Scope and limitation of activated carbons in aqueous organometallic catalysis

Article abstract of DOI:10.1016/j.jcat.2010.12.005

The effect of activated carbons has been studied in the palladium-catalyzed cleavage reaction of allylalkylcarbonate under aqueous biphasic conditions. A number of parameters were investigated, such as the type of carbon, the structure of the water-soluble ligand, the water conditions, and the metal complex loading. It was found that the intrinsic properties of carbons had a strong influence on the reaction rates. The best performances were obtained with the AC-WV carbon possessing the largest part of mesopores and lowest content of oxygen-surface groups. The results were rationalized by considering that AC-WV acted as a mass-transfer promoter increasing the interfacial area and collisions between the reactive species in the pore volume. The hypothesis of a confinement effect of the catalyst and reactants within the pores via adsorption-desorption processes was supported by isothermal studies and ³¹P{¹H} NMR investigations.

Full text of DOI:10.1016/j.jcat.2010.12.005

Journal of Catalysis 278 (2011) 208–218
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Scope and limitation of activated carbons in aqueous organometallic catalysis
Nicolas Kania a,b,d, Narasimhan Gokulakrishnan a,b,d, Bastien Léger a,b,d, Sophie Fourmentin a,c
,
Eric Monflier a,b,d, Anne Ponchel a,b,d,
⇑
a Univ Lille Nord de France, F-59000 Lille, France
b UArtois, UCCS, F-62300 Lens, France
c ULCO, UCEIV, EA 4492, F-59140 Dunkerque, France
d CNRS, UMR 8181, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of activated carbons has been studied in the palladium-catalyzed cleavage reaction of ally-
lalkylcarbonate under aqueous biphasic conditions. A number of parameters were investigated, such as
the type of carbon, the structure of the water-soluble ligand, the water conditions, and the metal complex
loading. It was found that the intrinsic properties of carbons had a strong influence on the reaction rates.
The best performances were obtained with the AC-WV carbon possessing the largest part of mesopores
and lowest content of oxygen-surface groups. The results were rationalized by considering that AC-WV
acted as a mass-transfer promoter increasing the interfacial area and collisions between the reactive spe-
cies in the pore volume. The hypothesis of a confinement effect of the catalyst and reactants within the
pores via adsorption–desorption processes was supported by isothermal studies and 31P{1H} NMR
investigations.
Received 8 October 2010
Revised 6 December 2010
Accepted 7 December 2010
Available online 14 January 2011
Keywords:
Mass transfer
Activated carbons
Palladium
Phosphanes
Ó 2010 Elsevier Inc. All rights reserved.
TPPTS
Tsuji–Trost reaction
Biphasic catalysis
Sustainable chemistry
1. Introduction
process (now OXEA company), in which the rhodium associated
with meta-trisulfonated triphenylphosphane (TPPTS; P(m-C6H4SO3-
Homogeneous catalysis plays a key role in carrying out chemi-
cal reactions in the most efficient way, thanks to the well-defined
structure–property relationship between the molecular catalyst
and the substrate. However, although the homogeneous catalysts
are generally highly active and selective, these systems may suffer
from difficulties associated with the separation of the catalyst from
the reaction medium and their subsequent recycling.
These problems can be elegantly solved by using biphasic aque-
ous organometallic catalysis, in which the molecular metal complex
is immobilized in the aqueous phase by water-soluble ligands. The
catalyst can be easily recovered in active form at the end of reaction
by decantation of the aqueous and organic phases, and the produc-
tion costs are significantly lower [1,2]. One class of ligands, which
attracts significant attention in aqueous organometallic catalysis,
are phosphanes bearing sulfonated polar groups (–SOꢀ3 Na+) to en-
sure a sufficient solubility in water. Notably, several industrial-scale
plants based on aqueous biphasic technologies, using metal com-
plexes stabilized by water-soluble phosphanes, are now operating.
The most striking example is the Ruhrchemie/Rhône-Poulenc
Na)3] catalyzes the hydroformylation of propene or butene [3]. An-
other industrial example concerns the hydrodimerization of
butadiene and water by the Kuraray company. The process includes
two catalytic steps in biphasic aqueous medium using palladium or
rhodium, stabilized by meta-monosulfonated triphenylphosphane
[TPPMS, PPh2(m-C6H4SO3Na)] [3]. Although appropriate for water-
soluble substrates, biphasic aqueous organometallic processes can-
not be easily generalized, especially when considering the case of
highly hydrophobic substrates for which mass-transfer limitations
occur.
To circumvent the problems of solubility in water, a variety of
strategies have been investigated, such as the utilization of co-sol-
vents [4], amphiphilic phosphanes [5,6], polymers [7], surfactants
[8–11], or cyclodextrins [12–16]. It is worth mentioning that some
drawbacks can be associated with these methods. Indeed, surfac-
tants and amphiphiles may give rise to foaming and persistent
emulsions during the workup, whereas co-solvents or polymers
can be difficult to remove from the products at the end of the
reaction.
The possibility of performing organometallic-catalyzed reac-
tions in aqueous media by using solid suspensions has also been
viewed as an alternative method. The most striking example is that
of supported aqueous-phase catalysis (SAPC) originally developed
⇑
Corresponding author at: UArtois, UCCS, F-62300 Lens, France. Fax: +33 3 2179
1755.
E-mail address: anne.ponchel@univ-artois.fr (A. Ponchel).
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.12.005

N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
209
by Davis et al. [17–20], whose principle is based on the use of a thin
film of water to dissolve and disperse the water-soluble catalytic
complex onto the surface of a hydrophilic porous support (mostly
silica-based materials). This strategy has proven fruitful for the
reported in the literature [35]. The synthesis of trisulfonated
biphenylphosphanes, i.e. P(BiPh)3TS, P(BiPh)2PhTS, P(BiPh)Ph2TS
where Ph and BiPh correspond to phenyl and biphenyl groups,
respectively, was performed by a procedure recently developed
in our laboratory [36]. The ligand purities were carefully controlled
by 1H, 13C, and 31P{1H} NMR analyses. In particular, 31P{1H} NMR
indicated that the products were mixture of phosphanes (ca.
98%) and oxides (ca. 2%). For NMR spectroscopy measurements,
D2O (99.92% isotopic purity) was purchased from Euriso-Top. All
other reactants were purchased from Aldrich Chemicals and Acros
Organics in their highest purity and used without further purifica-
tion. Distilled deionized water was used in all experiments.
hydroformylation of olefins [21], selective hydrogenation of
a,b-
unsaturated aldehydes [22], asymmetric hydrogenation [23], and
allylic alkylation [24]. The immobilized complex is assumed to
work at the aqueous–organic interface in the water film, thus
allowing a larger interfacial area. However, the main disadvantage
of SAPC lies in the fact that the degree of hydration of the support
must be carefully controlled in order to retain the catalyst on the
solid surface without any metal leaching [21]. Recently, a slightly
different approach was reported by Holmberg et al. [25] in the
Heck-type reaction between an arylboronic acid and styrene by
using an ordered mesoporous silica with a rhodium complex stabi-
lized by meta-disulfonated triphenylphosphane [TPPDS; PPh(m-
C6H4SO3Na)2]. The authors postulated that the reaction occurred
at the interface between the organic and water domains, namely
at the pore openings of the mesoporous silica and not at the sur-
face of the particles.
The utilization of carbon materials in biphasic aqueous organo-
metallic has not been the subject of extensive studies. This obser-
vation is surprising since aqueous suspensions of carbon particles
are known to enhance gas–liquid [26–29] and liquid–liquid
[30,31] mass transfers. Interestingly, carbon materials satisfy most
of the desirable requirements to be used in catalytic processes,
including a high porous volume, high surface area, thermal stabil-
ity, and chemical inertness [32]. Furthermore, the fact that acti-
vated carbons can be produced from agricultural and forestry
residues, or generally biomass residue wastes makes the applica-
tions even more appealing from the environmental point of view.
In this context, the combined use of an activated carbon and a rho-
dium complex coordinated by water-soluble phosphane was ini-
tially reported by Luft et al. [33]. They explained the results
obtained during the hydroformylation of 1-hexene by considering
a supported aqueous-phase catalytic mechanism.
2.2. Carbons and their treatments
The starting activated carbons were selected from commercially
available samples. The main part of the work was carried out using
the NucharÒWV-B-activated carbon (denoted AC-WV), which was
a gift from MeadWestvaco Corporation (Covington, USA). It was
produced from wood and activated by phosphorous acid. In some
cases, NoritÒSA-2 (denoted AC-NorA) and SorbonoritÒB3 (denoted
AC-NorB), produced by Norit (Amersfoort, Netherlands) from peat
and steam activated, were used for comparison.
To increase the number of acidic functions, the AC-WV carbon
was treated with different concentrations of nitric acid, from
10ꢀ4 to 16 M. In a typical experiment, 1.0 g of carbon was stirred
for 1 h with 100 mL of HNO3 of a known concentration at room
temperature or under reflux. After the treatment, the carbon was
recovered by filtration, thoroughly washed with deionized water
until the pH value of the washing water was the same as that of
deionized water. The resulting carbon particles were dried in an
oven at 100 °C overnight and finally ground to powder in an agate
mortar. The solids are denoted AC-WVx where x corresponds to the
initial concentration of nitric acid.
2.3. Characterization methods
Recently, the possibility of using an activated carbon as a mass-
transfer additive has been reported by our group [34]. Thus, we
have described that the addition of a small amount of activated
carbon could efficiently solve mass-transfer limitations in the Pd-
TPPTS-catalyzed-Tsuji–Trost reaction with water-insoluble alkylal-
lylcarbonates. It has been evidenced that the beneficial effect of the
studied carbon was strikingly connected to the solubility of the
substrates, i.e. the lower the solubility, the more important the
contribution to the mass transfer. For instance, a reaction rate
enhancement of 470 can be obtained with a very hydrophobic sub-
strate, i.e. the allyloctadecylcarbonate [34]. In the present paper,
we wish to examine in more details the effects of activated carbons
on the activity of phosphane-coordinated palladium complexes in
the cleavage of allylalkylcarbonates under biphasic conditions. A
number of parameters will be investigated, such as the type of car-
bon, the oxidizing treatment, the structure of the water-soluble li-
gands, the water conditions, and the metal complex loading. The
ability of activated carbons to promote the palladium-catalyzed-
Tsuji–Trost reaction will be discussed on the basis of adsorption
isotherm studies and 31P{1H} NMR experiments. In conclusion, a
rationale will be proposed to explain the promoting role of acti-
vated carbons during the course of the reaction.
The nitrogen adsorption/desorption isotherm was obtained at
ꢀ196 °C by using a Nova 2200 apparatus from Quantachrome Cor-
poration, after having degassed the sample overnight at 100 °C. The
specific area was calculated from the Brunauer–Emmett–Teller
(BET) equation using P/P0 values between 2.5 ꢁ 10ꢀ3 and
1.7 ꢁ 10ꢀ1, and the pore-size distribution was obtained from the
desorption branch using the BJH method. The total pore volume
was estimated at P/P0 = 0.95, while the micropore volume was
determined by the Dubinin-Radushkevich method. The average
pore size was estimated from the ratio of the total pore volume
to the BET surface area by the Gurvich’s rule, assuming that the
pores are of cylindrical geometry.
The Boehm’s titration method was used to determine the num-
ber of acidic and basic groups [37]. In a typical experiment, 1.0 g of
carbon sample was equilibrated for 24 h with 50 mL of NaHCO3,
Na2CO3, or NaOH solutions (0.05 mol Lꢀ1). Then, 5 mL of each fil-
trate was titrated with HCl (0.05 mol Lꢀ1). The numbers of acidic
sites were calculated from the assumption that NaOH neutralizes
carboxylic, phenolic, and lactonic groups, that Na2CO3 neutralizes
carboxylic and lactonic, and that NaHCO3 neutralizes only carbox-
ylic groups. Basic surface sites were determined similarly by using
HCl solution (0.05 mol Lꢀ1) as reactant.
ATR-FTIR experiments were carried out in the 1800–900 cmꢀ1
region with a spectral resolution of 2 cmꢀ1 on a Shimadzu IR Pres-
2. Experimental
tige-21 spectrometer equipped with
prism.
a MIRacleA Germanium
2.1. Chemicals
The zeta potential (f) of the carbon materials was determined in
water suspensions, using a Malvern Zetasizer Nano ZS (prepara-
tion: 1 mg of AC in 10 mL of deionized water). The f-potential
The sodium salt of meta-substituted trisulfonated triphenyl-
phosphane (TPPTS) was synthesized according to a procedure

210
N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
KF is the Freundlich equilibrium constant, and 1/n is an empirical
dimensionless constant related to the magnitude of the adsorption
driving force.
corresponds to the potential difference between the dispersion
medium and the electrical double layer of fluid attached to the dis-
persed particle. The measurements are based on a Laser Doppler
electrophoretic mobility of carbon particles via the Helmoltz–
Smoluchowsky equation:
2.5.2. 31P{1H} NMR study of the Pd(TPPTS)3 complex
Pd(TPPTS)3 in a D2O solution was synthesized according to a
modified literature procedure [38]. In a Schlenk tube, Pd(PPh3)4
(103 mg, 0.09 mmol) was dissolved in 2 g of previously degassed
toluene under nitrogen atmosphere. TPPTS (76 mg, 0.13 mmol)
was dissolved in 2 g of previously degassed D2O and then trans-
ferred by canula to the organic palladium solution. The biphasic
medium was stirred for 0.5 h at room temperature. After decanta-
tion, the aqueous phase was recovered. This solution contained the
expected palladium catalyst Pd(TPPTS)3 and an excess of about 10%
of free TPPTS (with regard to the initial amount of TPPTS). Studies
in the presence of activated carbon were conducted as follows: to
f ¼ ð
g
=
e
Þ ꢂ le
where le is defined as the ratio between the velocity of the carbon
particles and the magnitude under the applied electric field, is the
g
viscosity of the suspending liquid and
of water.
e
the dielectric conductivity
The 31P{1H} NMR spectra were recorded on a Bruker Avance 300
DPX instrument at 75.46 MHz, respectively. The 31P{1H} chemical
shift is given in ppm relative to the external reference that is
H3PO4 in D2O.
the required amount of AC-WV was introduced 500 lL of the above
2.4. Catalytic procedure
Pd(TPPTS)3 solution under nitrogen atmosphere. After 4 h of stir-
ring, the resulting solution was transferred by canula into a nitro-
gen pressurized 5 mm NMR tube and the 31P{1H} NMR spectrum
was recorded.
The water-soluble phosphane (94.5
lmol, 9 equiv.) and
Pd(OAc)2 (10.5 mol) were dissolved in water (2.0 g) in a Schlenk
l
tube and stirred with a magnetic bar for 4 h at room temperature
under nitrogen atmosphere. The yellow solution was transferred
by canula into another Schlenk tube containing the activated car-
bon (5 mg). The resulting catalytic suspension was stirred with a
magnetic bar at room temperature for 17 h. Then, a mixture con-
3. Results
3.1. Carbon characterizations
taining
allylundecylcarbonate
(0.64 mmol),
diethylamine
(1.28 mmol), and dodecane (0.32 mmol) used as internal standard
was dissolved in heptane (1.2 g) and transferred into the catalytic
suspension. The reaction was carried out at room temperature, and
the stirring rate was adjusted accordingly (1250 rpm), so that
external mass-transfer limitations were eliminated [34]. The reac-
tion was monitored by analyzing aliquots of the reaction mixture
(organic phase) with a Shimadzu GC-17A gas chromatograph,
Before examining the behavior of carbons in aqueous organo-
metallic catalysis, physicochemical characterizations have been
undertaken. Thus, the nitrogen sorption isotherms of the unmodi-
fied commercial carbons are reported in Fig. 1, whereas the corre-
sponding textural parameters are gathered in Table 1.
Among the three starting carbons, the MeadWestvaco AC-WV
carbon is the one which develops the highest surface area
(1690 m2 gꢀ1) and largest pore size (3.1 nm). It exhibits a compos-
ite type I–II isotherm, consistent with a combined micro- and mes-
oporosity, as shown on the BJH pore-size distribution with sizes in
the 2–10 nm range. In comparison, the plateaux in the N2 uptakes
observed for AC-NorA and AC-NorB suggest that the materials are
essentially microporous, with 60% and 75% of the total pore vol-
ume occupied by micropores, respectively. Concerning the pore
size, it can be noted that the higher the microporosity percentage,
equipped
with
a
methyl
silicone
capillary
column
(30 m ꢁ 0.32 mm) and a flame ionization detector.
2.5. Interactions with the activated carbon
2.5.1. TPPTS adsorption isotherm
The ability of AC-WV to adsorb TPPTS was evaluated using the
following procedure: in a round bottom flask and under nitrogen
atmosphere, AC-WV (25 mg) was suspended in 5 mL of TPPTS of
a known concentration and vigorously stirred for 4 h at room tem-
perature. After elimination of the solid phase by filtration, the
absorbance of the solution was analyzed at 250 nm by a UV–Vis
Lambda 19 Perkin-Elmer spectrophotometer. The experimental
data for the uptake of TPPTS were treated according to the Lang-
muir and Freundlich equations. The Langmuir model is based on
a monolayer adsorption with a uniform energy of adsorption and
characterized by the following equation:
1200
0.12
0.08
0.04
800
0
1
10
dp (nm)
KL ꢂ q0 ꢂ Ce
qe ¼
1 þ KL ꢂ Ce
where qe is the equilibrium amount of solute per unit mass of acti-
vated carbon (mg gꢀ1), q0 is the maximum adsorption per unit mass
of activated carbon (mg gꢀ1), Ce is the equilibrium concentration
(mg Lꢀ1), and KL is the Langmuir equilibrium constant.
In contrast to the theory of Langmuir, the Freundlich equation is
derived from an empirical model, in which the energy of adsorp-
tion varies exponentially with surface coverage, according to the
following equation:
400
0
0
0.25
0.5
0.75
1
1=n
qe ¼ KF ꢂ ðCeÞ
P/P0
where qe is the equilibrium amount of solute per unit mass of acti-
vated carbon (mg gꢀ1), Ce is the equilibrium concentration (mg Lꢀ1),
Fig. 1. Nitrogen adsorption isotherm and BJH pore-size distribution of AC-WV (j),
AC-NorA (d), and AC-NorB (ꢀ) carbons.

N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
211
Table 1
Textural characteristics of the starting carbon supports.
Carbon
SBET (m2 gꢀ1
)
Vtot (cm3 gꢀ1
)
Vmicro (cm3 gꢀ1
)
% Microa
Average pore diameter (nm)
AC-WV
AC-NorA
AC-NorB
1690
1207
1224
1.32
0.71
0.60
0.54
0.42
0.45
41
60
75
3.1
2.4
2.0
a
Relative percentage of microporosity defined as the ratio of the micropore volume to the total pore volume.
the smaller the average pore size. Although the pore size calcula-
tion based on the Gurvich’s rule is not the most accurate method,
it gives nevertheless an idea of the differences in pore size in the
materials. Thus, the pore size of AC-NorB is in the range of microp-
ores (2.0 nm), whereas for AC-NorA and AC-WV, the average pore
sizes are in the mesopore range (2.4 and 3.1 nm, respectively).
Another important concern is the determination of the surface
compositions by the Boehm’s titrations. As shown by results re-
ported in Table 2, the AC-WV carbon presents a variety of species
mainly acid, likely due to its origin and method of activation with
phosphoric acid. However, the total amount of sites remains low,
Further spectroscopic evidence of the presence of oxygen con-
taining organic functional groups after oxidative treatments has
been obtained using FTIR spectroscopy (Fig. 2). The spectrum of
the unmodified AC-WV shows weak bands at ca. 1570, 1200, and
1020 cmꢀ1 ascribed to the vibrational modes of the carbon matrix
[m(C@O), m(C@C) or m(CAO), and m(OACAO), respectively]. With
acid treatments, these bands are magnified. An additional band
at 1750 cmꢀ1 is also revealed, which indicates the formation of car-
boxylic and lactonic groups on the edges of layers planes. Accord-
ing to the literature [40], the carbon atoms located at the periphery
of the layer (so-called edge sites) exhibit higher reactivity toward
oxidation than those within the basal plane sites and surrounded
by other carbon atoms. Concurrently to the changes in the nature
and concentration of the surface functionalities, it is also shown
in Table 2 that the specific areas are affected by the oxidation treat-
ment of AC-WV with nitric acid. This tendency may reflect the
alteration of the carbon matrix by introduction of oxygen groups
i.e. 39 l
mol gꢀ1 corresponding to a surface density of 0.02 function
per nm2 (entry 1 in Table 2). This result is consistent with a carbon
surface almost free from oxygen functional groups. Note that this
assumption has been further supported by EDX analysis showing
that AC-WV mostly contains the carbon element in its matrix
(Fig. S1 in the Supplementary material). Thus, the presence of sili-
cium, aluminum, and phosphorus is detected in trace amount.
These elements could be the result of the synthesis process of
the activated carbon, like phosphorus coming from the activation
step by phosphoric acid. This analysis also corroborates with the
XPS surface analysis, whose spectrum indicates a surface composi-
tion of 94.6% carbon and 5.6% oxygen (Fig. S2 and Table S1 in the
Supplementary material). Thus, the resulting O/C surface ratio,
i.e. 0.59, is in agreement with the total acidity measured by
Boehm’s titration on the basis of the relationship established by
Devillers et al. [39]. In contrast, it is shown that AC-NorA and
AC-NorB possess more functionalized surfaces. Indeed, the total
number of surface sites (acid and basic) is 14 and 7 times higher,
respectively, than that measured with the unmodified AC-WV
sample (entries 2–3 in Table 2). More precisely, the titrations indi-
1200 cm-1 1020 cm-1
(f)
1750 cm-1 1570 cm-1
(e)
(d)
(c)
(b)
cate that AC-NorB is the most acidic support with 201 l
mol gꢀ1 of
(a)
total acid functions, whereas AC-NorA develops the highest basic
character. In line with what is usually observed, the oxidation of
AC-WV by nitric acid leads to an increasing number of oxygen con-
taining groups, in almost each category; the stronger the treat-
ment, the higher the number of acidic functions. For instance, the
most highly functionalized material (HNO3 16 M for 1 h under re-
1850
1650
1450
1250
1050
850
wavenumber (cm-1)
Fig. 2. Effect of the nitric acid treatments on the ATR-FTIR spectra of AC-WV: (a)
flux) displays a total acidity of 1421 l
mol gꢀ1 corresponding to 1.1
untreated support and treated with different concentrations of nitric acid, (b)
0.0001; (c) 0.001; (d) 0.01; (e) 0.1; and (f) 16 mol Lꢀ1
.
functions per nm2 (entry 8 in Table 2).
Table 2
Characteristics of the starting and modified carbons in terms of specific areas, Zeta potentials and Boehm’s titrations.
Entry Carbon
Treatment Temperature conditions CHNO3 (mol Lꢀ1
)
SBET (m2 gꢀ1
)
l )
n (mV)a Titration results ( mol gꢀ1
Carboxylic Lactonic Phenolic Acidic Basic Total
1
2
3
4
5
6
7
8
AC-WV
AC-NorA
AC-NorB
AC-WV.0001 Yes
AC-WV0.001 Yes
AC-WV0.01
AC-WV0.1
AC-WV16
No
No
No
–
–
–
–
–
–
1690
1207
1224
1540
1380
1286
917
ꢀ30.0
ꢀ34.1
ꢀ42.7
ꢀ31.4
ꢀ32.8
n.d.c
4
1
77
5
10
50
85
407
11
8
1
13
10
20
126
484
19
37
123
35
34
46
201
53
5
39
539
272
53
493
71
–
–
–
r.t.b
r.t.b
r.t.b
Reflux
Reflux
10ꢀ4
10ꢀ3
10ꢀ2
10ꢀ1
16
70
90
90
Yes
Yes
Yes
151
120
530
221
332
1421
221
332
1421
ꢀ34.7
–
–
776
ꢀ48.9
a
b
c
Zeta-potential measurements performed in water. The standard deviations, calculated by means of three replicates, were within 2 mV.
Room temperature.
Non-determined value.

212
N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
and removal of some carbon atoms [41]. The formation of surface
oxygen groups at the entrance to the pores may also obstruct N2
access to the micropores [42].
100
75
50
25
0
However, it should be also noted that the carbon structure was
not completely denatured after acid treatment. With this objective,
electron microscopy investigations have been performed on AC-
WV and AC-WV16 (Fig. S3 in the Supplementary material). Thus,
we observe from the Scanning Electron Microscopy (SEM) images
taken at the same magnification that the acid treatment provides
a notable decrease in the carbon particle size. Note that this last ef-
fect has also been confirmed by laser granulometry measurements,
which indicate medium particle sizes (d50) of 36 and 7 lm for AC-
WV and AC-WV16, respectively (Fig. S4 in the Supplementary
material). The effect of the oxidation treatment is also revealed
in the EDX spectrum with the appearance of an intense band K
a
1
at ca. 0.5 keV characteristic of the oxygen element (Fig. S5 in the
Supplementary material). A more detailed examination of the
structure of carbons is obtained at a higher magnification by Trans-
mission Electron Microscopy (TEM). In both cases, the TEM images
exhibit a similar internal structure made of a disordered entangle-
ment between carbon sheets.
0
10
20
30
40
50
time (h)
Fig. 3. Conversion of allylundecylcarbonate in the Pd(TPPTS)3-catalyzed cleavage
reaction as a function of time without (s) and with activated carbon in the reaction
medium: AC-NorB (j), AC-NorA (N), and AC-WV (d).
To evaluate the impact of the degree of the surface functionali-
zation on the stability of the carbon suspensions, zeta-potential
measurements have been carried out in water (Table 2). Indeed,
this technique is considered to be a useful indicator to evaluate
the balance between the repulsive and attractive forces in solid
suspensions. Thus, it is generally assumed that carbon particles,
with zeta potentials more positive than +30 mV or more negative
than ꢀ30 mV, form stable suspensions in water. Regarding the
unfunctionalized AC-WV, the zeta potential (in absolute value) is
rather lower than those obtained with AC-NorA and AC-NorB (30
vs 34.1 and 42.7 mV, respectively). This result can be related to
the fact that AC-WV possesses the lowest content of oxygen-sur-
face groups. This assumption concurs with the results obtained
after acid treatment, for which the zeta-potential values become
higher with the gradual increase in the number of oxygen acidic
groups in the backbone of the carbon structure. This effect can
be rationalized by an increase in the surface densities of charged
functional groups, giving rise to enhanced electrostatic repulsions
and enhanced stabilities of the carbon suspensions in water.
conversion are obtained. Thus, the conversions reach 13% and
79% after 4 h reaction in the presence of AC-NorA and AC-WV,
respectively, vs 2.3% in the presence of AC-NorB (entries 2–4 in Ta-
ble 3). Note that AC-WV is by far the most efficient promoter, with
a relative reaction rate (ratio of the initial activity with AC to that
measured without AC) achieving a factor of 83 by using 5 mg of
additive. This result suggests that the addition of carbon particles
leads to higher mass transfer as a result of higher interfacial area
between the aqueous and organic phases. Interestingly, a twofold
increase in the AC-WV loading results in a twofold increase in
the reaction rate (entry 5 in Table 3), whereas further addition of
carbon decreases the catalytic efficiency per unit surface (entries
6–8 in Table 3).
Two factors can account for the higher efficiency of AC-VW.
The first one is the pore-size distribution, which has to be ade-
quate to adsorb the allylundecylcarbonate substrate with a mini-
mum of steric hindrance [43]. Among the materials studied here,
it has already been reported that AC-WV possessed the broadest
pore-size distribution, with most of the pores in the 2–10 nm
range. Thus, for a similar range of amount of surface acid sites
3.2. Aqueous organometallic catalysis
3.2.1. Cleavage of allylundecylcarbonate
and specific surface area (200 l ,
molꢀ1 and 1200–1300 m2 gꢀ1
The carbon materials have been tested in the cleavage reaction
of allylundecylcarbonate under biphasic aqueous conditions and in
the presence of diethylamine as the allyl scavenger (Scheme 1).
The catalytic species are formed by in situ reduction of Pd(OAc)2
with 9 equiv. of phosphane. This procedure has the great advan-
tage to avoid the tedious steps of synthesis and purification of
the Pd(TPPTS)3 complex [38].
When comparing the allylundecylcarbonate conversion over
time, it is readily apparent that the reaction occurs very slowly
when the experiment is conducted without any addition of acti-
vated carbon (Fig. 3). This result confirms that the limiting step
of the reaction is the mass transfer. Whereas the addition of AC-
NorB is inadequate to improve the reaction rate, we observe that
AC-NorA and AC-WV, especially, appear to be valuable mass-trans-
fer agents in the Tsuji–Trost reaction since much higher levels of
respectively), our results suggest that carbons having the highest
proportion of mesopores remain the best carriers to promote the
Tsuji–Trost reaction (21.0 vs 3.5 mol hꢀ1 for the AC-WV0.01 and
AC-NorB, respectively). These textural features should facilitate
the adsorption and diffusion of the long-alkyl chain allylcarbonate,
from the organic phase to the aqueous phase that contains the
catalyst.
The second one is the quasi absence of oxygenated function at
the surface of the material. Indeed, experiments conducted in the
presence of the oxidized AC-WV samples with varying degree of
functionalization show that the overall reaction rate is drastically
reduced with increasing content of oxygenated groups (Fig. 4).
We assume that the incorporation of oxygen acidic groups makes
the surface more polar, thereby leading to a decrease in the
O
cat. [Pd]/L + AC
H2O/heptane, RT
Scheme 1. Palladium-catalyzed cleavage of allylalkylcarbonate (Tsuji–Trost reaction).
+
CO2
+ Et N
HNEt2
R
OH
+
R
O
O
2

N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
213
Table 3
Cleavage of the allylundecylcarbonate over Pd(TPPTS)3 using activated carbons.a
Entry
Additive
Weight (mg)
Initial reaction rateb
mol hꢀ1
Relative reaction ratec
Initial areal rated
mol hꢀ1 mol(Pd)ꢀ1 mꢀ2
)
Conv.e (%)
(l
)
(
l
l
1
2
3
4
5
6
7
8
9
10
11
12
13
14h
–
–
5
5
5
10
20
40
70
5
5
5
5
5
3.7
3.5
–
–
2.5
2.3
13
AC-NorB
AC-NorA
AC-WV
AC-WV
AC-WV
0.9
6.8
83
174
258
318
346
71
57
5.7
1.4
–
0.05
0.40
3.7
3.9
2.9
1.8
1.1
3.2
2.9
0.3
0.1
0
25.2
307
640
956
1176
1281
262
210
21.0
5.3
79
100
100
100
100
56f
50f
10.5
3.3
0g
AC-WV
AC-WV
AC-WV.0001
AC-WV0.001
AC-WV0.01
AC-WV0.1
AC-WV16
AC-WV16
0
493
5
–
–
100h
a
Reaction conditions: Pd, 10.5
lmol; TPPTS, 94.5 lmol; Water, 2.0 g; Allylundecylecarbonate, 0.64 mmol; Diethylamine, 1.28 mmol, C12H26 (internal standard),
0.32 mmol; Heptane, 1.2 g; Room temperature; Stirring rate, 1250 rpm.
b
Defined as the number of micromoles of allylundecylcarbonate converted per hour and calculated from the conversion measured at 30 min.
Defined as the ratio of the initial reaction rate in the presence of AC to that without AC.
Defined as the number of micromoles of allylundecylcarbonate converted per hour per micromoles of palladium and per surface area (on the basis of the specific surface
c
d
area) and calculated from the conversion measured at 30 min.
e
After 4 h of reaction.
After 2 h of reaction.
After 7 h of reaction.
f
g
h
Addition of
a-randomly methylated cyclodextrin (100 mg) in the reaction mixture of the run 13 after the 7 h of test. a-CDs are water-soluble cyclic oligosaccharides
formed of 6 glucopyranose units exhibiting a hydrophobic cavity that can form inclusion complexes with of a wide variety of organic molecules. This property can be
exploited in aqueous biphasic catalysis to solve mass-transfer limitations [12].
rapid increase in the conversion, reaching 100% in 2 h (entry 14
in Table 3). This last result proves that the inhibition of the reaction
observed in the presence of AC-WV16 alone is due to the alteration
300
of the mass-transfer capacity of this functionalized carbon.
3.2.2. Influence of the water content
The effect of water content has been also examined (Fig. 5). It is
found that, by using AC-WV, the initial activities remain low over a
wide range of water contents, i.e. until 85 wt.% and that no maxi-
mum is observed for water contents close to the pore filling
(10 wt.% of water). Thus, our results show that only large amounts
of water can ensure proper conditions for catalysis, and this behav-
ior is strongly different from what is usually described in SAPC. In-
deed, it is usually assumed that the highest activities are achieved
200
100
0
350
300
250
200
0
500
1000
1500
content of acid site (µmol g-1)
Fig. 4. Effect of the surface functionalization on the initial rate of allylundecylcar-
bonate cleavage over Pd(TPPTS)3 with (4) AC-NorA, (h) AC-NorB, and (d) AC-WV
series treated by nitric acid.
150
mass-transfer capacity toward the highly hydrophobic substrate,
i.e. the allylundecylcarbonate. This interpretation is in line with
studies in the literature dealing with the relationship between sur-
face chemistry of activated carbons and adsorption of organic pol-
lutants [44,45]. Note that the oxidative treatments do not provide
any irreversible degradation of the catalytic complex. Indeed,
whereas no catalytic activity is measured in the presence of
AC-WV16 after 7 h of reaction (entry 13 in Table 3), the catalytic
activity can be recovered by adding a typical water-soluble mass-
pore filling
10 wt. %
100
50
0
0
25
50
75
100
water content (wt. %)
transfer agent, i.e. the methylated-
[12], in the reaction medium. Thus, immediately after the addition
of RaMe- -CD at the end of the 7-h test of Run 13, we observe a
a-cyclodextrin (RaMe-a-CD)
Fig. 5. Effect of the amount of water on the initial reaction rate in the Pd (TPPTS)3-
catalyzed cleavage reaction of allylundecylcarbonate.
a

214
N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
with a low degree of pore filling and that slight deviations from the
optimum of water content inevitably lead to reduced activities
[21].
In our case, optimal conditions require large amounts of water to
counterbalance the preferential adsorption of the organic substrate.
Thus, it is proposed that, at these high water loadings, the carbon
allows the mass transfer of the organic substrate by adsorption and
the mixing with the aqueous phase containing the water-soluble
metallic complex, leading to more effective interfacial areas.
electronic, steric, or interfacial factors. To determine whether all
water-soluble phosphane exhibit activity enhancements in the
presence of the carbon additive, the behavior of different palla-
dium/water-soluble phosphane combinations has been studied in
the cleavage reaction of allylundecylcarbonate. In order to ensure
high water solubility, the ligands are based on the backbone of
TPPTS, with the substitution of one or more 3-sulfonatophenyl
groups by 4-sulfonatobiphenyl groups. The results are reported
in Table 4.
The promoting effect of AC-WV in the cleavage reaction is influ-
enced by the chemical structure of the ligand itself. Thus, the rela-
tive reaction rates, measured in the presence or absence of carbon,
for P(BiPh)3TS, P(BiPh)2PhTS, and P(BiPh)Ph2TS are ca. 75, 4, and
1.7 times lower than that measured by using TPPTS. The efficiency
of AC-WV seems to depend on the number of sulfonatobiphenyl
groups in the phosphane unit; thus, the higher the sulfonatobiphe-
nyl group content, the less important the beneficial contribution of
AC-WV. These observations can be related to an alteration of the
mass-transfer properties of the carbon, when the metal complex
is stabilized by phosphanes of higher molecular weights.
3.2.3. Influence of the palladium loading
In order to gain deeper insight into the understanding of the
catalytic process assisted by AC-WV, the influence of the complex
loading, i.e. Pd(OAc)2 and TPPTS, has been further examined. It
should be emphasized that the total content of diethylamine re-
mains here constant.
As shown in Fig. 6, two different domains are observed. Above
5
l
mol of palladium (zone I), the cleavage reaction rates are inde-
pendent of the palladium concentration, with a value close to
310
mol hꢀ1. The resulting linear relationship confirms that the
l
overall rate of the Tsuji–Trost reaction is limited by the diffusion
of the allylundecylcarbonate at the interfaces and not by the kinet-
ics of the reactions of the catalytic cycle. Below 5 lmol (zone II), we
4. Discussion
observe that the activity rises sharply with decreasing palladium
loading. This effect is caused by the leaching of palladium from
the aqueous to organic phase, associated with a ligand exchange
reaction occurring between the coordinated-TPPTS and the nucle-
ophilic diethylamine (Scheme 2). Thus, the equilibrium displace-
ment becomes prominent when the HNEt2/TPPTS ratio is very
high (above ca. 30). Similar observations have already been re-
ported in palladium-catalyzed Heck reactions [46].
Starting from the previous results, we think that the palladium
catalytic system, in the presence of activated carbons, could be
controlled by different adsorption–desorption processes. On this
basis, it can be hypothesized that, in addition to the adsorption
of the organic substrate, the palladium complex coordinated by
sulfonated phosphanes could also interact with the carbon surface.
Remembering that Pd(TPPTS)3 in the presence of AC-WV remains
the best combination to promote the cleavage reaction, we have fo-
cused the investigations on this particular system.
3.2.4. Use of other water-soluble phenylphosphane ligands
Evidence for the affinity of TPPTS toward the carbon surface has
been demonstrated from isothermal adsorption measurements
performed in aqueous solutions at room temperature. The experi-
mental sorption capacity is shown in Fig. 7, and the parameters of
The nature of the ligand is known to have a large influence on
the overall rates of catalytic reactions carried out under aqueous
biphasic conditions. This is usually attributed to a combination of
400
350
300
250
0
5
10
15
15
10
20
NHEt2 /TPPTS
7.5
30
Fig. 6. Effect of the palladium content and diethylamine/phosphane ratio on the initial reaction rate of the cleavage of allylundecylcarbonate over Pd(TPPTS)3 in the presence
of AC-WV.
Pd(TPPTS)3
+ 3 HNEt2
Pd(HNEt2)3
+ 3 TPPTS
Scheme 2. Ligand exchange between the TPPTS-coordinated palladium and diethylamine.

N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
215
Table 4
Catalytic performance of PdL3 systems stabilized by different water-soluble phosphanes in the presence or absence of AC-WV.
d
Ligand
Structure
Reaction rate without ACa
mol hꢀ1
Reaction rate with ACb
mol hꢀ1
Conv.c
(%)
Relative reaction rate
83
(l
)
(l
)
SO Na
TPPTS
3.7
307
48
3
P
P
3
P(BiPh)3TS
2.3
14
2.5
0.4
45
1.1
21
SO Na
3
3
SO Na
P(BiPh)2PhTS
288
3
NaO3S
NaO3S
P
2
SO Na
P(BiPh)Ph2TS
<1
44
6.9
ꢃ50
3
P
2
a
Reaction conditions: AC-WV, 5.0 mg; Pd, 10.5
lmol; Ligand, 94.5 lmol; Water, 2.0 g; Allylundecylecarbonate, 0.64 mmol; Diethylamine, 1.28 mmol, C12H26 (internal
standard), 0.32 mmol; Heptane, 1.2 g; Room temperature; Stirring rate 1250 rpm.
b
Defined as the number of micromoles of allylundecylcarbonate converted per hour and calculated from the conversion measured at 30 min.
Conversion (%) in the presence of AC-WV after 1 h.
Defined as the ratio of the initial reaction rate in the presence of AC to that without AC.
c
d
400
Table 5
Langmuira and Freundlichb parameters calculated from the adsorption of TPPTS on
AC-WV.
Experimental
Freundlich
Langmuir
Isotherm equation Parameters
r2
300
200
100
0
180:6ꢂCe
Langmuir
q0 = 172.4 mg gꢀ1
,
0.9386
qe
¼
1þ1:047Ce
KL = 1.047 L mgꢀ1
0:379
Freundlich
1/n = 0.379, KF = 69.4
0.9753
qe ¼ 69:4 ꢂ ðCeÞ
1/n
(mg gꢀ1)(L mgꢀ1
)
a
1
qe
1
1
0 ꢂKL
1
Langmuir linearization:
¼
þ q
ꢂ C
.
q0
Freundlich linearization: log qe ¼ log KFe þ ð1=nÞ ꢂ log Ce.
b
whereas the peak at d2 = 34.598 ppm is attributed to trace amounts
of TPPTS oxide. In agreement with previously reported data in the
literature, the d1 chemical shift corresponds to an average value
resulting from fast exchanges, occurring between the TPPTS-coor-
dinated Pd and the free TPPTS [48–50]. By adding 5 mg of AC-VW
into the Pd(TPPTS)3 solution, the spectrum reveals a narrowing and
a low-field shift of the d1 signal, concurrently to the decrease in the
peak area (Fig. 8b). These effects are still more visible with 10 mg
of AC-WV (Fig. 8c). These changes observed in the NMR spectra
suggest that the addition of AC-WV causes an equilibrium dis-
placement between the coordinated and un-coordinated TPPTS
species, associated with a slower exchange rate. This phenomenon
could only be explained by assuming the spontaneous adsorption
of the free TPPTS and palladium complex via the TPPTS ligand on
the support.
In line with the interpretation of Kalck et al. [48], the progres-
sive shift of the d1 signal toward lower fields indicates the progres-
sive increase in the ratio of TPPTS-coordinated Pd(0) to that of free
TPPTS. We explain this behavior by the increasing removal of
TPPTS from the aqueous solution due to the adsorption on the car-
bon surface. Thus, for the highest AC-WV loading, we have esti-
mated that ca. 42% of the initial amount of phosphane (i.e. free
or coordinated) was adsorbed onto the carbon support, corre-
sponding to a coverage capacity of 800 mg gꢀ1 (Table 6).
0
10
20
30
Ce (mg L-1)
Fig. 7. Adsorption isotherms of TPPTS onto AC-WV: (j) experimental uptake and
model adsorption isotherms obtained from the Freundlich (solid line) and Langmuir
(dotted line) parameters.
the Langmuir and Freundlich models are given in Table 5. Based on
the correlation coefficients of the linear regressions, the uptake
data seem to be best described by the Freundlich equation. Thus,
the modeling curve (solid line) constructed from the KF and 1/n val-
ues fits the experimental data in a very satisfactory way. A result-
ing 1/n value of 0.379 is found, suggesting that the adsorption of
TPPTS takes place easily on AC-WV. Indeed, it has already been
proposed that the Freundlich isotherms could be classified in five
common types of curves depending on the 1/n value and that a va-
lue between 0.1 and 0.5 was indicative of favorable adsorption
conditions [47].
Furthermore, the ability of the Pd(TPPTS)3 metallic complex to
be adsorbed onto AC-WV has also been proved by 31P{1H} NMR
spectroscopic studies, performed in D2O under nitrogen atmo-
sphere (Fig. 8 and Table 6). Pd(TPPTS)3 was synthesized using a
procedure described by Hermann et al. that involves ligand ex-
change reactions between Pd(PPh3)4 and TPPTS [38].
To complete this study, a last NMR investigation has been per-
formed in order to study the evolution of the adsorption phenom-
enon under conditions close to those encountered during the
aqueous biphasic catalytic process. Thus, competitive 31P{1H}
NMR titrations of TPPTS in D2O have been carried out in the pres-
ence of allylundecylecarbonate as the competitor. Experiments are
performed in the following manner: 10 mg of TPPTS and 10 mg of
AC-WV (both in 1 mL of D2O) are added to a defined volume of
Without addition of AC-WV, the spectrum reveals two main sig-
nals (Fig. 8a). The broad peak at d1 = 20.459 ppm is ascribed to the
Pd(TPPTS)3 complex in equilibrium with ca. 10% of free TPPTS,

216
N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
Fig. 8. Effect of the AC-WV support on the 31P{1H} NMR spectrum of Pd(TPPTS)3 in D2O at 25 °C: (a) without activated carbon, (b) in the presence of 5 mg of AC-WV, and (c) in
the presence of 10 mg of AC-WV.
Table 6
31P{1H} NMR characteristics of Pd(TPPTS)3 adsorption onto AC-WV.a
AC-WV (mg)
Free TPPTS (%)
TPPTS-coordinated Pd(0) (%)
Adsorbed TPPTS (%)
Adsorbed TPPTS (mg gꢀ1
)
0
5
10
10
3
ꢃ0
90
79
58
0
18
42
0
720
800
a
The relative amount of each type of TPPTS phosphane, i.e. free and coordinated-Pd(0), is deduced from the integrated peak intensity and the chemical shift of the d1 peak
in the 31P{1H} NMR spectra. The amount of free TPPTS is calculated by assuming that the chemical shift is the weighted average of the characteristic peaks of Pd(TPPTS)3 and
free TPPTS at d 23.0 and ꢀ5.6 ppm, respectively.
organic phase (allylundecylcarbonate dissolved in heptane). The
mixture is shaken for 4 h at room temperature under nitrogen
atmosphere, and the solution was transferred via canula into the
NMR tube prior to the acquisition of the spectrum (Fig. 9).
completely recovered. This demonstrates that the TPPTS ligand can
be partly retained onto the carbon matrix, even under competitive
conditions in the presence of the carbonate substrate.
Taken together, our experimental results are consistent with
the fact that the reaction mechanism is governed by a subtle inter-
play of interactions involving different adsorption equilibria. In-
deed, we have shown that AC-WV could adsorb the palladium/
water-soluble TPPTS catalytic system and that this adsorption pro-
cess was not affected to a large extent by the presence of the or-
ganic substrate. These results support the idea that the addition
of carbon in the reaction medium plays a role not only in the
improvement of mass-transfer of the hydrophobic reactant, but
also in the regulation of the amount of water-soluble palladium
complex in the pore volume. This phenomenon leads to more
effective interfacial areas that promote the molecular collisions
in the pore volume. Thus, we assume that the relatively apolar cav-
ities of AC-WV act as suitable nanometer-range pockets that allow
the confinement of all reactive species present in both organic and
aqueous phases, as described in Scheme 3. It is worth mentioning
that the system proved to be recoverable and reusable at least five
times without any appreciable loss of activity, as shown recently
by our group [34].
In comparison with the reference spectrum of TPPTS in D2O
(Fig. 9a), it is readily apparent that the presence of AC-WV causes
a significant decrease in the relative intensity of the free TPPTS sig-
nal (d3 = ꢀ6.3 ppm) due to its adsorption on the support (Fig. 9b).
In these conditions, the percentage of adsorbed TPPTS is estimated
to be ca. 40%.1 Addition of allylundecylcarbonate to the aqueous sus-
pension AC-WV and TPPTS gives rise to an intensity increase in d3,
showing that the affinity of TPPTS for the carbon surface is affected
by a small amount of the hydrophobic competitor (Fig. 9c). Thus, the
percentage of adsorbed TPPTS is reduced from 40% to 28%. Note that
this last value remains almost constant with further increments of
allylundecylcarbonate (Fig. 9d and e). These results can be explained
by a partial exchange reaction occurring between the surface-
adsorbed TPPTS and allylundecylcarbonate of the organic phase. We
assume that this phenomenon is connected to different adsorption
equilibrium constants. Interestingly, it should be also emphasized that
the removal of TPPTS from the surface by allylundecylcarbonate is not
complete, since the intensity of the original signal d3 is never
The proposed mechanism can also account for the lower rela-
tive reaction rates previously reported with the bulkiest water-
soluble ligands. Indeed, this behavior seems to indicate an
alteration of the mass-transfer ability of the carbon when the
metal complex is stabilized by phosphanes of higher molecular
1
The percentage of adsorbed TPPTS is calculated on the basis of the following
formula: TPPTSads (%) = 100 ꢁ {(A0 ꢀ A)/A0}, where A0 is the peak area of free TPPTS in
water (spectrum a, Fig. 9) and A is the peak area of free TPPTS in the presence of AC-
WV.

N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
217
Fig. 9. Effect of addition of allylundecylcarbonate on the 31P{1H} NMR spectrum of TPPTS (10 mg) in D2O (1 mL) at 25 °C using H3PO4 as the external reference: (a) without
AC-WV or carbonate, (b) with AC-WV and without carbonate, (c) with AC-WV and 10 mol of carbonate, (d) with AC-WV and 30 mol of carbonate, and (e) with AC-WV and
40 mol of carbonate.
l
l
l
method allows the easy separation of the products from the cata-
lytic suspension and catalyst recyclability without any detectable
loss of activity. A number of parameters, which governs the cata-
lytic properties of this triphasic organic-aqueous-AC system, have
been identified to increase the overall reaction rate, including ade-
quate pore size, carbon-rich surface, hydrophobic reactants, and high
water content. It was found that the mechanism could not been as-
cribed to a SAPC-type mechanism. Indeed, we have established
that different adsorption equilibria with the hydrophobic substrate
and the organometallic complex were involved in the interactions
with the carbon surface. This hypothesis was clearly supported by
isothermal adsorption studies and 31P{1H} NMR experiments. Fi-
nally, it is suggested that carbons combining a high surface area,
a high proportion of mesopores and a relatively apolar nature are
key factors to allow the confinement and mobility of the reactants
and catalyst within the porous network of the solid. These conclu-
sions lay the foundation for future investigations to improve the
performances of metal-catalyzed processes performed under bi-
phasic conditions. In particular, experiments are currently under
way to develop the use of hydrophobic mesoporous carbons with
high surface areas and large pore volumes.
Scheme 3. Proposed mechanism of action of activated carbon in aqueous organo-
metallic catalysis.
weights. This effect could result from a loss of accessible surface
area for the hydrophobic allylundecylcarbonate during the cata-
lytic course, due to a higher surface coverage by the palladium
complex species or extended pore blocking by steric congestions.
Acknowledgments
Dr. N. Kania is grateful to the Agence de l’Environnement et la
Maitrise de l’Energie and the Région Nord-Pas de Calais for a PhD
grant (2006–2009). Dr Gokulakrishnan is thankful for financial
support by the Institut de Recherche en Environnement Industriel
(IRENI, 2009). MeadWestvaco Corporation (Covington, USA) is
gratefully acknowledged for the generous gift of the activated
carbon.
5. Conclusion
In conclusion, we have described a simple and environmentally
friendly strategy for enhancing the Tsuji–Trost overall reaction
rates, based on the combined use of a water-soluble palladium
complex and activated carbon as mass-transfer promoter. The

218
N. Kania et al. / Journal of Catalysis 278 (2011) 208–218
[23] K.T. Wan, M.E. Davis, J. Catal. 148 (1994) 1.
Appendix A. Supplementary material
[24] S. dos Santos, Y. Tong, F. Quignard, A. Choplin, D. Sinou, J.P. Dutasta,
Organometallics 17 (1998) 78.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2010.12.005.
[25] P. Handa, M. Stjerndahl, K. Holmberg, Micropor. Mesopor. Mater. 100 (2007)
146.
[26] R.L. Kars, R.J. Best, A.A.H. Drinkenburg, Chem. Eng. J. 17 (1979) 201.
[27] E. Alper, B. Wichtendahl, W.D. Deckwer, Chem. Eng. Sci. 35 (1980) 217.
[28] A.A.C.M. Beenackers, W.P.M. van Swaaij, Chem. Eng. Sci. 48 (1993) 3109.
[29] S. Nedeltchev, A. Schumpe, Chem. Eng. Sci. 60 (2005) 6504.
[30] C. Moreno-Castilla, Carbon 42 (2004) 83.
[31] L. Peereboom, B. Kenigsknecht, M. Hunter, J.E. Jackson, D.J. Miller, Carbon 45
(2007) 579.
[32] F. Rodriguez-Reinoso, Carbon 36 (1998) 159.
References
[1] B. Cornils, W.A. Herrmann (Eds.), Aqueous-Phase Organometallic Catalysis,
Wiley-VCH, Weinheim, 2004.
[2] K.H. Shaughnessy, Chem. Rev. 109 (2009) 643.
[3] B. Cornils, J. Mol. Catal. A: Chem. 143 (1999) 1.
[33] C. Disser, C. Muennich, G. Luft, Appl. Catal. A 296 (2005) 201.
[34] N. Kania, B. Léger, S. Fourmentin, E. Monflier, A. Ponchel, Chem. Eur. J. 16
(2010) 6138.
[35] R. Gärtner, B. Cornils, H. Springer, P. Lappe, DE patent 3235030 (1982).
[36] M. Ferreira, H. Bricout, F. Hapiot, A. Sayede, S. Tilloy, E. Monflier,
ChemSusChem 1 (2008) 631.
[4] P. Purwanto, H. Delmas, Catal. Today 24 (1995) 135.
[5] H. Ding, B.E. Hanson, T. Bartik, B. Bartik, Organometallics 13 (1994) 3761.
[6] M.S. Goedheijt, B.E. Hanson, J.N.H. Reek, P.C.J. Kamer, P.W.N.M. van Leeuwen, J.
Am. Chem. Soc. 122 (2000) 1650.
[7] M. Bortenschlager, N. Schöllhorn, A. Wittmannand, R. Weberskirch, Chem. Eur.
J. 13 (2007) 520.
[8] E. Paetzold, G. Oehme, J. Mol. Catal. A: Chem. 152 (2000) 69.
[9] C.C. Miyagawa, J. Kupka, A. Schumpe, J. Mol. Catal. A: Chem. 234 (2005) 9.
[10] H. Fu, M. Li, H. Chen, X. Li, J. Mol. Catal. A: Chem. 259 (2006) 156.
[11] S.L. Desset, D.J. Cole-Hamilton, D.F. Foster, Chem. Commun. (2007) 1933.
[12] C. Binkowski, J. Cabou, H. Bricout, F. Hapiot, E. Monflier, J. Mol. Catal. A: Chem.
215 (2004) 23.
[13] P. Blach, D. Landy, S. Fourmentin, G. Surpateanu, H. Bricout, A. Ponchel, F.
Hapiot, E. Monflier, Adv. Synth. Catal. 347 (2005) 1301.
[14] F. Hapiot, L. Leclercq, N. Azaroual, S. Fourmentin, S. Tilloy, E. Monflier, Curr.
Org. Synth. 15 (2008) 162.
[37] H.P. Boehm, Carbon 32 (1994) 759.
[38] W.A. Herrmann, J. Kellner, H. Riepl, J. Organomet. Chem. 389 (1990) 103.
[39] S. Hermans, C. Diverchy, O. Demoulin, V. Dubois, E.M. Gaigneaux, M. Devillers,
J. Catal. 243 (2006) 239.
[40] S. Biniak, M. Pakula, G.S. Szymaanski, A. Swiatkowski, Langmuir 15 (1999)
6117.
[41] A.A. El-Hendawy, Carbon 41 (2003) 713.
[42] H. Valdés, M. Sánchez-Polo, J. Rivera-Utrilla, C.A. Zaror, Langmuir 18 (2002)
2111.
[43] M. Ahmednaa, W.E. Marshall, R.M. Rao, Bioresour. Technol. 71 (2000) 103.
[44] C.O. Ania, B. Cabal, C. Pevida, A. Arenillas, J.B. Parra, F. Rubiera, J.J. Pis, Water
Res. 41 (2007) 333.
[45] M.A. Lillo-Ródenas, D. Cazorla-Amorós, A. Linares-Solano, Carbon 3 (2005)
1758.
[15] F.X. Legrand, M. Sauthier, C. Flahaut, J. Hachani, C. Elfakir, S. Fourmentin, S.
Tilloy, E. Monflier, J. Mol. Catal. A: Chem. 303 (2009) 72.
[16] A.A. Dabbawala, J.N. Parmar, R.V. Jasra, H.C. Bajaj, E. Monflier, Catal. Commun.
10 (2009) 1808.
[46] M. Beller, T.H. Riermeier, Tetrahedron Lett. 37 (1996) 6535.
[47] R.L. Tseng, F.C. Wu, J. Hazard. Mater. 155 (2008) 277.
[48] F. Monteil, P. Kalck, J. Organomet. Chem. 482 (1994) 45.
[49] C. Amatore, E. Blart, J.P. Genêt, A. Jutand, S. Lemaire-Audoire, M. Savignac, J.
Org. Chem. 60 (1995) 6829.
[17] J.P. Arhancet, M.E. Davis, J.S. Merola, B.E. Hanson, Nature 339 (1989) 454.
[18] J.P. Arhancet, M.E. Davis, J.S. Merola, B.E. Hanson, J. Catal. 121 (1990) 327.
[19] J.P. Arhancet, M.E. Davis, B.E. Hanson, J. Catal. 129 (1991) 94.
[20] J.P. Arhancet, M.E. Davis, B.E. Hanson, J. Catal. 129 (1991) 100.
[21] G. Fremy, E. Monflier, J.F. Carpentier, Y. Castanet, A. Mortreux, J. Catal. 162
(1996) 339.
[50] G. Papadogianakis, J.A. Peters, L. Maat, R.A. Sheldon, J. Chem. Soc. Chem.
Commun. (1995) 1105.
[22] E. Fache, C. Mercier, N. Pagnier, B. Despeyroux, P. Panster, J. Mol. Catal. 79
(1993) 117.