Triphasic liquid systems: Generation and segregation of catalytically active Pd nanoparticles in an ammonium-based catalyst-philic phase

Article abstract of DOI:10.1039/b608414a

A triphasic liquid system fabricated from isooctane, aqueous base, and trioctylmethylammonium chloride/decanol promoted the formation of Pd-nanoparticles in the size range of 2-4 nm which remained immobilised in the onium phase, catalysed organic reactions, and could be recycled. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b608414a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
\
Triphasic liquid systems: generation and segregation of catalytically
active Pd nanoparticles in an ammonium-based catalyst-philic phase
Alvise Perosa,*a Pietro Tundo,a Maurizio Selvaa and Patrizia Cantonb
Received (in Cambridge, UK) 14th June 2006, Accepted 1st September 2006
First published as an Advance Article on the web 22nd September 2006
DOI: 10.1039/b608414a
A triphasic liquid system fabricated from isooctane, aqueous
base, and trioctylmethylammonium chloride/decanol promoted
the formation of Pd-nanoparticles in the size range of 2–4 nm
which remained immobilised in the onium phase, catalysed
organic reactions, and could be recycled.
Onium salts, that are phase transfer agents, can be used to produce
and stabilize metal nanoparticles (NPs) in situ,1,2 and to control
their size,3 starting from appropriate precursors. Reetz and co-
workers2 have amply demonstrated that nanostructured R4N+X2-
stabilized metal clusters can be prepared by a variety of methods.
Recently, peculiar liquid systems formed by three separate liquid
Fig. 1 Envisaged triphasic system.
phases have attracted our attention, in particular with a view
towards the possible effects on catalysis.4 The interest was
This communication describes how a ternary system of this kind
can be constructed, and used for catalytic reactions, through five
subsequent steps: the preparation of a suitable ternary liquid
system, formation of the active catalytic species, screening on a
model organic reaction, catalyst recovery and recycle tests, and
finally the characterisation of the active catalytic species by TEM.
At first, the conditions for the reproducible formation of a
model triphasic system were determined starting from previous
experience.5 The aim was to generate a stable and reproducible
ternary system. Aliquat 3361 (tricaprylmethylammonium chloride,
A336)—a well known industrially produced phase transfer agent
used in previous studies—was chosen as the third liquid catalyst-
philic phase. Commercial A336, however is a technical grade
compound: the alkyl chains of are a mixture of C8 and C10, and it
contains varying amounts of water (up to 5%), and of long chain
alcohols (1-decanol and 1-octanol, 5–10%), as well. It was soon
apparent that different A336 batches showed a large degree of
variability in their composition, and on their ability to form a third
phase in the presence of isooctane, and of aqueous base (KOH).
To eliminate this variable, trioctylmethylammonium chloride
(TOMAC) was synthesised from trioctylamine and dimethylsulfate
by adapting a reported procedure.8 The appearance of TOMAC
was that of a light amber coloured viscous liquid at room
temperature, effectively an ionic liquid. It was immiscible in
isooctane and in water, of intermediate density with respect to
these two solvents, with which it therefore formed a triphasic
system, placing itself between the other two.
stimulated by knowledge of the formation of triphasic liquid
systems made by mixtures of an organic solvent such as isooctane,
water, and an onium salt (ammonium or phosphonium), and their
use in heterogeneously catalysed organic reactions.5 Increases in
rates and selectivities could be achieved for a number of reactions
catalysed, for example, by supported metals such as Pd, Pt, and by
Raney-Ni. The onium salt was preferably liquid at the reaction
temperature (20–50 uC), had a strong affinity for the catalyst, and
formed a separate phase. The interaction between the liquid onium
salt, and the catalyst was shown to be the factor controlling the
improved catalytic effect.6 In addition, the strong affinity of the
onium salt for the catalyst made it a catalyst-philic phase, that
simplified the separation of products from catalysts and from by-
products once the reaction was complete.4
We reasoned that it would be interesting to take advantage of
triphasic liquid systems, to generate, immobilise, and stabilise
catalytically active metal NPs, starting from precursor metal
complexes. The outcome was envisaged as shown in Fig. 1.
This system offers a number of potential advantages in view of
process intensification: (i) the active catalytic NPs can be prepared
directly in the reactor, (ii) it is easy to separate the organic products
from the catalyst by decanting, (iii) the catalyst morphology can be
tuned, (iv) the metal NPs are stabilized by the third phase, and (v)
inorganic reagents (i.e. base) and by-products remain in the
aqueous phase, not to mention that the amount of catalyst-philic
phase can be as small as convenient. All these issues are the basis
for cleaner, cheaper catalytic systems,7 and characterize some of
the principles of green chemistry.
With the aim to generate and immobilize Pd-black in the third
TOMAC phase, a suitable palladium precursor was sought, by
testing PdCl2, Pd(OAc)2, Pd(PPh3)4, and PdCl2(BzCN)2.
Preliminary screening tests carried out with commercial A336
had indicated that Pd(PPh3)4 was the best precursor: it formed Pd-
black which resided in the A336 phase. The other complexes also
formed Pd metal, in these cases however the particles were
dispersed in more than one phase.
aDipartimento di Scienze Ambientali dell’Universita` Ca’ Foscari, and
Consorzio Interuniversitario ‘‘la Chimica per l’Ambiente’’ INCA
Dorsoduro 2137, 30123, Venezia, Italy. E-mail: alvise@unive.it;
Fax: +39 041 2348584; Tel: +39 041 2348958
bDipartimento di Chimica Fisica dell’Universita` Ca’ Foscari, Via Torino
155/B, 30172, Venezia-Mestre, Italy
4480 | Chem. Commun., 2006, 4480–4482
This journal is ß The Royal Society of Chemistry 2006

View Article Online
The same tests were therefore carried out with the triphasic
system made using TOMAC prepared in-house.{ Unlike the
screening tests done with commercial A336, these yielded Pd-black
very slowly. Only by adding 0.1 ml of 1-decanol to the standard
triphasic mixture—thereby reproducing a standard commercial
A336 composition—was a rapid formation of Pd-black in the
intermediate TOMAC/decanol phase observed. The effect of
decanol is presently beyond the scope of this work, but it may
indicate the formation of a water-in-oil microemulsion promoted
by the addition of the alcohol that may act as co-surfactant,9 and
intervenes in the precipitation of metallic palladium.10
While a relatively large amount (1.0 mL) of TOMAC was used
throughout this study in order to visually monitor the phase
separation, the amount of catalyst-philic phase can be smaller.
The system was tested for catalysis by using as the model
reaction the hydrodehalogenation (HDX) of 4-chloro-propiophe-
none (0.50 mmol) with hydrogen at atmospheric pressure to yield
propiophenone (eqn (1)).5b
ð1Þ
Fig. 3 Appearance of the triphasic system: (a) after run 1, (b) after run 2.
At room temperature, the hydrodehalogenation reaction was
sluggish for the first 80 min, this induction period lasted until Pd-
black started to precipitate. After that, quantitative and selective
removal of chlorine was observed in two hours (Fig. 2). No
competitive carbonyl reduction was observed, consistent with the
results obtained in the presence of Pd supported on charcoal as
catalyst, reported in earlier papers.5
When stirring was interrupted, the triphasic system settled and
the Pd-black catalyst remained segregated in the middle TOMAC/
decanol phase (Fig. 3a).
The reaction product (PhCOEt) was removed along with the
isooctane top layer, and the remaining phases (TOMAC/decanol
and aqueous base) were recycled by adding additional reagent and
organic solvent. The recycle procedure was repeated four times
(Fig. 4). No induction period was observed in run 2, and the
reaction was complete after 80 min, still maintaining complete
(.99%) selectivity for HDX.
Fig. 4 Recycling tests (conversion and selectivity after 2 h).
Run 2 was faster than run 1, likely due to a higher concentration
of Pd-black that had finished forming during run 1. This was
substantiated by the fact that after run 2, the triphasic system
showed a clearer phase separation than after run 1, and both the
upper organic and lower aqueous layers appeared perfectly clear,
with the catalyst segregated in between (Fig. 3b).
The TOMAC–Pd catalytic ensemble could be used repeatedly
without significant loss of activity, and without the induction
period of the first run. After 2 h, conversion of the reagent reached
85, 85, and 74%, in runs 3, 4, and 5 respectively (Fig. 4). However,
the Pd particles tended to aggregate, to become visibly larger and
less dispersed than in the beginning, and to slowly precipitate at the
TOMAC/water interface. This phenomenon was accompanied by
a slight decrease in selectivity in runs 3–5 where up to 8% of
phenylpropanol—deriving from concurrent carbonyl reduction—
was observed. The diminished selectivity was accounted for by the
structure sensitivity of the substrate towards the metal particles. In
fact, palladium particles with sizes in the 2–4 nm range are
composed predominantly of corner atoms, while larger particles
have a higher proportion of their surface composed of face
atoms.11 For the former, used in the first runs, the hydrodechlor-
ination occurred selectively, since the substrate approached only
single-atom active sites. For the larger particles, approach to the
Fig. 2 Plot of the hydrodehalogenation reaction of eqn (1).
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 4480–4482 | 4481

View Article Online
polarity and molar fraction of all the components) imply that these
systems must be constructed and fine-tuned for each particular
case.
Issues still open include whether the nanoparticles, or rather a
soluble form of palladium in equilibrium with them, are the active
catalytic species.2 We feel this may not be a problem since we have
demonstrated that Pd is not present in catalytically significant
amounts in the organic phase. Another issue is the tendency of the
Pd particles to aggregate. Although the catalytic efficiency is only
slightly reduced with each recycling sequence, it is probable that
this aggregation is responsible for the progressive deactivation.
This aggregation is likely due to the observed progressive loss of
decanol—i.e. of its nanoparticle stabilising effect—that slowly
leaches to the organic phase.5g,15
Fig. 5 HREM of the specimen after the first run.
face atoms on the flat surface promoted concurrent reduction of
both functional groups.5b,12
In view of efficiently segregating the Pd in the catalyst-philic
phase, it was important to determine whether any metal leaching
to the organic phase was occurring. Leaching was monitored after
run 2 and run 4, by filtering the organic phase, adding
4-chloroacetophenone as reagent, and stirring the organic phase
under hydrogen.13 No traces of acetophenone, the hydrodehalo-
genation product, were observed in either case. This indicated that
none of the hydrodehalogenation-catalysing Pd nanoparticles
leached to isooctane. Any leaching of Pd in the inorganic phase
was not monitored since it did not affect the outcome of the
reactions, the aqueous phase was in fact always retained in the
recycling experiments.
Notes and references
{ Experimental note. The standard triphasic catalytic system was made by
loading 1.0 ml of TOMAC, 9.0 ml isooctane, 0.10 ml 1-decanol and 2.0 ml
KOH(aq) (20%) (for the Heck reaction the latter was substituted by 5.0 ml
of 6% NaHCO3(aq)) in a three necked flask. To this were added
0.025 mmol of Pd(PPh3)4, and the reaction stirred under air for 1 h until the
middle phase turned completely black. The substrate was added (1.0 mmol
4-chloropropiophenone for HDX, or 1.0 mmol PhI and 1.5 mmol ethyl
acrylate for the Heck reaction) and the reaction products monitored by
GC-MS.
TEM measurements were performed using a JEM 3010 (JEOL) electron
microscope operating at 300 kV; the lens parameters were Cs = 0.6 mm, Cc
= 1.3 mm, giving a point resolution of 0.17 nm at Scherzer defocus. A few
milligrams of the specimen were sonicated for 5 min in order to disrupt
possible agglomerates. A 5 mL droplet of suspension was transferred onto
an amorphous carbon film, coating a 200 mesh copper grid, dried at room
temperature, and then put into the microscope.
The scope of the same triphasic catalytic system was then
extended by testing its activity for the Pd-catalysed Heck coupling
reaction of phenyl iodide with ethyl acrylate in the presence of
NaHCO3 as the base (eqn (2)).14 At 80 uC after 1 h, 95%
conversion to the coupling product ethyl cinnamate was observed.
In this case the system showed a clear separation of the phases at
the end, and no leaching of Pd into the organic phase was detected.
1 M. T. Reetz, W. Helbig, S. A. Quaiser, U. Stimming, N. Breuer and
R. Vogel, Science, 1995, 267, 367–369.
2 M. T. Reetz and J. G. de Vries, Chem. Commun., 2004, 1559–1563 and
references therein.
3 P.-F. Ho and K.-M. Chi, Nanotechnology, 2004, 15, 1059–1064.
4 P. Tundo and A. Perosa, Chem. Soc. Rev., 2006, DOI: 10.1039/
bB503021h.
ð2Þ
5 (a) M. Selva, P. Tundo and A. Perosa, J. Org. Chem., 1998, 63,
3266–3271; (b) A. Perosa, M. Selva and P. Tundo, J. Org. Chem., 1999,
64, 3934–3939; (c) P. Tundo, S. Zinovyev and A. Perosa, J. Catal., 2000,
196, 330–338; (d) S. S. Zinovyev, A. Perosa, S. Yufit and P. Tundo,
J. Catal., 2002, 211, 347–354; (e) A. Perosa, P. Tundo and S. S. Zinovyev,
Green Chem., 2002, 4, 492–494; (f) S. S. Zinovyev, A. Perosa and
P. Tundo, J. Catal., 2004, 226, 9–15; (g) A. Perosa, P. Tundo, M. Selva,
S. S. Zinovyev and A. Testa, Org. Biomol. Chem., 2004, 2, 2249–2252.
6 (a) P. Tundo and A. Perosa, React. Funct. Polym., 2003, 54, 95–101; (b)
P. Tundo, A. Perosa and S. S. Zinovyev, J. Mol. Catal. A: Chem., 2003,
204–205, 747–754.
The size of the Pd particles was then determined by TEM
measurements. Fig. 5 shows two HREM images of the catalyst-
philic phase from the HDX reaction (after run 1), taken at
different magnification (400 k and 1 M). Rounded crystalline Pd
particles, with no defects, are present with size ranging from 2 to 4
nm. The particles are well dispersed, isolated, and with no
formation of agglomerates. Some of the particles show lattice
fringes corresponding to Pd–Pd distances; in some case the lattice
fringes are not visible due to the covering by TOMAC.
7 V. Gewin, Nature, 2006, 440, 378–379.
8 K. Sato, M. Aoki, M. Ogawa, T. Hashimoto and R. Noyori, J. Org.
Chem., 1995, 61, 8310–8311.
This communication shows that it is possible to use multiphasic
solvent systems to generate highly dispersed Pd nanoparticles that
are active, and selective, for catalysis. Since separation of catalytic
particles from a reaction mixture is a practical problem, the system
was designed to segregate these particles in a catalyst-philic
phase—thus allowing reuse of the catalyst. It represents an
approach towards the development of catalytic systems that are in
a different phase from the reaction species, thereby simplifying
purification of the product, catalyst re-use, and reducing the
amount of energy used and of waste produced. Caution should of
course be used in this respect since the number of variables
affecting the formation of the third catalyst-philic phase (e.g. the
¨
9 Z. Do¨vyap, E. Bayraktar and U. Mehmetog˘lu, Enzyme Microb.
Technol., 2006, 38, 557–562.
10 A. Martino, M. Stoker, M. Hicks, C. H. Bartholomew, A. G. Sault and
J. S. Kawola, Appl. Catal., A, 1997, 161, 235–248.
11 R. K. Herz, W. D. Gillespie, E. E. Petersen and G. A. Samorjai,
J. Catal., 1981, 67, 371–386.
12 R. L. Augustine, Heterogeneous catalysis for the synthetic chemist,
Marcel Dekker Inc., New York, ch. 3, 1996.
13 R. A. Sheldon, M. Wallau, I. W. C. E. Arends and U. Schuchardt, Acc.
Chem. Res., 1998, 31, 485–493.
14 V. Calo`, A. Nacci, A. Monopoli, S. Laera and N. Cioffi, J. Org. Chem.,
2003, 68, 2929–2933.
15 Small amounts of 1-decanol were detected in the organic phases by
GC-MS.
4482 | Chem. Commun., 2006, 4480–4482
This journal is ß The Royal Society of Chemistry 2006