Design of highly active and chemoselective bimetallic gold-platinum hydrogenation catalysts through kinetic and isotopic studies

Article abstract of DOI:10.1016/j.jcat.2009.04.004

Kinetic model for the chemoselective hydrogenation of nitroaromatic compounds on Au/TiO₂ has been established by combining the Hougen-Watson formalism and isotopic studies. It has been found that, with this catalyst, the controlling step corres

Full text of DOI:10.1016/j.jcat.2009.04.004

Journal of Catalysis 265 (2009) 19–25
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Design of highly active and chemoselective bimetallic gold–platinum
hydrogenation catalysts through kinetic and isotopic studies
Pedro Serna, Patricia Concepción, Avelino Corma *
Instituto de Tecnología Química, UPV-CSIC, Universidad Politécnica de Valencia, Avda. de los Naranjos s/n, 46022 Valencia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Kinetic model for the chemoselective hydrogenation of nitroaromatic compounds on Au/TiO2 has been
established by combining the Hougen–Watson formalism and isotopic studies. It has been found that,
Received 21 January 2009
Revised 3 April 2009
Accepted 3 April 2009
Available online 8 May 2009
with this catalyst, the controlling step corresponds to the dissociation of H
dination. Taking this into account, a new bimetallic Au@Pt/TiO solid has been prepared that, when opti-
mised, increases the rate of H dissociation while preserving high chemoselectivity. The resultant catalyst
2
on gold atoms of low coor-
2
2
is the most effective catalyst reported, up to now, for hydrogenating nitroaromatic compounds in the
presence of other sensitive functionalities.
Keywords:
Hydrogenation of nitroaromatics
Bimetallic gold catalysts
Ó 2009 Elsevier Inc. All rights reserved.
Chemoselective hydrogenations
kinetics of hydrogenation reactions
1
. Introduction
2
ciation of H . Thus, in order to design a more active catalyst one
should know if the controlling step of the reaction corresponds
to the adsorption of the substituted nitroaromatic compound
The discovery of gold catalysts for the production of substituted
anilines and related derivatives has opened up new possibilities
around the chemistry of nitrogenated compounds [1,2]. Regarding
the hydrogenation of nitroaromatic compounds, the use of synthe-
sis procedures to create well-dispersed metal, and especially gold
nanoparticles, onto the appropriated support allows designing pro-
cesses where inefficient stoichiometric reducing agents such as so-
dium hydrosufide [3], iron [4], tin [5] or zinc in ammonium
through the nitro group, to the rate of H
atoms or to the surface reaction between the adsorbed nitro com-
pound and the dissociated H . It is clear that depending on the
rate-controlling step, one will take different actions to modify
the characteristics of the catalyst to improve its activity.
2
dissociation on gold
2
In the present work, a kinetic and isotopic study has been car-
ried out, which has allowed finding that the rate-controlling step
hydroxide [6], or catalysts based on Pb–Pt/CaCO
3
with iron salts
for the hydrogenation of substituted nitroaromatics on Au/TiO
corresponds to the dissociation of H on the gold nanoparticles.
Taking this into account, a new optimised bimetallic Au–Pt/TiO
catalyst has been prepared. Thanks to the presence of Pt, the
amount of dissociated H available on the catalyst surface in-
creases, resulting in a solid that is almost one order of magnitude
more active than the corresponding monometallic gold material,
and can work in solvent-free media. Meanwhile, the high chemose-
lectivity of gold is preserved.
2
in solution or H PO –Pt/C with vanadium salts in solution [7,8]
3
2
2
can be substituted by more efficient and greener catalytic systems.
In this sense, gold can be especially sensitive to discriminate differ-
ent functional groups in poly-substituted compounds, offering an
alternative for manufacturing complex chemicals and fine chemi-
cals [1,9–20]. Unfortunately, in the case of the hydrogenation of
substituted nitroaromatics to the corresponding anilines, the activ-
2
2
2
ity of the Au/TiO catalyst could be too low for practical applica-
tions [2]. The challenge was then to increase the catalyst activity
while preserving its high chemoselectivity.
2
. Materials and methods
When attempting to design a more efficient catalyst, it is highly
desirable to know the molecular mechanism and the catalytic ac-
tive sites involved. A contribution to this has been recently pre-
sented [21], in which the activation of the nitroaromatic
compound was shown to take place at Au/Ti boundaries, and the
reduction process occurs on the gold nanoparticle after the disso-
2
.1. Preparation of catalysts
Gold catalysts were synthesised by a deposition-precipitation
technique of the gold nanoparticles onto the surface of TiO
gussa, P-25). The deposition-precipitation procedure was carried
out by adding to the support an aqueous solution of HAuCl
0.01 M) containing a proper amount of gold. The resulting mixture
must be maintained for 2 h at 343 K under vigorous stirring,
2
(De-
4
(
*
Corresponding author. Fax: +34 96 3877809.
E-mail address: acorma@itq.upv.es (A. Corma).
0
021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.04.004

2
0
P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
controlling the pH of the solution with NaOH at a specific set point.
Then, the resulting powder is filtered, washed with distilled water
to remove chlorides, dried in an oven at 373 K, and finally calcined
in air at the desired temperature. Depending on the pH of deposi-
tion, loading of gold and calcination temperature, gold particles of
different sizes and morphologies can be obtained, leading to vari-
able levels of activity [22].
2.3. Isotopic HD exchange experiments
Hydrogen/deuterium (H/D) exchange experiments were carried
out in a flow reactor at 298, 348 and 393 K, according to the proce-
dure depicted in [23]. The feed gas consisted of 2 ml/min H
min D and 6 ml/min argon, and the total weight of catalyst was
125 mg. Reaction products (H , HD and D ) were analysed with a
2
, 2 ml/
2
2
2
The kinetic work was done with the Au/TiO
2
sample provided
mass spectrometer (Omnistar, Balzers). Prior to catalytic test, the
samples were activated in flowing Argon (10 ml/min) at room tem-
perature (298 K) for 30 min. Then, the gas feed was changed to the
reactant gas composition, while the temperature was maintained
at 298 K for additional 30 min. At this point, the temperature
was increased to 348 K at 10 K/min, maintained for 30 min at that
temperature and further increased to 393 K at 10 K/min.
by the World Gold Council (standard reference catalyst), which con-
tains 1.5 wt% of gold, and this has been synthesised according to
the described deposition-precipitation procedure at a pH 7, fol-
lowed by calcination in air flow at 673 K. In addition, other samples
were prepared to obtain catalysts with different levels of activity
(
see synthesis conditions in Table S1).
Finally, the bimetallic Au@Pt/TiO
impregnating a proper amount of Pt onto the Au/TiO
vided by the World Gold Council, using H PtCl as precursor, fol-
lowed by a treatment in H flow at 723 K.
2
catalysts were prepared by
sample pro-
2
3
. Kinetic experiments and models
2
6
2
2
To study the kinetic behaviour of the Au/TiO system on the
hydrogenation of nitroaromatic compounds, we have selected
nitrobenzene as a model reactant. Indeed, the kinetics of nitroben-
zene hydrogenation into aniline has been widely studied in the lit-
erature with metal catalysts such as Pd [24], Pt [25,26], Cu [27] or
Ni [28,29], but no references were found for gold-based materials.
According to Hougen–Watson/Langmuir–Hinshelwood principles,
for describing the mechanism of reactions occurring onto the sur-
face of heterogeneous catalyst [30,31], different rate expressions
can be obtained when assuming different reaction steps as the
rate-limiting step.
2
.2. Kinetic experiments
Kinetic measurements were performed in liquid phase using a
batch reactor, and the evolution of composition with reaction
time was analysed by taking samples at different times on
stream. The experiments were carried out in the same reactor
(
2 ml home-made reinforced glass vial), keeping the volume of
the reaction mixture (1 ml) and the stirring rate (1000 rpm) con-
stant, and using always the same amount of catalyst (7.1 mg of
the 1.5 wt% Au/TiO sample provided by the World Gold Council).
2
In a typical experiment, the catalyst is placed into the reactor to-
gether with the reaction mixture. The system is then purged with
In the case of the hydrogenation of nitroaromatics on Au/TiO
catalysts, it has been previously reported that the nitro groups
are activated at Au/Ti interphases [21,32], while H dissociatively
2
2
H
2
, before heating up the solution, in order to completely remove
the oxygen from the system. Under atmospheric pressure of H
the reactor is heated up to the desired temperature, and finally
the H pressure is fixed and maintained during all the experiment
adsorbs on certain positions of the gold crystals [33]. Considering
this previous knowledge, a reduced set of reasonable kinetic equa-
tions may be initially proposed as the most probable candidates to
match the behaviour of the gold catalyst. These models, which
have been summarised in Table 1, involve that both reactants, i.e.
2
,
2
at a selected set point value. Several samples were taken at differ-
ent times to have a minimum of three results at conversion levels
below 20%. The conversion-time results were fitted to a straight
line through the origin, and initial rates were obtained from the
slope.
nitrobenzene and H
fore reacting, with the adsorption of H
tive. However, it remains to be known if nitrobenzene and H
share one unique type of active site or, on the contrary, they adsorb
2
, are adsorbed onto the surface of Au/TiO
2
be-
2
molecules being dissocia-
2
Table 1
2
Plausible Hougen–Watson/Langmuir–Hinshelwood models to describe the kinetic behaviour of the nitrobenzene hydrogenation on Au/TiO catalysts.
k ꢀ KNB ꢀ KH2 ꢀ CNB ꢀ P
H
(1) Control of the surface reaction
2
r ¼ À
r ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiÁ
3
1
þ KNB ꢀ CNB
þ
K
H
ꢀ P
H
2
2
NB and H
(2) Control of the surface reaction
NB and H react after adsorbing on active sites with different nature
Control of the surface reaction
NB and H react after adsorbing on active sites with different nature
(3) In addition, H also adsorbs on the same active site as NB
2
react after adsorbing on active sites with the same nature (competitive adsorption)
k ꢀ KNB ꢀ KH2 ꢀ CNB ꢀ P
H
2
À
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiÁ
2
ð1 þ KNB ꢀ CNBÞ ꢀ 1 þ
K
H
ꢀ P
H
2
2
2
2
k ꢀ KNB ꢀ K
H
ꢀ CNB ꢀ P 2
H
2
r ¼ ꢀ
r ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiꢁ
2
À
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiÁ
2
0
1
þ KNB ꢀ CNB
þ
K
ꢀ P
H
ꢀ
1 þ
K
H
ꢀ P
H
2
H
2
2
2
k ꢀ KNB ꢀ K
ð1 þ KNB ꢀ CNBÞ ꢀ 1 þ
H
ꢀ CNB ꢀ PH2
H
2
À
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Á
2
(4) In addition, NB also adsorbs on the same active site as H
2
K
ꢀ P
H
2
þ K0 ꢀ CNB
2
NB
k
H2 ꢀ PH2
ð1 þ KNB ꢀ CNB
r ¼
r ¼ À
*
(5) Control of the adsorption/dissociation of H
2
NB and H react after adsorbing on active sites with the same nature, but the H dissociation is the limiting
2
2
Þ
2
step of the reaction
k
NB ꢀ CNB
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiÁ
(6) Control of the adsorption/dissociation of NB
1
þ
K
H
ꢀ P
H
2
2
2
NB and H react after adsorbing on active sites with the same nature, but the NB adsorption is the limiting
step of the reaction
k = kinetic constant of the surface reaction; kH2 = kinetic constant of the H
2
dissociation; kNB = kinetic constant of the nitrobenzene adsorption; KH2 = adsorption constant of
H
2
in the equilibrium; KNB = adsorption constant of nitrobenzene in the equilibrium.
*
*
C
2
NB = concentration of nitrobenzene; PH2 = pressure of H .

P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
21
on different sites (with a different nature, e.g. on the top of gold
nanoparticles or at Au/Ti interphases), and moreover which step
of the process (surface reaction, adsorption of any reactant, etc.)
is limiting the global reaction rate. According to the mathematical
formulation of the models in Table 1, it is possible to predict the
evolution of the initial reaction rate with the initial concentration
correspond to a competing adsorption between H
2
and nitroben-
zene in one unique type of active sites, or model 5 controlled by
adsorption/dissociation of H at higher CNB values, could properly
fit the experimental results within the most interesting range of
NB (0.4–1.2 mmol/l). Up to this point, three-reaction models, i.e.
2
C
1, 4 and 5, could fit the experimental results, and further work is
required to discriminate among them.
2
of nitrobenzene and initial pressure of H for each of the potential
mechanisms. Figure S1 in Supplementary Online Material graphi-
cally shows these trends, which will facilitate discrimination
among the different potential mechanisms by comparing the real
experimental results with the theoretical curves.
2
4.1.2. Influence of H pressure
Eqs. (1) and (4) could be easily distinguished from 5 by evaluat-
ing the effect of hydrogen pressure (PH2) on the initial reaction rate
We have considered a specific range of working conditions for
the kinetic study taking into account feasible industrial conditions
0
(r ) [30,31], since a non-lineal increase of the reaction rate (asymp-
totic or parabolic) should be observed for the surface reaction con-
trol (models 1 and 4), whereas a linear increase should be observed
[
34,35], previous literature on the field [7,24–29] and our knowl-
edge about the behaviour of Au/TiO catalysts [1,21,32]. This space
of research is shown in Table S2, where the range of process vari-
ables studied, i.e. concentration of substrate (CNB), H pressure
H2) and reaction temperature (T), has been presented.
2
2
if the dissociation of H is the limiting step (model 5). Therefore,
new experiments were performed at variable initial hydrogen
pressures (from 2 to 12 bar), keeping the concentration of nitro-
benzene (0.418 mmol/l), the reaction temperature (393 K) and
the amount of catalyst (7.1 mg) constant. When initial reaction
rates at different PH2 are plotted in Fig. 2, a linear increase in the
initial reaction rate can be seen when increasing the H pressure
2
within the range of pressures studied here. This kinetic behaviour
indicates that, at least for high concentrations of nitrobenzene, the
2
(
P
4
. Results and discussion
.1. Kinetic study
.1.1. Influence of the concentration of nitrobenzene on the reaction
4
adsorption/dissociation of H
see curves of reference for each model 5 in Figure S1).
We have decided to confirm the conclusion previously achieved
through a classical Hougen–Watson/Langmuir–Hinshelwood ki-
netic approach, by measuring the rate of H –D exchange with
the previous catalyst as well as with a series of Au/TiO catalysts
showing different activities for the hydrogenation of nitroaromat-
ics. Thus, several Au/TiO samples were prepared under variable
2
is the controlling step of the reaction
4
(
rate
From the kinetic equations presented in Table 1, it is possible to
infer that the effect of the initial nitrobenzene concentration on the
initial hydrogenation rate should allow an easy discrimination be-
tween models 2–3–6 and models 1–4–5, since the former group of
expressions involve a continuous increase in the initial reaction
rate with CNB, while the latter group involve a decrease in the ini-
tial rate of hydrogenation over a certain initial concentration of the
nitroaromatic compound. Thus, initial reaction rates were mea-
sured at different CNB levels under ‘‘constant” H pressure and tem-
2
perature (Fig. 1). It can be observed that for highly diluted reaction
mixtures the initial reaction rate rapidly increases when the con-
2
2
2
2
synthesis conditions (see Section 3), leading to catalysts with dif-
ferent activity levels. We could assume that if a direct correlation
between the rate of hydrogenation and the rate of H
is observed, it will be a confirmation that for Au/TiO
hydrogenation of nitroaromatics is indeed controlled by H
ciation. Results from Fig. 3 show a linear correlation between the
rate of H –D exchange for the different Au/TiO catalysts and their
activity to reduce 3-nitrostyrene, confirming that the reaction is
controlled by H dissociation on gold, and that model 5 represents
2
–D
catalysts the
disso-
2
exchange
2
2
centration of nitrobenzene increases up to
a
maximum
(r
0
= 0.31 mmol/h at CNB = 0.198 mmol/l), and then decreases when
2
2
2
increasing further the concentration of nitrobenzene. This kinetic
behaviour for CNB > 0.198 mmol/l, which is the preferred range of
nitrobenzene concentration from a process point of view, cannot
be reproduced by models 2, 3 and 6 (see the reference curves for
each model in Figure S1). On the contrary, models 1 and 4, which
2
the kinetic behaviour (see Supplementary Material for the elucida-
tion of the related equation from the elemental reaction steps).
Fig. 1. Influence of the concentration of nitrobenzene on its initial hydrogenation
rate at a constant reaction temperature (393 K), constant H
constant amount of the Au/TiO catalyst (7.1 mg). Initial reaction rates were always
calculated at conversion levels inferior to 15%.
2
pressure (8 bar) and
Fig. 2. Influence of the H
fixed reaction temperature (393 K), fixed concentration of nitroaromatic compound
(0.418 mmol/l) and fixed amount of the Au/TiO catalyst (7.1 mg).
2
pressure on the nitrobenzene hydrogenation rate at a
2
2

2
2
P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
Fig. 3. Relationship between the conversion of nitroaromatic compound (hydro-
genation of 3-nitrostyrene) and the isotopic H/D exchange results for a series of Au/
TiO catalysts with different levels of activity.
2
Fig. 5. Influence of the reaction temperature on the kinetic parameters k and KNB
model 5), taking into account the linearised expression of Arrhenius and Van’t Hoff
equations, respectively.
(
Table 3
Influence of the reaction temperature on the kinetic parameters of the nitrobenzene
hydrogenation with Au/TiO
2
catalysts.
Linearised Arrhenius/Van’t Hoff equation
E
a
(KJ/mol)
Aa
Slope
Ordinate at origin
r2
k
K
ꢂ3700,6
ꢂ709.4
6.3
1.0
0.96
0.95
30.8
5.9
561.8
2.8
NB
a
ꢂ1 ꢂ1
ꢂ1
The same unities as k (mmol bar
h
) and KNB (l mmol ).
(
see Table 2). The true activation energy for the kinetic constant
and the enthalpy of the nitrobenzene adsorption were calculated
according to Arrhenius and Van’t Hoff equations (see Fig. 5),
respectively, and the results are given in Table 3. The true activa-
tion energy measured in this work (31 KJ/mol) is in agreement
with the values proposed by other authors on gold-catalysed
hydrogenations (36 KJ/mol for the reduction of crotonaldehyde
Fig. 4. Catalytic results during the hydrogenation of nitrobenzene with Au/TiO
catalysts at different conditions of H pressure, nitrobenzene concentration and
reaction temperature, taking into account the linearised expression of model 5.
2
with Au=TiO catalysts [36], 36 KJ/mol for reduction of butadiene
2 3
with Au/Al O catalysts [37]), and is close to the activation energies
determined for the H –D exchange on Au/Al O [38]. On the other
2 2 2 3
2
2
hand, the slightly endothermic character of the nitrobenzene
adsorption (ꢁ6 KJ/mol) has to be highlighted upon, as evidenced
by the increase of KNB as the reaction temperature increases.
Although this phenomenon, involving an increase of the adsorbed
4
.1.3. Calculation of kinetic parameters
A final design of 24 kinetic assays at different CNB, PH2 and T val-
ues was completed to calculate the parameters for model 5 (Tables
S3 and S4). Kinetic results have been fitted to model 5, conve-
niently linearised (see Fig. 4). It can be observed that the correla-
tion between the experimental and calculated results is good,
indicating that the proposed model is suitable to predict the
ꢂNO
2
species with the temperature, is not very usual in gas-phase
reactions, the endothermic behaviour is sometimes observed when
working with solid/liquid systems, since the adsorbant–adsorbate
interactions can be slightly influenced by additional solvent–
adsorbate and solvent–adsorbent interactions [39].
behaviour of the Au/TiO
compounds. Then, from the values of the slope (KNB=k ) and the
ordinate at the origin (1/k ), the kinetic rate constant for H dis-
2
2
catalysts for hydrogenating nitroaromatic
0:5
4.2. Designing a catalyst of higher activity
H2
0:5
H2
sociation (kH2) and the adsorption equilibrium constant of nitro-
benzene (KNB) are obtained at different reaction temperatures
From the performed kinetic study, we can extract relevant con-
clusions about the Au/TiO system mode of action, and how to in-
2
Table 2
Catalytic results at different conditions of H
2
pressure, nitrobenzene concentration and reaction temperature for the hydrogenation of nitrobenzene with Au/TiO
2
catalysts (r2
corresponds to the linear regression coefficient).
2
ꢂ1 ꢂ1
2
ꢂ1
T (K)
Linearised Model 5 (Eq. 7)
k
H2 ꢃ 10 (mmol bar
h
)
K
NB ꢃ 10 (l mmol
)
Slope
Ordinate at origin
r2
343
368
393
423
3.4
2.4
1.9
1.8
10.1
5.8
4.5
0.96
0.97
0.97
0.99
0.97
3.0
5.1
33.7
42.1
46.1
50.3
3.6
7.7

P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
23
Table 4
Catalytic results during the hydrogenation of 3-nitrostyrene using different Au and Pt catalysts. Feeding composition: 90.5% Toluene, 8.5%3-nitrostyrene, 1% o-xilene (% mol).
Catalyst
T
r
(K)
P
H2 (bar)
% Aua (mol)
% Pt a (mol)
Time (h)
% Conversion
% Selectivityb
TOFc
1
0
1
0
1
0
.5% Au/TiO
2
393
313
358
358
358
358
8
2
8
8
8
8
0.23
–
0.23
–
0.23
–
–
0.31
–
0.31
0.0155
0.0155
6
6.5
9
0.25
0.52
6
98.5
95.1
97.8
95.1
94.5
25.3
95.9
93.1
96.2
69.7
93.4
48.4
173
60
70
1380
550
2994
.2% Pt/TiO
.5% Au/TiO
2
2
.2% Pt/TiO
.5%Au@0.01%Pt/TiO
2
2
.01% Pt/TiO
2
a
mol metal/mol nitroaromatic ꢃ 100.
b
c
Selectivity to 3-vinylaniline.
ꢂ1 ꢂ1
TOF measured as mol converted mol metal
h .
crease the production during the chemoselective hydrogenation of
nitroaromatic compounds. For instance, taking into account the ki-
netic curves (see Figs. 1 and 2) one could work with low concentra-
alyst, which has been shown to be much more chemoselective.
Although the amount of Pt introduced should be high enough to
have an impact on the H dissociation, it should be relatively low
2
tions of nitrobenzene and high H
2
pressures. However, this would
to avoid the unselective effects of Pt for hydrogenating substituted
nitroaromatics. To do this, we have prepared a series of new cata-
require increasing the ratio of solvent/reactant. On the contrary,
from a practical point of view, it would be desirable to minimise
the use of solvents during the manufacture of substituted anilines
in order to (1) increase the productivity for a given reactor size, (2)
reduce the number of separation operations and (3) reduce the for-
mation of potential wastes. Also, low H pressures would be pre-
2
ferred to simplify the design of equipments and to reduce
operation costs. Thus, since the reaction is limited by the amount
lysts based on the 1.5% Au/TiO
amounts of Pt from 50 to 2000 ppm. Results in Fig. 6 show that
while activity of these Au@Pt/TiO catalysts for the hydrogenation
2
catalyst doped with increasing
2
of 3-nitrostyrene, measured as TOF, increases when increasing Pt
content, the chemoselectivity to reduce the nitro group decreases
for Pt contents over 100 ppm (see Table S5 for a detailed distribu-
tion of reaction products). Moreover, this result has been compared
in Table 4 with a catalytic system containing 100 ppm of Pt but no
gold. The low-hydrogenation rate observed in the latter case indi-
cates that even if such small amount of Pt is key to rapidly disso-
2
of H activated on the surface of the catalyst on low-coordinated
atoms at the surface of the gold nanoparticles [33], one could pre-
pare catalysts with a larger number of smaller gold crystallites and
with proper crystal shapes. While this is certainly a possibility, we
thought that, within our means, it would be difficult to selectively
increase the number of defects on the exposed gold domains.
Therefore, we have taken a second approach to increase the rate
ciate H
hydrogenation of the nitro group. Then, the final 1.5%Au@
0.01%Pt/TiO catalyst maintains excellent levels of selectivity at
2
, the presence of gold remains necessary to promote the
2
high conversion levels (see Table 4), while its high activity allows
completing the hydrogenation of 3-nitrostyrene in 30 min, instead
of the 9 h needed by the Au/TiO system (see Fig. 7 for a compari-
2
son between the kinetic behaviour of different catalysts at their
respective optimised reaction conditions). In addition, we have
of H
Recently, we have shown [40] that by using a more active metal
such as Pt, and when properly supported and activated on TiO , it is
possible to perform the chemoselective hydrogenation of 3-nitro-
2
dissociation on the catalyst, as will be described below.
2
styrene at milder reaction temperatures and H
Au/TiO
2
pressures that with
catalysts. Unfortunately, TOF values (mol converted mol
checked that, contrary to the Au/TiO
2
catalyst, the hybrid
2
1.5%Au@0.01%Pt/TiO system can efficiently work in solvent-free
2
ꢂ1 ꢂ1
metal
3
h
) provided by the reported Pt/TiO
2
solid at 313 K and
media (Fig. 8), which is important for a sustainable process.
Unfortunately, the use of only 0.01% Pt for enhancing the activ-
ity of the gold catalyst, while maintaining high chemoselectivity,
makes really hard to accurately characterise the distribution of Pt
species on the surface of the support, applying either spectroscopic
2
bar of H are low, and the selectivity of this catalyst is partially
lost when increasing the reaction temperature to increase TOF
see Table 4 and Table S5). To overcome this situation, we thought
to introduce Pt, which can rapidly dissociate H , on an Au/TiO cat-
(
2
2
2
Fig. 6. Influence of the Pt content in the 1.5% Au/TiO catalyst on the TOF values and selectivity towards 3-vinylaniline during the hydrogenation of 3-nitrostyrene.

2
4
P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
or microscopic techniques. However, we have observed that such
increase in the activity using the 1.5%Au@0.01%Pt/TiO catalyst
only occurs when the sample is reduced at high temperatures
the same effect is not observed, for instance, by reducing the cat-
2
(
alyst at 373 K or using an organic reducing agent). In addition, we
have also checked that no increase in the activity is produced by
simply reducing the Au/TiO sample (without Pt) at high tempera-
2
tures. Thus, we can hypothesise that the formation of bimetallic
Au@Pt particles during the thermal treatment, favoured at high-
reduction temperatures by migration of single Pt atoms along the
surface of the support, is essentially responsible for the special cat-
alytic behaviour of the sample. Accordingly, a plausible scenario
has been graphically depicted in Fig. 9 in order to easily under-
2
stand the 1.5%Au@0.01%Pt/TiO mode of action. Unfortunately,
we cannot provide any direct evidence of this phenomenon, and
further efforts for accurately characterising the catalyst surface
should be carried out.
Fig. 7. Evolution of 3-nitrostyrene conversion with reaction time using different
supported Au and Pt catalysts (optimised reaction conditions for each type of
material, see Table 4).
5
. Conclusion
Kinetic and isotopic studies for the hydrogenation of nitroaro-
matics have allowed to find that the rate-controlling step is the
dissociation of H on gold. The information obtained in this part
2
of the work has allowed to design a new bimetallic catalyst with
an optimised Pt content. The presence of Pt on the bimetallic cat-
alyst increases the rate of H
is almost one order of magnitude more active than the monometal-
lic Au/TiO , while maintaining high chemoselectivity during the
2
dissociation resulting in a solid that
2
hydrogenation in the presence of other sensitive functional groups.
Acknowledgments
We thank the Spanish government (Project MAT 2006-14274-
C02-01), and PROMETEO Project of Generalitat Valenciana.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.04.004.
Fig. 8. Evolution of the 3-nitrostyrene conversion (squares) and selectivity to 3-
vinylaniline (circles) with reaction time, using a 1.5% Au/TiO
2
catalyst (World Gold
References
Catalyst, white symbols) and the 1.5%Au@0.01%Pt/TiO system (black symbols)
2
under solvent-free reaction conditions. Reaction conditions: 358 K,
hydrogen, 100 mg of catalyst.
8 bar of
[
1] A. Corma, P. Serna, Science 313 (2006) 332.
[2] A. Grirrane, A. Corma, H. García, Science 322 (2008) 1661.
[
[
[
[
[
3] R.F. Kovar, F.E. Armond, US Patent 3,975,444 to the US Air Force, 1976.
4] M. Suchy, P. Winternitz, M. Zeller, WO Patent 91/02278 to Ciba-Geigy, 1991.
5] J. Butera, J. Bagli, WO Patent 91/09023 to American Home Products, 1991.
6] A. Burawoy, J.P. Critchley, Tetrahedron 5 (1959) 340.
7] U. Siegrist, P. Baumeister, H.-U. Blaser, M. Studer, Chem. Ind. Dekker 75 (1998)
2
07.
8] H.-U. Blaser, U. Siegrist, H. Steiner, M. Studer, in: R.A. Sheldon, H. van Bekkum
Eds.), Fine Chemicals Through Heterogeneous Catalysis: Aromatic Nitro
[
(
Compounds, Wiley, New York, 2001, p. 389.
[
9] A. Corma, P. Serna, H. García, J. Am. Chem. Soc. 129 (2007) 6358.
[
[
10] F. Porta, L. Prati, M. Rossi, G. Scari, J. Catal. 211 (2002) 464.
11] M.D. Hughes, Yin-Jun Xu, P. Jenkins, P. McMorn, P. Landon, D.I. Enache, A.F.
Carley, G.A. Attard, G.J. Hutchings, F. King, E.H. Stitt, P. Johnston, K. Griffin, C.J.
Kiely, Nature 437 (2005) 1132.
[
[
12] T.A. Nijhuis, T. Visser, B.M. Weckhuysen, Angew. Chem. Int. Ed. 44 (2005) 1115.
13] A.S.K. Hashmi, J.P. Weyrauch, M. Rudolph, E. Kurpejovic, Angew. Chem. Int. Ed.
4
3 (2004) 6545.
14] A.S.K. Hashmi, L. Schwarz, J.-H. Choi, T.M. Frost, Angew. Chem. Int. Ed. 39
2000) 2285.
[
(
[
[
15] S. Carrettin, J. Guzman, A. Corma, Angew. Chem. Int. Ed. 44 (2005) 2242.
16] C. González-Arellano, A. Corma, M. Iglesias, F. Sánchez, Chem. Commun. (2005)
3451.
[
[
17] S. Naito, M. Tanimoto, J. Chem. Soc. Chem. Commun. (1988) 832.
18] C. Milone, R. Ingoglia, L. Schipilliti, C. Crisafulli, G. Neri, S. Galvagno, J. Catal.
Fig. 9. Plausible scenario to understand the high activity and chemoselectivity
shown by the bimetallic 1.5%Au@0.01%Pt/TiO , as suggested by catalytic observa-
tions. Nitroaromatic compounds are selectively adsorbed at Au–Ti boundaries,
while activated hydrogen species are rapidly provided by the Pt sites.
236 (2005) 80.
2
[19] Y. Ito, M. Sawamura, T. Hayashi, J. Am. Chem. Soc. 108 (1986) 6405.
[20] C. González-Arellano, A. Corma, M. Iglesias, F. Sánchez, Chem. Commun. (2005)
1990.

P. Serna et al. / Journal of Catalysis 265 (2009) 19–25
25
[
21] M. Boronat, P. Concepción, A. Corma, S. González, F. Illas, P. Serna, J. Am. Chem.
Soc. 129 (2007) 16230.
[31] J.B. Butt, Reaction Kinetics and Reactor Design, second ed., Marcel Dekker,
1999.
[
[
[
22] M. Haruta, Catal. Today 36 (1997) 153.
23] E. Bus, J.T. Miller, J.A. van Bokhoven, J. Phys. Chem. B. 109 (2005) 14581.
24] B. Amon, H. Redlingshöfer, E. Klemm, E. Dieterich, G. Emig, Chem. Eng. Proc. 38
[32] A. Corma, P. Concepcion, P. Serna, Angew. Chem. Int. Ed. 46 (2007)
7266.
[33] A. Corma, M. Boronat, S. González, F. Illas, Chem. Commun. (2007) 3371.
[34] Kirk and Othmer, Encyclopedia of Chemical Technology, vol. 2., second ed.,
(1967), p. 411.
[35] T. Kahl, K.-W. Schröder, Ullmann’s Encyclopedia of Industrial Chemistry, sixth
ed., Wiley-VCH, Weinheim, Germany, 2002.
[36] R. Zanella, C. Louis, S. Giorgio, R. Touroude, J. Catal. 223 (2004) 328.
[37] D.A. Buchanan, G. Webb, J. Chem. Soc. Faraday Trans. 71 (1975)
134.
(
1999) 395.
[
[
25] V. Höller, D. Wegricht, I. Yuranov, L. Kiwi-Minsker, A. Renken, Chem. Eng.
Technol. 23 (2000) 3.
26] E. Lamy-Pitara, B. N’Zemba, J. Barbier, F. Barbot, L. Miginiac, J. Mol. Catal. A
Chem. 142 (1999) 39.
27] L. Petrov, K. Kumbilieva, N. Kirkov, Appl. Catal. 59 (1990) 31.
28] F. Turek, R. Geike, R. Lange, Chem. Eng. Process. 20 (1986) 213.
29] J. Relvas, R. Andrade, F.G. Freire, F. Lemos, P. Araújo, M.J. Pinho, C.P. Nunes, F.R.
Ribeiro, Catal. Today 33–135 (2008) 828.
30] F.F. Froment, K.B. Bischoff, Chemical Rector Analysis and Design, second ed.,
John Wiley, Berlin, 1990.
[
[
[
[38] E. Bus, J.T. Miller, J.A. van Bokhoven, J. Phys. Chem. B. 109 (2005) 14581.
[39] J.M. Corkill, J.F. Goodman, J.R. Tate, Trans. Faraday Soc. 62 (1966) 979.
[40] A. Corma, P. Serna, P. Concepción, J.J. Calvino, J. Am. Chem. Soc. 130 (2008)
8748.
[