Mesoporous soot-supported palladium as a heterogeneous catalyst for the Heck reaction in ionic liquids

Article abstract of DOI:10.1016/j.mencom.2008.11.017

The Heck reaction between bromobenzene and styrene in the ionic liquid catalyzed by ultrafine particles of palladium metal supported onto hydrophylized mesoporous soot follows a true heterogeneous mechanism.

Full text of DOI:10.1016/j.mencom.2008.11.017

Mendeleev
Communications
Mendeleev Commun., 2008, 18, 334–335
Mesoporous soot-supported palladium as a heterogeneous
catalyst for the Heck reaction in ionic liquids
a
b
b
a
Leonid A. Aslanov, Yury A. Kabachii, Sergey Yu. Kochev, Boris V. Romanovsky,*
b
c
a
a
Petr M. Valetsky, Vladimir V. Volkov, Alexander V. Yatsenko and Valery N. Zakharov
a
b
c
Department of Chemistry, M. V. Lomonosov Moscow State University, 119991 Moscow, Russian Federation.
Fax: +7 495 939 3570; e-mail: bvromanovsky@mail.ru
A. N. Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, 119991 Moscow,
Russian Federation
A. V. Shubnikov Institute of Crystallography, Russian Academy of Sciences, 119991 Moscow, Russian Federation
DOI: 10.1016/j.mencom.2008.11.017
The Heck reaction between bromobenzene and styrene in the ionic liquid catalyzed by ultrafine particles of palladium metal
supported onto hydrophylized mesoporous soot follows a true heterogeneous mechanism.
The C–C coupling of an aryl halide with an olefin (the Heck
reaction) mediated by supported metal Pd is commonly believed
distributions of both Pd particles and support mesopores are
similar, which is indicative of predominant location of Pd
nanocrystals inside of the support pores.
1–7
to proceed as a quasi-homogeneous process. This mechanism
suggests that the thermodynamically non-stable Pd nanoclusters
immobilized onto a porous solid play only a role of precursors
of active species, which result from the interaction between
reaction media and metal nanophase. Once formed, these active
Pd-containing species (small clusters or metal complexes) are
transferred to the liquid phase where they perform homogeneous
catalytic cycles. This suggestion implicates that at the end of
reaction the palladuim species return to the support pores and
The comparison of turnover frequencies (TOF) obtained
using NaOAc and Bu NH bases showed that amine added is
2
about twice as efficient as additive than acetate. This result
7
is in contrast to the data of Köhler et al. where the Heck reac-
tion performed in N-methylpyrrolidone showed quite opposite
efficiency for these bases. Such a different behaviour of ionic
and molecular solvent in the biphase system composed of a
liquid with dissolved reagents and a solid containing the active
moiety can be attributed to the different properties of these two
media. First, effective diffusion coefficients in ILs are lower by
1.5–2 orders of magnitude than that in both polar and apolar
solvents.10 In addition, the ILs exhibit medium range ordering
that has dramatic consequence for the diffusion mechanism in
these solvents.11 Therefore, ILs act as low-polarity media with
non-polar compounds, while the transport of polar and ionic
species is strongly hampered. Second, of great importance is
the non-polarity of tetraalkylammonium ILs, even more so for
those ILs with bistriflimide anions bearing two electron
acceptor Me groups.10 These ILs cannot, in principle, act as
ligands to stabilize small metal nanoclusters or complexes by
this way preventing them from aggregation. As to molecular
solvents, their polarity is the key factor for stabilization of
colloid Pd nanoparticles, clusters etc.9 Note that the Pd/C
catalyst displays no activity in the Heck reaction when using an
5
re-precipitate onto Pd particles larger than ‘mother’ ones. This
re-precipitation accounts for the Ostwald aging of metal nano-
phase and, as a consequence, results in an irreversible loss in
the catalytic efficiency, which makes it impossible to reuse a
catalytic material. There are a few papers on the possibility of
the direct involvement of Pd atoms located at the vertexes and
8
edges of nanocrystals into catalytic cycles.
The problem of whether the Heck reaction in the presence of
supported Pd is heterogeneous or quasi-homogeneous exhibits a
clear practical bearing connected with the not easy separation
of leached Pd from the reaction mixture and almost inevitable
losses of the precious metal.
The multistage catalytic cycle of the Heck reaction is highly
susceptible not only to the properties of both reagents and a
catalyst but also to the reaction medium. Hence, very thorough
adjustment of reaction conditions, and essentially of a solvent,
9
5
is required for every particular Heck arylation. In this connec-
apolar molecular solvent. A poor solvation ability, if any, of
tion, ionic liquids (ILs) have evident advantages over commonly
used polar solvents not only because of being ‘green solvents’
but also as exhibiting rather unique combination of physico-
chemical characteristics (melting point, viscosity, polarity etc.).
These important characteristics of ILs as solvents can be readily
the IL used in our study strongly prevents the leaching of Pd
from the supported catalytic material upon reaction. This makes
it next to impossible for the Heck reaction to proceed via a quasi-
homogeneous mechanism and vice versa favours the classical
Langmuir–Hinshelwood to be operative.
‘
task-specified’ by combining the cation and anion constituents
†
The 5 wt% Pd/C was prepared by the reduction of PdII hydroxide
1
0
of an ionic liquid.
Here, we report on the catalytic properties of Pd nanoclusters
immobilized onto mesoporous acetylene soot using a tetraalkyl-
freshly precipitated upon the hydrophilized acetylene soot A100
2
–1
(
200 m g ). The catalytic reaction between bromobenzene (PhBr) and
styrene (PhC H ) giving stilbene (Ph C H ) was carried out at 140 °C
under Ar using NaOAc, Bu NH or Bu N as bases and tributylhexyl-
†
2
3
2
2
2
ammonium-based IL as a solvent.
2
3
After thermooxidative and acid hydrophilisation of starting
soot, its surface exhibits a mosaic structure consisted of spatially
separated graphite-like domains (nanographenes) over an amorphous
carbon substrate. Metallic palladium presents in Pd/C as nano-
crystals with an average diameter of 4–5 nm. Note that the size
ammonuim bistriflimide IL (mp 19 °C) as a solvent. The starting carbon
support, as well as the Pd/C catalyst, was characterized by AAS, N
BET, TEM, HEED, SAXS and H NMR. The qualitative and quanti-
2
1
tative analysis of reactants and products was performed by GLC and
MALDI-TOF MS.
©
2008 Mendeleev Communications. All rights reserved.
– 334 –

Mendeleev Commun., 2008, 18, 334–335
Meanwhile, the 24 h time gap between the first and second
1
00
100
80
60
40
20
1
2
(
a)
(b)
catalysis cycles, as well as the lowering of temperature down
to 25 °C, reduces noticeably the preactivation effect. This
result is consistent with the catalyst autoadjustment mechanism
mentioned above. Indeed, on changing the conditions of inter-
action between catalysts and reaction media, the steady-state
concentration of active sites also shifts to its initial value.
The above results point out that performing the Heck reaction
in the tetraalkylammonium IL affords the re-use of catalytic
system with no significant loss in its efficiency.
8
6
4
2
0
0
0
0
1
2
3
3
0
2
4
6
8
10 12
0
3
6
9
12 15
Reaction time, t/h
Reaction time, t/h
The Heck reaction mediated by Pd supported on mesoporous
soot as a classical heterogeneous process by no means excludes
the sintering of catalyst active component via the surface diffu-
sion of metal atoms from small clusters to larger particles
Figure 1 (1) Conversion of bromobenzene, yields of (2) stilbene and (3)
biphenyl vs. reaction time; reaction conditions: (a) 0.1 mol% Pd/C catalyst
not preactivated, see the text); (b) 0.1 mol% Pd/C catalyst (preactivated,
(
see the text).
enhanced by the presence of Bu NH. Here, the features of soot-
2
based support are very likely the key factor hampering this
diffusional transport of Pd within its pores. In contrast to an
activated carbon support, the mosaic surface of which exhibits
high interdomain barriers, the mesoporous acetylene soot displays
almost ‘metallic’ conductivity. Such a peculiarity of soot-based
support equalizes the electrochemical potential of spatially
separated Pd nanoclusters and in this way it endows their
assembly with the property of bulk metal. The lack of chemical
potential gradient explains why there is no palladuim transfer
between the particles of different sizes.
The assumption that the Heck reaction catalyzed by supported
Pd using a finely ‘tuned’ IL as a reaction medium may proceed
as a true heterogeneous one is supported, though indirectly, by
the results of a kinetic study. The typical kinetic curves are
given in Figure 1(a). On the one hand, the reaction starts with no
visible induction period that would be characteristic of quasi-
12
homogeneous process, provided the accumulation of the active
Pd-containing species in a liquid phase. On the other hand, the
kinetic curves obtained in tetraalkylammonium IL exhibit
the well-pronounced reflection points, which are typical of an
autoacceleration accompanying the main reaction regardless the
progressive reagent consumption. If the fresh sample Pd/C is
first heated at 140 °C for 10 h in the presence of dibutylamine
using the tetraalkylammonium IL as a solvent and then the
starting PhBr, PhC H and Bu NH (1:1:1.2 molar ratio) are intro-
Two important results obtained in this investigation provide
the plain evidence of a true heterogeneous mechanism of the
Heck reaction in the presence of soot-supported palladium.
First, after preactivation of Pd/C in IL at 140 °C for 5 h and
separation of solid catalytic material from the liquid phase
this latter does not catalyze the Heck arylation and there is no
palladium traces in this liquid. Second, the Heck reaction com-
pleted, both the morphology and average diameter of palladium
particles were found to be quite similar to those in the as-
prepared Pd/C sample. We believe that this finding could be
considered as providing a strong evidence for the lack of
solvation–reprecipitation and hence the metal loss effects that
2
3
2
duced in the reaction system, the value of TOF increases by a
factor of 2, the reflection point on the kinetic curve disappears
and the reaction rate decreases gradually on the consumption of
PhBr [Figure 1(b)]. A blank experiment on heating Pd/C in IL
in the absence of Bu NH showed no catalyst activation.
2
Therefore, the slow stage of metal Pd activation initiated by
an amine base that occurs in parallel with the main C–C cross-
coupling is likely to explain an unusual profile of kinetic
curve [Figure 1(a)] resulted from the experiment omitting the
step of preliminary catalyst activation. Note that such an effect
of supported Pd activation in the presence of amine base was
not reported earlier. In this connection, the question arises whether
the same activation mechanism as was shown for the quasi-
homogeneous mechanism is taking place with amine base.
5
were well documented on using molecular NMP solvent.
This work was supported by the Federal Target Program ‘R&D
Priority Directions in the Science and Technology for 2007–2012’
(
code 2007-3-1.3-28-01-534) and the Russian Foundation for
Basic Research (grant no. 08-03-0044).
References
Significantly, neither NaOAc nor Bu N activate the supported
3
Pd/C catalyst but Bu NH only.
Supposing the low-coordinated palladium atoms to operate
as active centers staying bonded to the nanocrystal surface, the
1
2
S. S. Pröckl, W. Kleist and K. Köhler, Tetrahedron, 2005, 61, 9855.
A. Bifis, M. Zecca and M. Basato, J. Mol. Catal. A: Chem., 2001, 173,
2
2
49.
C. R. LeBlond, A. T. Andrews, Y. Sun and J. R. Sowa, Org. Lett., 2001,
, 1555.
3
mechanism of palladuim activation by Bu NH can be depicted
2
3
in the following way. First, these reactive Pd(0) atoms undergo
oxidative addition to dibutylamine N–H bond and then by
surface migration shift to more stable positions on the face of
nanocrystal. It is easy to ascertain using a 3D model of the FCC
metal Pd framework that such a migration results in an increase
in the nanocrystal deficiency, i.e., the generation of new active
centers until the steady state is established. The steady-state
concentration of these centers may be substantially different
from the initial one according to the auto adjustment theory in
4
5
M. T. Reetz and E. Westerman, Angew. Chem., 2000, 112, 165.
R. G. Heidenreich, J. G. E. Krauter, J. Pietsch and K. Köhler, J. Mol.
Catal. A: Chem., 2002, 182–183, 499.
K. Köhler, R. G. Heidenreich, J. G. E. Krauter and J. Pietsch, Chem.
Eur. J., 2002, 8, 622.
6
7
8
9
F. Zhao, K. Kuracami, M. Shirami and M. Arai, J. Catal., 2000, 194,
4
79.
J. Le Bars, U. Sprecht, J. S. Bradley and D. G. Blackmond, Langmuir,
999, 15, 7621.
I. P. Beletskaya and A. V. Cheprakov, Chem. Rev., 2000, 100, 3009.
1
1
3
heterogeneous catalysis.
On addressing the mechanism of Pd nanocrystal activation,
it should be taken into account that the efficiency of a catalytic
10 L. A. Aslanov, M. A. Zakharov and N. L. Abramycheva, Ionnye zhidkosti
v ryadu rastvoritelei (Ionic Liquids as Solvents), Izdatel’stvo Moskovskogo
Universiteta, Moscow, 2005, p. 271 (in Russian).
11 C. Chiappe, Monatsch. Chem., 2007, 138, 1035.
12 V. Parvulescu and J. Hardacre, Chem. Rev., 2007, 107, 2615.
13 G. K. Boreskov, Geterogennyi kataliz (Heterogeneous Catalysis), Nauka,
Moscow, 1988 (in Russian).
system preactivated in the presence of Bu NH is retained after
2
the first catalytic cycle provided that the second reaction cycle
starts immediately after the first one. On adding a new portion
of reagents to a mixture of IL + Pd/C, in which the complete
conversion of PhBr is attained, the reaction does restart with
the same rate.
Received: 1st August 2008; Com. 08/3198
335 –
–