Preparation of Au-Pd/C catalysts by adsorption of metallic species in aqueous phase for selective oxidation

Article abstract of DOI:10.1016/j.cattod.2010.04.010

Au/C and Au-Pd/C catalysts were prepared on SX PLUS activated carbon using an adsorption method in which the precursor(s)-support interactions in aqueous solution were sought to be optimized. pH windows where maximum adsorption occurs were identified for four bimetallic cases, varying the incorporation order of the adsorbed metallic precursors and the oxidation state of the first metal introduced. All samples were characterized by XPS and SEM/EDXS, and the above-mentioned parameters were found to have an influence on the surface microstructure of the samples and on the activity of the bimetallic catalysts obtained. This preparation method leads to highly active catalysts for the selective oxidation of glucose, with the activity correlated with high surface Pd/C ratios.

Full text of DOI:10.1016/j.cattod.2010.04.010

Catalysis Today 157 (2010) 77–82
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Preparation of Au–Pd/C catalysts by adsorption of metallic species in aqueous
phase for selective oxidation
Sophie Hermans∗, Aurore Deffernez, Michel Devillers
Institut IMCN, Université catholique de Louvain, 1/3 Place Louis Pasteur, B-1348 Louvain-la-Neuve, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 13 May 2010
Au/C and Au–Pd/C catalysts were prepared on SX PLUS activated carbon using an adsorption method
in which the precursor(s)–support interactions in aqueous solution were sought to be optimized. pH
windows where maximum adsorption occurs were identified for four bimetallic cases, varying the incor-
poration order of the adsorbed metallic precursors and the oxidation state of the first metal introduced.
All samples were characterized by XPS and SEM/EDXS, and the above-mentioned parameters were found
to have an influence on the surface microstructure of the samples and on the activity of the bimetallic
catalysts obtained. This preparation method leads to highly active catalysts for the selective oxidation of
glucose, with the activity correlated with high surface Pd/C ratios.
Keywords:
Adsorption
Selective oxidation
Heterogeneous catalysis
Carbon
Palladium
Gold
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
with the opposite charge. This implies an electrostatic mechanism
of adsorption. This concept was used by Regalbuto and co-workers
Gold-based catalysts have proven important for environmental
and fine chemicals applications [1]. Supported gold catalysts are
active in a variety of reactions, including in liquid phase [2]. Three
main groups have demonstrated the activity of carbon-supported
Au to oxidize alcohols [3–5], carbohydrates [6,7] and aldehydes [8].
These recent developments have led researchers to investigate the
effect of adding other metals to gold within supported catalysts.
Combinations of gold with platinum group metals have displayed
potential for industrial applications, and in particular Pd–Au cata-
lysts supported on carbon for liquid phase oxidations [9,10].
It is known that the activity of supported catalysts depends on
the size and distribution of the metal particles, which depends on
the preparative method and the support [11]. Many techniques
allow the preparation of highly dispersed gold catalysts, such as
using precursors containing Au and the support metal compo-
nent, favoring strong support–Au interactions, using pre-formed
colloidal Au. However, most of these do not use carbon as support
and are not suitable for bimetallic materials. In addition, acidic,
neutral and/or basic surface groups (due to hetero-atoms, mainly
O, N and S) present on the surface of activated carbon influence its
surface characteristics and adsorption behavior [12]. Consequently,
charged surface sites exist in aqueous phase, depending on the pH,
and relying amongst other things on the point of zero charge (PZC).
Below that value the carbon surface is positively charged, while
above that value the surface is negatively charged, attracting ions
[13] to obtain highly dispersed and highly loaded Pt/carbon mate-
rials. We believe that it is important to go one step further and to
study also the influence of the metallic species charge present in
solution on their adsorption behavior.
In this paper, our goal is to maximize the interactions
between the metallic precursor(s) and the activated carbon sup-
port in aqueous solution, by studying the adsorption of Au and
Pd species on activated carbon SX PLUS in order to prepare
Au(5 wt.%)–Pd(5 wt.%)/C catalysts. To do so, adsorption curves for
Au and Pd on C were established in order to identify the pH win-
dows where a maximal amount of metal adsorbs. The influence
of the incorporation order of the two metallic precursors and the
oxidation state of the first metal introduced when adsorbing the
second one were also studied. In order to test the simple elec-
trostatic view to explain the maximum adsorption windows, the
respective charges of the support surface and of the metallic species
in solution were determined. The metal speciation in solution was
calculated from the stability constants of the various compounds
involving the H2O, OH− and/or Cl− ligands. Finally, the Au–Pd sam-
ples (prepared within the identified optimal adsorption windows)
were reduced to bimetallic Au–Pd/C catalysts and tested in glu-
cose selective oxidation in comparison with their monometallic
analogues.
2. Experimental
2.1. Starting materials
∗
An SX PLUS activated carbon supplied by NORIT (abbreviated as
Corresponding author. Tel.: +32 10 472810; fax: +32 10 472330.
E-mail address: Sophie.Hermans@uclouvain.be (S. Hermans).
C in this paper, SBET ≈ 900 m2 g−1, particle size = 50–100 ␮m, total
0920-5861/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2010.04.010

78
S. Hermans et al. / Catalysis Today 157 (2010) 77–82
acidity = 32.0 mmol/100 g) was used as support. The metal pre-
cursors were Pd(OAc)2 (Aldrich, 98%) and HAuCl4·3H2O (Aldrich,
99.9+%).
solids recovered by filtration at the end of the adsorption experi-
ments.
2.5. Control experiments
2.2. Preparation of starting monometallic Pd and Au samples
In order to check the possibility of bulk precipitation, control
experiments were carried out at ambient temperature as follows,
using the highest concentration used for the preparations, without
support. 50 mg of Pd(OAc)2 was solubilized in 2 mL concentrated
HNO3. Then 50 mL of water was added and the solution stirred until
complete dissolution. Solid Na2CO3 was added cautiously in order
to increase the pH. This experiment was repeated in order to obtain
solutions at various fixed pH. The solutions were left to stand for
24 h before observation.
Two carbon-supported Pd samples (activated and non-
activated) and one activated Au/C sample containing 5 wt.% of metal
were prepared, based on earlier studies [12,14]. The first Pd sam-
ple was prepared by impregnation (24 h contact time) in aqueous
solution at pH 4 from Pd(OAc)2 solubilized with nitric acid and is
noted Pd(OAc)2/C. The second Pd sample was prepared similarly
but with an additional reduction step by formalin at pH 10 (1 h at
80 ◦C), and is simply named Pd/C. The Au/C sample was prepared
by impregnation (24 h) in aqueous solution of HAuCl4 at pH 10, and
reduction by formalin at 80 ◦C (pH 10, 1 h).
2.6. Physico-chemical characterization techniques
X-ray photoelectron spectroscopy (XPS) was performed on an
SSI-X-probe (SSX-100/206) spectrometer from Fisons. The sam-
ples were stuck onto troughs with double-sided adhesive tape,
then placed on an insulating home-made ceramic carrousel with
a nickel grid 3 mm above the samples, to avoid differential charg-
ing effects. A floodgun set at 8 eV was used for charge stabilization.
The energy scale was calibrated by taking the Au 4f7/2 binding
energy at 84 eV. The C1s binding energy of contamination car-
bon set at 284.8 eV was used as internal standard value. Data
treatment was performed with the CasaXPS program (Casa Soft-
ware Ltd). The analyses of palladium and gold were based on the
Pd3d and Au4f photopeaks and the following constraints were
used for decomposition: intensity ratios I(Pd 3d5/2)/I(Pd 3d3/2) = 1.5
and I(Au 4f7/2)/I(Au 4f5/2) = 1.33, FWHM ratios = 1, ꢀ(Pd 3d5/2 − Pd
3d3/2) = 5.26 eV and ꢀ(Au 4f7/2 − Au 4f5/2) = 3.67 eV. Scanning Elec-
tron Microscopy (SEM) analyses were performed on a field effect
gun Digitan Scanning Electron Microscope (DSM 982 Gemini from
LEO), equipped with an energy dispersive X-ray system (EDAX
Phoenix equipped with a CDU LEAP detector). The powder sam-
ples were pressed onto conducting double-face adhesive tape fixed
onto 0.5ꢀꢀ aluminum specimen stubs from Agar Scientific. The quan-
tification of Pd and Au in solution was made by atomic absorption
analysis, using a Perkin-Elmer 3110 spectrometer equipped with
an air–acetylene flame atomizer. Calibration curves (from 1 to
10 mg/L for Pd and from 1 to 15 mg/L for Au) were realized with
standard solutions obtained by dilution of commercial palladium
(1006 ␮g/mL, Acros) and gold (1003 ␮g/mL, Aldrich) solutions.
2.3. Determination of supports surface charge
The point of zero charge (PZC) was determined by a modified
literature method [15], by bringing 1 g of solid sample in contact
with 10 mL of water at a fixed pH, and measuring the resulting
pH after 24 h. This operation is repeated over a range of initial pH
values (3–12), and the PZC corresponds to the plateau obtained
when plotting the measured pH versus the initial fixed pH.
2.4. Establishment of adsorption curves
The general procedure is to bring into contact a sample of sup-
port suspended in water with a solution of metal precursor for 24 h
and then to determine the amount of non-adsorbed metal in the
filtrate, and to repeat this for a range of pH values. The pH was
fixed by the addition of Na2CO3, NaOH or HNO3. The monometallic
cases were described previously [12,14]. For the bimetallic Au–Pd
samples (noted M2 − M1/C where M1 is the first metal introduced),
the influence of the two metal precursors incorporation order and
of the oxidation state of the first metal introduced were stud-
ied. Four bimetallic adsorption curves were established. When the
starting monometallic sample was not reduced before incorpo-
rating the second metal, the experimental procedure consisted in
the following: an aqueous solution of given first metal concentra-
tion (calculated by considering a monolayer of metallic precursor
molecules on the SX PLUS carbon surface and neglecting the area
developed by the micropores [12], i.e. 0.025 g of HAuCl4·3H2O or
0.056 g of Pd(OAc)2 solubilized with nitric acid) at fixed pH was
added to an aqueous suspension of 0.238 g carbon pre-conditioned
at the same pH value in order to get a final volume of 100 mL
and the suspension was agitated for 24 h. Then, 50 mL of a solu-
tion containing the second metal precursor pre-conditioned at
the same pH value was added and stirred for an additional 24 h
before filtration. Depending on the incorporation order, the two
adsorption curves obtained between pH 1 and 14 will be called
Pd(OAc)2–HAuCl4/C and HAuCl4–Pd(OAc)2/C. If a reduction step
of the first metal (M1) was added, the second metallic precursor
(0.05 g of HAuCl4·3H2O or 0.053 g of Pd(OAc)2) was solubilized in
50 mL water before being stabilized at a given pH and transferred to
an aqueous suspension of 0.238 g monometallic sample (reduced
Pd(5 wt.%)/C or Au(5 wt.%)/C) pre-conditioned at the same pH value
and the mixture was stirred during 24 h before filtration. Again,
depending on the incorporation order, these two bimetallic cases
will be named Pd(OAc)2–Au/C and HAuCl4–Pd/C. The adsorption
filtrates containing the non-adsorbed Pd and/or Au were analyzed
by atomic absorption spectrometry, and the amount of metals
adsorbed on the support at each different pH value was deduced
by subtraction. In this study, we call “adsorption samples” all the
2.7. Catalysis
Bimetallic catalysts and their monometallic analogues were
synthesized by impregnation in aqueous solution, using the pH
values where maximum adsorption occurs, and maintaining it
throughout all synthetic steps. The procedure is similar to the
establishment of the adsorption curves, with an additional step
consisting in activating in situ the catalysts with formalin or NaBH4.
Catalytic tests were performed in a thermostatised double-walled
glass reactor equipped with an automatic titration device for pH
control, as described previously [12,14]. The materials were com-
pared, in the oxidation of glucose, on the basis of the yield in
gluconic acid (only detected secondary product is fructose).
3. Results and discussion: adsorption curves
3.1. Determination of support surface charge
The PZC value obtained for C was 8.5 [12]. Two other PZC values
were determined: (i) after adsorption of Pd acetate on C (sam-
ple named Pd(OAc)2/C) and (ii) after its reduction, giving sample

S. Hermans et al. / Catalysis Today 157 (2010) 77–82
79
Table 1
XPS results for monometallic Pd(OAc)2/C, Pd/C [12] and Au/C samples.
M/C (×100)a
Sample (M = Pd, Au)
Calc.
Exp.
Pd(OAc)2/C
Pd/C
Au/C
0.6
0.6
0.3
13.1
4.0
0.86
a
M/C atomic intensity ratios: calc. = values calculated from the bulk composition
in the case of 100% adsorption and without taking into account the carbon contri-
bution of the ligands, exp. = experimental values retrieved from XPS data analysis.
named Pd/C: values of 4.7 and 9.6 were obtained, respectively. In
the first case, the presence of remaining acetates, and the fact that
the sample was prepared at pH 4, might explain the obtained acidic
value. In the second case, the fact that the Pd/C sample presents a
PZC value superior to the PZC value of C could be explained by the
reduction by formalin of some of the surface acidic functions during
the activation step.
3.2. Characterization of starting monometallic Pd and Au samples
Samples Pd/C [12], Pd(OAc)2/C and Au/C were characterized by
XPS and SEM (Table 1 and Fig. 1). The Pd/C surface atomic ratios
measured for the two Pd samples are very high (0.04 and 0.13)
in comparison with the value calculated on the basis of the bulk
composition (0.006), suggesting a high concentration of palladium
on the surface, probably in a relatively dispersed form. In oppo-
site, the Au/C surface ratio of the Au/C sample (0.0086) is more
similar to the calculated value (0.003), meaning that either gold
penetrates more in the porous network or is present as larger aggre-
gates. This was visualized by electron microscopy. The SEM images
of the Pd(OAc)2/C sample (Fig. 1a) show a precipitate of Pd covering
the carbon surface while the electron micrographs of the Pd/C sam-
ple (Fig. 1b) show a homogeneous distribution of relatively small Pd
particles (∼10–15 nm in size). This might explain the huge differ-
ence of Pd/C ratios measured by XPS (respectively, 0.13 and 0.04),
as the precipitate must be mainly located at the external surface
while the quite smaller particles might have penetrated the porous
network more hence be less detected by XPS. The SEM images of
the Au/C sample (Fig. 1c) reveal an agglomeration of metallic parti-
cles on the carbon surface, with particles sizes up to 200 nm. In all
three cases, the microstructure obtained is far from ideal, as far as
particles sizes are concerned, especially in the case of gold, where
obtaining small particles is known to be crucial for the catalytic
activity.
When studying the adsorption of Au on C, it was observed that
the whole amount of metal introduced in solution was adsorbed
on the support over the whole pH range [14]. The XPS Au/C surface
ratios were weak, with a maximum value of 0.012 at pH 9 [14]. The
distribution of Au(III) species in solution was determined from the
stability constants [16,17]: all gold species envisaged (Cl− and OH−
ligands) were negatively charged [14]. This explains the absence
of clear maximum in the Au/C adsorption curve. The weak Au/C
ratios determined by XPS could be related to the agglomeration
of Au in the form of big particles. This might be explained by the
concept of aurophilicity [18,19], which favors the close packing of
adsorbed gold ions. Moreover, the reducing nature of carbon limits
Fig. 1. SEM images obtained for starting monometallic samples: (a) Pd(OAc)2/C
(50 000×), (b) Pd/C (200 000×) [12], (c) Au/C (50 000×).
the deposition of gold in a dispersed state, as observed for Au/C
catalysts prepared by deposition–precipitation [18]. This last effect
must be the most important, as reduction of gold by the surface
must occur readily and give the observed big particles.
Table 2
Summary of observations for bimetallic adsorption samples.
Sample
Support PZC
Maximum adsorption
Optimum (highest M/C XPS ratios)
SEM
Pd(OAc)2–Au/C
8.5 (C)
8.5 (C)
4.7 (Pd(OAc)2/C)
9.6 (Pd/C)
pH 3–12
pH 3–12
pH 1–9
pH 1–9
pH 5
pH 10
pH 9
None
Big particles
Pd(OAc)2–HAuCl4/C
HAuCl4–Pd(OAc)2/C
HAuCl4–Pd/C
Pd precipitates + big Au particles (>200 nm)
Pd precipitates + big Au particles (>200 nm)
Small Pd particles (<10 nm) + big Au particles (>50 nm)

80
S. Hermans et al. / Catalysis Today 157 (2010) 77–82
Fig. 3. Pd/C and Au/C surface ratios determined by XPS for bimetallic adsorption
samples: Pd(OAc)2–Au/C, as representative example.
Au at pH below 9, while no Pd leaching is observed (HAuCl4–Pd/C,
Fig. 2d).
Some solid samples arising from each bimetallic adsorption
curve have been characterized by XPS. The results obtained with
the Pd(OAc)2–Au/C samples are illustrated as an example in Fig. 3.
The first three adsorption curves reveal pH values where the Pd/C
and Au/C atomic intensity ratios are maximum: at pH 5, 10 and 9 for
the Pd(OAc)2–Au/C, Pd(OAc)2–HAuCl4/C and HAuCl4–Pd(OAc)2/C
curves, respectively. The adsorption curve giving the highest Pd/C
ratios (between 0.058 and 0.252), in comparison with the calcu-
lated value (0.0125) on the basis of the bulk composition in the
case of 100% adsorption, is Pd(OAc)2–Au/C. In opposite, it is the
HAuCl4–Pd(OAc)2/C curve which gives the highest Au/C ratios (up
to 0.018) when compared with the calculated value (0.0028). The
analysis by XPS of the fourth case (HAuCl4–Pd/C) did not give a pH
range where Pd/C and Au/C ratios were optimized, and the absolute
values of Pd/C ratios were much smaller.
The adsorption samples arising from the first three curves and
giving the highest Pd/C and Au/C ratios by XPS have been exam-
ined by SEM/EDXS. The images of the sample prepared at pH 5 in
the case of the Pd(OAc)2–Au/C curve (Fig. 4a) show big particles
(>100 nm) and the EDXS analysis gives the signals corresponding
to carbon and gold. In the two other cases (Pd(OAc)2–HAuCl4/C
and HAuCl4–Pd(OAc)2/C), SEM analysis of, respectively, the sam-
ples prepared at pH 10 and 9 revealed precipitates of Pd and big
particles of Au (>200 nm) covering the carbon surface (Fig. 4b and
c). EDXS analyses within the microscope confirmed the nature of
each type of particle observed. The difference between these two
cases is the density of Pd precipitate, which is higher when Pd is
introduced first. We have also analyzed by SEM/EDXS (Fig. 5) two
adsorption samples (prepared at pH 5 and pH 10) belonging to
the fourth adsorption curve (HAuCl4–Pd/C). No major difference
between these two samples was observed. Detailed examination
of the images shows a distribution of small Pd particles (<10 nm)
and big Au particles (>50 nm) in the two cases.
In order todeterminewhether an electrostatic adsorption model
can be invoked, the distribution of Pd and Au species in solution was
determined from stability constants [16,17,20] using the experi-
mental conditions (pCl = 2.77) used for establishment of the two
adsorption curves without reduction. Fig. 6 shows the speciation
diagram for Pd(II): neutral [Pd(OH)2(H2O)2] is the major species for
pH 4–12. Au speciation is identical to the distribution obtained pre-
viously [14]: Au species are negatively charged over the whole pH
range. Thus, the metals charge in solution does not explain why we
do not obtain the same trend for the two bimetallic curves without
reduction: if the Au precursor is introduced first, the total amount
of gold remains adsorbed at all pH values, while if the Pd precursor
is incorporated first, the maximum adsorption window of gold is
reduced (pH < 9). The surface of the Pd(OAc)2/C sample is negatively
charged at pH superior to 4.7 as evidenced by PZC determination.
Fig. 2. Bimetallic adsorption curves in aqueous phase: (a) Pd(OAc)2–Au/C, (b)
Pd(OAc)2–HAuCl4/C, (c) HAuCl4–Pd(OAc)2/C, (d) HAuCl4–Pd/C, (ꢀ) amount of Pd
adsorbed on C, (♦) amount of Au adsorbed on C.
3.3. Bimetallic adsorption curves
Four bimetallic adsorption curves have been determined, dif-
ferentiated by the incorporation order and the oxidation state of
the first metal adsorbed (Table 2). The adsorption curve named
Pd(OAc)2–Au/C gives a maximum adsorption window between
pH 3 and 12 for Pd, and reveals that no leaching of Au occurs
(Fig. 2a). This pH range is identical for the adsorption curve
Pd(OAc)2–HAuCl4/C (Fig. 2b). By opposition, the window of Au
adsorption is reduced between pH 1 and 9 when introducing the Pd
precursor first (HAuCl4–Pd(OAc)2/C, Fig. 2c). Finally, adsorbing the
Au precursor on a Pd(5 wt.%)/C sample gives a total adsorption of

S. Hermans et al. / Catalysis Today 157 (2010) 77–82
81
Fig. 5. SEM images obtained for the bimetallic samples HAuCl4–Pd/C prepared: (a)
at pH 5 (50 000×), (b) at pH 10 (100 000×).
adsorbed. Characterization by SEM/EDXS of the bimetallic samples
giving the best Pd/C and Au/C XPS ratios showed that the incorpora-
tion order also had an influence on the surface microstructure. Gold
is present as big particles (Au + C detected by EDXS) while Pd dis-
plays better dispersed smaller particles (undetectable by EDXS), but
these particle sizes are far from ideal. It appears that an electrostatic
mechanism does not allow the windows of maximum adsorption to
be explained. A deposition–precipitation process within the win-
dow where neutral species are present must be invoked. Control
experiments were carried out, by preparing a range of Pd solutions
at various fixed pH (between 1 and 14), by dissolving Pd(OAc)2
at the highest concentration used in the preparation. These solu-
tions were left to stand without support for 24 h and homogeneous
bulk precipitation indeed occurred. This precipitation phenomenon
might be slow in dilute solutions without support but is acceler-
ated by the presence of the support in suspension. This also means
that speciation data calculated from a restricted set of equilibrium
Fig. 4. SEM images obtained for the bimetallic samples adsorbed at optimized pH:
(a) Pd(OAc)2–Au/C at pH 5 (20 000×), (b) Pd(OAc)2–HAuCl4/C at pH 10 (20 000×),
(c) HAuCl4–Pd(OAc)2/C at pH 9 (20 000×).
When the Pd precursor is introduced first and if we consider that
its major form is the neutral complex [Pd(OH)2(H2O)2], the charge
state of the support is not modified but the available surface area
for the adsorption of gold has decreased (if one considers that gold
would preferably sit on the bare support rather than on palla-
dium). So, the adsorption of gold at pH superior to the PZC value
is not favored due to the identical charge of the support and Au
metallic species, and due to the lower available surface area. The
HAuCl4–Pd/C adsorption curve could be explained similarly: the
charge of the Pd/C surface and the Au species are both negative
at pH superior to 9.6, and the Au adsorption is reduced in this pH
window. By opposition, the adsorption of the acidic gold precur-
sor in first position could decrease the PZC value of the support.
However, the presence of Au has no effect on the amount of Pd
Fig. 6. Distribution of Pd(II) species as a function of pH: H2O, Cl− and OH− as ligands,
pCl = 2.77.

82
S. Hermans et al. / Catalysis Today 157 (2010) 77–82
order of the two metals and the oxidation state of the first metal
introduced influence both the structure and the performance of the
catalysts. Only metallic gold was detected in all catalysts by powder
XRD. This observation can be related to the relative sizes of Au and
Pd particles, which as described above, are very big in the case of
gold and relatively small in the case of Pd. The bimetallic material
activated with NaBH4, showing the highest Pd/C and Au/C ratios
and a precipitate covering the carbon surface, is the most active in
the oxidation of glucose into gluconic acid.
5. Conclusion
This work aimed at studying the adsorption of Au and Pd pre-
cursor(s) on activated carbon in aqueous phase. Adsorption curves
were realized to identify maximum adsorption windows, followed
by characterization of the adsorption samples. The results of four
bimetallic adsorption curves were slightly different: the precur-
sors’ incorporation order influences the amount of metal adsorbed
on the support and the microstructure of the samples obtained.
The presence of Pd on the support has an effect on the amount of
Au adsorbed, while the presence of Au does not seem to influence
the quantity of adsorbed Pd. SEM images of the sample presenting
the highest XPS Pd/C and Au/C ratios revealed a Pd precipitate in
contact with big Au particles. The results allowed to discard a sim-
ple electrostatic model because the maximum adsorption ranges
correspond to the existence of neutral Pd species and negative Au
species, while the support is either positively or negatively charged
but were consistent with a surface deposition–precipitation pro-
cess. The Au–Pd/C catalysts prepared by this method are very active
in glucose oxidation, and more so than their monometallic ana-
logues, indicating a synergetic effect between the two metals. They
are characterized by high Pd/C surface ratios and relatively small
metallic Pd particles. The metals incorporation order and oxidation
state as well as the activating agent used also influence the catalytic
performance.
Fig. 7. Comparison of catalytic results for Pd(5 wt.%)/C and Au(5 wt.%)–Pd(5 wt.%)/C
materials in glucose oxidation: (ꢀ) Pd(OAc)2/C activated with formalin, (᭹)
Pd(OAc)2/C activated with NaBH4, (ꢁ) HAuCl4–Pd(OAc)2/C activated with formalin,
(ꢁ) HAuCl4–Pd(OAc)2/C activated with NaBH4.
constants based on hypothetical Pd species are not accurate, and
even more when considering the preparation of heterogeneous
catalysts as the solid supports interferes in the kinetics for some
species formation. Precipitation on the surface of the support would
result in irregular aggregates and big particles. This phenomenon
explains the large windows of adsorption, the discrepancy between
surface and metallic species charges when adsorption occurs, the
SEM images showing precipitates and the irregular XPS ratios. In
the case of gold, self-reduction in solution, or when contacting the
carbon surface, must also be considered. These two events would
participate in the growth of big particles at the surface.
4. Preliminary catalytic results
Bimetallic catalysts were prepared by impregnation in aque-
ous solution by keeping the pH in the maximum adsorption
windows, followed by chemical activation. The Au–Pd/C cata-
lysts prepared here give performances superior to their Pd/C and
Au/C (totally inactive) monometallic analogues (Fig. 7), demon-
strating the synergetic effect between these two metals. The
optimization of the synthetic conditions allowed the preparation
of catalysts which are active and selective, even if the particles
are big, especially in the case of gold. Thus the present syn-
thetic method is not competitive regarding particles sizes but
gives results in terms of performance. The bimetallic catalysts
described here give similar results than with the most performant
Bi(5 wt.%)–Pd(5 wt.%)/C catalysts for glucose oxidation [21] (sam-
ple Ac·PdBi/CSX+ prepared by deposition of carboxylate precursors
on the same carbon SX PLUS giving YGLU/mPd (t = 4 h) = 17.52% mg−1
or sample noted ref·PdBi/CSX+ prepared by a patented procedure
giving YGLU/mPd (t = 4 h) = 21.48% mg−1 to compare with YGLU/mPd
(t = 4 h) = ∼9.5% mg−1 with Pd–Au/C samples here). In the same test-
ing conditions [21], a commercial trimetallic Pd–Pt–Bi/C catalyst
from Degussa gave YGLU/mPd (t = 4 h) = 19.2% mg−1. However, the
Bi-based formulation suffered from a major drawback that is over-
come here: leaching of promoter element in solution.
Acknowledgments
The authors acknowledge financial support from the FNRS and
the Belgian State (Belgian Science Policy, IAP Project INANOMAT
No. P6/17), NORIT for supplying carbon SX PLUS, E.M. Gaigneaux,
M. Genet, and J.-F. Statsijns for discussions and technical assistance.
References
[1] M. Haruta, Gold Bull. 37 (2004) 27–36.
[2] L. Prati, G. Martra, Gold Bull. 32 (1999) 96–101.
[3] L. Prati, F. Porta, Appl. Catal. A 291 (2005) 199–203.
[4] S. Demirel, P. Kern, M. Lucas, P. Claus, Catal. Today 122 (2007) 292–300.
[5] G.J. Hutchings, Gold Bull. 37 (2004) 3–11.
[6] Y. Onal, S. Schimpf, P. Claus, J. Catal. 223 (2004) 122–133.
[7] M. Comotti, C. La Pina, R. Matarrese, M. Rossi, A. Siani, Appl. Catal. A 291 (2005)
204–209.
[8] S. Biella, L. Prati, M. Rossi, J. Mol. Catal. A: Chem. 197 (2003) 207–212.
[9] N. Dimitratos, A. Villa, D. Wang, F. Porta, D.S. Su, L. Prati, J. Catal. 244 (2006)
113–121.
The obtained Au–Pd/C catalysts present weak Au/C and rela-
tively high Pd/C surface ratios (measured by XPS) and metallic Pd on
the surface, suggesting a relatively good dispersion of Pd on the sup-
port. The catalysts presenting the highest Pd/C ratios give the best
catalytic activity in glucose selective oxidation. These observations
suggest that high activities are related to high Pd surface concen-
tration and to an efficient reduction into Pd metal. In addition,
SEM/EDXS characterization indicates that the least active catalysts
display separated small Pd particles and big Au particles, while the
more active materials reveal either a homogeneous distribution of
smaller Pd and Au particles, or a precipitate covering the carbon
surface. The surface structure of these bimetallic materials influ-
ences thus their catalytic activity. It implies that the incorporation
[10] L. Prati, A. Villa, F. Porta, D. Wang, D.S. Su, Catal. Today 122 (2007) 386–390.
[11] M. Haruta, Catal. Today 36 (1997) 153–166.
[12] A. Deffernez, S. Hermans, M. Devillers, J. Phys. Chem. C 111 (2007) 9448–9459.
[13] X. Hao, L. Quach, J. Korah, W.A. Spieker, J.R. Regalbuto, J. Mol. Catal. A: Chem.
219 (2004) 97–107.
[14] A. Deffernez, S. Hermans, M. Devillers, Stud. Surf. Sci. Catal. 162 (2006) 79–86.
[15] J. Park, J.R. Regalbuto, J. Colloid Interface Sci. 175 (1995) 239–252.
[16] L.G. Sillen, Stability Constants of Metal–Ion Complexes: Section I, The Chemical
Society, London, 1964.
[17] S.A. Wood, Geochim. Cosmochim. Acta 55 (1991) 1759–1767.
[18] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, Imperial College Press,
London, 2006.
[19] H. Schmidbaur, A. Schier, Chem. Soc. Rev. 37 (2008) 1931–1951.
[20] I.V. Mironov, L.D. Tsvelodub, Russ. J. Inorg. Chem. 45 (2000) 633–637.
[21] M. Wenkin, R. Touillaux, P. Ruiz, B. Delmon, M. Devillers, Appl. Catal. A 148
(1996) 181.