Pt- and Pd-catalysed limonene hydrogenation in high-density carbon dioxide

Article abstract of DOI:10.1007/s00706-009-0196-5

This paper discusses the development of catalytic (Pt, Pd) hydrogenation in high-pressure CO₂ to convert limonene into valuable chemicals. It was noticed that varying the catalyst results in product regioselectivity. Platinum as catalyst favours the formation of p-menthane stereoisomers in equimolar quantities whereas palladium as catalyst furnishes trans-p-menthane and cis-p-menthane in 2:1 ratio. These results are in agreement with hydrogenation mechanisms for Pt- and Pd-catalysed reactions. Platinum is a more active catalyst than palladium, but higher activity results in lower chemical stability of the catalyst. Palladium as catalyst usually catalyses limonene isomerisation in the first stage of the process. Pressure tuning of the processes affects termination of the reaction. The flow rate of the reaction mixture through the stationary catalyst bed affects the composition of products; partially hydrogenated limonene is obtained at low flow rates and completely hydrogenated products at high flow rates.

Full text of DOI:10.1007/s00706-009-0196-5

Monatsh Chem (2009) 140:1361–1369
DOI 10.1007/s00706-009-0196-5
ORIGINAL PAPER
Pt- and Pd-catalysed limonene hydrogenation in high-density
carbon dioxide
Ewa Bogel-Łukasik
Manuel Nunes da Ponte
•
Rafał Bogel-Łukasik
•
Received: 25 June 2009 / Accepted: 5 October 2009 / Published online: 29 October 2009
Ó Springer-Verlag 2009
Abstract This paper discusses the development of cata-
lytic (Pt, Pd) hydrogenation in high-pressure CO2 to
convert limonene into valuable chemicals. It was noticed
that varying the catalyst results in product regioselectivity.
Platinum as catalyst favours the formation of p-menthane
stereoisomers in equimolar quantities whereas palladium as
catalyst furnishes trans-p-menthane and cis-p-menthane
in 2:1 ratio. These results are in agreement with hydro-
genation mechanisms for Pt- and Pd-catalysed reactions.
Platinum is a more active catalyst than palladium, but
higher activity results in lower chemical stability of the
catalyst. Palladium as catalyst usually catalyses limonene
isomerisation in the first stage of the process. Pressure
tuning of the processes affects termination of the reaction.
The flow rate of the reaction mixture through the stationary
catalyst bed affects the composition of products; partially
hydrogenated limonene is obtained at low flow rates and
completely hydrogenated products at high flow rates.
phase greatly benefit from enhanced mass and heat transfer
[2], better product selectivity [3, 4], and potential in-situ
catalyst regeneration, thus improving catalyst lifetime
[4, 5]. Also, facile product recovery can be achieved by
simple depressurisation of the reaction mixture and final
products at any point in the reaction process can be easily
obtained at low cost by appropriately reducing the system
pressure. Carbon dioxide is by far the most examined
supercritical fluid, mostly because of its low price, its
advantages as a ‘‘generally regarded as safe (GRAS)’’
solvent, and its convenient critical temperature, which
allows it to be used as a supercritical fluid just above room
temperature. The advantages of supercritical fluid appli-
cations in various reactions including hydrogenation have
already been reported in the literature [6–9].
Hydrogenation of terpenes was reported for the first time
in 1912, by Ipat’ev [10]. Nowadays this hydrogenation
process is widely carried out for industrial purposes [11].
In the hydrogenation reaction the mass-transfer limitation
imposed by the low solubility of hydrogen in the liquid
terpene can be overcome by using high-pressure carbon
dioxide, either as supercritical solvent or as a ‘‘liquid
Keywords High-pressure chemistry Á
Supercritical fluids Á Terpene Á Catalysis
expander’’. In general, supercritical CO is highly soluble
2
Introduction
in most terpenes at pressures close to the critical pressure.
The liquid phase in a biphasic mixture can contain
In recent years, application of supercritical-fluid technol-
ogy in multiphase catalytic reactors has become a potential
means of overcoming major drawbacks of conventional
processes [1]. Reactions occurring in a supercritical fluid
80 mol% or more of CO , as indicated by phase equilib-
2
rium data for terpenes ? CO [12–15]. In fact, it is often
2
found that, even when high-pressure CO does not dissolve
2
an organic liquid well, it can itself dissolve in large
amounts in the liquid phase, causing it to expand sub-
stantially in volume [6]. In that case high-density CO acts
2
E. Bogel-Łukasik (&) Á R. Bogel-Łukasik Á M. N. da Ponte
REQUIMTE, Departamento de Qu ´ı mica,
Faculdade de Ci eˆ ncias e Tecnologia,
Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
e-mail: ewa@dq.fct.unl.pt
as a ‘‘liquid expander’’.
Under biphasic conditions, enhancement of the solubility
of hydrogen in the liquid is achieved by the high quantities
of carbon dioxide [7] dissolved in the ‘‘expanded fluid’’.
123

1
362
E. Bogel-Łukasik et al.
For this reason the requirement of single supercritical phase
formation is avoided [16]. Indeed, it results in minimisation
of limitations of mass transport toward the catalyst bed.
The hydrogenation of terpenes (mainly citral [17–19]
Furthermore, the effect of flow rate of a biphasic reaction
mixture on limonene hydrogenation catalysed by Pd in high-
pressure CO was examined [24]. The idea of the investi-
2
gations presented in this work relies on reuse of the catalyst
to check its ability to perform the next reaction until it loses
the capability to catalyse the hydrogenation of limonene.
The hydrogenation reactions were performed in one (fluid)
or two phases (gas ? liquid) using platinum or palladium
catalysts at a fixed hydrogen pressure of 4 MPa. Carbon
dioxide employed in the catalysed processes is responsible
for partial poisoning of platinum catalyst because of for-
mation of CO via the reverse water gas shift reaction [25];
this was also observed in the reactions we performed.
and pinene [20, 21]) in supercritical CO2 (scCO ) has
2
received much attention in recent years. Burgener et al. [17]
noticed that the kinetics of the conversion of citral strongly
depend on the phase behaviour of the reaction mixture.
Chatterjee et al. [18] attributed the lower conversion for
systems containing gas and liquid phases to the gas–liquid
mass-transfer limitation. Liu et al. showed that phase
behaviour affects reactivity and product distribution in
compressed CO . They also noticed maximum total con-
2
version at the pressure at which the reaction system changes
from the two-phase to the single-phase state [19]. Milewska
et al. [20] reported that hydrogenation under biphasic con-
ditions could sometimes be faster than under monophasic
conditions, because the concentration of substrate (as
Results and discussion
Hydrogenation of limonene at high-pressure in carbon
dioxide proceeds with good yields of very distinct relative
concentrations of reactants, either under biphasic or in
supercritical conditions. Scheme 1 presents the pathways
of the reactions analysed in this work.
opposed to H ) is lower under monophasic conditions. It
2
was confirmed that fast hydrogenation of pinene occurs at
low CO2 pressures, but the difference in reaction rate
between biphasic and supercritical systems was less with
platinum than with palladium as catalyst [21].
Hydrogenation of the terpene was performed using the
precious metals as catalysts. To use the catalyst efficiently
in the developed processes [26] the catalyst was exploited
permanently until it lost its ability to hydrogenate limo-
nene. The hydrogenation of limonene was carried out at a
constant pressure of hydrogen (4 MPa), with a range of
different pressures of carbon dioxide: 8.5 or 7 MPa for two
phases, and 12.0 MPa for monophase. It should be noted
To follow the multi-phase study of hydrogenation of
terpenes in the presence of CO , hydrogenation of limonene
2
(
as monoterpene, to our knowledge never reduced under
supercritical conditions) was regarded as an attractive topic
for investigation [22, 23]. A short communication on
selective hydrogenation of limonene in high-pressure car-
bon dioxide catalysed by Pt driven by phase equilibrium has
already been published [22]. Later research focussed on the
hydrogen pressure-dependence of the product distribution.
This work was performed with palladium catalyst under
that these CO pressures are obtained by subtraction of the
2
initial pressure of hydrogen from the final total pressure in
the cell. They cannot indeed be taken as partial pressures,
because the mixtures behave in a highly non-ideal way,
3
-1
biphasic conditions at 3.3 cm min
flow rate [23].
Scheme 1
1
23

Pt- and Pd-catalysed limonene hydrogenation
1363
especially when they are supercritical fluids. Also, as we
proved earlier, changes of carbon dioxide pressure (from
The results presented in this work start with the reaction
carried out under biphasic conditions at higher flow rate
3
-1
8
.5 to 12 MPa) determine the number of phases, and
(7.3 cm min ) with fresh catalyst. Under these conditions,
p-menthane isomers in equimolar quantities are the final
products and the intermediate (p-menth-1-ene) disappears
after 80 min of the reaction (Fig. 2). This contrast with the
selectivity is controlled by the phase behaviour [22].
The hydrogenation was performed in an apparatus
which allows tuning of the flow rate of the reaction mixture
through the stationary catalyst bed. The low flow rate gave
the possibility of simulating processes carried out in a
semi-continuous mode, because the low flow rate of the
mixture disturbed the feed that occurred in the view cell
minimally. On the other hand at high flow rates of the
reaction mixture through the stationary catalyst bed, the
feed composition changed rapidly, and that process can be
regarded as a semi-batch reaction.
3
-1
reaction carried out at 1.3 cm min , where partially
hydrogenated limonene was the final product (Fig. 1a). This
3
-1
indicates that at low flow rate (1.3 cm min ) the reaction
is executed in the semi-continuous mode whereas at
3
-1
7.3 cm min the feed composition changes rapidly as in
the batch-like reaction.
These results obtained can be explained on the basis of
the literature reports of Horiuti and Polanyi [27] and
Augustine et al. [28], who suggested the mechanism for Pt-
catalysed hydrogenation of cycloolefins without the pres-
Reactions with platinum catalyst
ence of scCO . At lower hydrogen availability, both
2
Distribution of products
adsorption of terpene and half-hydrogenated state
formation stages are reversible. In that case the product-
determining step is the hydrogenation of the half-hydro-
genated state, which, because the thermodynamics of the
process, leads to formation of partially hydrogenated
product. In contrast, at high hydrogen availability in the
vicinity of the catalyst, the adsorption of limonene on
catalyst surface is the product-determining step. This leads
to formation of the p-menthane stereoisomers in 1:1 ratio,
driven by the different steric constraints on contact of each
face of the double bond with the catalyst surface. The data
presented here are in substantial agreement with results
from the hydrogenation of 4-tert-butylmethylcyclohexene
The product distribution from the hydrogenation of limo-
nene catalysed by Pt depends on the flow rate of the reaction
mixture (gas ? liquid) through the stationary catalyst bed.
The Pt-catalysed hydrogenation of limonene at minimal
3
-1
1
.3 cm min
flow rate has already been presented,
although the conclusions are recalled to provide a back-
ground for further analysis of the results obtained [22].
Initially, the reaction was executed in a biphasic system
with fresh catalyst, which was next reused under mono-
phasic conditions. Under biphasic conditions (Fig. 1a), the
reaction equilibrium was reached after 1 h and a dominant
product was p-menth-1-ene, the intermediate, in which the
external double bond was hydrogenated. In single super-
critical phase (Fig. 1b) with reused catalyst, the reaction
took only 10 min and the trans and cis isomers of p-men-
thane in 1:1 ratio were the final products [22].
occurring via the 1,2-r entity explained by the Horiuti–
2
Polanyi mechanism [27].
Tuning two phases into one phase at high flow rate with
the reused catalyst opens a possibility of achieving high
Fig. 1 Mol percentages of trans-p-menthane (filled circles), cis-p-
menthane (open circles), p-menth-1-ene (filled squares), and limo-
nene (open squares). Products from hydrogenation of limonene on 1
a a biphasic (liquid ? gas) system at 12.5 MPa total pressure and b
under single supercritical conditions at 16 MPa total pressure, as a
function of time
3
-1
wt% Pt at minimal 1.3 cm min flow rate, 50 °C, and 4 MPa H
2
in
123

1
364
E. Bogel-Łukasik et al.
rate presented in Fig. 1. The hydrogenation was carried
out with platinum as catalyst, which has aptitude for
poisoning [25]. The poisoning of the platinum catalyst
arises mostly because of the carbon monoxide formed in
the course of the water gas shift reaction [25]. It has
already been stated that pressure affects catalyst behav-
iour and that the biphasic conditions may protect Pt
catalyst from poisoning by CO [17, 25, 29, 30]. The Pt
catalyst reused in a single-phase reaction was no longer
active in the next reaction (limonene was not hydroge-
nated). This indicates that the reaction performed in the
single phase promotes poisoning of the catalyst compared
with biphasic conditions. Scanning electron microscopy
(
SEM) images of the catalyst particles illustrate the
Fig. 2 Mol percentages of trans-p-menthane (filled circles), cis-p-
menthane (open circles), p-menth-1-ene (filled squares), and limo-
nene (open squares). Products from hydrogenation of limonene on 1
progressive leaching of the metal. SEM analysis of the
catalyst particles (Fig. 4) confirmed our suggestions on
this subject. Additional support was also given by anal-
ysis of the reaction solution after reaction in single
supercritical phase; this showed that the solution con-
tained traces of platinum.
3
-1
wt% Pt in a biphasic (liquid ? gas) system at 50 °C, 7.3 cm min
MPa H , and 12.5 MPa total pressure, as a function of time
,
4
2
Hydrogenation of limonene in the single phase is
probably faster than formation of CO, which poisons the
catalyst, because the catalyst still promotes reaction with-
out suspending the process during the 10 min when
reaction occurred (Fig. 1b).
Reactions with palladium catalyst
The 1 wt% palladium catalyst was used to hydrogenate
limonene in the following sequence of reactions: biphasic
(12.5 MPa), monophasic (16 MPa), biphasic (12.5 MPa),
and, as the last one, under biphasic (11 MPa) ‘‘at 1.5 MPa
lower total pressure’’ conditions, all at fixed minimal flow
3
-1
Fig. 3 Mol percentages of trans-p-menthane (filled circles), cis-p-
menthane (open circles), p-menth-1-ene (filled squares), and limo-
nene (open squares). Products of the hydrogenation of limonene on 1
rate (1.3 cm min ). Fresh palladium catalyst was used in
the first reaction, and was then reused in the others. The
catalyst employed in four consecutive reactions was still
active. Palladium is not leached from the catalyst particles
after their use in reactions under low or under high CO2
pressure. Leaching of palladium was also not observed in
SEM images of the catalyst particles (Fig. 5) taken during
the sequence of reactions. The lack of leaching was also
confirmed by the absence of Pd in the reaction mixtures
analysed after each reaction.
3
-1
wt% Pt in a single phase at 50 °C, 7.3 cm min , 4 MPa H
6 MPa total pressure, as a function of time
2
, and
1
equimolar yield of fully hydrogenated p-menthane stereo-
isomers after 15 min of reaction, as shown in Fig. 3. At the
beginning of reaction p-menth-1-ene is formed, but in the
course of the reaction it hydrogenates to the final products.
After 15 min, the reaction is already complete, resulting in
trans and cis-p-menthane in 1:1 ratio.
Activation of palladium catalyst
Summarising, platinum-catalysed hydrogenation of lim-
onene yields p-menth-1-ene, trans-p-menthane, or cis-p-
menthane, and the composition depends on the flow rate of
the reaction mixture.
Palladium, as less active noble metal [28], was exploited in
the successive reactions without pre-activation. Pre-equil-
ibration was avoided to eliminate the effect on product
distribution of hydrogen from activation of the catalyst.
Activation of the Pd catalyst is an important step which
affects the outcome of the reaction, as has been reported for
various metal catalysts [31, 32].
Poisoning of catalyst
Poisoning of the catalyst was observed and analysed on
the basis of the reactions occurring at the minimal flow
1
23

Pt- and Pd-catalysed limonene hydrogenation
1365
Fig. 4 SEM photographs of 1
wt% Pt catalyst: a particle of the
fresh catalyst covered by a film
of platinum; b cut particle of the
fresh catalyst with darker
internal part (carbon support)
and a thin layer of platinum on
the surface (bright halo around
the carbon); c particle of the
catalyst after reaction at
1
2.5 MPa with platinum (bright
parts) partly removed from the
surface of the support (dark
spots); d particle of catalyst
after being reused in reaction at
1
6 MPa with exiguous amount
of platinum (bright spots) on the
support (dark sphere)
Fig. 5 SEM photos of 1 wt%
Pd catalyst: a fresh catalyst
particle; b particle of the
catalyst after reaction at
1
2.5 MPa; c particle of the
catalyst reused for a second time
in consecutive reactions at 12.5
and 16 MPa; d particle of the
catalyst reused for a third time
in consecutive reactions at 12.5,
1
6, and 12.5 MPa; e particle of
the catalyst reused for a fourth
time in consecutive reactions at
1
2.5, 16, 12.5, and 11 MPa
The results obtained with the fresh catalyst clearly show
requires further activation, which occurs as a short period
of lack of limonene conversion at the beginning of the
reactions under the biphasic conditions presented in
Figs. 6c and 7. The induction period in the reaction pre-
sented in Fig. 6c is a factor of four shorter than for use of
the fresh catalyst (Fig. 6a). This means that activation
of the catalyst occurs continuously during the preceding
reaction.
that activation of the catalyst occurs during the first 50 min
of the reaction at 12.5 MPa overall pressure, as is pre-
sented in Fig. 6a. Once the catalyst is active the reaction
starts and within the next 40 min is finished. Furthermore,
the reaction conditions have a remarkable effect on the
catalyst. The catalyst is activated during the earlier reac-
tion but even after reaction in the single phase (Fig. 6b) it
123

1
366
E. Bogel-Łukasik et al.
Fig. 6 Mol percentages of trans-p-menthane (filled circles), cis-p-
menthane (open circles), p-menth-1-ene (filled squares), limonene
(liquid ? gas) system at 12.5 MPa total pressure with fresh catalyst,
b in a single phase at 16 MPa total pressure with the catalyst being
reused from the previous reaction, and c in a biphasic (liquid ? gas)
system at 12.5 MPa total pressure with the catalyst being activated in
two preceding processes
(
diamonds). Products of the hydrogenation of limonene catalysed by 1
open squares), terpinolene (open triangles), and c-terpinene (open
3
-1
2
wt% Pd at 50 °C, 1.3 cm min , and 4 MPa H , a in a biphasic
catalyst loses its ability to catalyse the second hydrogena-
tion as efficiently as when the fresh catalyst was employed
(Fig. 6a).
Hydrogenation of limonene over palladium catalyst
gives mostly p-menthane stereoisomers and p-menth-1-
ene. Furthermore, traces of terpinolene and c-terpinene
were also detected. These isomers of limonene appear
because of double bond migration via p-allyl-absorbed
species during palladium catalysed reactions, as reported in
the literature [33–35]. The next reaction (Fig. 6c) per-
formed with reused catalyst under biphasic conditions
(12.5 MPa) gives either the same major products, or the
isomerisation products which were observed in the previ-
ous reaction (Fig. 6b). On the other hand, the concentration
of intermediate after 20–30 min of reaction is much higher
than it is in the previous reaction; this might be because of
progressive diminution of the ability of the catalyst to
catalyse a second hydrogenation to cis and trans-p-
menthane.
Fig. 7 Mol percentages of trans-p-menthane (filled circles), cis-p-
menthane (open circles), p-menth-1-ene (filled squares), and limo-
nene (open squares). Products of the hydrogenation of limonene on 1
3
-1
wt% Pd in a biphasic (liquid ? gas) system at 50 °C, 1.3 cm min
MPa H , and 11 MPa total pressure, as a function of time
,
4
2
Distribution of products
Interestingly, when the partially activated catalyst was
reused twice the isomers of limonene emerge under
milder reaction conditions, although this was not detec-
ted in the course of the reaction with fresh catalyst
(compare Fig. 6a, c).
The mechanism of hydrogenation over palladium is not
generally as uncomplicated as hydrogenation with plati-
num and proceeds through isomerization of cycloolefins
[
28]. Mechanistically, the hydrogenation of cycloolefin
The reaction presented in Fig. 7 was executed in two
phases at reduced total pressure (11 MPa) with the catalyst
reused a fourth time. The vapour–liquid equilibrium data
for 12.5 MPa presented previously by us [22] shows for the
total pressure of 11 MPa that the limonene concentration
over palladium catalyst is regioselective. In all reactions,
trans-p-menthane is the predominant product, with the
yield higher than 60 mol%, whereas the yield of the cis
stereoisomer slightly exceeds 30 mol%. The greater for-
mation of trans isomers in reactions over palladium is an
effect of the preferred manner of adsorption of p-allyl
species of cycloolefin isomers on the catalyst surface, as
was proposed by Trost et al. [33].
-
3
significantly increases from 2.2 mol dm at 12.5 MPa to
-
3.5 mol dm at 11 MPa, with a simultaneous decrease in
3
-
3
the concentration of hydrogen from 3.2 to 1.3 mol dm
.
Remarkable is the fact that relatively small reduction of
total pressure dramatically modifies the kinetics of the
process that occurs, with slower consumption of the lim-
onene at 11 MPa (at least 60 min) compared with the
reaction at 12.5 MPa (30 min).
The reaction in the single phase (Fig. 6b) with reused
catalyst is regioselective, as is the biphasic reaction pre-
sented in Fig. 6a, even though the second hydrogenation to
the final products is slower. Accordingly, with time the
1
23

Pt- and Pd-catalysed limonene hydrogenation
1367
Palladium and platinum catalyst—comparison
Conclusion
The analysis of the reactions performed with platinum and
palladium catalysts leads to the following conclusions:
The hydrogenation of limonene using 1 wt% Pt and 1 wt% Pd
leads to the formation of monoterpenes. It was observed that
the catalyst affects the outcome of the reaction. Activation of
the palladium catalyst affects the launching of the reaction. It
was noticed that carrying out reactions under biphasic con-
ditions protects the catalyst from poisoning by CO formed in
the reverse water gas shift reaction. It was observed that
palladium as catalyst promotes regioselectivity and trans and
cis-p-menthane are formed in 2:1 ratio, whereas the platinum
catalyst produces them in 1:1 ratio. Besides fully saturated
products, intermediate p-menth-1-ene, and isomers of limo-
nene such as terpinolene and c-terpinene were also formed
during hydrogenation of terpene.
1
.
.
The high activity of platinum catalysts leads to lower
chemical stability of the catalyst, because of suscep-
tibility of the catalyst to poisoning by carbon
monoxide and because of leaching of the catalyst
metal.
2
Lack of conversion of limonene during the first period
of the reaction is attributed to the need for activation of
the catalyst. Use of the platinum catalyst, which is the
more active [28], does not result in deficiencies of the
reaction at the beginning of the process, in contrast
with palladium as catalyst. Palladium catalyst is
activated in the course of the reactions and catalyses
hydrogenation through isomerisation [33, 35].
Higher total pressure (one phase) of processes results
in notably faster completion of the reaction compared
with the two-phase reaction system.
Experimental
3
4
.
.
Materials
The flow rate of the reaction mixture through the
stationary catalyst bed affects reaction product distribu-
Hydrogen and carbon dioxide, with a stated purity of
9
9.998 mol%, were supplied by Air Liquide. R-(?)-Limo-
nene (purity: 98%), (?)-p-menth-1-ene (C97%), terpinolene
C97%), c-terpinene (C95%), and n-nonane (C99%) were
3
tion. In the reaction at minimal flow rate (1.3 cm min
-1
)
under biphasic conditions catalysed by platinum, the
adsorption of terpene and formation of the half-hydro-
genated state steps are reversible. Under these conditions,
the product-determining step is the hydrogenation of the
half-hydrogenated state. Yet, because of the conditions of
the process controlled by the thermodynamics, the major
product is the intermediate. In all the other cases (under
(
supplied by Fluka. cis-p-Menthane (C97%) and trans-
p-menthane (C97%) were supplied by Fluorochem.
Catalyst
Catalysts used in this work were prepared according to the
procedure presented by Tarasenko et al. [38]. Synthetic
spherically granulated activated carbon of the SCN type,
obtained by carbonization of a copolymer of vinylpyridine and
divinylbenzene, was used, with subsequent activation by steam
with a degassed solution oftetrachloropalladate acid, H PdCl ,
3
-1
monophasic conditions at 1.3 cm min
and under
3
biphasic or monophasic conditions at 7.3 cm min
-1
flow rate), p-menthane stereoisomers in 1:1 ratio are the
major products [27, 28].
5
.
The catalyst used for the reactions affects the outcome
of the process. The platinum catalyst gives trans and
cis-p-menthane in equimolar quantities whereas palla-
dium promotes regioselectivity and twice as much
trans-p-menthane as cis-p-menthane is formed. These
results are either in a good agreement with the data
obtained for the Pd-catalysed hydrogenation of limo-
2
4
or hexachloroplatinate(IV) acid, H PtCl , of concentration
2
6
appropriate to obtain a Pd or Pt load of 1 wt%. Next, nitrogen
was bubbled through the suspension up to total vaporisation of
the liquid, and the remaining solid was dried at 120 °C.
Scanning electron microscopy (SEM) with a Hitachi S-
2400 apparatus was used to characterise the catalyst sur-
nene without supercritical CO presented by Stolle
2
face. To avoid charging effects during observation, the
surfaces were previously sputter-coated with a gold layer.
The presence of the catalyst metals in the solution collected
after the course of reaction was tested by use of atomic
absorption spectroscopy (AAS) with a ThermoElectron
Atomic Absorption Spectroscope S Series.
et al. [36] or with the results of hydrogenation of
4
-tert-butylmethylcyclohexene catalysed by platinum
and palladium [28, 37]. The trans to cis-p-menthane
isomer ratio for Pt- and Pd-catalysed reactions can be
explained by the different mechanism of cycloolefin
hydrogenation. The hydrogenation of olefins over
palladium takes place via p-allyl-absorbed species
Sample analysis
[
33, 35] rather than the 1,2-r entity utilised in the
2
Horiuti–Polanyi mechanism [27] valid for Pt catalysed
reactions.
Quantitative and qualitative analysis of the samples was
performed by gas chromatography. Qualitative identification
123

1
368
E. Bogel-Łukasik et al.
of the terpenes was performed with a Fisons MD800 GC–MS
system, using a GS-Molesieve (30 m 9 0.541 mm) column.
Analyses were performed on a few representative reaction
samples. Helium was used as carrier gas at a flow rate of
mixture consisted of the expanded liquid terpene dissolved
in gaseous CO and H in biphasic reactions or in one gas
2
2
phase in single supercritical reactions. The mixture was
vigorously stirred in the cell, in order to promote phase
equilibrium, and was continuously withdrawn from the
bottom of the view cell. At this stage the flow rate of the
reaction mixture was measured by use of a flowmeter situ-
ated between the exit of the cell and the circulation pump.
Use of a Rheonik RHM 015 GNT flowmeter equipped with
an RHE 11 electronic transmitter enabled flow rate repeat-
ability better than 0.05% and accuracy better than 0.23% to
be obtained. The reactants were circulated through the cat-
alyst bed, and were sent back to the upper entrance of the
3
-1
1
.0 cm min . Oven temperature was programmed from
-1
5
0 °C, hold 1 min, to 250 °C at 5 ° min , hold for 5 min.
3
Injector and detector temperatures were 250 °C. 1 mm of
sample was injected in split mode, with a split ratio of 30.
Quantitative analysis was performed by GC (HRGC-3000C)
with a CP-Sil 8 CB column (Varian) and flame ionization
detection. Oven temperature program: 87–91 °C ramp at
-
1
-1
0
.5 ° min , and 91–240 °C ramp at 20 ° min . Injector and
detector temperatures were 250 °C. All liquid samples con-
tained n-nonane in n-hexane (1.5 mM) as external standard
for GC analysis. The response factors were: R-(?)-limonene,
3
sapphire-windowed cell. A 100 mm sampling loop placed
at the top of the tubular reactor was used to collect the
samples of the reaction mixture at regular intervals and to
1
.42; (?)-p-menth-1-ene, 1.26; cis-p-menthane, 1.44; trans-
p-menthane, 1.43; terpinolene, 1.44; and c-terpinene, 3.76.
vent CO to the atmosphere. Samples were collected in
2
volumetric flasks containing the internal standard and sol-
vent, and subsequently analysed by GC–MS and GC. In the
reactions performed, first, the fresh catalyst was used; the
catalyst was then re-used under various biphasic or mono-
phasic conditions.
Experimental set-up
The hydrogenations were performed in an apparatus con-
sisting of one sapphire-windowed cell connected by a pump
to a tubular reactor that encloses a catalyst bed. The appa-
ratus is illustrated schematically in Fig. 8. Hydrogenation of
limonene was carried out at 323.15 K under 4 MPa of
hydrogen pressure, to which carbon dioxide was added up to
either 11 and 12.5 or 16 MPa. In the first case, the reaction
mixture was biphasic and in the second case monophasic.
Acknowledgments This work was supported by the European
Commission in the framework of the Marie Curie Research Training
Network Supergreenchem (EC contract no.: MRTN-CT-2004-
5
04005) and by the Funda c¸ a˜ o para a Ci eˆ ncia e a Tecnologia (FCT,
Portugal) through grants SFRH/BD/26356/2006 and SFRH/BPD/
4577/2007.
3
3
Catalyst (0.4 g) and 1 cm limonene were used. The reaction
References
1
2
. Cole-Hamilton DJ (2006) Adv Synth Catal 348:1341
. Brunner G (1994) Gas extraction. Steinkopff and Springer,
New York
3
4
5
. Tacke T, Wieland S, Panster P (1996) In: von Rohr R, Trepp Ch
(
eds) High-pressure chemical engineering. Elsevier, New York
. Hitzler G, Smail FR, Ross SK, Poliakoff M (1982) Org Process
Res Dev 2:137
. Savage PE, Gopalan S, Mizan TI, Martino CJ, Brock EE (1995)
AIChE J 41:723
6
7
. Nunes da Ponte M (2008) J Supercrit Fluids 47:344
. Solinas M, Pfaltz A, Cozzi PG, Leitner W (2004) J Am Chem Soc
1
26:16142
8
9
. Wandeler R, Kunzle N, Schneider MS, Mallat T, Baiker A (2001)
Chem Commun 673
. Stephenson P, Kondor B, Licence P, Scovell K, Ross SK,
Poliakoff M (2006) Adv Synth Catal 348:1605
1
1
0. Ipat’ev V, Balachinskii GJ (1912) Russ Phys Chem Soc 43:1754
1. Albert RM, Traynor SG (1989) In: Zinkel DF, Russel J (eds)
Naval stores-production, chemistry and utilization. Pulp Chemi-
cal Association, New York
1
1
2. Akgun M, Akgun NA, Dincer S (1999) J Supercrit Fluids 15:117
3. Matos HA, Gomes de Azevedo EJS, Sim o˜ es PC, Carrondo MT,
Nunes da Ponte M (1989) Fluid Phase Equilib 52:357
Fig. 8 Flow chart of the apparatus used for hydrogenation of
limonene at high pressure: C, CO compressor; TR, tubular reactor;
2
14. Fonseca J, Sim o˜ es PC, Nunes da Ponte M (2003) J Supercrit
Fluids 25:7
VC, view cell; S, sampling loop; CP, circulation pump; PT, pressure
transducer; T, temperature controller; F, flowmeter
15. Budich M (1999) Ph.D. Thesis, Technische Universit a¨ t Hamburg-
Harburg, Germany
1
23

Pt- and Pd-catalysed limonene hydrogenation
1369
1
1
1
6. Wie M, Musie GT, Busch DH, Subramanian BJ (2002) J Am
Chem Soc 124:2513
7. Burgener M, Furrer R, Mallat T, Baiker A (2004) Appl Catal A
26. Leitner W (2004) Pure Appl Chem 76:635
27. Horiuti I, Polanyi M (1934) Trans Faraday Soc 30:1164
28. Augustine RL, Yaghmaie F, van Peppen JF (1984) J Org Chem
49:1865
29. Arunajatesan V, Subramaniam B, Hutchenson KW, Herkes FE
(2001) Chem Eng Sci 56:1363
2
68:1
8. Chatterjee M, Chatterjee A, Ikushima Y (2004) Green Chem
:114
6
1
2
9. Liu R, Zhao F, Fujita S, Arai M (2007) Appl Catal A 316:127
0. Milewska A, Banet Osuna AM, Fonseca IM, Nunes da Ponte M
30. Ichikawa S, Tada M, Iwasawa Y, Ikariya T (2005) Chem
Commun 924
(
2005) Green Chem 7:726
31. Singh UK, Vannice MA (2000) J Catal 191:165
32. Carter JL, Sinfelt JH (1978) US Pat. 4088603
33. Trost BM, Strege PE, Weber L, Fullerton TJ, Dietsche TJ (1978)
J Am Chem Soc 100:3407
34. Grau RJ, Zgolicz PD, Gutierrez C, Taher HA (1999) J Mol Catal
A Chem 148:203
2
2
1. Chouchi D, Gourgouillon D, Courel M, Vital J, Nunes da Ponte
M (2001) Ind Eng Chem Res 40:2551
2. Bogel-Łukasik E, Fonseca I, Bogel-Łukasik R, Tarasenko YA,
Nunes da Ponte M, Paiva A, Brunner G (2007) Green Chem
9
:427
2
2
2
3. Bogel-Łukasik E, Bogel-Łukasik R, Kriaa K, Fonseca I, Nunes da
Ponte M (2008) J Supercrit Fluids 45:225
4. Bogel-Łukasik E, Bogel-Łukasik R, Nunes da Ponte M (2009)
Ind Eng Chem Res 48:7060
5. Minder B, Mallat T, Pickel KH, Steiner K, Baiker A (1995) Catal
Lett 34:1
35. Gault FG, Rooney JJ, Kemball C (1962) J Catal 1:255
36. Stolle A, Brauns C, Nuchter M, Ondruschka B, Bonrath W,
Findeisen M (2006) Eur J Org Chem 3317
37. Siegel S, Dmuchovsky B (1962) J Am Chem Soc 84:3132
38. Tarasenko YA, Kopyl SA, Lapko VF, Lysenko AA, Tomizuka I
(2002) Electrochemistry 70:316
123