 106:1024–1031
9. Silvia FM, Ruth LM, Antje O, Edward LK, Malte B, Martin S
9. Ferrari AM, Xiao C, Neyman KM, Pacchioni G, Rösch N (1999)
Phyl Chem Chem Phys 1:4655–4661
(2012) Appl Catal A 411–412:105–113
0. Freund HJ, Pacchioni G (2008) Chem Soc Rev 37:2224–2242
0. Huaqing Z, Zhangfeng Q, Wenjuan S, Wenjie S, Jianguo W
2004) J Catal 225:267–277
1. Claus P, Berndt H, Mohr C, Radmik J, Shin E-J, Keane MA
2000) J Catal 192:88–97
(
(
1
3