Influence of the preparation method on the surface characteristics and activity of boron-nitride-supported noble metal catalysts

Article abstract of DOI:10.1021/jp060183x

In this article, we report how variations in the preparation method of boron-nitride-supported noble metal catalysts may influence the surface characteristics of the active phase and consequently the potential applications as catalysts for oxidation reactions. The deposition and the dispersion of the active phase are strongly influenced by the preparation process and in particular by the protic or aprotic solvent used as the dispersing phase; in this study, benzene, glyme, water, tetrahydrofuran, diglyme, 2-propanol, and glycol have been investigated. Characterization techniques, such as Brunauer-Emmett-Teller, X-ray diffraction, transmission electron microscopy, Fourier transform infrared spectroscopy, and thermogravimetric analysis, have been used to study the influence of the choice of a solvent phase on the particle size and dispersion of the metal deposited on the BN support. The modifications undergone by the support during the deposition of palladium in different solvents have also been studied. Through the use of the same deposition procedure, different noble metal coatings (Pt, Pd, Au, and Ag) have been prepared. The acidic and redox characteristics of the resulting samples were characterized by temperature-programmed reduction and adsorption microcalorimetry. The catalytic performances of these materials were tested in the total oxidation of methane in lean conditions (excess oxygen and presence of water).

Full text of DOI:10.1021/jp060183x

12572
J. Phys. Chem. B 2006, 110, 12572-12580
Influence of the Preparation Method on the Surface Characteristics and Activity of
Boron-Nitride-Supported Noble Metal Catalysts
Georgeta Postole,†,| Antonella Gervasini,‡ Claude Guimon,§ Aline Auroux,*,† and
Bernard Bonnetot
Institut de Recherches sur la Catalyse, Centre National de la Recherche Scientifique, 2 AV. A. Einstein,
69626 Villeurbanne Cedex, France, Dipartimento di Chimica Fisica ed Elettrochimica and Centro di
Eccellenza, Centro Interdisciplinare Materiali e Interfacce Nanostrutturati, UniVersita` di Milano,
Via Golgi 19, I-20133 Milano, Italy, LPCM, UniVersite´ de Pau, 2 AV. P. Angot, 64053 Pau Cedex 09, France,
“Ilie Murgulescu” Institute of Physical Chemistry of the Romanian Academy, Spl. Independentei 202,
060041 Bucharest, Romania, and Laboratoire Multimate´riaux et Interfaces, CNRS UMR 5615,
UniVersite´ Claude Bernard Lyon 1, 69622 Villeurbanne Cedex, France
ReceiVed: January 10, 2006; In Final Form: April 26, 2006
In this article, we report how variations in the preparation method of boron-nitride-supported noble metal
catalysts may influence the surface characteristics of the active phase and consequently the potential applications
as catalysts for oxidation reactions. The deposition and the dispersion of the active phase are strongly influenced
by the preparation process and in particular by the protic or aprotic solvent used as the dispersing phase; in
this study, benzene, glyme, water, tetrahydrofuran, diglyme, 2-propanol, and glycol have been investigated.
Characterization techniques, such as Brunauer-Emmett-Teller, X-ray diffraction, transmission electron
microscopy, Fourier transform infrared spectroscopy, and thermogravimetric analysis, have been used to study
the influence of the choice of a solvent phase on the particle size and dispersion of the metal deposited on
the BN support. The modifications undergone by the support during the deposition of palladium in different
solvents have also been studied. Through the use of the same deposition procedure, different noble metal
coatings (Pt, Pd, Au, and Ag) have been prepared. The acidic and redox characteristics of the resulting samples
were characterized by temperature-programmed reduction and adsorption microcalorimetry. The catalytic
performances of these materials were tested in the total oxidation of methane in lean conditions (excess
oxygen and presence of water).
1. Introduction
to a larger supply of reactive oxygen species and thus promoting
the catalytic oxidation. Jacobsen4 studied a Ba-Ru/BN catalyst
for ammonia synthesis and found that the activity of this catalyst
is significantly higher than that of traditional catalysts used in
ammonia synthesis. The author concluded that the stable activity
of the Ba-Ru/BN catalyst is related to the thermodynamic
stability of BN under the test conditions. Boron nitride has also
been used as support for vanadium oxide and showed promising
results in propane oxidation.5 Wu et al.6 have also presented
favorable findings for the selective hydrogenation of R,â-
unsaturated aldehydes into unsaturated alcohols by employing
BN-supported Pt-Sn catalysts.
Catalytic combustion of methane has drawn much attention
in recent years because methane is a well-known greenhouse
gas and the most stable and abundant alkane. Moreover, methane
combustion is an attractive source of energy because of the large
world reserves of CH4 as natural gas and its low levels of sulfur-
and nitrogen-containing impurities. The two main applications
of methane combustion where catalysis may play a role are in
gas turbines for generating power and in automobile exhausts
for removing small amounts of methane from the emissions of
methane-burning engines. In the first case, the catalyst is
operating at temperatures of up to 800 °C, while in the latter
case, the catalyst must exhibit activity below 300 °C.7-9 Since
both carbon dioxide and water are products of the combustion
reaction, it is important that the effect of these compounds on
the activity of the catalysts should be negligible. Although in
Due to its remarkable properties, boron nitride has shown
definite promise as a novel support in catalysis and as a
replacement for traditional oxide supports in different processes.
Boron nitride has not been studied extensively as a support;
however, there are recent reports in the literature about the good
performance of boron-nitride-based catalysts in some catalytic
processes. Several recent studies describe the use of BN-
supported noble metals for the deep oxidation of volatile organic
compounds (VOCs). Wu et al.1 have shown that a Pt/BN catalyst
can remain active for 80 h at 185 °C, while the activity of
traditional Pt/γ-Al2O3 at the same temperature was found to
decrease continuously with time. Lin et al.2 and Wu et al.3 have
studied Pt/BN as a catalyst for the deep oxidation of methanol
and benzene. They concluded that Pt particles form weaker
bonds with the BN surface, compared to γ-Al2O3. This weak
bonding facilitated the reduction of the Pt particles. The reduced
state of the Pt particles on Pt/BN caused a weaker oxygen
bonding on the surface of Pt in the presence of oxygen, leading
* Author to whom correspondence should be addressed. Fax: (33) 472
44 53 99. E-mail: aline.auroux@catalyse.cnrs.fr.
† Centre National de la Recherche Scientifique.
‡ Universita` di Milano.
§ Universite´ de Pau.
| “Ilie Murgulescu” Institute of Physical Chemistry of the Romanian
Academy.
Universite´ Claude Bernard Lyon 1.
10.1021/jp060183x CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/03/2006

BN-Supported Noble Metal Catalysts
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12573
the case of carbon dioxide there is no clear agreement as to its
effect on methane combustion,9 it is commonly agreed that water
has a negative effect on the combustion reaction over traditional
supported noble metal catalysts such as Pd/Al2O3.9-11
conditions, different metals have been considered, and platinum,
gold, and silver metal particles have been deposed, leading to
catalysts which were tested in the reaction of total oxidation of
methane in severe conditions (excess of oxygen and presence
of water). To our knowledge, no other groups have reported
studies about the use of BN-supported catalysts for methane
oxidation/combustion reactions.
It is therefore necessary to develop novel materials that are
able to withstand high temperatures and can serve as catalysts
or supports for these reactions even in humid atmospheres.
In this work, the deposition of palladium on boron nitride
supports has been carried out using protic or aprotic solvents
as dispersing phases. The route using aprotic solvents involves
the choice of palladium acetylacetonate as the precursor and
anhydrous hydrazine as the reducer. The choice of BN as the
support was also directed by the search for new hydrophobic
supports with high thermal conductivity, adapted to high-energy
oxidation reactions under severe conditions. Boron nitride is
one of the most interesting non-oxide ceramic materials because
of its low density, excellent resistance to chemical attacks, high
melting point, good thermal conductivity, and high stability with
respect to oxidation.12-14 Contrary to BN, materials traditionally
used as supports of active phases (oxides such as γ-Al2O3 and
SiO2) possess low thermal conductivity and many acidic and
basic sites, causing sintering of the supported metal on hot spots
and coverage of the catalyst by water.15,16 Moreover, the metal-
support interaction present in most oxide-supported metal
catalysts has a negative influence on the catalytic activity.15,17
2. Experimental Section
2.1. Materials. All the experiments were performed using
as the catalyst support a boron nitride powder from Starck
exhibiting a hexagonal nonisotropic structure formed of 5 µm
platelets with 3 µm thickness (as determined by scanning
electron microscopy (SEM)) and presenting a surface area of
136 m2 g-1. The palladium precursor was palladium(II) acetyl-
acetonate, provided by Aldrich and used as received. For the
other noble metals the precursors used were hexachloro-
platinic(IV) acid (Metalor), hydrogen tetrachloroaurate(III)
(Strem), and silver nitrate (Strem) for Pt, Au, and Ag, re-
spectively. The solvents or dispersing phases used were supplied
by SDS Company for the organic compounds such as benzene,
ethyleneglycol-dimethyl ether (monoglyme or glyme), tetrahy-
drofurane (THF), diethyleneglycol-dimethyl ether (diglyme),
2-propanol, and ethyleneglycol (glycol). They were all HPLC
grade without stabilizer, anhydrous (water amount lower than
0.03 wt %), and kept on active 5 Å molecular sieves. Anhydrous
hydrazine (water amount lower than 0.1 wt %) was prepared in
the laboratory following a method described in the literature.28
Concerning the metals deposed as active phases, previous
works13,18 have demonstrated that the particle size is an
important parameter driving the catalytic properties; large
particles with dimensions around 20 nm or aggregates have been
shown to be inefficient. Noble metal nanoparticles are usually
obtained by reducing a metal precursor with a strong reducer;
this reaction is often performed using borohydrides in protic
solvents.19 However these chemicals are very strong reducers,
and the particles obtained by this process have been shown to
present a broad size dispersion. Moreover, the presence of
borohydride hydrolysis residues such as sodium or potassium
borates can have a negative impact on the purity of the catalyst.
The reducing character of hydrazine is known to be softer than
that of borohydrides, and the solubility of hydrazine is rather
high in almost all solvents. For example, hydrated hydrazine
has been used to prepare magnetic nanocomposites20 from
organometallic complex salts by ligand displacement and cation
reduction to metal in a single operation.
2.2. Preparation of the Samples. In a typical preparation, 4
g of BN powder was added to 200 cm3 of dispersing phase in
a three-necked flask equipped with a refrigerant and under argon
flow, under a vigorous stirring during 15 min. Approximately
0.1145 g of palladium(II) acetylacetonate (0.376 mmol leading
to 40 mg of metallic palladium) were then added, while stirring
for another 15 min to obtain the precursor dissolution and a
homogeneous suspension of the support. A solution of 1 g of
anhydrous hydrazine (31.25 mmol) dissolved into 100 cm3 of
solvent was then added drop by drop into the suspension for
10 min. The reaction occurred instantaneously, leading to a dark
gray color due to the formation of metallic palladium. The
stirring was maintained for 1 h and then stopped. After the
catalyst was allowed to settle overnight, a clear liquid phase
was easily separated from the solid. The catalyst was then dried
under low pressure (10-2 h Pa during 10 h). Very thin powders
were obtained, with a light gray color due to the metal deposed.
The samples were labeled in relation to the dispersion liquid
used: Pd/BN_bz when prepared in benzene, Pd/BN_gly for
monoglyme, Pd/BN_wat for water, Pd/BN_thf for THF,
Pd/BN_digl for diglyme, Pd/BN_isop for 2-propanol, and
Pd/BN_glc for glycol. All samples were calcined up to 500 °C
Noble metals are extensively used as the active components
in many industrial catalytic processes and for the conversion
of environmentally unfriendly chemicals in gaseous emissions.
Among noble metals, Pt and Pd are the most commonly used
and studied catalysts.21,22 Platinum and palladium are known
to present low light-off temperatures in the oxidation of
hydrocarbons and other organic chemicals.22-24 The activity of
palladium is generally better than that of platinum for the
conversion of methane, but it is lower for the transformation of
other chemicals.24 Silver and gold have also been used with
good results as active phases in hydrocarbon oxidation
reactions.25-27
for 10 h under air flow, using a heating rate of 1 °C min-1
The same preparation procedure was followed, using ethyl-
eneglycol as solvent and the respective metal precursors, to
obtain Pt/BN_glc, Ag/BN_glc, and Au/BN_glc catalysts.
.
2.3. Characterization. The physicochemical properties of the
catalysts were determined using several techniques. The specific
surface areas of the samples were measured by N2 adsorption
and calculated from the Brunauer-Emmett-Teller (BET)
method. Prior to surface area determination, the samples were
outgassed at 100 °C overnight and then at 400 °C for 6 h. The
pore size distribution of each sample was determined from the
desorption branch of the N2 isotherm, using an ASAP 2010M
apparatus, from Micromeritics, and gave evidence for some
In this paper, the synthesis of palladium nanoparticles and
their deposition on high-surface-area boron nitride have been
performed using seven different dispersion phases. The sizes
of the particles and their dispersion have been studied using
X-ray diffraction (XRD) and transmission electron microscopy
(TEM). The support and catalysts have been fully characterized
to compare the roles played by the deposed metal in the catalytic
performances of the studied samples. In the second part of this
work, using the same support, BN, and the same deposition

12574 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Postole et al.
TABLE 1: Changes in Surface Area of the BN Support
microporosity in the parent BN. Therefore, deBoer’s t-plot
analysis was used to obtain the micropore surface area. For this
support, approximately 36 m2 g-1 of the surface area was
contained in micropores (i.e., radius less than 1 nm), and 100
m2 g-1 was in larger pores. A micropore volume of 0.0168 cm3
g-1 was found by the t-method. Due to the presence of
micropores, the BET method using nitrogen is perhaps not the
most appropriate technique for determining the surface area of
high-surface-area BN samples. However, to compare our results
with those from the literature, we used this method. Some
microporosity was also observed by Lindquist et al.29 who
reported a micropore surface area of 43 m2 g-1 and surface area
of larger pores of ∼307 m2 g-1 for a BN sample heated at 1500
°C for 24 h. Dibandjo et al.30 reported a BN surface area of
540 m2 g-1 using the BET method on their samples, and Han
et al.31 reported a surface area of 168 m2 g-1 for a porous BN.
after Swelling in a Solvent
SBET
mass loss H2O
(wt %)
sample
(m2 g-1
)
BN fresh
BN calcined
BN_gly calcined
BN_isop calcined
136
136
233
292
2
1.7
1.3
5.1
was pretreated in an argon flow at a rate of 20 cm3 min-1 at
400 °C (2 °C min-1). A reducing gas mixture composed of 5%
H2 plus 95% Ar was employed for the reduction, at a flow rate
of 20 cm3 min-1 and with a heating rate of 10 °C min-1 from
room temperature to 600 °C. The hydrogen consumption was
detected with a thermal conductivity detector (TCD).
2.4. Catalytic Tests. Catalytic tests of the methane oxidation
reaction were performed in a quartz tubular microreactor. The
samples were prepared by adding 0.05 g of catalyst to 1 g of
quartz (1:20 dilution). The powders tested and the quartz used
for dilution were ground and sieved separately to obtain particles
of 0.2-0.3 mm. The samples were pretreated for 4 h at 400 °C
in O2/He (20% v/v) flow. A set of mass flow controllers
(Bronkhorst, Hi-Tec) made it possible to determine accurately
the composition of the reactant mixture: 1500 ppm CH4 and
30 000 ppm O2 in helium. The reaction was performed under
wet atmosphere by passing the reactant mixture through water.
A catalytic test was also performed on the sample Pd/BN_glc
using a dry gas mixture to compare the behavior of the catalyst
in the presence of dry and water-saturated reactant gases. The
reaction was carried out at fixed space velocity (gas hourly space
velocity ) 42 000 h-1) and variable temperature. Seven different
temperatures in the interval of 400-700 °C were chosen, each
temperature being held for 90 min. The reactant and exit gas
streams of the reactor flowed through a gas cell (path length
2.4 m multiple reflection gas cell) in the beam of a Fourier
transform infrared (FTIR) spectrometer (Bio-Rad with -DTGS
detector). The spectrometer response permitted quantitative
analyses for CH4 (1360 cm-1), CO (2086 cm-1), and CO2 (721
cm-1). The conversion percentages and selectivities were
calculated from the measured concentrations of the reactants
and products.
X-ray diffraction spectra were recorded using a Bruker
(Siemens) D5005 apparatus. The concentrations of the sup-
ported metals were determined by chemical analysis using
inductively coupled plasma atomic emission spectroscopy (ICP-
AES). For that purpose, the samples were treated with a mixture
of H2SO4 + HF + HNO3 to dissolve them completely. The
morphologies of the different samples have been examined using
SEM, and the active phase dispersion and particle size have
been measured by TEM. Scanning electron microscopy images
were obtained from a JEOL 55 CF (CMEABG, UCB, Lyon),
and TEM pictures were recorded using a JEOL 2010 (IRC,
Lyon). The Pd particle sizes were calculated from the TEM
micrographs obtained after calcination, taking into account about
20 particles out of the aggregates to obtain an average value
for the dispersed active phase. X-ray photoelectron spectrometry
(XPS) analysis was performed using a Surface Science Instru-
ments 301 spectrophotometer with monochromatic Al KR
radiation to analyze the Pd dispersion and the possible formation
of boron oxides. The 1s binding energy of boron was considered
to survey the surface chemical species of the hexagonal boron
nitride (h-BN) samples. Reference binding energies (BEs) were
190.5 eV for BN and between 192.2 and 193.0 V for BOx.1,2,12
Thermogravimetry experiments were performed in a TG-DSC
111 apparatus from Setaram, in air up to 600 °C using a heating
rate of 5 °C min-1. Infrared spectra were recorded with a Vector
22 spectrometer from Bruker with KBr windows, working in a
spectral range of 4000-400 cm-1. The pellets were a mixture
of 100 mg of KBr, dried at 120 °C, and 7 mg of studied catalyst.
3. Results and Discussion
3.1. Behavior of the BN Support during the Chemical
Treatment. To verify the inertness of the BN support during
the metal deposition, samples of BN powder were dispersed in
the employed solvents, added with hydrazine, and then separated
and calcined as in the experimental procedure for the preparation
of the catalysts. From this procedure, the diffraction patterns
did not show any difference in the widths of the BN diffraction
lines, and the IR spectra did not reveal any hydrolysis. The
surface area was the only parameter that changed upon such
chemical treatment. The BET surface area and water loss (as
determined by TGA) of the support after chemical treatment
and calcination are given in Table 1. These changes in surface
area can be explained by an increase of microporosity originating
from specificities of the BN structure. The hexagonal BN
crystallographic structure consists of three BN groups stacked
into a hexagonal framework, leading to BN layers, similar to
graphite. These layers are bound together through anisotropic
bonds leading to a very oriented compound. It seems that when
BN powders are swelled in a solvent an important fraction of
the solvent molecules enter inside the BN layers and open
nanopores during the calcination process. The increase in surface
area can therefore be ascribed to the opening of nanopores by
Microcalorimetric studies of ammonia adsorption were per-
formed at 80 °C in a heat flow calorimeter (C80 from Setaram)
linked to a conventional volumetric apparatus and equipped with
a Barocel capacitance manometer for pressure measurements.
The samples were pretreated in a quartz cell by heating overnight
under vacuum at 400 °C with a heating rate of 0.6 °C min-1
.
The differential heats of adsorption were measured as a function
of coverage by repeatedly sending small doses of gaseous
ammonia onto the catalysts until an equilibrium pressure of
about 67 Pa was reached.32 The samples were then outgassed
for 30 min at the same temperature, and a second adsorption
was performed at 80 °C on each sample until an equilibrium
pressure of about 27 Pa was attained, to calculate the amount
of irreversibly chemisorbed ammonia at this pressure, which
provides an estimate of the number of strong acidic sites.
Temperature-programmed reduction (TPR) was performed in
a custom-made apparatus already described in the literature,33
equipped with a U-shaped quartz tube microreactor and sur-
rounded by a furnace controlled by a programmed heating
system. Prior to each test, a sample of about 150 mg of catalyst

BN-Supported Noble Metal Catalysts
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12575
Figure 1. Acidity of the support; differential heats of ammonia
adsorption vs coverage.
Figure 2. X-ray diffraction patterns of the Pd/BN_thf catalyst and
the corresponding BN support.
TABLE 2: Calorimetric Data for NH3 Adsorption at 80 °C
on BN Supports
ammonia uptake
a
b
c
ntotal
nreads
nirr
sample
BN fresh
BN calcined
BN_gly calcined glyme
BN_isop calcined 2-propanol
solvent
(µmol g-1
)
(µmol g-1
)
(µmol g-1
)
141
145
139
110
63
66
66
56
78
79
73
54
a Amount of NH3 absorbed under an equilibrium pressure of 27 Pa.
b Amount of NH3 reabsorbed after pumping at 80 °C under an
equilibrium pressure of 27 Pa. c Amount of irreversibly chemisorbed
ammonia.
the solvent trapped during the chemical treatment, which then
evolved during the calcination, while the crystallographic
parameters remain unchanged. Scanning electron microscopy
micrographs reveal that the layer structure of boron nitride is
changed only at the edge of the BN platelets, as the BN structure
in the platelets is very slightly affected by the edge effects of
the splitting. The weight loss measured on calcined samples
corresponds to the water uptake when the catalyst is kept without
any special care. The dispersing phases did not influence
noticeably the acidity of the BN supports (Figure 1). The initial
heat of adsorption of ammonia on the parent BN support was
measured to be 105 kJ mol-1; this is indicative of a relatively
weak acidity, compared to an oxide support such as γ-Al2O3.
Calcined BN and BN_gly samples display differential heat
curves similar to the fresh support, except for the initial heats
which are higher than 120 kJ mol-1. Meanwhile, the shape of
the curve of heat of interaction versus coverage for ammonia
adsorption over the BN_isop calcined powder is also similar to
the previous ones but shifted toward lower heats, except for
the initial heat value. A slightly lower number of acid sites were
titrated (Table 2), which is consistent with an increase of the
nanoporosity of the support.
3.2. Pd Catalysts Obtained using Various Dispersing
Phases. Several dispersing phases, protic or aprotic and with
different physical and chemical properties, were used to dissolve
the precursor, with the aim to avoid clustering of the nanopar-
ticles produced. A sample diffraction pattern is given in Figure
2, which represents the XRD spectra of uncalcined Pd/BN_thf
and the corresponding fresh support. These diffraction patterns
show that the BN support was not modified by the chemical
treatment, and no hydrolysis occurred during the palladium
deposition. Palladium is clearly characterized by its (111)
diffraction peak at 2θ ) 39.95° with a small difference from
the 40.2° theoretical value. This increase of the lattice parameter
Figure 3. X-ray diffraction patterns of the Pd deposed on BN using
different solvents before calcination.
could be related to the presence of a small amount of oxide or
oxygen in the metal. Figure 3 represents the Pd(111) diffraction
line for all uncalcined samples as synthesized in different
solvents. The only difference between the various spectra is in
the width of the (111) diffraction line of metallic palladium.
The size of the palladium particles can be calculated from
the X-ray diffraction pattern recorded using the Scherrer
formula34 applied to the full width at half maximum of the (111)
Pd diffraction line
(L ) Kλ/Bstruct cos θ)
In this formula, K describes the crystalline shape factor of the
particles (here K is 0.85 as proposed for cubic systems), λ is
the wavelength, and θ is the incidence angle. The term Bstruct
describes the structural broadening, which is the difference in
the integral profile between a standard reference (0.05° mea-
sured for silicon reference with the same intensity) and the
sample to be analyzed. Table 3 gathers the measured values
after standardization of the spectra using the (100) BN diffrac-
tion line intensity as an internal reference.
Important changes appeared after calcination of the freshly
prepared samples. The heating in air induced an oxidation of
palladium into the oxide, PdO, which led to an increase of the
particle size. The measured increases were more important than

12576 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Postole et al.
TABLE 3: Characterization of the Catalysts
XPSd
Pd amount
XPS ICP
total pore average pore
Pd particle size (nm)
SBET
Vmicropore
volumeb
diameterc
(nm)
BN BOx
a
sample
solvent
benzene
glyme
water
THF
(m2 g-1
)
cm3 g-1
(cm3 g-1
)
(%) (%) (at. %) (wt %) XRDe XRDf TEMg
BN fresh
136
73
0.0168
0.0034
0.0215
0.0504
nd
nd
nd
0.0966
0.0892
0.1128
0.1382
nd
nd
nd
4.7
7.4
4.4
4.0
nd
nd
nd
100
85
96
100
93
100
100
93
0
15
4
0
7
0
0
7
Pd/BN_bz
Pd/BN_gly
Pd/BN_wat
Pd/BN_thf
0.1
0.6
0.4
0.5
0.4
0.5
0.6
0.94
0.89
0.95
0.94
0.99
0.94
0.74
7
7
11
9
13
9
7
16
15
18
12
15
14
16
12
14
16
15
nd
15
nd
179
254
138
215
229
221
Pd/BN_digl diglyme
Pd/BN_isop 2-propanol
Pd/BN_glc
ethyleneglycol
nd
nd
nd
a Micropore volumes obtained by the t-method. b Total pore volume at P/P0 ) 0.98. c Average pore diameter according to the BJH method
(adsorption branch). d Relative concentrations of boron compounds (BN and BOx). e Size of the palladium particles calculated from X-ray diffraction
patterns before calcination. f Size of the palladium particles calculated from X-ray diffraction patterns after calcination. g Size of the palladium
particles calculated from TEM micrographs after calcination.
Figure 5. Fourier transform infrared spectra of the calcined BN
support, Pd/BN_digl, and Pd/BN_bz.
troscopy results show also that the worst dispersion of Pd is
achieved in benzene solvent. Moreover, the relative concentra-
tions of boron species at the surface were also determined by
XPS. We observed a slight oxidation of the surface of some
samples; the Pd/BN_bz sample expressed the highest relative
concentration of BOx. As can be seen in Table 3, all samples
present surface areas higher than that of the parent support,
except for the Pd/BN_bz catalyst. While the increase in SBET
can be related to the BN structure as suggested in subsection
3.1, the decrease in surface area observed for Pd/BN_bz could
be assigned to the poor solubility of hydrazine in benzene (see
the preparation method), which induced a polyphasic reaction
favoring a partial hydrolysis leading to the formation of oxide
derivatives. Accordingly, as mentioned above, XPS measure-
ments revealed the presence of B2O3 (relative concentration
15%) in this sample. In addition, FTIR spectra (Figure 5)
confirmed the partial hydrolysis and the presence of hydroxyl
groups. All the samples except Pd/BN_bz presented FTIR
spectra similar to the BN support. For this reason, only one
sample spectrum, Pd/BN_digl, was represented in Figure 5
together with BN and Pd/BN_bz.
Figure 4. Transmission electron microscopy images of the Pd/BN_gly
sample after calcination under air flow.
expected and were found to be irreversible, after reduction of
PdO in a flow of 5% H2/Ar. The size of the particles remained
unchanged after the oxidation-reduction cycle.
The TEM micrograph presented in Figure 4 shows clearly
that particles are stacked into larger clusters; this clustering
occurred during the palladium oxidation-reduction cycle. It is
well-known that palladium oxidation was found to be dependent
on the dispersion; the lower the dispersion, the lower the Pd
oxidation extent. Moreover, the rate of the oxidation is strongly
dependent on temperature.21,35,36 In our case, the nanoparticles
are gathered at the support grain border, and they aggregate
after heating and oxidation of palladium, leading to larger
particles. The small size particles (about 7 nm) observed initially
aggregated into particles of sizes of about 15 nm, whereas larger
particles were less affected. With the same metal amount the
larger particles were less numerous than the small ones, as
observed on the TEM micrograph. The average particle size
was enhanced by the sintering, as shown by Table 3, where the
values obtained from XRD and TEM are consistent. In addition,
Table 3 gives the amounts of Pd, determined either by ICP (wt
%) or XPS (at. %) methods. We found that the binding energies
of Pd 3d5/2 vary between 337.5 and 338.4 eV, which can be
assigned to the presence of Pd(II). X-ray photoelectron spec-
DeBoer’s t-plot analysis was used to obtain the micro-
pore surface area. For samples Pd/BN_bz, Pd/BN_gly, and
Pd/BN_wat, the micropore surface areas were 8, 47, and 109
m2 g-1, respectively, which explains the changes in surface areas
observed by the BET method for these samples compared to
BN fresh support (Table 3). Table 3 displays also the micropore
volume, the total pore volume, and average pore diameter. The
pore size diameters were obtained from the adsorption branch
of the isotherm using the corrected form of the Kelvin equation
by means of the Barrett-Joyner-Halenda (BJH) method.

BN-Supported Noble Metal Catalysts
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12577
TABLE 5: Calorimetric Data for NH3 Adsorption at 80 °C
on the BN-Supported Pd Catalysts
ammonia uptake
a
b
c
ntotal
nreads
nirr
sample
solvent
(µmol g-1
)
(µmol g-1
)
(µmol g-1
)
Pd/BN_bz
Pd/BN_gly
Pd/BN_wat
Pd/BN_thf
Pd/BN_digl diglyme
Pd/BN_isop 2-propanol
benzene
glyme
water
198
183
195
193
172
160
179
42
69
92
68
73
74
49
156
114
103
125
99
86
130
THF
Pd/BN_glc
glycol
a Amount of NH3 absorbed under an equilibrium pressure of 27 Pa.
b Amount of NH3 reabsorbed after pumping at 80 °C under an
equilibrium pressure of 27 Pa. c Amount of irreversibly chemisorbed
ammonia.
Figure 6. Derivative of the pore volume as a function of the pore
diameter for some Pd/BN samples.
Figure 8. Differential heats of ammonia adsorption vs coverage for
calcined BN support and Pd/BN catalysts prepared in different solvents.
can react partially with palladium oxide to form a more stable
compound.
Figure 7. Thermoprogrammed reduction profiles of Pd/BN samples
under H2/Ar flow.
Stoichiometrically, the theoretical H2 consumption for the
reduction of PdO to Pd should be 88 µmol g-1 for Pd/BN_bz,
84 µmol g-1 for Pd/BN_gly, 89 µmol g-1 for Pd/BN_wat, and
70 µmol g-1 for Pd/BN_glc, respectively. The hydrogen uptake,
calculated from the area under the experimental curve, is 31
µmol g-1 for Pd/BN_bz, 19 µmol g-1 for Pd/BN_gly, 17 µmol
g-1 for Pd/BN_wat, and 16 µmol g-1 for Pd/BN_glc, respec-
tively. The degree of PdO reduction for each catalyst is listed
in Table 4, confirming the results obtained by TEM: The
palladium nanoparticles deposed on the support surface had a
slight tendency to form stacked aggregates during the thermal
oxidation-reduction cycles.
The acidic character of the catalysts was estimated by
ammonia adsorption microcalorimetry. It is well-known that
besides determining quantitatively the surface acidity of solids
(the acidity playing an important role in all catalytic processes)
adsorption microcalorimetry allows the determination of the
strength distribution of the sites. The results obtained are
presented in Table 5 and Figure 8. The initial values of the
differential heats of ammonia adsorption indicate the existence
of stronger acid sites on Pd/BN_gly, Pd/BN_wat, and Pd/BN_thf
in comparison with the calcined BN support. The influence of
the preparation method on the acid strength distribution of the
catalysts and consequently on the surface state modification of
the BN support in the dispersing phase is evident.
TABLE 4: Temperature-Programmed Reduction Results
Obtained for the Catalysts Studied Using 5% H2 + Ar
sample
Pd reduced (%)
T
M (°C)
Pd/BN_bz
Pd/BN_gly
Pd/BN_wat
Pd/BN_glc
35
23
19
23
160
76
99
72
Figure 6 displays the derivative of the pore volume as a
function of pore diameter, for some of the samples studied. In
comparison to the BN support, the microporosity was increased
for all samples except for Pd/BN_bz, confirming the above
results and the pore blocking by some boron oxide. The
micropore volume measured by the t-method is consistent with
the surface area determination; the sample prepared using
benzene, which is the most oxidized, presents the smallest
surface area and also the smallest pore volume (Table 3).
Temperature-programmed reduction (TPR) experiments were
also performed on some of the samples presenting the most
different surface areas (Pd/BN_bz, Pd/BN_gly, Pd/BN_wat, and
Pd/BN_glc). The H2 TPR profiles of Pd/BN catalysts recorded
up to 300 °C are shown in Figure 7. No hydrogen uptake peaks
were observed at higher temperatures. For all the tested samples,
the H2 TPR profiles show only one peak of reduction, with a
maximum below 100 °C, except Pd/BN_bz which displayed a
maximum at 160 °C (Table 4). The reduction temperature for
Pd/BN_bz is higher than those for the other samples and could
be related to the presence on the support of boron oxide, which
The catalytic performances of samples were tested in methane
oxidation. Table 6 presents the CH4 conversions calculated from
the residual methane concentration at each reaction temperature.
The first test on Pd/BN_glc was performed using a dry mixture

12578 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Postole et al.
TABLE 6: Methane Conversion Using Different
BN-Supported Pd Catalysts
CH4 conversion (%)
wet conditions
dry conditions
T
BN
(°C) Pd/BN_glc Pd/BN_gly Pd/BN_glc Pd/BN_wat calcined
400
500
550
600
625
650
700
0
0.1
0.2
0.7
14.6
38.8
86.3
99.6
0.2
0.8
1.3
16.7
36.4
74.1
99.9
0.6
0.9
4.1
29.8
64.1
98.0
100.0
0
0
0
1.8
3.5
5.3
nd
nd
0
1.3
9.4
99.3
71.9
of reactant gases. The results obtained show a low methane
conversion (around 70% conversion at 700 °C). After the
catalyst was cooled to room temperature and 2 h rest under dry
reactive gas atmosphere, a second run was performed on the
same sample using a water-saturated gas mixture. The enhance-
ment of the conversion at 700 °C was important (around 90%
methane conversion), which led us to evaluate the conversion
capabilities of the other catalysts only under wet conditions.
The BN support was totally inactive below 625 °C. However,
a slight conversion of about 9% was observed at 650 °C; at
this temperature, only CO and H2O were obtained as products.
The methane conversion increased significantly at 700 °C, with
CO2 formation; however, CO remained present in important
amounts, probably due to homogeneous reactions.
The palladium catalysts supported on boron nitride showed
roughly similar activities, as can be seen in Figure 9. The
Pd/BN catalysts were inactive in methane oxidation at temper-
atures lower than 400 °C. At temperatures ranging between 400
and 650 °C, important concentrations of CO were detected in
the products stream. The main oxidation products became CO2
and H2O only above 650 °C, as can be seen for example in
Figures 10 and 11 where CO and CO2 productions are plotted
Figure 10. Comparison of the productions of CO and CO2 during CH4
oxidation for the Pd/BN_wat catalyst and the BN support
complete oxidation and to shift the onset of oxidation toward a
lower temperature.
Concerning the use of BN as a support for noble metals in
the CH4 combustion reaction, we observed light-off temperatures
(XCH ) 50%) higher than those reported in the literature for
4
the traditional catalysts (Pd/Al2O3 and Pd/ZrO2).1,23,37-41 How-
ever, we have to point out that our experiments were carried
out under very severe conditions with a O2/CH4 ratio equal to
20 and in the presence of water. We observed a very good
stability with time for Pd/BN catalysts, while the activity of
Pd/Al2O3 catalysts under similar conditions decreased steadily
over the first 20 hours.14
3.3. Supported Noble Metal (Pd, Pt, Ag, and Au) Cata-
lysts. The above study of the performance of catalysts using
palladium as the active phase on boron nitride supports has been
completed by a comparison between different noble metals
deposed using essentially the same preparation conditions (same
BN support, dispersing phase and reducing agent). The ef-
ficiency of the catalysts obtained by deposition of small
percentages of noble metals on the support is related to the size
of the particles deposed and to the aggregation of the metal
grains in larger clusters. Ethylene glycol was used as the
dispersing phase, and the reduction of the precursors into metal
was performed using hydrazine as described previously. How-
ever, the precursors used were not only acetylacetonates but
also chlorinated derivatives. The use of inorganic precursors
has induced important and unexpected differences in terms of
amounts deposed and dispersion, related to the low solubility
of halogenated compounds. The physicochemical characteristics
of the catalysts obtained by the deposition of noble metals Pd,
Pt, Ag, and Au on a BN support are presented in Table 7.
The metal particle sizes for each calcined sample were
obtained from XRD data, using the Scherrer formula as
described previously. In fact, after calcination of the catalyst in
air, palladium was oxidized into PdO, as shown by the
diffraction line at 2θ ) 38.84°. For the Au catalyst, the amount
of gold was clearly lower than expected, which can be explained
by the low solubility and the differences in reactivity of the
precursor employed.
versus temperature. The onset temperatures (XCH ) 15%)
4
decreased in the order BN > Pd/BN_gly > Pd/BN_glc >
Pd/BN_wat, which can be related to the calculated activation
energies (238 kJ mol-1 for Pd/BN_gly, 200 kJ mol-1 for
Pd/BN_glc, and 160 kJ mol-1 for Pd/BN_wat).
Across the entire temperature range, Pd/BN catalysts present
higher conversions than boron nitride alone. Similarities between
the activities at 700 °C suggest that homogeneous gas-phase
reactions could be also involved in methane oxidation at this
temperature. Above 625 °C, we conjecture that the high amount
of CO formed is mainly due to the support, while the main role
of the palladium deposed as the active phase is to improve the
The catalysts were tested in the reaction of methane oxidation
under the same conditions as described for Pd/BN samples. The
results are presented in Table 8. The light-off temperatures (50%
methane conversion) increased from Ag/BN_glc (597 °C) to
Au/BN_glc (613 °C) to Pt/BN_glc (614 °C) to Pd/BN_glc (634
°C) to BN (672 °C); the same trend is followed by the onset
temperatures, which are 20-30 °C lower than the light-off
temperatures. The main products of methane oxidation using
boron-nitride-supported noble metal catalysts were CO and CO2.
Figure 9. Activity for methane oxidation over boron-nitride-supported
palladium catalysts: (a) in the presence of water, (b) in dry conditions

BN-Supported Noble Metal Catalysts
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12579
Figure 11. Concentrations of CH4 (red), CO (blue), and CO2 (green) vs time on stream during CH4 oxidation over Pd/BN_gly.
TABLE 7: Physicochemical Properties of M/BN Catalysts,
use under severe conditions (high temperature, humid atmo-
sphere). Several catalysts were prepared from commercial BN
of about 130 m2 g-1 of surface area on which a low percentage
of noble metal was deposed. The surface area of the support
was significantly enhanced (about 250 m2 g-1) by the chemical
treatment, which led to the metal deposition, in most cases
without hydrolysis or oxidation. The enhancement is due to the
edge effect on the platelets, creating nanoporosity and a higher
surface area. To test the conditions of deposition of the noble
metal and the efficiency of the resulting catalysts, several
preparation procedures have been applied, leading to significant
differences in the metal particle sizes. Through the use of
palladium as the reference metal and acetylacetonate as the
precursor reduced by hydrazine, the various dispersing media
tested have led to important differences in the particle size, from
7 to 13 nm, and the best results have been obtained using
ethylene glycol as the dispersing phase. Unfortunately the
palladium nanoparticles deposed on the support seemed to gather
and even collapse into larger aggregates during the thermal
oxidation-reduction cycles that were applied to the samples.
Through the use of the same support and the same deposition
conditions, major differences were observed when different
noble metals were used, illustrating the complexity of the rules
that govern metal deposition from precursors and dispersing
phases. Silver and gold seem to yield the most promising results.
Catalytic tests of methane oxidation performed under severe
conditions have shown a good stability of the catalysts and a
higher activity in the presence of water than that in dry
atmosphere but excessively high conversion temperatures.
with M ) Pd, Pt, Ag, Au
SBET
metal
(wt %)
metal particle size
(nm)
catalyst
(m2 g-1
)
Pd/BN_glc
Pt/BN_glc
Ag/BN_glc
Au/BN_glc
221
220
250
249
0.74
0.73
1.16
0.34
16
12
31
16
TABLE 8: Methane Conversion and the Ratio between CO
and CO2 Formation on Different M/BN Catalysts (Wet
Conditions)
T (°C)
400 500 550
600
625
650
700
CH4 Conversion (%)
Pd/BN_glc
Pt/BN_glc
Ag/BN_glc
Au/BN_glc
BN calcined
0.2
0.3
0
0.3
0
0.8
0.6
0
1.3
3.6
4.0
1.6
0
16.7 36.4 74.1
30.1 66.7 97.9 100.0
53.0 89.0 99.2
29.1 68.8 98.4
99.9
99.8
99.9
99.3
0.3
0
0
1.3
9.4
CO/CO2
0
0.7
3.5
0
Pd/BN_glc
Pt/BN_glc
Ag/BN_glc
Au/BN_glc
BN calcined
0
0
0
0
0
0
0
0
0
0
16.3 12.6
3.2
0.6
0.5
0.8
5.6
0.1
0.1
0.1
0.1
0.5
4.6
7.0
35.3
0
2.5
1.8
3.7
0
0
At low temperatures the most abundant reaction product
obtained was CO for all the catalysts studied, with Pt/BN_glc
presenting the best selectivity for CO2 above 550 °C. The
selectivity toward CO was decreased by the addition of noble
metals as active phases. The CO/CO2 ratios measured over BN
and BN-supported metal catalysts during methane oxidation are
reported in Table 8. It can easily be seen that if the methane
conversions on the Pt/BN_glc and Au/BN_glc catalysts at 600
°C are similar (Table 8), then the production of CO2 is more
pronounced when the active phase is platinum.
Acknowledgment. The authors thank W. Desquesnes and
co-workers for isotherm measurements and M. Aouine for TEM
pictures. G.P. thanks the ECO-NET project of the French
Government for financial support.
Surprisingly, the silver-based catalyst exhibited the highest
performance despite a larger particle size. However this could
be explained by the larger amount of silver deposed on the
support; rescaling the conversion values for the same amount
of noble metal led to a conversion at 600 °C of 46% for Ag,
41% for Pt, and 85% for Au respectively, which is more
consistent with the published results.21
References and Notes
(1) Wu, J. C. S.; Lin, Z. A.; Pan, J. W.; Rei, M. H. Appl. Catal., A
2001, 219, 117.
(2) Lin, C. A.; Wu, J. C. S.; Pan, J. W.; Yeh, C. T. J. Catal. 2002,
210, 39.
(3) Wu, J. C. S.; Fan, Y. C.; Lin, C. A. Ind. Eng. Chem. Res. 2003,
42, 3225.
(4) Jacobsen, C. J. H. J. Catal. 2001, 200, 1.
(5) Taylor, S. H.; Pollard, A. J. J. Catal. Today 2003, 81, 179.
(6) Wu, J. C. S.; Chen, W. C. Appl. Catal., A 2005, 289, 179.
(7) Lee, J. H.; Trimm, D. L. Fuel Process. Technol. 1995, 42, 339.
(8) Su, S. C.; Carstens, J. N.; Bell, A. T. J. Catal. 1998, 176, 125.
4. Conclusions
Boron nitride can be considered as a good candidate support
for the preparation of high-performance catalysts, especially for

12580 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Postole et al.
(9) Burch, R.; Urbano, F. J.; Loader, P. K. Appl. Catal., A 1995, 123,
173.
(10) Ribeiro, C. F.; Chow, M.; Dalla-Betta, R. A. J. Catal. 1994, 146,
537.
(11) Cullis, C. F.; Willatt, B. M. J. Catal. 1984, 86, 187.
(12) Postole, G.; Caldararu, M.; Ionescu, N. I.; Bonnetot, B.; Auroux,
A.; Guimon, C. Thermochim. Acta 2005, 434, 150.
(13) Perdigo´n-Melo´n, J. A.; Auroux, A.; Guil, J. M.; Bonnetot, B. Stud.
Surf. Sci. Catal. 2002, 143, 227.
(14) Postole, G.; Gervasini, A.; Auroux, A.; Bonnetot, B. Chem. Eng.
Trans. 2004, 4, 271.
(15) Yao, H. C.; Sieg, M.; Plummer, H. K., Jr. J. Catal. 1979, 59, 365.
(16) Epling, W. S.; Hoflund, G. B. J. Catal. 1999, 185, 5.
(17) Carstens, J. N.; Su, S. C.; Bell, A. T. J. Catal. 1998, 176, 136.
(18) Perdigo´n-Melo´n, J. A.; Auroux, A.; Bonnetot, B. J. Therm. Anal.
Calorim. 2003, 72, 443.
(19) Fie´vet, F. Polyol Process. In Fine Particles, Synthesis, Character-
ization and Mechanisms of Growth; Sugimoto, T., Ed.; Marcel Dekker:
New York, 2000; p 460.
(20) Yang, C.; Xing, J.; Guan, Y.; Liu, J.; Liu, H. J. Alloys Compd.
2004, 385, 283.
(21) Gelin, P.; Primet, M. Appl. Catal., B 2002, 39, 1.
(22) Zwinkels, M. F. M.; Ja¨rås, S. G.; Menon, P. G.; Griffin, T. A. Catal.
ReV.-Sci. Eng. 1993, 35 (3), 319.
(23) Choudhary, T. V.; Banerjee, S.; Choudhary, V. R. Appl. Catal., A
2002, 234, 1.
(24) Centi, G. J. Mol. Catal. A: Chem. 2001, 173, 287.
(25) Haruta, M. Catal. Today 1997, 36, 153.
(26) Kundakovic, Lj.; Flytzani-Stephanopoulos, M. J. Catal. 1998, 179,
203.
(27) Iglesias-Juez, A.; Ferna´ndez-Garcia, M.; Martinez-Arias, A.; Schay,
Z.; Koppa´ny, Zs.; Hungria, A. B.; Fuerte, A.; Anderson, J. A.; Conesa, J.
C.; Soria, J. Top. Catal. 2004, 30/31, 65.
(28) Delalu, H.; Sa¨ıd, J.; Cohen-Adad, R. J. Chim. Phys. Phys. Chim.
Biol. 1980, 77, 277.
(29) Lindquist, D. A.; Borek, T.; Kramer, S.; Narula, C.; Johnstone,
G.; Schaeffer, R.; Smith, D.; Paine, R. T. J. Am. Ceram. Soc. 1990 73,
757.
(30) Dibandjo, P.; Bois, L.; Chassagneux, F.; Cornu, D.; Le´toffe´, J. M.;
Toury, B.; Babonneau, F.; Miele, P. AdV. Mater. 2005 17, 571.
(31) Han, Wei-Qiang; Brutchey, R.; Tilley, T. D.; Zettl, A. Nano Lett.
2004, 4, 173.
(32) Auroux, A. Top. Catal. 1997, 4, 71.
(33) Jouguet, B.; Gervasini, A.; Auroux, A. Chem. Eng. Technol. 1995,
18, 243.
(34) Patterson, A. L. Phys. ReV. 1939, 56, 978.
(35) Datye, A. K.; Bravo, J.; Nelson, T. R.; Atanasova, P.; Lyubovsky,
M.; Pfefferle, M. Appl. Catal., A 2000, 198, 179.
(36) Lieske, H.; Lietz, G.; Hanke, W.; Voelter, J. Z. Anorg. Allg. Chem.
1985, 527, 135.
(37) Yang, S.; Maroto-Valiente, A.; Benito-Gonzales, M.; Rodriguez-
Ramos, I.; Guerrero-Ruiz, A. Appl. Catal., B 2000, 28, 223.
(38) Zwinkels, M. F. M.; Jaras, S. G.; Menon, P. G. Catal. ReV.-Sci.
Eng. 1993, 35, 319.
(39) Oh, S. H.; Mitchell, P. J.; Siewert, R. M. J. Catal. 1991, 132,
287.
(40) Muto, K. I.; Katada, N.; Niwa, M. Appl. Catal., A 1996, 134,
203.
(41) Lyubovsky, M.; Pfefferle, L. Appl. Catal., A 1998, 173, 107.