Direct Methyl Formate Formation from Methanol over Supported Palladium Nanoparticles at Low Temperature

Article abstract of DOI:10.1002/cctc.201200325

Palladium nanoparticles of different sizes and supported on TiO₂ (2wt.%) were synthesized by using the water-in-oil microemulsion method. The control of palladium nanoparticles was investigated in terms of the nature of organic surfactant and solvent. Attention was paid to the reduction of palladium ions in solution during synthesis. Materials were characterized by N₂-BET at low temperature, XRD, XPS, H₂ chemisorption, and TEM and were tested in the gas phase oxidation of methanol. The results confirmed the effect of microemulsion composition on the size of palladium nanoparticles. The direct formation of methyl formate from methanol was observed. Supported palladium catalysts produced methyl formate at low temperature (<150°C) with a very high selectivity, and, in some cases, a selectivity of approximately 100%. At higher temperature, methyl formate is not formed at all and the total oxidation to CO₂ occurs. A linear correlation between palladium particle size and methanol conversion was observed. Conversion and selectivity were correlated with the acidity of the catalysts. Very small particles were more active but less selective. A mechanism was proposed to explain the formation of methyl formate from methanol.

Full text of DOI:10.1002/cctc.201200325

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201200325
Direct Methyl Formate Formation from Methanol over
Supported Palladium Nanoparticles at Low Temperature
Roberto Wojcieszak,* Eric M. Gaigneaux, and Patricio Ruiz[a]
Palladium nanoparticles of different sizes and supported on
TiO2 (2 wt.%) were synthesized by using the water-in-oil micro-
emulsion method. The control of palladium nanoparticles was
investigated in terms of the nature of organic surfactant and
solvent. Attention was paid to the reduction of palladium ions
in solution during synthesis. Materials were characterized by
N2-BET at low temperature, XRD, XPS, H2 chemisorption, and
TEM and were tested in the gas phase oxidation of methanol.
The results confirmed the effect of microemulsion composition
on the size of palladium nanoparticles. The direct formation of
methyl formate from methanol was observed. Supported palla-
dium catalysts produced methyl formate at low temperature
(<1508C) with a very high selectivity, and, in some cases, a se-
lectivity of approximately 100%. At higher temperature,
methyl formate is not formed at all and the total oxidation to
CO2 occurs. A linear correlation between palladium particle
size and methanol conversion was observed. Conversion and
selectivity were correlated with the acidity of the catalysts.
Very small particles were more active but less selective. A
mechanism was proposed to explain the formation of methyl
formate from methanol.
Introduction
The preparation and use of advanced catalytic materials is at
the heart of new catalytic processes and forms the basis for
the development of advanced catalytic technologies. Catalytic
nanoparticles present high catalytic performances because of
their high surface area/volume ratio. However, these high per-
formances are not related exclusively to the high surface area
of the nanoparticles. They can be related to the new properties
developed by virtue of the nanocharacter of the catalytic parti-
cles. The novel features acquired on the nanodimensional
regime cannot be a mere result of scaling factors of the sur-
face area and volume ratio. Some structures of the nanosize di-
mension show new behaviors that are not demonstrated by
the bulk solids of the same composition. If the size of the
structure is decreased, the surface area/volume ratio increases
considerably and the surface properties predominate over the
bulk ones. New effects can appear and play an important role
that can be related to quantum mechanics. Novel physical (op-
tical, magnetic, and electronic) and chemical (catalytic, adsorp-
tion, activation of chemical species, dynamics processes on the
surface, kinetics, etc.) properties and/or reactivity that is char-
acteristic of the nanosize dimension of the catalytic particles
are expected to emerge.[1,2]
and of the operating reaction conditions. At present, there is
a lack of literature about the correlation of chemical and physi-
cal properties of nanoparticles with their catalytic activity. This
work is the first one of a series, and its objective is to correlate
the chemical, structural, surface, and textural properties of the
catalysts with the activity and size of the particles. It is expect-
ed to put in evidence that the catalytic performances of the
studied nanoparticles are directly related to the size and prop-
erties of metal nanoparticles. In a second paper, the relation-
ship between the catalytic properties and the adsorbed spe-
cies will be studied by using operando DRIFT spectroscopy.
The study of the physical properties of the synthesized cata-
lysts is in progress.
To reach our objectives, particular strategies of preparation
of supported metal catalytic nanoparticles are mandatory. The
generation of metal nanoparticles requires mainly the develop-
ment of advanced methods. One of the difficulties is that
metal colloidal particles with a narrow size distribution are not
stable.[3–5] Nanoclusters must be stabilized against aggregation
into bigger particles and, eventually, into the bulk material.
Metal nanoparticles can be obtained by mixing of two water-
in-oil microemulsions: one containing a salt or a complex of
the metal and the other containing a reducing agent.[6] The
synthesis of nanoparticles in microemulsions allows one to
obtain the monodisperse size of the particles and, in some
cases, to control the size of the particles by variation of the
size of the microemulsion droplet radius and of the precursor
concentrations. Metal cations constrained in the water droplets
can react with the reducing agent to form the metal nanopar-
ticles. Hydrazine is an appropriate reducing agent because it
can be used at low temperature and it decomposes into am-
monia and/or nitrogen if reduction takes place. Reverse micro-
The principal objective when working with catalytic nano-
particles must be to correlate chemical and physical properties
and reactivity with each other. These new properties and reac-
tivity depend on the nature of the metal particles and supports
[a] R. Wojcieszak, Prof. E. M. Gaigneaux, P. Ruiz
Institute of Condensed Matter and Nanosciences (IMCN)
Division “Molecules, Solids and Reactivity (MOST)”
Universitꢀ catholique de Louvain
Croix du Sud, 2/17, 1348 Louvain la Neuve (Belgique)
E-mail: robert.wojcieszak@uclouvain.be
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 339

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
octanol, 1-decanol, and n-octane (99.5%) were purchased from
Fluka. Hydrazine hydrate solution (80%) was purchased from Fluka
(99%). All reagents were used without further purification.
emulsion is an isotropic liquid medium with nanometer-sized
water droplets dispersed in a continuous oil phase and stabi-
lized by surfactant molecules at the water/oil interface. For the
preparation of well-dispersed metal nanoparticles in the 1–
10 nm size range, the use of surfactants is absolutely needed.[6]
Nonionic surfactants and anionic surfactants are effective in
the preparation of catalysts with a narrow size distribution. In
nonionic surfactants, the metal particle size in the range of 3–
8 nm, according to the structure of surfactants and organic sol-
vents, can be obtained.[7] The carbon number of the organic
solvent n-alcohol is also an important factor to control the size
of metal particles. The metal particle size increases with the in-
crease in the carbon number of n-alcohol. The smallest metal
particles were obtained if 1-butanol was used as an organic
solvent.[7] Thus, very small metal particles could be obtained
by using the microemulsion method. One disadvantage of mi-
croemulsion is the presence of organic surfactants on the
metal or support surface after reduction. It is argued that
these organic compounds can hide active metal from the reac-
tant during the heterogeneous catalytic reaction.
Catalysts were prepared as follows:
1) Precursor solution: An appropriate quantity of PdCl2 was dis-
solved in distilled water (5 mL) in the presence of NaCl (99.5%;
Merck). Then, the solution was evaporated at 608C until only
a volume of 1 mL remained.
2) Microemulsion: The microemulsion was formed by using an or-
ganic solvent (50 mL), a precursor solution (1 mL), and the appro-
priate quantity of the organic surfactant. The microemulsion was
heated to 508C, and then TiO2 (2 g) was added to the reactant
flask under magnetic stirring. After 30 min, the appropriate quanti-
ty of hydrazine was injected. The reduction was performed in
a thermostated water bath at 508C. The reduction time of 30 min
was measured from the hydrazine injection to the microemulsion.
The reaction was performed under N2 atmosphere. The reactant
solution changed color from light red to black, which indicated the
reduction of palladium. The solution was then filtered, washed
with acetone (99%, VWR) and distilled water for 1 h, and dried at
1008C for 30 min. Reaction conditions are listed in Table 1. These
catalysts were denoted as Pd/TiO2(X/Y), in which X corresponds to
the organic surfactant and Y to the organic solvent used in the
preparation of microemulsion.
Herein, palladium nanoparticles were prepared with a range
of the metal particle size by using the microemulsion method
with a chemical reduction in the synthetic solution. The sup-
ported palladium nanoparticles were prepared by using the
water-in-oil microemulsion method. The different organic sol-
vents, such as n-alcohols, n-octane, and cyclohexane, were
used to form the oil phase. The
To compare with the catalysts, the bare TiO2 was also treated with
hydrazine under the same conditions as those for the catalysts. It
is denoted as TiO2–N2H4.
anionic (sodium bis(2-ethylhex-
Table 1. Amount of H2O, organic solvent, and surfactant as well as surfactant/Pd, H2O/surfactant, and N2H4/Pd
ratios used in the synthesis.
yl)sulfocuccinate [AOT]) and
nonionic (ethylene glycol mono-
lauryl ether [Brij30]) surfactants
were used as stabilizing agents.
Hydrazine was used as the re-
duction agent.
Pd/TiO2(organic surfactant/organic solvent) H2O Organic solvent Surfactant/Pd H2O/surfactant N2H4/Pd
[mL] [mL]
[mol/mol]
[mol/mol]
[mol/mol]
Pd/TiO2(AOT/cyclohexane)
Pd/TiO2(AOT/1-butanol)
Pd/TiO2(AOT/1-octanol)
Pd/TiO2(AOT/n-octane)
Pd/TiO2(Brij30/1-decanol)
2
2
2
2
2
50
50
50
50
50
40
40
40
40
40
6
6
6
6
6
100
100
100
100
100
An important aspect to con-
sider is the reaction tempera-
ture. To preserve the nanosize
character of the particles, a low-
temperature reaction was stud-
Characterization of physical and chemical properties
ied, namely, the oxidation of methanol to methyl formate in
the presence of oxygen. Supported palladium nanoparticles
are very active and used as selective catalysts in the oxidation
of methanol to methyl formate in the presence of oxygen, but
only at low temperature. At higher temperature, methyl for-
mate is not formed at all and the formation of CO2 is promot-
ed. Catalysts were characterized by using several physicochem-
ical processes (chemical analysis, N2 physisorption, H2 chemi-
sorption, XRD, XPS, and TEM).
Chemical analysis
Inductively coupled plasma-atomic emission spectroscopy analysis
was used to determine the chemical composition of the prepared
catalysts, and it was performed on a Thermo Jarrell ASH IRIS Ad-
vantage analyzer.
H2 chemisorption
H2 chemisorption at RT was used for the determination of disper-
sion and average size of palladium particles supported on TiO2.
The amounts of chemisorbed hydrogen were determined at 358C
on a Micrometrics device. The following method was applied in
the measurements: preliminary evacuation and heating at 3008C
for 1 h under a He flow; hydrogen introduction at 3008C for 1 h
and then evacuation at 3008C for 2 h, cooling down to 358C,
again evacuation at 358C for 1 h; before hydrogen was introduced,
the absence of leak was ascertained and finally the measurement
of hydrogen uptake was performed at 358C. A double adsorption
Experimental section
Catalysts preparation
Palladium(II) chloride (from Aldrich) was used as the palladium pre-
cursor and TiO2 (Merck; BET area 7.6 m2 gꢀ1) as the support. We
used low surface titania to obtain surface deposition of metal
nanoparticles. The organic surfactants AOT and Brij30 (99.9%) were
purchased from Sigma. Organic solvents cyclohexane, 1-butanol, 1-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 340

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
method was used: 1) a first adsorption isotherm was measured,
which includes both physisorption and chemisorption; 2) after
a 2 h outgassing at 358C, a second isotherm was measured, which
includes physisorption only. Both isotherms are determined in the
pressure range of 0.5–700 mmHg. The difference between first and
second isotherms gave the H2 chemisorption isotherm.[8] By assum-
ing a spherical shape, an average crystallite size d was calculated
by using the irreversible adsorption isotherm of hydrogen, accord-
ing to the following equation: d=6000/S1, in which S is the cata-
lyst surface area (of the fraction of reduced palladium) and 1 is the
palladium density.[8] S was calculated by assuming an adsorption
stoichiometry H/Pd of 1 and a surface area of 7.93 ꢁ2 per palladium
atom.
XPS spectroscopy was also used to estimate the size of palladium
nanoparticles by using the Davis method.[10] This method is very
useful for the study of very small metal particles (1–5 nm). In this
method, the particle size is obtained from the experimental inten-
sity ratio for two core levels with different kinetic energies arising
from the same dispersed phase. The advantage of this method is
a certain independence from the physical properties of the sam-
ples, such as metal loading and specific surface area of the sup-
port. Calculations are described in detail in References [10] and
[11].
Electron microscopy
TEM images were recorded on a LEO 922 Omega TEM operating at
an accelerating voltage of 200 kV after placing a drop of the cata-
lyst suspension on a carbon-coated copper grid. The samples were
removed from the reactor and stored in small glass containers
until TEM inspection. The catalyst suspension was obtained by ul-
trasonically dispersing the sample in butanol.
X-ray diffraction
XRD analysis was performed on a Siemens D5000 X-ray diffractom-
eter with the CuKa radiation (l=1.5418 ꢁ). The 2q range was
scanned between 2 and 658 (scanning rate: 0.018sꢀ1). The phases
were identified by using the ICDD–JCPDS database.
The particle size was estimated from the XRD line broadening anal-
ysis by using the Scherrer formula. This analysis has been used
widely for characterizing supported metal crystallites. A perfect
crystal would extend in all directions to infinity; thus, no crystal is
perfect because of its limited sizes. Such a deviation from perfect
crystallinity leads to broadening of the diffraction peak. The Scher-
rer formula relating the crystallite size (d) to the peak width is d=
Kl/bcosq, in which K is the Scherrer constant, l is the X-ray wave-
length, and b is the integral breadth of a reflection (in rad) located
at 2q.[8]
Catalytic tests
The partial oxidation of methanol was performed in a metallic
fixed-bed microreactor made of an Inconel tube of 1 cm internal
diameter (PID Eng&Tech) operating at atmospheric pressure. The
catalytic bed was composed of the catalyst powder (100 mg) se-
lected within the granular fraction of 200–315 mm and diluted in
glass spheres (600 mg) that were confirmed inactive. The catalyst
was placed in the reactor. Before the catalytic test, the catalyst was
reduced in a flow of pure H2 (Praxair; flow rate: 50 cm3 minꢀ1) at
3008C for 1 h, and then heated (heating rate: 108Cminꢀ1); next,
the sample was cooled to 3008C. After the reduction, a gas mix-
ture containing 5 vol% of methanol and 2.5 vol% of O2 was pre-
pared by passing a He stream through methanol (>99%; Merck)
placed in a saturator maintained at 5.48C. The reactant mixture
N2 physisorption
The textural properties of the support and reduced materials were
studied by means of N2 adsorption/desorption at ꢀ1968C with
a Micromeritics ASAP2000 apparatus. Low-temperature N2 adsorp-
tion isotherms enable the BET calculation of the specific surface
area and the BJH calculation of the pore volume and pore
diameter.
was passed through the reactor with
a total flow rate of
100 cm3 minꢀ1. The standard catalytic test was performed at various
temperatures: 75, 100, 125, 150, 200, 225, 250, and 3008C. The
sample was heated or cooled at a rate of 108Cminꢀ1. Methanol,
oxygen, and the reaction products were each analyzed for 30 min
with a Varian CP-3800 GC equipped with a thermal conductivity
detector and operating at a programmed temperature. CP-Pora-
PLOT Q (25 m; 0.53 mm) and CP-Molsieve 5 ꢁ (25 m; 0.53 mm) col-
umns were used for the separation of the reaction products. The
following molecules were analyzed during the reaction: H2, O2,
CH4, CO, CO2, methanol, methyl formate, and H2O. Any presence of
formaldehyde and formic acid was detected.
X-ray photoelectron spectroscopy
The XPS analysis was performed by using an SSI X probe (SSX 100/
206) spectrometer from Surface Science Instruments. The analysis
chamber was operated under ultrahigh vacuum conditions with
a pressure close to 5ꢂ10ꢀ9 Torr (1 Torr=0.133 kPa), and the sample
was irradiated with a monochromatic AlKa (1486.6 eV) radiation
(10 kV; 22 mA). Charge compensation was obtained by using an
electron flood gun adjusted to 8 eV and by placing a nickel grid
3.0 mm above the sample. The pass energy for the analyzer was
150 eV, and the spot size was approximately 1.4 mm2. For these
measurements, the binding energy (BE) values were referred to the
Cꢀ(C,H) contribution of the C1s peak at 284.8 eV. The surface
atomic concentrations were calculated by correcting the intensities
with theoretical sensitivity factors based on Scofield cross sec-
tions.[9] Peak decomposition was performed by using curves with
an 85% Gaussian type, a 15% Lorentzian type, and a Shirley non-
linear sigmoid-type baseline. The following peaks were used for
the quantitative analysis: O1s, C1s, Ti2p, N1s, and Pd3d. Based on
the XPS analysis, the XPS surface ratio of a given element is de-
fined as the atomic concentration of the element (%) divided by
the atomic concentration of Ti (%).
Catalyst activity was calculated on the basis of methanol conver-
sion (X), yield (Y), and selectivity (S) of methyl formate by using the
following formulas [Eqs. (1)–(3)]:
Conversion of CH3OH :
À
Á
Moles of CH3OHin ꢀ Moles of OHout
ð1Þ
ð2Þ
ð3Þ
X ð%Þ ¼
ꢁ 100
Moles of CH3OHin
Yield of HCOOCH3 :
2 ꢁ Moles of HCOOCH3products
Y ð%Þ ¼
ꢁ 100
Moles of CH3OHin
Selectivity of HCOOCH3 :
Moles of HCOOCH3products
Moles of ðHCOOCH3 þ CO þ CO2Þ
S ð%Þ ¼
ꢁ 100
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 341

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
In the case of palladium reduction in the presence of organ-
ic surfactants, the formation of stabilized palladium complexes
is expected. Moreover, the reaction between the surfactant
and hydrazine can occur. In the presence of anionic AOT sur-
factant, this reaction results in the polymerization of AOT
molecules.[15]
Results
Inductively coupled plasma analysis and N2 physisorption
The results of inductively coupled plasma (ICP) analysis are
given in Table 2. The mass content of palladium is between 1.6
and 2%, which was lower than the nominal expected (2 wt.%).
Structural and morphological
characteristics of palladium
nanoparticles
Table 2. Pd content (ICP), specific surface area (BET) of the prepared catalysts and the support, Pd particle size
from TEM, H2 chemisorption, and XPS study.[a]
XRD analysis shows the pres-
ence of TiO2 in the anatase
structure. The main difficulty in
studying the small metal nano-
particles (<3 nm) is to deter-
mine their exact size. The com-
plications arise because particles
are usually not spherical and
their distribution on the support
is not homogeneous. Moreover,
some techniques are not useful
because of the size limitation. In
Pd/TiO2(organic surfactant/organic solvent) Pd
Specific surface area
Particle size [nm]
H2 chemisorption XPS
[wt.%] [m2 gꢀ1
]
TEM
Pd/TiO2(AOT/cyclohexane)
Pd/TiO2(AOT/1-butanol)
Pd/TiO2(AOT/1-octanol)
Pd/TiO2(AOT/n-octane)
Pd/TiO2(Brij30/1-decanol)
TiO2–N2H4
1.7
8.3
8.0
7.9
8.4
8.0
7.6
<2.0
ꢂ8.0
<2.0
2–4
3–10
–
ꢂ1.0
6.0
1.3 (1.5)
1.8
1.6
2.2
1.6
–
7.7 (5.9)
1.0 (1.3)
1.5 (1.8)
3.9 (2.2)
–
ꢂ1.0
n.r.[b]
8.0
–
[a] Particle size values after reduction with N2H4/H2 are indicated in parentheses; [b] n.r.=no results obtained
for this sample.
However, such small differences are likely due to the prepara-
tion method. After chemical reduction with hydrazine, palladi-
um should precipitate onto the TiO2 surface. Some palladium
stabilized by the organic surfactants is expected to stay in the
solution. However, the reasonable results obtained from ICP
are comparable with those obtained from other methods, such
as precipitation and ion exchange.
XRD study, the applicability is restricted by several factors such
as the weight fraction or the size of crystallites. XRD analysis
indicated that palladium in the metallic state was detected for
all catalysts (Figure 1). However, the very low resolution of
these reflections did not enable the determination of the aver-
age particle size from the well-known line broadening analysis
as proposed by Scherrer.
The specific surface area did not change significantly after
chemical reduction as shown by the BET results in Table 2. The
BET surface area of commercial TiO2 is approximately
7.6 m2 gꢀ1
.
Chemical reduction with hydrazine
The hydrazine reduction has been used to reduce various
metal cations with high efficiency.[7,12–14] The hydrazine reduc-
tion depends on several parameters such as the nature of the
metal, hydrazine concentration, pH, reduction time, and reduc-
tion temperature. In the case of noble metals, such as palladi-
um, this reaction occurs very easily at low temperature. The re-
duction proceeds according to the following reactions
[Eqs. (4)–(5)]:
Pd2þ þ 2 N2H4 þ 2 OHꢀ ! Pd0 þ N2 þ 2 NH3 þ 2 H2O
2 Pd2þ þ N2H4 ! 2 Pd0 þ N2 þ 4 Hþ
ð4Þ
ð5Þ
Figure 1. XRD patterns of the catalysts prepared by using the microemulsion
method.
After hydrazine injection at 508C, the solution became gray
and quickly turned dark, which indicated the metallic palladi-
um formation. The reaction occurred very quickly and passed
However, TEM analysis did enable the observation of very
small palladium particles. Palladium nanoparticles were not
visible on TEM images for Pd/TiO2(AOT/cyclohexane) (Fig-
ure 2a) and Pd/TiO2(AOT/1-octanol) catalysts, although the sur-
face analysis (XPS; see below) confirmed the presence of palla-
through the formation of
a
very unstable complex
[Pd(N2H4)3]2+. This complex could be formed via the substitu-
tion of water ligands by hydrazine ligands, which quickly de-
composed to metallic palladium and gaseous nitrogen.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 342

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
particle size have been studied in the past decades.[16]
The XPS study can determine the chemical composition and
oxidation degrees of components present on the catalyst sur-
face. A very low contamination with carbon was observed in
all samples. Moreover, the results confirmed that the organic
surfactant was removed from the catalyst surface after filtra-
tion and washing steps.
In all catalysts, the XPS analysis showed a peak correspond-
ing to the presence of nitrogen on the catalyst surface after
the hydrazine reduction. This peak could originate from nitro-
gen from hydrazine adsorbed on the surface. The value of
400 eV for N1s was assigned to nitrogen originating from hy-
drazine adsorption on the oxide surface.[13,17–19] The amount of
hydrazine introduced in the synthesis was the same in all prep-
arations. The quantity of nitrogen adsorbed was the highest
for the catalyst prepared with AOT/1-butanol and the lowest
for the catalyst prepared with AOT/cyclohexane (Table 3). For
the catalysts prepared with Brij30/1-decanol, AOT/n-octane,
and AOT/1-octanol, the amounts of nitrogen are intermediate
(Table 3). The increase in the adsorbed nitrogen quantities is
related to differences in the acidity of the surface. Moreover,
a very good linear correlation was found between the nitrogen
adsorbed on the surface and the palladium particle size.[11]
The XPS spectra of the Pd 3d region (Figure 3a) and the
Auger M45N45N45 peak (Figure 3b) for the Pd/TiO2(AOT/n-
octane) catalyst is shown in Figure 3. The results confirm the
presence of palladium in the metallic state (Pd0) at a BE of
334 eV. Oxidized palladium (Pd+) was also detected, which
confirmed that in all catalysts palladium is not reduced fully
(BE=337 eV).
Figure 2. TEM images of a) Pd/TiO2(AOT/cyclohexane), b) Pd/TiO2(AOT/1-bu-
tanol), and c) Pd/TiO2(AOT/n-octane) catalysts after hydrazine reduction.
dium on the surface. After the reduction with hydrazine, the
palladium particles formed were very small (<2 nm).
Metal nanoparticles were evidenced for the Pd/TiO2(AOT/1-
butanol) catalyst. The particles had a mean size of approxi-
mately 6–7 nm. For this sample, palladium was homogeneous-
ly dispersed on the support surface (Figure 2b). In the case of
the Pd/TiO2(AOT/n-octane) catalyst, TEM images showed some
rare and isolated particles of approximately 4–5 nm size (Fig-
ure 2c). However, even if the very small palladium particles (1–
2 nm) were not visible on TEM in this case, they could be seen
on the surface.
The XPS measurement was also performed after the catalytic
test. The results are given in Table 4. The most significant
result is the decrease in the Pd/Ti XPS atomic ratio after the
catalytic test. This decrease suggests the occurrence of the
growth of palladium particles during the catalytic test.
To estimate the average palladium particle size, the Davis
method was applied by using the intensity ratio of two core
levels with different kinetic energies of the dispersed phase.
The results are given in Table 2.[11] The two core levels used to
determine the size of palladium particles are 3d, with the BE
approximately 334 eV (Figure 3a) and the Auger M45N45N45
peak at 1161 eV (Figure 3b). In the case of catalysts reduced
only by hydrazine, the calculations confirmed the presence of
Surface properties
H2 chemisorption was performed on the catalysts reduced as
explained in the section on H2 Chemisorption. The results are
given in Table 2. A very high dispersion (100%) was found for
the catalysts prepared with anionic AOT and 1-octanol and cy-
clohexane as oil phases (Pd/TiO2(AOT/1-octanol) and Pd/
TiO2(AOT/cyclohexane)).
The
anionic surfactant could adsorb
on the palladium surface and
protect the particles from
growth. These very small parti-
cles are evidenced from H2 ad-
sorption data. Discrepancies be-
tween TEM and H2 adsorption
data (Table 2) could be due to
the resolution limit of the elec-
tron microscopy used. The dis-
parities between H2 chemisorp-
tion and physical methods for
the determination of the metal
Table 3. Molar concentrations (%) of the catalysts: PdT total after hydrazine reduction and Pd0/Pd+ XPS atomic
ratios for catalysts reduced by hydrazine, reduced by hydrazine and hydrogen, and after catalytic test.[a]
Catalyst
PdT
N
O
Ti
XPS atomic ratio Pd0/Pd+
N2H4
N2H4/H2
Catalytic test
Pd/TiO2(AOT/cyclohexane)
Pd/TiO2(AOT/1-butanol)
Pd/TiO2(AOT/1-octanol)
Pd/TiO2(AOT/n-octane)
Pd/TiO2(Brij30/1-decanol)
TiO2–N2H4
2.3
1.5
1.1
2.8
1.6
–
0.32
0.64
0.47
0.45
0.39
2.76
33.8
42.5
42.2
39.4
40.9
42.1
19.3
19.9
19.6
17.2
19.7
19.3
2.0
2.2
1.9
1.9
2.1
–
2.2
2.3
1.9
1.9
2.1
–
5.6
5.1
3.4
11.2
2.2
–
[a] Errors of 0.1%.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 343

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Table 4. Pd/Ti atomic ratios for catalysts reduced by hydrazine, reduced
by hydrazine and hydrogen, and after catalytic test calculated from XPS
peaks of Pd3d and Ti2p.
Catalyst
Pd/Ti
N2H4/H2
N2H4
Catalytic test
Pd/TiO2(AOT/cyclohexane)
Pd/TiO2(AOT/1-butanol)
Pd/TiO2(AOT/1-octanol)
Pd/TiO2(AOT/n-octane)
Pd/TiO2(Brij30/1-decanol)
0.12
0.07
0.05
0.17
0.08
0.05
0.04
0.03
0.05
0.05
0.04
0.03
0.04
0.06
0.06
or by an oxidation of palladium after washing. Because the cat-
alysts were activated in H2 at 3008C for 1 h before all catalytic
tests, the XPS study on the catalysts reduced by N2H4 and H2
was also performed (Table 4). The same degree of reduction
was found as for the hydrazine reduction or after the subse-
quent hydrogen method. In contrast, after the catalytic test,
a significant increase in the degree of reduction was observed
for all AOT-prepared catalysts (from 67% to ꢂ80–90%). In the
case of the catalysts prepared with AOT/n-octane, the increase
in the Pd0/Pd+ XPS atomic ratio after the test is clearly marked
(from 1.9 to 11.20). However, this increase was not observed
for the Pd/TiO2(Brij30/1-decanol) catalyst. In this case, the re-
duction degree remained nearly unchanged after the reduction
with N2H4 and H2 and after the catalytic test.
Catalytic activity
During the catalytic tests, only methyl formate, CO2, CO, and
H2O products were detected. TiO2 alone, previously treated or
not in aqueous hydrazine media, was inactive in the oxidation
of methanol at low temperature (<1508C) (Table 5). Some ac-
tivity was, however, observed at high temperature (>1508C).
The only product was CO2, which indicated the unselective
complete oxidation of methanol.
Figure 3. XPS spectra of Pd/TiO2(AOT/n-octane): a) region 3d and b) Auger
M45N45N45 peak at 1161 eV.
the very small palladium particles with a diameter of 0.9 nm
for the Pd/TiO2(AOT/1-octanol) catalyst. Small particles of 1.3
and 1.5 nm size were also obtained for Pd/TiO2(AOT/cyclohex-
ane) and Pd/TiO2(AOT/n-octane) catalysts, respectively. After re-
duction with N2H4/H2 or after the reaction, XPS reveals a sinter-
ing of palladium particles. However, as calculated from XPS
data, the changes in the particle size are small. After the cata-
lytic test the Pd/Ti XPS atomic ratio remains unchanged, which
confirmed the absence of parti-
All Pd/TiO2(X/Y) catalysts showed high activity in the oxida-
tion of methanol (Figure 4). In some cases, catalysts are already
highly active at very low temperature (508C). In the tempera-
ture range 100–1508C, conversion of methanol is practically
100%. Under these conditions, the direct formation of methyl
formate was observed for all catalysts (Figure 5). In addition,
methyl formate was produced with a very high selectivity
(ꢂ90–100%) (Table 5). Only small amounts of CO2 and H2O, as
cle growth because of the high
Table 5. Conversion, selectivity, and yield of samples prepared by using the microemulsion method.
stability of the Pd/TiO2 systems.
Catalyst
Conversion [%]
Yield [%]
Selectivity [%]
508C 808C
508C
808C
508C
808C
Degree of palladium reduction
MF[a]
CO2
MF
CO2
The results obtained by using
XPS are given in Table 4. For hy-
drazine-reduced catalysts, the
average degree of reduction is
approximately 67%. This could
be due to an incomplete reduc-
tion of palladium by hydrazine
Pd/TiO2(AOT/cyclohexane)
Pd/TiO2(AOT/1-butanol)
Pd/TiO2(AOT/1-octanol)
Pd/TiO2(AOT/n-octane)
Pd/TiO2(Brij30/1-decanol)
TiO2–N2H4
37
3
8
9
16
0
78
67
20
45
54
0
29
3
8
9
15
0
40
55
20
41
43
0
78
100
100
100
92
22
–
–
–
8
52
82
100
92
79
–
48
18
–
8
21
–
–
–
[a] MF=methyl formate.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 344

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 4. Methanol conversion versus temperature of reaction for Pd/TiO2
catalysts prepared by using the microemulsion method.
Figure 6. Methyl formate yield versus palladium particle size at 808C for Pd/
TiO2 catalysts prepared by using the microemulsion method.
ane) catalyst), but its selectivity is lower (78% and CO2 is
formed). The lowest conversion is observed for the biggest
particles (5.9 nm in the case of the Pd/TiO2(AOT/1-butanol) cat-
alyst). This catalyst also showed the highest selectivity to
methyl formate (100%) and no CO2 was formed at 508C. This
trend was observed for all catalysts: the smaller the particle
size, the higher the activity but lower the selectivity to methyl
formate.
Discussion
Microemulsion synthetic method
The role of water/surfactant molar ratio
Figure 5. Methyl formate yield versus temperature for Pd/TiO2 catalysts pre-
pared by using the microemulsion method.
The water-in-oil microemulsion is thermodynamically stable. At
the macroscopic scale, a microemulsion seems to be a homo-
geneous solution, but at the molecular scale, it appears hetero-
geneous.[6] The internal structure of a microemulsion, at
a given temperature, is determined by the ratio of its constitu-
ents. The main properties of the emulsion are governed by the
water/surfactant molar ratio. This parameter determines the
size of the final particles. Notably, three kinds of water exist in
the microemulsions. The bulk water localized in the center of
the water core is generally not considered for the water/surfac-
tant molar ratio below 10. Two other kinds of water are consid-
ered in the determination of the water-in-surfactant microe-
mulsions: the bounded water and the trapped water. The
bounded water represents the molecules of water bounded to
the hydrophilic part of surfactants, whereas the trapped water
represents water trapped at the interface in the form of dimers
or monomers.[20] Considering only the presence of bounded
and trapped waters, the water/surfactant molar ratio was 6.
Under these conditions, a very small size of palladium nano-
particles was obtained. The H2 chemisorption method con-
firmed the presence of small (<2 nm) particles of palladium on
the TiO2 surface. This result was also confirmed by using XPS.
Notably, the formation of particles generally proceeds in two
secondary products, were observed. At 50 and 808C, the Pd/
TiO2(AOT/1-octanol) catalyst showed a 100% selectivity to
methyl formate. For other catalysts, the selectivity decreased
with temperature but it still remained high at 808C (between
79 and 92%). Only the Pd/TiO2(AOT/cyclohexane) catalyst
showed low selectivity at 808C (52%).
At high temperature (>1508C), the conversion of methanol
is complete, but methyl formate is not formed at all. At above
1508C, the total methanol conversion was observed. Under
these conditions, only CO2 is observed. Traces of CO were ob-
served at high temperature (3008C).
The maximum methyl formate yield (55% at 808C) was
found for the Pd/TiO2(AOT/1-butanol) catalyst. In this case, the
selectivity to methyl formate reached 82% (67% conversion) at
808C and 100% (5% of conversion) at 508C. The lowest
methyl formate yield observed for the Pd/TiO2(AOT/1-octanol)
catalyst was 20% at 808C. However, at this temperature, the
selectivity to methyl formate reached 100% for this catalyst.
An interesting correlation of the average palladium particle
sizes with the methyl formate yield was observed (Figure 6).
The highest conversion was observed for the smallest palladi-
um particles (1.3 nm in the case of the Pd/TiO2(AOT/cyclohex-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 345

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
steps: the first is the nucleation process inside the droplet and
the second is the aggregation process to form the final parti-
cles.[21] Under the conditions of preparation of nanoparticles
used herein, the aggregation step was minimized, which
allows us to conclude that as the size of the particles was very
low (<10 nm), the effect of agglomeration processes during
the synthesis of metal nanoparticles was very low.
chemical state of the TiO2 support. An important observation
is that the amount of hydrazine adsorbed on the catalyst sur-
face, observed in XPS studies, was different. As hydrazine is
a basic compound, this is a clear indication that the nanoparti-
cles have different acidic–basic properties. Higher the metal
particle size [5.9 nm; Pd/TiO2(AOT/1-butanol)], higher is the
amount of hydrazine adsorbed, higher the acidity, and higher
the selectivity to methyl formate. Lower the metal particle size
[1.5 nm; Pd/TiO2(AOT/cyclohexane)], lower is the amount of
hydrazine adsorbed, lower the acidity, and lower the
selectivity.
The role of surfactants
The presence of organic surfactants controls the rate of the
particle growth; they slow down the nuclei growth, which
favors the formation of homogeneously distributed small-sized
metal particles. If all nuclei formed grow at the same time,
a monodispersed system could be obtained. In addition, the
size of particles depends on the number of nuclei formed at
the very beginning of the reduction. This number is related to
the number of water cores containing enough metal ions to
form stable nuclei.[20] This can explain the difference in particle
size for the different surfactants used in microemulsion cata-
lysts. The biggest particle size (ꢂ8 nm) for the Pd/TiO2(AOT/1-
butanol) catalyst could be due to the polarity of 1-butanol
being different from that of higher alcohols (1-octanol and 1-
decanol) and other solvents (n-octane and cyclohexane). Be-
cause the polarity of n-alcohol became stronger with the de-
crease in the carbon number, 1-butanol had a stronger affinity
for the charged palladium particles. The stabilization of the
palladium nanoparticles would be little affected by the AOT
surfactant.
The results obtained by using the Davis method also al-
lowed us to understand the differences observed between
XRD and TEM studies; for instance, in the case of samples with
nanoparticles polydispersed in size, the Scherrer equation ap-
plied to XRD data gives an indication of the size of the biggest
particles present whereas the Davis method applied to XPS
data gives an indication of the size of the smallest particles
present. TEM analysis revealed all particles, provided enough
particles were observed and accounted for in the measure-
ment statistics. After reduction with H2N2/H2 catalysts (Table 2),
the XPS study also confirmed that in all catalysts palladium
was not reduced fully. The presence of the oxidized palladium
on the surface could be due to the diffusion of oxygen from
the bulk of the palladium particles toward its surface. The pal-
ladium surface is always formed by an oxidized layer, even
after reduction with hydrogen. The Pd/g-Al2O3 catalyst is
always in the PdO form, with its surface in an intermediate
state between surface lacunary PdO and crystalline PdO spe-
cies. This phenomenon has been demonstrated experimentally
by using Raman spectroscopy.[23] It cannot be excluded that
the presence of the oxidized palladium is also due to the air
exposition under washing and drying steps. The increase in
the reduction degree of palladium observed after the catalytic
test was due to the reducing role of methanol.
The choice of an appropriate surfactant is crucial for the
final size of metal particles. The surfactants used represented
two groups: anionic (AOT) and nonionic (Brij30). One expects
that different models of stabilization will be responsible for dif-
ferences in the final particle size or shape. Different effect was
observed in the case of the anionic (AOT) and nonionic (Brij30)
surfactants. In the case of the anionic AOT, two opposite ef-
fects were present. The polar head group of AOT increased the
charge on the oxygen in a similar way as in Brij30, but the
counterion Na+ from AOT exerted an opposite effect. The in-
teraction between palladium and these surfactants was differ-
ent. The balance between both effects in AOT resulted in
a lower hydration of metal cations and, accordingly, a higher
rate constant of complex formation in AOT microemulsion
than in Brij30 microemulsion. In the microemulsion formed
with the nonionic surfactant, the absence of counterion and its
contribution did not reduce the effect of the polar head
group.[22]
A very good correlation was observed between methanol
conversion and nitrogen concentration on the catalyst surface
determined by using XPS (Figure 7). The amount of nitrogen
on the surface depends on the acidity of the surface. The basic
character of hydrazine provided its adsorption on the surface
after the hydrazine reduction. The nitrogen concentration on
the catalyst surface is a function of the acidity of the surface.
The lowest acidity was observed for the Pd/TiO2(AOT/cyclohex-
ane) catalyst, which also showed the highest activity but the
lowest selectivity to methyl formate. Contrarily, the Pd/
TiO2(AOT/1-butanol) catalyst showed the lowest activity and
the highest acidity at 508C.
The most active catalyst is Pd/TiO2(AOT/cyclohexane), which
presents after the catalytic test a Pd0/Pd+ XPS atomic ratio of
5.6. This catalyst is not very selective at low temperature, pres-
ents the lowest acidity, and has the smallest size of palladium
particles (1.5 nm). The highest methyl formate yield was ob-
served for the Pd/TiO2(AOT/1-butanol) catalyst, which showed
a similar Pd0/Pd+ XPS atomic ratio (5.1) but a metal particle
size of 5.9 nm and higher acidity. At higher temperature, the
yield and selectivity of Pd/TiO2(AOT/cyclohexane) catalysts is
higher. Pd/TiO2(AOT/cyclohexane) catalysts form higher
Physical and chemical properties of the supported catalysts
The modification of TiO2 through palladium precipitation–re-
duction does not change the adsorption isotherm, and the BET
surface area practically remains unchanged. Similarly, there was
no change in the BET specific surface area when the pure sup-
port was treated in the presence of N2H4. Moreover, no
changes were observed in the XPS spectra of fresh and hydra-
zine-treated supports. Thus, hydrazine did not change the
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 346

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
The reaction resulting in methyl formate formation involves
a C-O-C coupling. There is still no consensus on the role of the
nature of the active phase, the role of the support and the re-
actant molecules, or the rate-determining step of this reaction.
The only agreement that comes out from the literature data is
that the surface reaction sequence changes in a wide range
with the reaction conditions. Methanol oxidation reactions
lead to formaldehyde (HCHO), dimethoxymethane, and methyl
formate products. Oxidative routes to HCHO are practiced on
silver-based and iron-molybdate catalysts.[24,25] Methyl formate
is produced through (nonoxidative) methanol dehydrogena-
tion on CuO or carbonylation with liquid bases,[24,26–27] and di-
methoxymethane is produced in a two-step process involving
methanol oxidation to HCHO followed by the acetalization of
HCHO–CH3OH mixtures with liquid or solid acids.[24]
The decomposition of methanol on the TiO2 surface starts
with the formation of methoxide species upon adsorption at
room temperature. On the TiO2 surface, the scission of both Cꢀ
O and OꢀH bonds is equally probable.[28,29] If methanol is ad-
sorbed on the surface of palladium, its OH bond weakens, for-
mation of PdꢀO bonds prevails, chemisorbed methoxy species
CH3O are formed, and then it undergoes successive dehydro-
genation to HCHO and to chemisorbed CO and H.[28] In the
first step, methanol adsorbed on palladium promotes the for-
mation of CO at the interface of the TiO2 surface with the pal-
ladium particles. At that same time, the dehydrogenation of
water occurs, which gives OH and O adsorbed species. In the
case of the Pd/SiO2 catalyst, methanol adsorbs and reacts at
258C on palladium and the dissociation of methanol through
OꢀH and CꢀO bond breaking is observed. The latter scission is
detected only at 3808C. At room temperature, adsorbed meth-
anol decomposes easily on the palladium crystallites to give
CO coordinated to the metal surface.[28] Formyl species have
also been found at this temperature, but they promptly disap-
pear if the catalyst is heated and the formation of CO and H2
prevails.[28,30] Methyl formate can be formed through the con-
densation of adsorbed methoxides with HCHO to form me-
thoxymethanol intermediates, which then dehydrogenate to
methyl formate.[31] The esterification of formic acid intermedi-
ates formed by HCHO oxidation is also possible, but the high
selectivities to methyl formate reported here indicate that fast
reactions occur through condensation of HCHO to methyl
formate.
Figure 7. Methanol conversion (at 508C) versus nitrogen concentration for
Pd/TiO2 catalysts determined by using XPS.
amounts of CO2 at low and high temperatures, which indicates
that the high particle size and acidity are important to produce
high selectivity to methyl formate at low temperature. The Pd/
TiO2(AOT/n-octane) catalyst is moderately active with a selectiv-
ity of 100% at 508C and of 90% at 808C. The Pd0/Pd+ XPS
atomic ratio for this catalyst is the highest (11.2), and the size
of the particles is approximately 1.8 nm, comparable to that of
the Pd/TiO2(AOT/cyclohexane) catalyst, which suggests that
a more reduced catalyst promotes selectivity at low and high
temperatures. The Pd/TiO2(Brij30/1-decanol) catalyst (2.2 nm)
showed a similar conversion but produces CO2 at 508C. For
this catalyst, the Pd0/Pd+ XPS atomic ratio is low (2.2) and the
size of the particles and acidity are intermediate compared to
those of other catalysts, which confirms that the reduction
state of palladium is important. Pd/TiO2(AOT/cyclohexane), Pd/
TiO2(AOT/1-octanol), and Pd/TiO2 (n-octane) catalysts have sim-
ilar sizes of metal particles (1.5, 1.3, and 1.8, respectively); how-
ever, the Pd/TiO2(AOT/1-octanol) catalyst is the most selective;
even at high temperature, this catalyst gives the highest reduc-
tion state (3.4) among the three catalysts, which confirms that
the reduction of palladium is important for selectivity.
As selectivity with bigger particles is higher, this mechanism
is likely to occur for these particles. A high acidity is necessary
to promote an adequate pathway to produce methoxide and
HCHO and then methyl formate. Contrary to that, for small
particles, the high selectivity is low probably because the acidi-
ty is too low, which promotes mainly the transformation of ad-
sorbed CO and H species to CO2. Importantly, the change in
the acidity of the catalysts depends only on the size of the par-
ticle and that this change in acidity implies important change
in the catalytic properties. The catalytic activity and selectivity
directly depend on the size of the particles. Selectivity and ac-
tivity can be modulated only by changing the size of the parti-
cles. However, works using the operando DRIFT mode are in
progress to understand the mechanism of methyl formate for-
Catalytic activity and mechanism of the reaction
Notably, the nature of the active sites, for a given metal phase,
depends strongly on the support and the method of prepara-
tion, which, in turn, may modify the catalytic performances. In
our case, two methods of preparation were used: the water-in-
surfactant microemulsion method and the chemical reduction
with hydrazine in water without use of any organic surfactants
or solvents. In the first method, very small palladium particles
have been obtained. These catalysts are very active and selec-
tive to methyl formate at low temperature and produce mainly
CO2 at high temperature. The differences in catalytic activity
and selectivity are due to the different size and reduction
degree of palladium particles and acidity of the catalysts.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 347

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[1] C. Burda, X. Chen, R. Narayanan, M. El-Sayed, Chem. Rev. 2005, 105,
1025–1102.
[2] Nanoparticles and the Environment: F. Banfield, A. Navrotsky, in Review
in Mineralogy and Geochemistry, MSA, 2001, vol. 44.
[3] F. Bonet, V. Delmas, S. Grugeon, R. H. Urbina, P. Y. Silvert, K. Tekaia-Elhsis-
sen, Nanostruct. Mater. 1999, 11, 1277–1284.
mation and the role of the size of particles in selectivity at low
temperature.
Conclusions
[4] J. D. Aiken, III, R. G. Finke, J. Mol. Catal. A 1999, 145, 1–44.
[5] J. Turkevich, P. C. Stevenson, J. Hillier, Discuss. Faraday Soc. 1951, 11,
55–75.
[6] M. Boutonnet, J. Kizling, P. Stenius, G. Maire, Colloids Surf. 1982, 5, 209–
225.
[7] T. Hanaoka, T. Hatsuta, T. Tago, M. Kishida, K. Wakabayashi, Appl. Catal. A
2000, 190, 291–296.
[8] G. Bergeret, P. Gallezot in Handbook of Heterogeneous Catalysis, Vol. 2
(Eds.: G. Ertl, H. Knozinger, J. Weitkamp), VCH, Weinheim, 1997,
pp. 439–464.
The palladium nanoparticles supported on TiO2 were obtained
by using a simplified water-in-oil microemulsion method. The
results showed that the average palladium particle size could
be controlled (from 1 to 8 nm) by changing the chemical com-
position of the microemulsion. The smaller nanoparticles were
obtained with the sodium bis(2-ethylhexyl)sulfocuccinate sur-
factant compared with ethylene glycol monolauryl ether. The
differences in particles sizes were ascribed to the different ion-
icities of the surfactants and solvent polarity.
[9] J. H. Scofield, J. Electron Spectrosc. Relat. Phenom. 1976, 8, 129–137.
[10] S. M. Davis, J. Catal. 1989, 117, 432–446.
The direct methyl formate production at low temperature
was observed. This was not observed yet over Pd/TiO2 in this
reaction. The very active catalysts are less selective to methyl
formate and produce more CO2. The differences in catalytic ac-
tivities of catalysts could be due to the different sizes of palla-
dium particles, acidity, and reduction degree of palladium. A
high reduction state of palladium seems to be important for
selectivity at low and high temperatures. A good correlation
was observed between methanol conversion and selectivity
and acidity of the surface provided by the decomposition of
basic hydrazine. Catalysts with the lowest acidity showed the
highest activity, but the lowest selectivity to methyl formate.
Catalysts with the highest acidity showed the lowest activity. A
selectivity of 100% to methyl formate with a high conversion
of methanol was observed at low temperature.
[11] R. Wojcieszak, M. Genet, P. Eloy, P. Ruiz, E. M. Gaigneaux, J. Phys. Chem.
C 2010, 114, 16677–16684.
[12] R. Wojcieszak, M. Zielinski, S. Monteverdi, M. M. Bettahar, J. Colloid Inter-
face Sci. 2006, 299, 238–248.
[13] D. V. Goia, E. Matijevic, New J. Chem. 1998, 22, 1203–1215.
[14] R. Wojcieszak, S. Monteverdi, M. M. Bettahar, Colloids Surf. A 2008, 317,
116–122.
[15] L. Xi, Y. M. Lam, Y. P. Xu, L. J. Li, J. Colloid Interface Sci. 2008, 320, 491–
500.
[16] J. Fraissard, Stud. Surf. Sci. Catal. 1990, 57, B201–B264.
[17] R. Wojcieszak, S. Monteverdi, J. Ghanbaja, M. M. Bettahar, J. Colloid Inter-
face Sci. 2008, 317, 166–174.
[18] A. Jasik, R. Wojcieszak, S. Monteverdi, M. Ziolek, M. M. Bettahar, J. Mol.
Catal. A 2005, 242, 81–90.
[19] La Surface Data Base. XPS database. www.lasurface.com.
[20] C. Destrꢃe, J. B. Nagy, Adv. Colloid Interface Sci. 2006, 123–126, 353–
367.
[21] S. Eriksson, U. Nylꢃn, S. Rojas, M. Boutonnet, Appl. Catal. A 2004, 265,
207–219.
The mechanism of formation of methyl formate from metha-
nol via the formation of intermediate methoxides and formal-
dehyde could explain the results. We showed that not only the
size of palladium particles is important at the nanometric scale
but chemical properties such as acidity and reduction step also
play an important role in the selectivity to methyl formate.
[22] C. Cabaleiro-Lago, L. Garcꢄa-Rꢄo, P. Hervꢃs, J. Pꢃrez-Juste, Colloids Surf. A
2007, 309, 286–291.
[23] O. Demoulin, M. Navez, E. M. Gaigneaux, P. Ruiz, A. S. Mamede, P. Grang-
er, E. Payen, Phys. Chem. Chem. Phys. 2003, 5, 4394–4401.
[24] H. Liu, E. Iglesia, J. Phys. Chem. B 2005, 109, 2155–2163.
[25] J. M. Tatibouꢅt, Appl. Catal. A 1997, 148, 213–252.
[26] E. Tronconi, A. S. Elmi, N. Ferlazzo, P. Forzatti, G. Busca, P. Tittarelli, Ind.
Eng. Chem. Res. 1987, 26, 1269–1275.
[27] M. Ai, J. Catal. 1982, 77, 279–288.
[28] S. A. Kirillov, P. E. Tsiakaras, I. V. Romanova, J. Mol. Struct. 2003, 651–653,
365–370.
Acknowledgements
[29] S. P. Bates, M. J. Gillan, G. Kresse, J. Phys. Chem. B 1998, 102, 2017–
2026.
[30] G. C. Cabilla, A. L. Bonivardi, M. A. Baltanꢆs, J. Catal. 2001, 201, 213–
220.
[31] J. Lichtenberger, D. Lee, E. Iglesia, Phys. Chem. Chem. Phys. 2007, 9,
4902–4906.
The authors acknowledge the Univeristꢀ catholique de Louvain
and the Fonds National de la Recherche Scientifique (FNRS) of
Belgium. The authors involved in the “INANOMAT” IUAP network
acknowledge the service Public Fꢀdꢀral de Programmation Politi-
que Scientifique (Belgium).
Received: May 17, 2012
Keywords: microemulsion · nanoparticles · palladium · XPS
Revised: July 10, 2012
Published online on November 19, 2012
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 339 – 348 348