Ionic liquid layer on Pd/C catalyst: Membrane-like effect on the selectivity for multistep hydrogenation reactions

Article abstract of DOI:10.1016/j.molcata.2012.05.019

In this work, it has been investigated a membrane-like effect of ionic liquid (IL) in the multistep catalytic hydrogenation of trans-cinnamaldehyde (CALD) and 1,5-ciclooctadiene (1,5-COD). A commercial Pd/C catalyst was impregnated with 1-hexyl-3-methyl-imidazolium at 0.5-10 wt%. SEM, EDS, XRD, TG and BET data suggested that the IL penetrates the carbon pore structure coating the catalyst, especially at 5 and 10% concentration. The presence of the IL layer strongly affected the selectivities of the hydrogenations of CALD and 1,5-COD. In the case of CALD, the IL favored the formation of the fully hydrogenated product (hydrocinnamyl alcohol). This result is discussed in terms of a membrane-like effect, in which the small H₂ molecule diffuses more easily from toluene through the IL layer as compared to the larger cinnamaldehyde. This effect results in the production of comparatively higher concentrations of the Pd-H species, which favors the double hydrogenation. On the other hand, the IL layer during the 1,5-COD hydrogenation favored the intermediate product COE (cyclooctene). In this case, a membrane-like effect was also used to discuss the result, where the IL layer has the effect to expel the intermediate COE hindering its contact with the Pd surface.

Full text of DOI:10.1016/j.molcata.2012.05.019

Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Ionic liquid layer on Pd/C catalyst: Membrane-like effect on the selectivity for
multistep hydrogenation reactions
a
b
a
a
a
Evelisy C.O. Nassor , Juliana C. Tristão , Eduardo N. dos Santos , Flávia C.C. Moura , Rochel M. Lago ,
a,∗
Maria Helena Araujo
Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Minas Gerais, Belo Horizonte – MG 31270-901, Brazil
a
b
Universidade Federal de Vi c¸ osa, Campus Florestal, Florestal – MG 35690-000, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 April 2012
Received in revised form 18 May 2012
Accepted 27 May 2012
Available online 2 June 2012
In this work, it has been investigated a membrane-like effect of ionic liquid (IL) in the multistep catalytic
hydrogenation of trans-cinnamaldehyde (CALD) and 1,5-ciclooctadiene (1,5-COD). A commercial Pd/C
catalyst was impregnated with 1-hexyl-3-methyl-imidazolium at 0.5–10 wt%. SEM, EDS, XRD, TG and
BET data suggested that the IL penetrates the carbon pore structure coating the catalyst, especially at 5
and 10% concentration. The presence of the IL layer strongly affected the selectivities of the hydrogena-
tions of CALD and 1,5-COD. In the case of CALD, the IL favored the formation of the fully hydrogenated
product (hydrocinnamyl alcohol). This result is discussed in terms of a membrane-like effect, in which
the small H2 molecule diffuses more easily from toluene through the IL layer as compared to the larger
cinnamaldehyde. This effect results in the production of comparatively higher concentrations of the Pd-H
species, which favors the double hydrogenation. On the other hand, the IL layer during the 1,5-COD hydro-
genation favored the intermediate product COE (cyclooctene). In this case, a membrane-like effect was
also used to discuss the result, where the IL layer has the effect to expel the intermediate COE hindering
its contact with the Pd surface.
Keywords:
SCILL catalysts
Hydrogenation
Ionic liquids
Membrane-like effect
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
catalysts with an ionic liquid layer [19]. This method known as SCILL
has been patented in cooperation with Süd-Chemie [20]. Since
then, several reports involving the SCILL concept have appeared in
the literature, for example, Ni/SiO2 catalyst coated with [BMIM][n-
The use of ionic liquids (ILs) in catalysis has been extensively
investigated in the last years [1–13]. Although commercially avail-
able, ILs are still relatively expensive compared to traditional
solvents [14–17]. Besides, some of them show evidence of low
biodegradability and (eco)toxicological properties [11]. Hence, the
immobilization of ILs onto a support is an attractive strategy to min-
imize the amount of ILs used, maintaining their catalytic properties.
Additionally, supported ILs have the advantage of easy separa-
tion and recyclability, as well as the potential to facilitate the
development of continuous processes [14,15,18]. Therefore, the
concept of immobilized ionic liquids entrapped, for instance, on
the surface and pores of various solid materials offers an attrac-
tive cost-effective alternative. Furthermore, due to the high relative
viscosity of many RTILs (Room Temperature Ionic Liquids), mass
transfer limitations are usually an issue and the established thin
ionic liquid layer minimizes this problem [17].
C8H17OSO ] was used for the hydrogenation of 1,5-COD [19] with
3
an increase in the selectivity for cyclooctene; Ru/Al O coated with
2
3
[C10mim]NTf was investigated for the hydrogenation of limonene
2
and the results have shown an increase in the selectivity for the
intermediate p-menthene [21]. Furthermore, Pd and Pt supported
on SiO and Al O catalysts coated with [BMIM][N(CN) ] have been
2
2
3
2
the subject of several investigations related to H2 adsorption and
interaction of IL with the metals [22–25]. In a recent study, SiO2
and Al O supported Pd catalysts coated with [BMIM][N(CN) ] have
2
3
2
also been investigated for the hydrogenation of citral resulting in
increased selectivity for citronelal [26]. SCILL concept has been also
successfully tested in the consecutive hydrogenation of butadiene
to butene and butane, in a gas phase reaction, it was observed that
the IL strongly favors formation of butane over a Pd-catalyst [27,28].
Although hydrogenations with Pd/C catalysts are well known
and have been the subject of a number of recent reports [29–32],
the strategy of impregnating these catalysts with IL to modulate
selectivity has been little explored.
Improvement of the selectivity of heterogeneous catalysts has
been achieved recently by Jess and coworkers, by covering solid
In this work, we have investigated a potential membrane-like
effect of a thin IL layer to create a special surface environ-
ment, which modifies the selectivity of hydrogenations on a Pd/C
∗ Corresponding author. Tel.: +55 3134097557; fax: +55 3134095700.
E-mail addresses: maria.araujo@pq.cnpq.br, mharaujo1993@yahoo.com.br
M.H. Araujo).
(
1
381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.05.019

E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
75
amounts of the ILs, i.e. 0.5, 1, 2.5, 5 and 10 wt% for HMIm.Br, 1 and
wt% for BMIm.BF . Typically 0.150 g of Pd/C catalyst was added
5
4
to solution of the ionic liquid in 30 mL of methanol and the solvent
was slowly evaporated (ca. 2 h) under magnetic stirring at room
temperature.
2.3. Hydrogenation reactions
The trans-cinnamaldehyde
(Aldrich,
99%)
and
1,5-
cyclooctadiene (Aldrich, 99%) hydrogenations were carried
out in a 100 mL stainless steel bomb (Parr 4560) at 80, 100 or
◦
1
20 C, H2 pressure of 20 bar and mechanic stirring at 300 rpm.
In a typical run, the bomb was loaded with Pd/C-IL (0.017 g),
trans-cinnamaldehyde (0.53 g) or 1,5-cyclooctadiene (0.433 g) and
toluene (60 mL). The reaction was followed by collecting liquid
samples through a dip tube. The samples were analyzed by gas
chromatography using dodecane as internal standard in a Shi-
madzu GC-2010 apparatus equipped with split/splitless injector,
flame ionization detector and a capillary column RTX® – 5MS (5%
phenylmethylsilicone) 30 m × 0.25 mm × 0.25 ␮m.
Scheme 1. General simplified reaction scheme showing possible pathways for the
multistep hydrogenation of A to produce B and C on a Pd/C catalyst coated with IL.
catalyst. The systems studied were composed of two liquid phases,
i.e. toluene and the IL 1-hexyl-3-methyl-imidazolium bromide
(
HMIm.Br), and the probe substrates were trans-cinammaldehyde
and 1,5-cyclooctadiene. For multistep reactions involving the
formation of intermediates (e.g. a consecutive hydrogenation
A → B → C) the kinetics and selectivity might be strongly influenced
by the equilibrium and mass transfer of the species between the
two phases, i.e. toluene and IL (Scheme 1). The diffusion of the
substrate and H2 through the IL layer to reach the Pd surface can
have a significant effect on the reaction selectivity by controlling
the relative concentration of the surface species H-Pd, APd and BPd
2.4. Materials characterization
HMIm.Br was analyzed by 1H and 13C { H} Nuclear Mag-
1
◦
netic Resonance (Bruker CXP 200 MHz, CDCl , 25 C), Infrared
3
Spectroscopy (Perkin Elmer Spectrum GX FT-IR System) in trans-
mittance mode using KBr windows, and Thermogravimetric
Analysis (TGA-60 Shimadzu Instruments) under the following
(
Scheme 1). Also, in a multistep reaction, in which C is a secondary
◦
operational conditions: 5 mg powder sample heated to 800 C
product formed from B, the interaction of B with the IL layer might
also influence the selectivity. If the IL interacts well with the inter-
mediate B, this species probably stays longer near the Pd surface,
favoring formation of the final product C. On the other hand, if the IL
interacts poorly with B, the IL layer tends to expel the intermediate,
increasing its selectivity.
Two important model substrates were used for the hydro-
genations, i.e. trans-cinnamaldehyde and 1,5-cyclooctadiene. These
molecules have been selected due to their different polarities and
interactions with the phases toluene/IL. Cinnamaldehyde (CALD)
and all the hydrogenation products, hydrocinnamaldehyde (HALD)
and hydrocinnamyl alcohol (HALC) are fairly polar molecules and
should interact better with the polar IL compared to toluene. On the
other hand, the hydrogenation products of 1,5-COD, i.e. ciclooctene
−
1
in dynamic N atmosphere (100 mL min ) with heating rate of
1
2
◦
−1
.
0 C min
SCILL catalysts based on Pd/C covered with different amounts
of the ILs HMIm.Br and BMIm.BF4 were analyzed by X-ray diffrac-
◦
−1
tion XRD (Rigaku, Geigerflex) with Co-K␣ radiation from 4 min
◦
from 10 up to 80 . Thermogravimetric analysis (TGA-60 Shi-
madzu Instruments) was carried out under dynamic N atmosphere
2
−
1
◦
−1 ◦
(
100 mL min ) with heating rate 10 C min
up to 800 C. The
morphology of the catalysts was analyzed by scanning electron
microscopy SEM (Jeol JSM 840A). Surface area and porosity mea-
surements (Autosorb 1 Quantachrome) were determined by B.E.T.
method, using 22 cycles of N2 adsorption/desorption using HK and
BJH method for pore volume calculations. Infrared Spectroscopy
(Perkin Elmer Spectrum GX FT-IR System) was performed in trans-
(
COE) and ciclooctane (COA) are much more hydrophobic and
mittance mode using KBr pellets.
should interact better with the solvent toluene.
2
. Experimental
3. Results and discussion
2.1. Synthesis of ionic liquid
3.1. Preparation and characterization of the catalysts
1
-hexyl-3-methyl-imidazolium bromide, HMIm.Br, was synthe-
The coated Pd catalysts were prepared by impregnation with
different amounts of HMIm.Br (1-hexyl-3-methyl-imidazolium
bromide), i.e. 0.5, 1, 2.5, 5 and 10 wt%, on a commercial Pd (5%)/C.
Powder XRD analyses of the prepared IL-coated catalysts
sized by direct alkylation of N-methylimidazol with an excess of
-bromohexane, using acetonitrile as solvent, on a round bottom
flask at 80 C for 48 h under nitrogen atmosphere. The product
was crystallized by adding toluene dropwise in an ice bath until
a solid starts to precipitate. The mixture was let still for 2 h and
the solid was separated by filtration and dried under vacuum
1
◦
0
◦
showed Pd metal (Pd , PDF 1-1201) and a broad peak at ca. 27
related to the carbon support (see Supplementary material). As IL
concentration increased, the intensity of the Pd0 diffraction peaks
gradually decreased due to the formation of a coating on the catalyst
surface.
−
2
◦
(
[
5 × 10 mbar) at 60 C, resulting in an almost pure ionic liquid
33] (see Supplementary material).
Under the employed conditions IR (Supplementary material)
of the pure Pd/C catalyst did not show any significant absorp-
tion. As the IL was impregnated on the catalyst surface, some of
2
.2. Catalysts preparation
−
the HMIm.Br absorptions were observed, e.g. at 3400–3600 cm
1
The catalyst 5% Pd/C (E-101) was obtained from Evonik. The ionic
−
1
liquids used were HMIm.Br (synthesized) and BMIm.BF4 (Aldrich
8.5%, for catalysis). The SCILL catalysts were prepared by sim-
ple impregnation of the commercial Pd/C catalyst with different
hydroxyl groups or water adsorbed, 2880–3160 cm
zolium ring and ␥(CH) aromatic stretches, C C at 1465 cm (low
intensity), C C and C N stretching at 1573 and 1627 cm−1, C
for imida-
−
1
9
H

7
6
E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
Fig. 1. SEM images of Pd/C and Pd/C-IL catalysts.
and N H deformations at 1300–1400 cm−1, C C and C N stretches
The pore filling degree was estimated considering the total
−1
3
−1
in 800–1200 cm
.
pore volume (0.36 cm g ) of the catalyst Pd/C and the IL volume
−
3
The Pd/C-IL catalysts were also characterized by TG analyses
Supplementary material). The Pd/C catalyst did not show any sig-
nificant weight loss up to 800 C under N2 atmosphere. On the
impregnated. Considering that the IL with density 1.23 g cm com-
pletely fills the pore structure of the catalyst, the impregnation of
Pd/C with 1, 5 and 10% should produce a pore filling degree of ca.
3, 13 and 25%, respectively.
(
◦
other hand, all Pd/C-IL materials showed a weight loss between
◦
2
20 and 350 C related to the thermal decomposition of HMIm.Br.
Based on these weight losses, IL contents of ca. 1, 5 and 10 wt% were
estimated for the different materials.
3.2. Hydrogenation reactions
SEM images were obtained for the commercial Pd/C catalyst and
for the Pd/C-IL (Fig. 1). The commercial Pd/C catalyst showed an
irregular surface with visible pore structure. Upon impregnation of
only 1% of IL the surface becomes slightly smoother and some of
the very small pores seemed to disappear. The addition of IL at 5
and 10% completely coats the porous surface in some areas.
BET measurements for the catalyst Pd/C showed a surface area
The selective hydrogenation of cinnamaldehyde (CALD) has a
great scientific and technological interest [34–37]. The addition of
one molecule of hydrogen may lead to cinnamyl alcohol (CALC)
or hydrocinnamaldehyde (HALD) (Scheme 2). Both products of the
primary hydrogenation can be further hydrogenated to yield 3-
phenyl-1-propanol (hydrocinnamyl alcohol, HALC). Nevertheless,
under the examined reaction conditions, CALC was not observed
and the only two products were HALD and HALC, as observed in pre-
vious work for palladium catalysts [29–32]. The obtained reaction
profiles for the different catalysts can be seen in Fig. 2.
2
−1
3
−1
of 717 m g with pore volume of ca. 0.36 cm g . As expected,
the impregnation of IL at 1 and 10% reduced the BET surface area to
2
−1
5
83 and 425 m g , respectively. However, pore volume measure-
ments from the N adsorption for the catalysts Pd/C-IL did not show
It can be observed that neat Pd/C converts near 100% of
the cinnamaldehyde in only 15 min producing mainly hydrocin-
namaldehyde (HALD) with ca. 60% yield. After 30 min the reaction
apparently approaches the end as no significant change on prod-
uct distribution takes place. This result is related to the difficulty of
Pd catalyst to hydrogenate C O bond of the intermediate HALD
molecule as reported in theoretical and experimental previous
work [29–32,38,39]. Our data, as well as previous ones [39,40],
suggest that the product HALC is formed directly by a consecu-
tive double addition of hydrogen to the CALD adsorbed over the Pd
surface.
2
consistent results. For example, the impregnation of 1% IL decreased
3
−1
the pore volume from 0.36 to 0.30 cm g . Unexpectedly, when the
Pd/C was impregnated with 10% IL the pore volume was greater
3
−1
than 0.30, i.e. 0.34 cm g . During the BET measurements of the
samples Pd/C-IL, especially 10% IL, it was necessary very long times
to reach N adsorption equilibrium. This anomalous effect could be
2
related to the dissolution of N2 into IL layer and diffusion into the
pores. Therefore, the amounts of N during adsorption and desorp-
2
tion cannot be used consistently to obtain representative surface
area and pore structure of the catalysts Pd/C-IL.

E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
77
Scheme 2. trans-Cinnamaldehyde (CALD) hydrogenation routes.
These results can be rationalized in terms of a reaction scheme
depicted in Eqs. (1)–(7). Cinnamaldehyde (CALD) and H2 are
absorbed on the Pd surface (Eqs. (1) and (2)). The activated hydro-
gen is transferred to the C C bond and HALD is desorbed (Eq. (3)).
The results shown in Fig. 2 suggest that this desorbed HALD from
the Pd catalyst is not significantly converted into HALC(soln). There-
fore, under the reaction conditions employed over pure Pd/C Eqs.
ZnCl2 [44,45] favors the hydrogenation of carbonyl groups of ␣,␤-
unsaturated aldehydes, e.g. citral and CALD, on Pd catalysts. This
promoting effect is explained in terms of an interaction of the car-
bonyl group with the Lewis acid center (A+
O C), which facilitates
the hydride transference from the Pd surface to the carbon of the
carbonyl group. Based on these results one can suggest that the
IL can interact with the C O of the HALD surface intermediate
species and promote its hydrogenation into HALC by Pd. In fact,
the acid Lewis properties of IL have been reported in some works
(
4) and (5) do not occur significantly and Eqs. (6) and (7) are the
main path for the production of HALC.
CALD(soln) → CALD(IL) → CALD∗(Pd)
H2(soln) → H2(IL) → H2∗(Pd) → H∗(Pd)
CALD∗(Pd) + H2∗(Pd) → HALD(soln)
HALD(soln) → HALD∗(Pd)
HALD∗(Pd) + H∗(Pd) → HALC(soln)
CALD∗(Pd) + H∗(Pd) → HALD∗(Pd)
HALD∗(Pd) + H∗(Pd) → HALC(soln)
[
46,47]. For example, the ILs derivate from dialkylimidazolium bro-
(1)
(2)
(3)
(4)
(5)
(6)
(7)
mides and trifluoroacetates showed good activity as catalyst for
the Diels–Alder reactions indicating that this kind of dialkylimi-
dazolium salts can act as Lewis acids [48]. Also, the ionic liquid
is a medium with a high dielectric constant, which enhances the
polarization on the carbon–oxygen bond and helps to promote the
hydride attack to the carbon of the CO double bond by an effect
analogous to the Lewis acid.
Therefore, a possible role of the IL layer is to function as an active
membrane where the more facile diffusion of H2 and a potential
effect of the IL as Lewis acid or dielectric medium will favor the
total hydrogenation of cinnamaldehyde.
The catalyst coated with small amounts of IL, i.e. 0.5 Pd/C-IL,
showed similar results to the uncoated one. The catalysts contain-
ing 5 or 10% of IL (i.e. Pd/C-5IL or Pd/C-10IL) led to lower conversion
rates as it can be observed by the smoother decrease in CALD
curves (see also Table 1 in Supplementary material). These lower
reaction rates are likely related to the lower solubility of H2 and
trans-cinnamaldehyde in the IL layer, which reduces the effective
concentration of the reactants at the surface of the catalyst. How-
ever, it is remarkable the complete change in product distribution
caused by the presence of IL layer, i.e. the selectivity/yield for HALC
is much higher in the presence of IL. These results clearly show
that the presence of the IL favors the total hydrogenation product
HALC and this effect is more pronounced for higher amounts of IL.
Although, the origin of the effect of IL on the reaction selectivity
observed in this work is not clear and more detailed investigations
are necessary, one can speculate some of the possible effects of
the IL on the reaction. The diffusion limitations caused by the IL
layer is possibly more important for the larger molecule of cin-
namaldehyde (Eq. (1)) compared to the small H2 molecules (Eq.
The Pd/C catalyst with and without IL was also employed
in the multistep hydrogenation of 1,5-cyclooctadiene (1,5-COD).
Selective hydrogenation of 1,5-COD to monoolefin COE is of great
interest especially in the industrial synthesis of polymers and
olefins [49]. As depicted in Scheme 3, this reaction may result in the
formation of four main products: 1,4-cyclooctadiene (1,4-COD) and
1,3-cyclooctadiene (1,3-COD) due to isomerization, cyclooctene
(COE) due to the selective hydrogenation and cyclooctane (COA)
due to the exhaustive hydrogenation.
As it can be observed in Fig. 3 for the uncoated catalyst (Pd/C),
near 100% of the 1,5-COD is converted in only 15 min producing
approximately 90% of the intermediate product COE. It can also be
observed the 1,5-COD isomerization to produce very small amounts
of 1,4-COD and 1,3-COD, which are consumed by hydrogenation
to COE. Further hydrogenation of COE rapidly takes place, mainly
after the dienes (1,5-, 1,4- or 1,3-COD) concentrations are reduced.
Similar results are described in the literature for the hydrogenation
of 1,5-COD in the presence of different Pd catalysts [50,51]. The
reason for this great selectivity for COE is that dienes (e.g. COD) bind
much more strongly to metals than the corresponding monoenes.
Thus, the catalytic sites on the metal surface at the beginning of
the reaction are saturated by dienes. As hydrogen is added to the
first C C double-bond, the resulting monoene (COE) is immediately
displaced by a diene [52].
(
2)) [21]. As result, the relative concentration of H2 on the Pd sur-
face is higher, also increasing the Pd-H concentration. The presence
of higher concentrations of Pd-H would favor the HALC according
to Eq. (7).
Also, the presence of acidic species which interact with the car-
bonyl group of the aldehydes is known to favor its hydrogenation.
For example, several works suggested that the presence of Lewis
acid, such as FeCl2 [41], SnCl , FeCl , MnCl , or CoCl [42,43] and
Only when the diene concentration is comparatively low, the
monoene in higher concentration competes for the catalyst sites. A
2
3
2
2

7
8
E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
1
00
Pd/C
100
80
1,5-COD
COA
8
6
4
2
0
0
0
0
0
HALD
HALC
60
Pd/C
COE
4
2
0
0
1
,3/1,4-COD
CALD
100
0
0
50
100 150 200 250 300
COA
0
20
40
60
80
120
1
00
1,5-COD
1
00
Pd/C-0.5IL
HALD
8
0
0
0
0
8
6
4
2
0
0
0
0
0
6
4
2
Pd/C-0.5IL
HALC
COE
1,3/1,4-COD
CALD
0
0
50
100 150 200 250 300
COE
0
20
40
60
80
100
120
1
,5-COD
1
00
1
00
Pd/C-5IL
HALC
8
0
0
0
0
8
6
4
2
0
0
0
0
0
6
4
2
Pd/C-2.5IL
HALD
1,3/1,4-COD
COA
CALD
80
0
0
50
100
150
200
250
COE
300
0
20
40
60
100
120
1
00
1
,5-COD
100
Pd/C-10IL
HALC
8
0
0
0
8
6
4
2
0
0
0
0
0
6
4
Pd/C-5IL
1,3/1,4-COD
COA
HALD
20
0
CALD
0
50
100 150 200 250 300
0
20
40
60
80
100
120
1
00
Pd/C-10IL
Time / min
1,5-COD
8
0
0
0
0
Fig. 2. Molar distribution (%) of reagent and products as a function of time in CALD
hydrogenation (100 C, H2: 20 bar, HMIm.Br).
COE
◦
6
4
2
1
,5-COD
1,4-COD
1,3-COD
1,3/1,4-COD
H2
0
H2
COA
500
H2
0
100
200
300
400
600
Time / min
H2
◦
Fig. 3. Conversion of 1,5-cyclooctadiene and selectivities (80 C, H2: 20 bar,
COE
COA
HMIm.Br).
Scheme 3. Reaction products during catalytic 1,5-COD hydrogenation.

E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
79
coefficients should be proportional to the tabled octane/water par-
tition coefficient (Kow). The Kow for 1,5-COD and COE are 3.2 and.
3
.9, respectively. These values suggest that both substrates, but spe-
cially COE, should have low solubility in polar media such as the
−
1
IL. Indeed, the water solubility for 1,5-COD (64.1 mg L ) is greater
−
1
than COE’s one (13.5 mg L ). Therefore, once formed and desorbed
to the IL, the COE should be expelled to toluene. The diffusion back
into the IL layer for further hydrogenation is an unfavorable process.
These steps are depicted in Scheme 4.
Therefore, in the case of the hydrogenation of 1,5-COD the IL
layer has a membrane like effect to expel the intermediate COE and
avoid its return to the catalyst surface after the first hydrogenation.
4
. Conclusions
Scheme 4. Schematic representation of the possible processes taking place on the
IL coated Pd catalyst surface (bold lines are supposed to be the more important
pathways).
The IL HMIm.Br (1-hexyl-3-methylimidazolium bromide) coat-
ing on a commercial Pd/C catalyst seems to show a membrane-like
effect during the hydrogenation of trans-cinnamaldehyde (CALD)
and 1,5-cyclooctadiene (1,5-COD).
similar behavior is observed for Pd/C and Pd/C-0.5IL. On the other
hand, the behavior completely changes for IL contents higher than
2
.5%, i.e. Pd/C-2.5IL – 5IL and 10IL. One clear effect observed when
The small H2 molecule diffuses more easily from toluene
through the IL layer than the larger cinnamaldehyde molecule pro-
ducing higher concentrations of the Pd-H species, which leads to
the total hydrogenated product hydrocinnamyl alcohol. The role of
the IL as a weak Lewis acid or medium with high dielectric con-
stant is also considered as a possible explanation for the activation
of the carbonyl group. In contrast with the cinnamaldehyde hydro-
genation, the effect of the IL layer for the hydrogenation of 1,5-COD
was to favor the selective hydrogenation to form the intermediate
product COE (cyclooctene). In this case, the IL layer has the effect of
expel the intermediate COE avoiding its contact with the Pd surface.
This membrane-like effect observed for the IL having an apolar
solvent as a counterpart can potentially be exploited to change the
selectivity of many multi-step hydrogenation processes.
IL was added to the Pd/C is a gradual decrease on the reaction rate,
which is expected due to a diffusion barrier on the catalyst surface.
This effect is very pronounced for the catalyst Pd/C-10IL.
It is also interesting to observe for the catalysts Pd/C-2.5IL and
IL a good 1,5-COD conversion but fail to convert COE to COA. As a
5
consequence, the COE concentration remains high, which is desir-
able from the point of view of the selective hydrogenation of COD
to COE. The results displayed in Fig. 3 suggest that the Pd/C cata-
lyst is very efficient to convert COE in COA. On the other hand, in
the presence of 2.5IL the hydrogenation of COE is very low even at
low diene concentration. This behavior seems to be repeated for
Pd/C-5IL and Pd/C-10IL, but as these catalysts are slower, the 1,5-
COD conversion did not reached the maximum value during the
observed time span. These results suggest a strong effect of the IL
layer on the hydrogenation selectivity.
Acknowledgments
Besides the chemical steps, the hydrogenation of 1,5-COD
should involve a complex equilibrium with several phase trans-
fer steps. 1,5-COD should diffuse from toluene to the IL phase and
adsorb on Pd surface (Eq. (8)). This step is probably critical, since the
apolar 1,5-COD molecule is very soluble in toluene but has limited
solubility in the more polar IL HMIm.Br. In fact, as the IL content
increases, the 1,5-COD conversion rate gradually decreases, likely
due to this diffusion/solubility limitations into the IL. Once in the IL,
The authors acknowledge the financial support from FAPEMIG,
PRPq/UFMG, INCT-Catálise, CNPq and CAPES.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
molcata.2012.05.019.
1
,5-COD adsorbs on Pd and is hydrogenated to form the intermedi-
ate COE on the Pd surface (Eq. (9)). The adsorbed intermediate (COE)
can desorb from Pd to the IL phase (Eq. (10)) and then follow two
different paths: (i) diffusion to toluene (Eq. (11)) or (ii) adsorption
back onto Pd for hydrogenation to produce COA (Eq. (12)).
References
[
1] D.-Q. Xu, Z.-Y. Hu, W.-W. Li, S.-P. Luo, Z.-Y. Xu, J. Mol. Cat. A: Chem. 235 (2005)
137–142.
CODtol → CODIL → (CODIL)ads-Pd
(8)
[
[
2] T. Welton, Chem. Rev. 99 (1999) 2071–2083.
3] R. Sheldon, Chem. Commun. 23 (2001) 2399–2407.
(
(
CODIL)ads-Pd + H → (COEIL)ads-Pd
(9)
(10)
(11)
[4] D. Zhao, M. Wu, Y. Kou, E. Min, Catal. Today 74 (2002) 157–189.
2
[
[
[
[
[
5] J. Dupont, R.F. de Souza, P.A.Z. Suarez, Chem. Rev. 102 (2002) 3667–3692.
6] C.E. Song, Chem. Commun. 9 (2004) 1033–1043.
7] D. Zhao, M. Wu, Y. Kou, E. Min, Catal. Today 1–33 (2002) 2654.
8] Y. Gua, G. Lia, Adv. Synth. Catal. 351 (2009) 817–847.
9] H. Olivier-Bourbigou, L. Magna, D. Morvan, Appl. Catal. A 373 (2010) 1–56.
COEIL)ads-Pd → COEIL
COEIL → COEtol
[
[
10] M. Jakuttis, A. Schnweiz, S. Werner, R. Franke, K.-D. Wiese, M. Haumann, P.
Wasserscheid, Angew. Chem. Int. Ed. 50 (2011) 4492–4495.
11] H. Hagiwara, T. Nakamura, T. Hoshi, T. Suzuki, Green Chem. 13 (2011)
COEIL → (COEIL)ads-Pd + H → (COAIL)ads-Pd → COAIL → COAtol
2
1133–1137.
[
[
[
12] M. Haumann, A. Riisager, Chem. Rev. 108 (2008) 1474–1497.
13] T. Welton, Coord. Chem. Rev. 248 (2004) 2459–3247.
14] M.I. Burguete, H. Erythropel, E. Garcia-Verdugo, S.V. Luis, V. Sans, Green Chem.
(12)
Different from the reaction of cinnamaldehyde, where the IL
favors the final hydrogenation product, for the 1,5-COD hydrogena-
tion the IL favors the formation of the intermediate hydrogenation
product, i.e. COE. This result can be explained in terms of the parti-
tion coefficients among the toluene and IL phases. These partition
1
0 (2008) 401–407.
[15] B. Gadenne, P. Hesemann, J.J.E. Moreau, Chem. Commun. 15 (2004) 1768–1769.
[
16] J-P. Mikkola, P. Virtanen, H. Karhu, T. Salmi, D.Y. Murzin, Green Chem. 8 (2006)
97–205.
1
[
17] J-P.T. Mikkola, P.P. Virtanen, K. Kordás, H. Karhu, T.O. Salmi, Appl. Catal. A 328
(2007) 68–76.

8
0
E.C.O. Nassor et al. / Journal of Molecular Catalysis A: Chemical 363–364 (2012) 74–80
[
[
[
[
[
[
18] Y. Shen, Y. Zhang, Q. Zhang, L. Niu, T. You, A. Ivaskab, Chem. Commun. 33 (2005)
193–4195.
19] U. Kernchen, B. Etzold, W. Korth, A. Jess, Chem. Eng. Technol. 30 (8) (2007)
85–994.
20] A. Jess, W. Korth, B. Etzold (Patent of Süd Chemie), WO 2007/124896 A1 (DE 10
006 019 460 A1, 31.10.2007).
21] E. Bogel-Łukasika, S. Santos, R. Bogel-Łukasika, M.N. da Ponte, J. Supercrit. Fluids
4 (2010) 210–217.
22] M. Sobota, M. Happel, M. Amende, N. Paape, P. Wasserscheid, M. Laurin, J.
Libuda, Adv. Mater. 23 (2011) 2617–2621.
23] M. Sobota, M. Schmid, M. Happel, M. Amende, F. Maier, H-P. Steinrück, N. Paape,
P. Wasserscheid, M. Laurin, J.M. Gottfried, J. Libuda, Phys. Chem. Chem. Phys.
[34] F. Benvenuti, C. Carlini, M. Marchionna, A.M.R. Galletti, G. Sbrana, J. Mol. Catal.
A: Chem. 145 (1999) 221–228.
[35] F. López-Linares, G. Agrifoglio, A. Labrador, A. Karam, J. Mol. Catal. A: Chem. 207
(2004) 117–122.
[36] Y. Kume, K. Qiao, D. Tomida, C. Yokoyama, Catal. Commun. 9 (2008) 369–
375.
[37] Y. Zhang, S. Liao, Y. Xu, D. Yu, Appl. Catal. A 192 (2000) 247–251.
[38] M. Jahjah, B. Caron, S. Menuel, E. Monflier, L. Djakovitcha, C. Pinel, Arkivoc vii
(2011) 406–415.
[39] F. Delbecq, P. Sautet, J. Catal. 152 (1995) 217–236.
[40] S. Morrissey, I. Beadham, N. Gathergood, Green Chem. 11 (2009) 466–474.
[41] M.A. Aramendá, V. Borau, C. Jiménez, J.M. Marinas, A. Porras, F.J. Urbano, J. Catal.
172 (1997) 46–54.
4
9
2
5
1
2 (2010) 10610–10621.
[
24] H.-P. Steinrück, J. Libuda, P. Wasserscheid, T. Cremer, C. Kolbeck, M. Laurin, F.
Maier, M. Sobota, P.S. Schulz, M. Stark, Adv. Mater. 23 (2011) 2571–2587.
25] J. Arras, E. Paki, C. Roth, J. Radnik, M. Lucas, P. Claus, J. Phys. Chem. C 114 (2010)
[42] U. Schröder, L. De Verdier, J. Catal. 142 (1993) 490–498.
[43] Y. Nagase, H. Hattori, K. Tanabe, Chem. Lett. 12 (10) (1983) 1615–1618.
[44] Y. Chen, Y. Zu, Y. Fu, X. Zhang, P. Yu, G. Sun, T. Efferth, Molecules 15 (2010)
9486–9495.
[
1
0520–10526.
[26] N. Wörz, J. Arras, P. Claus, Appl. Catal. A 391 (2011) 319–324.
[27] A. Jess, C. Kern, W. Korth, Erdoel-Erdgas-Kohle 128 (2012), OG 38-45.
[28] A. Jess, C. Kern, W. Korth, Oil Gas Europ. Mag. 38 (2012) 38–45.
[29] A. Cabiac, T. Cacciaguerra, P. Trens, R. Durand, G. Delahay, A. Medevielle, D. Plée,
B. Coq, Appl. Catal. A 340 (2008) 229–235.
[45] Z. Duan, Y. Gu, Y. Deng, Catal. Commun. 7 (2006) 651–656.
[46] P. Wasserscheid, W. Keim, Angew. Chem. Int. Ed. 39 (2000) 3772–3789.
[47] D. Li, F. Shi, J. Peng, S. Guo, Y. Deng, J. Org. Chem. 69 (2004) 3582–3585.
[48] J. Howwarth, K. Hanlon, D. Fayne, P. McCormac, Tetrahedron Lett. 38 (17) (1997)
3097–3100.
[
[
[
[
30] A. Cabiac, G. Delahay, R. Durand, P. Trens, B. Coq, D. Plée, Carbon 45 (2007)
[49] F.C.C. Moura, R.M. Lago, E.N. dos Santos, M.H. Araujo, Catal. Commun. 3 (2002)
541–545.
[50] M. Schwarze, J. Keilitz, S. Nowag, R.Y. Parapat, R. Haag, R. Schomäcker, Langmuir
27 (2011) 6511–6651.
[51] A. Schmidt, R. Schomäcker, Ind. Eng. Chem. Res. 46 (2007) 1677–1681.
[52] F.C.C. Moura, E.N. dos Santos, R.M. Lago, M.D. Vargas, M.H. Araujo, J. Mol. Catal.
A: Chem. 226 (2005) 243–251.
3
–10.
31] F. Zhao, Y. Ikushima, M. Chatterjee, M. Shirai, M. Arai, Green Chem. 5 (2003)
6–79.
7
32] L. Zhang, J.M. Winterbottom, A.P. Boyes, S. Raymahasay, J. Chem. Technol.
Biotechnol. 72 (1998) 264–272.
33] S.O. Nwosu, J.C. Schleicher, A.M. Scurto, J. Supercrit. Fluids 51 (2009)
1
–9.