Regioselective catalytic hydrogenation of citral with ionic liquids as reaction modifiers

Article abstract of DOI:10.1039/b822992a

Silica and polyaniline supported palladium catalysts prepared by different techniques (incipient-wetness impregnation, deposition precipitation) using Pd(OAc)₂ or H₂PdCl₄ as precursors were studied in the liquid-phase hydrogenation of citral under addition of several ionic liquids ([BMIM][NTf₂], [BMIM][PF₆], [BMPL][NTf ₂], [BMIM][DCA], [BMPL][DCA], [B3MPYR][DCA]) either as catalyst coating or as additive. By an incipient-wetness technique, the catalysts were coated with a mixture of IL in acetone. The catalysts were characterized by nitrogen physisorption, whereby a decrease of surface area and pore volume was detected by the IL coated catalysts. Furthermore, ICP-OES, TEM and IR spectroscopy were performed to analyze metal content, particle size and coverage of the catalyst with ionic liquid. Citral hydrogenation was performed at 323 K and under 2.0 MPa H₂ in an autoclave with off-line GC analysis of the product mixtures. Beside stirrer speed, catalyst mass and citral concentration, the type and quantity of ionic liquid were also varied to elucidate their influence on activity and selectivity of the Pd/SiO₂ catalysed citral hydrogenation. The results show that treatment of the catalyst with ionic liquids - independent of catalyst coating or additive - leads to a selectivity enhancement of the desired product, citronellal. With [PF₆] ^- or [NTf₂]^- as the IL anion, maximum selectivities were (60 ± 2)% at 70% conversion. In particular, dicyanamide (DCA) containing ionic liquids allow, under optimised conditions, the quantitative one-pot synthesis of citronellal, at least if the Pd/SiO ₂ catalyst was coated with 29 wt% [B3MPYR][DCA]. By using polyaniline supported Pd catalysts and [BMIM][DCA] as additive, the consecutive hydrogenation towards dihydrocitronellal was less pronounced and the influence of metal precursor, support material and preparation technique of the catalyst could be excluded. Hydrogenation of pure citral on [BMIM][DCA] coated palladium catalysts offers a solvent-free, green route to citronellal in reasonable selectivities (S = 86%).

Full text of DOI:10.1039/b822992a

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Regioselective catalytic hydrogenation of citral with ionic liquids as
reaction modifiers
J u¨ rgen Arras, Martin Steffan, Yalda Shayeghi, Dominik Ruppert and Peter Claus*
Received 24th December 2008, Accepted 5th February 2009
First published as an Advance Article on the web 27th February 2009
DOI: 10.1039/b822992a
Silica and polyaniline supported palladium catalysts prepared by different techniques
(
incipient-wetness impregnation, deposition precipitation) using Pd(OAc)
precursors were studied in the liquid-phase hydrogenation of citral under addition of several ionic
liquids ([BMIM][NTf ], [BMIM][PF ], [BMPL][NTf ], [BMIM][DCA], [BMPL][DCA],
B3MPYR][DCA]) either as catalyst coating or as additive. By an incipient-wetness technique, the
2
or H
2
PdCl
4
as
2
6
2
[
catalysts were coated with a mixture of IL in acetone. The catalysts were characterized by nitrogen
physisorption, whereby a decrease of surface area and pore volume was detected by the IL coated
catalysts. Furthermore, ICP-OES, TEM and IR spectroscopy were performed to analyze metal
content, particle size and coverage of the catalyst with ionic liquid. Citral hydrogenation was
performed at 323 K and under 2.0 MPa H in an autoclave with off-line GC analysis of the
2
product mixtures. Beside stirrer speed, catalyst mass and citral concentration, the type and
quantity of ionic liquid were also varied to elucidate their influence on activity and selectivity of
the Pd/SiO
2
catalysed citral hydrogenation. The results show that treatment of the catalyst with
ionic liquids—independent of catalyst coating or additive—leads to a selectivity enhancement of
-
-
the desired product, citronellal. With [PF
6
] or [NTf ] as the IL anion, maximum selectivities were
2
(
60 ± 2)% at 70% conversion. In particular, dicyanamide (DCA) containing ionic liquids allow,
under optimised conditions, the quantitative one-pot synthesis of citronellal, at least if the Pd/SiO
catalyst was coated with 29 wt% [B3MPYR][DCA]. By using polyaniline supported Pd catalysts
2
and [BMIM][DCA] as additive, the consecutive hydrogenation towards dihydrocitronellal was less
pronounced and the influence of metal precursor, support material and preparation technique of
the catalyst could be excluded. Hydrogenation of pure citral on [BMIM][DCA] coated palladium
catalysts offers a solvent-free, green route to citronellal in reasonable selectivities (S = 86%).
1
4,15
1
. Introduction
materials like polyaniline as the catalyst support.
By choosing
palladium, a consecutive hydrogenation could be considered,
whereby dihydrocitronellal and 3,7-dimethyloctan-1-ol are ob-
tained at higher conversion. In industrial applications, the
selective hydrogenation of citral towards citronellal is catalysed
by palladium on carbon which is usually performed in a bubble
reactor with structured packings. Yields of citronellal >90%
Regioselective hydrogenations of multifunctional substrates are
widely used for the study of factors determining activity
and selectivity on supported metal catalysts. a,b-Unsaturated
aldehydes contain at least one carbonyl in conjugation an olefinic
bond, which results in a heterogeneously catalysed competitive
hydrogenation towards saturated aldehydes and unsaturated
alcohols. Citral belongs to this class of compounds and
has additionally an isolated double bond. The liquid phase
hydrogenation of citral leads to valuable flavouring compounds
like citronellal, geraniol, nerol and citronellol (Scheme 1).
1
16
17
were obtained by using methanol as solvent and trimethylamine
2–5
16
as a base (citral:CH
3
OH:N(CH
3
)
3
= 70:27:3 w/w).
Ionic liquids (ILs) are low-melting salts with negligible vapour
pressure and their application is a field of growing interest.
18
In catalysis, ILs have been introduced in several ways, e.g.
as solvent, as stabilizer for metallic nanoparticles and as
supported ionic liquid phase (SILP catalysts). The latter
6,7
8
Several active metals, such as palladium, ruthenium,
19
9,10
11
12
13
platinum,
gold, and bimetallic rhodium-tin, nickel-tin
20
13
or silver-indium supported on carbon or metal oxides, were
studied in citral hydrogenation conducted in liquid phase,
whereby different consecutive products were obtained. Selec-
tivity can also be controlled by introducing electron-conducting
offers the opportunity to combine the advantages both of
homogeneous and heterogeneous catalysis, as described for
21
22
olefin hydrogenation, olefin hydroformylation and methanol
23
carbonylation. Hereby, transition metal complexes are dis-
solved in a thin layer of ionic liquid, which is immobilized on
microporous silica.
The application of ionic liquids as coating for nickel/silica-
catalysts (SCILL: solid catalyst with ionic liquid layer) was
shown for the liquid-phase hydrogenation of cyclooctadiene,
Ernst-Berl-Institut, Chemical Technology II, Department of Chemistry,
Technische Universit a¨ t Darmstadt, Petersenstr. 20, 64287 Darmstadt,
Germany. E-mail: claus@ct.chemie.tu-darmstadt.de; Fax: +49 6151 16
4
788; Tel: +49 6151 16 5369
7
16 | Green Chem., 2009, 11, 716–723
This journal is © The Royal Society of Chemistry 2009

View Online
Scheme 1 Reaction network of citral hydrogenation, 3,7-dimethyl-2,7-octadienal (1), 3,7-dimethyl-2-octenal (2), dihydrocitronellal (3), 3,7-dimethyl-
-octen-1-ol (4).
2
leading to an enhanced selectivity towards cyclooctene in
(ACC) supported palladium catalysts, ILs were used as a
24
comparison to the IL-free catalysts. It was assumed that the
chosen IL ([BMIM][n-C ]) shows a promoting effect.
nanoparticle stabilizer. The selectivity towards citronellal and
dihydrocitronellal was controlled by varying the anion of
8
H
17OSO
3
-
-
Furthermore, leaching of the IL into the organic solution was
not detected.
the ILs ([PF
present.
6
]
or [BF ] ), but product mixtures were still
4
In the selective hydrogenation of citral, ILs were used as bulk
solvents, whereby a commercial Pd/C catalyst showed maximum
Recently, we could show that applying a dicyanamide con-
taining ionic liquid ([BMIM][DCA]) on supported palladium
catalysts leads to high selectivities towards citronellal in citral
selectivity towards citronellal by using [EMIM][NTf ] and a
2
25
30
quantitative yield of citronellal was obtained. Although citral
hydrogenation in toluene with Pd/C gave mixtures of citronellal
and further side and consecutive products, the reaction rate was
faster with organic solvents due to a higher hydrogen solubility
hydrogenation. By treating a conventional Pd/C catalyst with
minor amounts of ionic liquids prior to reaction, a quantitative
citronellal yield was obtained. The ionic liquid covers the
palladium nanoparticles, as shown in Fig. 1.
2
6,27
than for ionic liquids.
recycled and product extraction is still necessary.
The selectivity pattern of monometallic Pd/Al
lic Ni-Sn/Al in citral hydrogenation could not be mod-
ified by replacing the organic solvent (n-hexane) with IL
The catalytic system was multiply
2
O or bimetal-
3
2
O
3
13
(
[BMIM][NTf ]). Product selectivities at equal citral conver-
2
sion level were in the same range, whereby the reaction rate was
lowered in presence of bulk IL.
The concept of supported ionic liquid catalysts, named
SILCA, was extensively studied in the selective hydrogenation
2
8,29
of citral.
During preparation of the activated carbon cloth
Fig. 1 Supposed structure of an IL coated solid catalyst.
Green Chem., 2009, 11, 716–723 | 717
This journal is © The Royal Society of Chemistry 2009

View Online
The aim of this investigation was to study reaction variables
catalyst mass, citral concentration, stirrer speed) on the course
incipient-wetness. The catalyst was then dried at 343 K (3 h)
to remove traces of the organic solvent. The amount of ionic
(
of citral hydrogenation by using ionic liquids as catalyst coating
or as additive for supported catalysts. Furthermore, influences of
catalyst precursor and preparation route were studied and pure
citral was also applied for hydrogenation. Focusing on dicyan-
amide ionic liquids, further cations ([BMPL] and [B3MPYR] )
were investigated to validate their influence on selectivity and
activity. Catalyst characterization by IR spectroscopy and pulse
CO chemisorption will give information about the interaction
of the ionic liquid with the silica supported palladium catalyst.
liquid was varied in a range between 0.04 and 1 gIL/gcatalyst.
2.2. Catalyst characterization
+
+
2.2.1. Nitrogen physisorption. For the determination of
textural properties, nitrogen physisorption was carried out
on a Quantachrome Autosorb at 77 K. Prior to analysis,
the catalyst samples (100 mg) were dried at 423 K for 2 h.
Adsorption isotherms were recorded with increased relative
pressures (multipoint-method), the specific surface area, total
pore volume and pore diameter of the catalysts were calculated
2
. Experimental
31
32
according to the BET and BJH methods, respectively.
2
.1. Catalyst preparation
2
.2.2. Transmission electron microscopy (TEM). HRTEM
was performed using a 300 kV JEOL JEM-3010 microscope
equipped with a LaB -cathode. Samples were prepared by
suspending ground catalyst powder in methanol and dripping
the suspension onto a 3.2 mm copper mesh coated with a
continuous carbon film.
2
.1.1. Pd/SiO via incipient-wetness. Pd/SiO
2
2
(wPd
=
5
wt%) was obtained via incipient-wetness impregnation of the
silica support (9.5 g, silica gel 100, Merck, pre-dried at 473 K
for 2 h) with a 1.1 g portion of Pd(OAc) (Evonik-Degussa)
dissolved in 40 mL acetone (Merck), whereby the support was
impregnated in five steps. After drying (RT overnight, then at
53 K for 2 h), the catalyst was reduced in a flow of hydrogen at
73 K for 1 h.
6
2
2
.2.3. Pulse CO chemisorption. In order to identify the
3
3
number of surface palladium, pulse chemisorption of the
catalysts was investigated on a TPD/R/O 1100 with carbon
monoxide as the probe molecule. The samples (1 g) were
charged in a quartz reactor and reduced at 473 K prior to
analysis. The chemisorption experiment was performed at room
temperature and the sample was pulsed with CO until saturation
was completed. Mean particle diameters were calculated using
the following equation:
2
.1.2. Pd/PANI via deposition-precipitation. Polyaniline
(
11.5 g, M
w
= 65 kg/mol, Aldrich) was suspended at room
temperature in water (100 mL) and the pH was raised to 11
by adding aqueous sodium carbonate (5 mL, 10wt%) dropwise.
Then, an aqueous solution of dihydrogentetrachloropalladate
(
2.5 g, Heraeus, 20wt% Pd) was diluted with water to 20 mL and
added to the suspension. The solution was heated to 363 K and
the pH was held constant at 7 with aqueous sodium hydroxide
Ê v
Ë a
ˆ
¯
Pd
Pd
6
fStochw Pd
Á
˜
(
10wt%). After 15 min, it was cooled to 353 K; aqueous sodium
dPd
=
(1)
hydroxide (5.8 mL, 10wt%) and formaldehyde (1.2 mL, 37wt%
in water) were then added for reducing. After filtering, washing
with water (5 ¥ 10 mL) and drying, a black powder was obtained.
Ê n
ˆ
CO
M Pd
Á
˜
Ë m
¯
cat
where vPd represents the atomic volume and aPd the atomic
surface area of palladium (data and equation obtained from
ref. 33), f Stoch the stoichiometry factor of CO on palladium and
2
.1.3. Preparation of IL coated catalysts. The ionic
liquid (0.3 g) ([BMIM][NTf
] = 1-butyl-3-methylimidazolium
bis(trifluoromethylsulfonyl)amide, [BMPL][NTf
] = N-butyl-
N-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide,
BMIM][PF 1-butyl-3-methylimidazolium hexafluoro-
2
2
Ê n
Ë m
ˆ
CO
the chemisorbed amount of CO per catalyst mass.
Á
˜
[
6
]
=
cat
¯
phosphate, [BMIM][DCA] = 1-butyl-3-methylimidazolium di-
cyanamide, [BMPL][DCA] = N-butyl-N-methylpyrrolidinium
dicyanamide, [B3MPYR][DCA] = N-butyl-3-methylpyridinium
dicyanamide, Merck, Scheme 2) was dissolved in acetone
2
.2.4. Infrared spectroscopy. Infrared spectroscopic stud-
ies were carried out on a Paragon 1000 PC (Perkin Elmer)
spectrometer in transmission mode. Prior to analysis, the solid
ionic liquid coated sample was diluted with KBr and pressed to
a tablet. In order to observe the influence of palladium, a Pd-
free sample was prepared by incipient wetness impregnation of
(
Merck, 2 mL) and impregnated over the Pd/SiO catalyst via
2
[
BMIM][DCA] dissolved in acetone on silica, further designated
as “[BMIM][DCA]/SiO ”.
2
2
.3. Citral hydrogenation
The hydrogenation runs were performed in a stainless steel
autoclave (Parr, 300 mL) equipped with a heating jacket and
magnetic stirring (impeller type). At first, the reactor was loaded
with catalyst (mcat = 0.05–0.5 g) prior to addition of the reaction
mixture (20 mL citral, Merck; 80 mL n-hexane, Roth; 5 mL
n-tetradecane, Merck). Then, the reactor was closed and the
Scheme 2 Cation (top) and anion (bottom) structures of ionic liquids
used in this work.
7
18 | Green Chem., 2009, 11, 716–723
This journal is © The Royal Society of Chemistry 2009

View Online
Table 1 Results of nitrogen physisorption on Pd/SiO
IL coating (wIL = 17 wt%)
2
catalysts with
2
Table 2 Pulse CO chemisorption results of Pd/SiO catalysts coated
with or without [BMIM][DCA]
S
BET
2
V
pore
nCO/mcat
Pd [wt%] [mmolCO/gcat
a
Entry IL
Measurement
[m /g] [mL/g]
Entry Catalyst
w
]
CO/Pd
1
2
3
4
5
6
—
prior to hydrogenation 317
prior to hydrogenation 230
prior to hydrogenation 171
prior to hydrogenation 236
0.96
0.66
0.75
0.69
0.66
0.83
1
2
Pd/SiO
Pd/SiO
2
2
5
57.8
9.6
0.12
0.04
[BMIM][NTf
[BMIM][PF ]
6
[BMPL][NTf
[BMPL][NTf
—
2
]
-[BMIM][DCA] 2.5
a
˚
3
˚
2
-1
2
]
]
constants: vPd = 14.7 A , aPd = 7.93 A , MPd = 106.42 g mol , f Stoch =
2
after hydrogenation
after hydrogenation
234
295
1 (data from ref. 33).
magnetic stirrer was tuned to its desired frequency (n = 500–
200 rpm). After flushing with argon (2.0 MPa, three times),
1
the reaction suspension was heated up to 323 K. Charging
the reactor with hydrogen (1–2 MPa) determined the start
of the reaction. Periodically taken samples were analyzed by
temperature programmed gas capillary chromatography (HP
6
890) using an Agilent DB-Wax column (l = 30 m, d = 0.25 mm,
i
t
f
= 0.25 mm); n-tetradecane was used as internal standard.
If ionic liquids were used as additive, the reactor was charged
with an amount of it prior to addition of the substrate solution.
3
. Results and discussion
3
.1. Catalyst characterisation
3
.1.1. Nitrogen physisorption. In good accordance to pre-
30
vious findings with dicyanamide based ionic liquids, the
textural properties of the catalyst were influenced by coating
it with ionic liquids based on the [NTf
2
], and [PF ] anions,
6
respectively (Table 1). The IL-free catalyst showed a BET
2
surface area of 317 m /g, and a pore volume of 0.96 mL/g,
whereas these parameters were lowered for all IL coated catalysts
prior to hydrogenation. As indicated by a post-hydrogenation
Fig. 2 TEM characterization of Pd/PANI.
whereas the IL-free Pd/SiO had 5 wt%. Under the assumption
2
analysis with the [BMPL][NTf
2
] coated catalyst, the pore sizes
that the shape of the palladium nanoparticles remains stable
remained on the same level. A slight decrease in surface area and
pore volume was observed with an IL-free catalyst after citral
hydrogenation. It can be assumed that the deposition of the
ionic liquid into the pores remains stable, possibly by covering
the catalytically active palladium nanoparticles.
after ionic liquid impregnation, i.e. particle diameters are equal
for the IL-free and IL-coated catalyst, the uptake of the
latter one should be approximately 30 mmol CO/g , which
cat
is three times higher than the observed uptake of ~10 mmol
CO/g . Additionally, the relatively low solubility of CO in
cat
34
[
BMIM][DCA] compared to molecular solvents may also
3
.1.2. Transmission electron microscopy (TEM). TEM
decrease the CO uptake on palladium.
analysis of the prepared Pd/PANI catalyst also revealed a high
distribution of palladium nanoparticles, as depicted in Fig. 2.
A thorough measurement of the particle sizes gave a mean
diameter of (3.1 ± 0.8) nm, where a spherical particle shape
3.1.4. Infrared spectroscopy. Infrared spectroscopic stud-
ies were performed on IL-free and [DCA]-coated catalysts and
-
1
characteristic nitrile vibrations between 2100–2300 cm for
the dicyanamide anion are shown on Table 3. The observed
nitrile absorptions for [BMIM][DCA]/SiO2 are comparable
to IR studies on neat [EMIM][DCA] and [BMPL][DCA],
was assumed. Pd/SiO
size between 5 and 10 nm, as described elsewhere.
2
exhibits an average palladium particle
30
3
.1.3. CO pulse chemisorption of Pd/SiO
modified by [BMIM][DCA]. Pulse chemisorption results
showed a decrease in CO uptake if the IL-free Pd/SiO
~58 mmol CO/gcat) was treated with [BMIM][DCA] (~10 mmol
2
and Pd/SiO
2
35
respectively. In presence of palladium (entry 2), a blue shift
-1
to higher energy of at least 15 cm is observed which can be
2
interpreted as an electronic interaction between palladium and
(
-
the anion. A coordination of N(CN)
2
and palladium via the
CO/gcat) and imply that the chosen ionic liquid modifies
the adsorption characteristics of CO on palladium (Table 2).
Assuming a linear coordination of CO on palladium, the mean
36
central nitrogen, which was described for Pd{N(CN)
2
}, could
be formed.
particle diameter is 9 nm for the IL-free Pd/SiO and in good
agreement with observations by TEM.
2
3
.2. Citral hydrogenation with ionic liquids as catalyst additives
30
The decrease of ~48 mmol CO/gcat can be generally attributed
to the lower palladium loading (2.5 wt%) of the SCILL catalyst,
Compared to other metals like nickel and rhodium, palladium
as the active metal favours the hydrogenation of the conjugated
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 716–723 | 719

View Online
Table 3 Infrared study of [BMIM][DCA] supported silica materials
n + nas
s
Entry
(C–N) n
1
(C∫N) n (C∫N) Ref.
2
1
2
3
[BMIM][DCA]/SiO
Pd/[BMIM][DCA]/SiO
2
2235 2135
2260 2151
2305 2190
2195
2210
2235
this work
this work
36
2
II
Pd {N(CN)
2 2
}
C=C bond of citral instead of its C=O bond; density-functional-
theory (DFT) calculations predict a preferred adsorption of
3
7
a,b-unsaturated aldehydes on Pd(111) than for simple alkenes.
4
Moreover, the h adsorption mode of the conjugated C=O group
on palladium is favoured due to its low four-electron repulsion
leading to a poor catalyst towards C=O hydrogenation.
Fig. 3 S,X-plot for various Pd/SiO
2
masses in citral hydrogenation;
37
open symbols = IL-free catalyst, close symbols = [BMIM][DCA] as
catalyst additive; conditions: c0,citral = 1.1 mol/L, T = 323 K, pH2
.0 MPa, n = 1200 rpm, t = 6 h.
=
3
.2.1. Influence of catalyst mass. Several reaction condi-
2
R
tions, i.e. catalyst mass, stirring frequency, and citral concen-
tration, were varied to determine the influence of the added
fore, stirrer speed does not affect the catalysts performance and,
thus, external film diffusion can be excluded.
[
BMIM][DCA] in comparison to the IL-free Pd/SiO (Table 4).
2
Nearly full citral conversion was obtained with a catalyst mass
of 500 mg and the selectivity for dihydrocitronellal (S = 70%)
indicates a high degree of consecutive reaction. By adding
For an industrial catalyst application, the chosen citral
concentration (1.1 mol/L) is low. Thus, two runs—either with
or without [BMIM][DCA] as additive—were carried out with
pure citral (5.6 mol/L, entries 7 and 8); n-tetradecane was
added for GC analysis and the hydrogenation was performed
with 500 mg catalyst. Under IL-free conditions, the citronellal
selectivity at 48% conversion is moderate due to a higher
amount of isomerization of citral to 1 and other by-products
like dihydrocitronellal. The lower conversion level can be
attributed to higher citral/catalyst ratio. Because no solvent
was added, citral may also adsorb on the acidic sites of the
support material and unpreferred side-reactions took place.
With the addition of [BMIM][DCA], selectivity of citronellal
could be also enhanced at a comparable conversion level. The
[
BMIM][DCA] to the substrate solution and performing the
hydrogenation under the same conditions, a decrease of citral
conversion was observed (X = 41%), while citronellal was
obtained with very high selectivity (S = 97%).
These relationships between IL-free and IL-as-additive con-
ditions were also observed by a less amount of catalyst (mcat
=
5
0 mg), but the final conversions were lowered due to a higher
citral/catalyst ratio. Again, by adding [BMIM][DCA] to the
substrate solution, selectivity towards citronellal was enhanced
(
S = 100%) and a poor conversion was obtained (X = 4%).
For IL-free conditions, citronellal selectivity could be ob-
served in a range between 10 and nearly 100% conversion,
showing a consistent transition between the catalyst masses
ionic liquid might be covering the acidic sites of SiO , which
2
formerly catalyse the isomerisation. The considerable space-
time-yield towards citronellal for the [BMIM][DCA] coated
(
Fig. 3). Hence, the selectivities at a fixed conversion are
independent of the chosen catalyst mass and reaction time.
Furthermore, these results are in accordance with experiments
at 250 mg catalyst mass showing similar selectivity patterns both
-1
catalyst (13.4 kgCAL (kgPd/SiO2 h) ) is comparable to results
17
described by Br o¨ cker et al., although the citral conversion did
not reach 100%. Nevertheless, the use of volatile co-solvents
like methanol and trimethylamine is not necessary with the
introduced SCILL(DCA) concept enabling a greener route to
citronellal.
30
for IL coated and IL-free catalyst.
3.2.2. Influence of stirrer speed and citral concentration.
The influence of stirrer speed on the catalyst properties was
examined at 500 and 1200 rpm (Table 4, entries 5 and 6), whereby
3.2.3. Influence of palladium precursor and support material.
Determining influences of the choice of palladium precursor,
support material and preparation route, a Pd/PANI catalyst was
investigated—for the first time—in the citral hydrogenation. The
2
50 mg Pd/SiO was introduced. Independent of the reaction
2
conditions (IL-free or IL-additive), the conversion levels and
selectivities were congruent for both stirring frequencies. There-
Table 4 Citral hydrogenation with Pd/SiO
2
catalysts added with [BMIM][DCA] in n-hexane (T = 323 K, pH2 = 2.0 MPa, t = 6 h) under various
R
conditions
Entry
m
cat [mg]
m
IL [mg]
c0,citral [mol/L]
n [rpm]
X [%]
S
CAL [%]
S
DHC [%]
S
ISO [%]
SOP [%]
1
2
3
4
5
6
7
8
500
500
50
—
500
—
50
—
250
—
250
1.1
1.1
1.1
1.1
1.1
1.1
5.6
5.6
1200
1200
1200
1200
500
500
1200
1200
98
41
53
4
90
23
48
52
22
97
51
100
40
100
43
69
—
12
—
36
—
9
1
—
8
—
—
—
11
4
8
3
30
—
24
—
37
8
50
250
250
500
500
86
1
7
20 | Green Chem., 2009, 11, 716–723
This journal is © The Royal Society of Chemistry 2009

View Online
catalyst showed higher citronellal selectivity, with 78% at 50%
conversion, than Pd/SiO
X = 50%). As depicted in Fig. 4, the course of hydrogenation
gave initial citronellal selectivity of 80% and at final conversion,
dihydrocitronellal was the main product (S = 96%).
Table 5 Reaction rate constants and reaction rates under first-order
kinetics of Pd/PANI catalysed citral hydrogenation with or without
BMIM][DCA] as additive
2
under IL-free conditions (S = 50% at
[
0
r [mmol
-1 -1 b
-
1
a
2
Entry Catalyst
IL
—
k [min ]
L
s ]
r
1
2
Pd/PANI
0.05619 ± 0.00324 10.4 ± 0.06 0.994
Pd/PANI [BMIM]- 0.02293 ± 0.00038 0.43 ± 0.07 0.999
[
DCA]
a
b
c = c
0
exp(-kt
R
),
r
0
0
= kc .
of citronellal to dihydrocitronellal took place and the product
composition remained stable for two hours. The consumption
of citral followed first-order kinetics, which was determined
by an exponential curve fit (Table 5), whereas on Pd/PANI
-
1
-1
the initial reaction rate r
0
was higher (1.04 mmol L s )
than for the respective [BMIM][DCA] coated Pd/PANI catalyst
-
1
-1
(
0.43 mmol L s ).
Despite the higher intrinsic selectivity for that catalyst, which
Fig. 4 c,t-plot for Pd/PANI-catalysed citral hydrogenation; condi-
tions: mcat = 250 mg, c0,citral = 1.1 mol/L, T = 323 K, pH2 = 2.0 MPa,
may be attributed to another preparation route, IL treatment
leads again to a selectivity enhancement showing that this
concept is independent of the palladium catalyst source. As
determined by nitrogen physisorption, the BET surface area of
n = 1200 rpm, t = 6 h.
R
In contrast, the addition of [BMIM][DCA] leads to a high
citronellal selectivity of at least 90% (Fig. 5). Even when the
substrate citral was consumed, no consecutive hydrogenation
2
2
PANI (37 m /g) is lower than for the applied SiO (344 m /g),
2
not affecting the selectivity pattern. Moreover, these experiments
show that the concept is also suitable for unconventional support
materials like polyaniline.
3
.3. Citral hydrogenation with ionic liquids as catalyst coating
.3.1. Variation of fluorous anions based ionic liquids. Sev-
3
eral ionic liquids based on fluorous containing anions were de-
posited with a loading of 17% on a Pd/SiO catalyst to determine
2
its influence on selectivity and activity in the hydrogenation of
citral (Table 6, entries 1–4). For comparison, an experiment with
IL-free Pd/SiO was performed, and throughout the reactions
2
the molar Pd/citral ratio was held constant, so a higher mass of
ionic liquid coated catalyst was introduced.
Except for the [BMIM][PF
version level for coatings of [BMIM][NTf
respectively, were in the same order as for the IL-free catalyst.
The lower hydrogen solubility in ionic liquids compared to
organic solvents does not affect the conversion level, indicating
no mass transport limitation. The final citronellal selectivity
6
] coated catalyst, the citral con-
2
] and [BMPL][NTf ],
2
38
Fig. 5 c,t-plot for Pd/PANI-catalysed citral hydrogenation with
[
BMIM][DCA] as additive; conditions: mcat = 250 mg, m[BMIM][DCA]
55 mg, c0,citral = 1.1 mol/L, T = 323 K, pH2 = 2.0 MPa, n = 1200 rpm,
= 6 h.
=
39
1
t
R
Table 6 Citral hydrogenation with Pd/SiO
2
and IL coated catalysts in n-hexane (T = 323 K, pH2 = 2.0 MPa, c0,citral = 1.1 mol/L, n = 1200 rpm,
t
R
= 6 h) with various ILs. Unless noted otherwise, selectivities referring to 80% citral conversion
Entry
IL
m
cat [mg]
w
IL [wt%]
X [%]
S
CAL [%]
S
DHC [%]
S
OP [%]
Ref.
a
b
b
b
1
2
3
4
5
6
7
8
9
—
0.2
—
17
17
17
10
7
29
10
24
79
74
63
75
100
100
99
100
100
45
59
59
62
81
82
98
90
99
19
15
10
14
11
10
1
36
26
31
24
8
8
1
33
38
38
30,38
this work
this work
this work
this work
this work
a
b
b
b
b
b
b
b
b
b
[BMIM][NTf
[BMIM][PF
[BMPL][NTf
[BMIM][DCA]
[B3MPYR][DCA]
[B3MPYR][DCA]
[BMPL][DCA]
[BMPL][DCA]
2
]
0.24
0.24
0.24
0.28
0.27
0.35
0.28
0.33
a
6
]
2
a
]
3
—
7
1
a
b
final citral conversion after t
R
= 4 h, selectivities for final conversion.
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 716–723 | 721

View Online
after 4 h reaction time was increased for all IL coated
was determined under the assumption that the GC standard
n-tetradecane is insoluble in ionic liquids. As a general trend,
the solubility of citronellal in dicyanamide based ionic liquids
catalysts compared to the IL-free reaction (S = 45%), whereby
[
BMPL][NTf
2
] showed the best performance (S = 62%).
6
Minimum amounts of fluoride present in [BMIM][PF
] may
was lower than for citral. In comparison to [BMIM][NTf ], di-
2
also decrease the conversion level. It is known that citral
hydrogenation is sensitive towards halides as contaminants,
which was shown for chloride contamination by Hardacre
cyanamide based ionic liquids gave lower solubilities indicating
better desorption of citronellal from the ionic liquid layer into
the organic phase.
To understand in detail the interaction of DCA based ionic
liquids with active metal palladium and the reactants citral and
hydrogen, respectively, further examinations are in progress.
25
et al. The higher average pore diameter of [BMIM][PF
Pd/SiO determined by N physisorption (167 nm instead of
14 nm for Pd/SiO ) could be interpreted by partially etching
of silica by the fluoride possibly forming SiF
6
] coated
2
2
1
2
4
.
In contrast to the dicyanamide based ionic liquids, the
selectivity towards citronellal was enhanced to a lesser extent.
4
. Conclusion
Dicyanamide containing ionic liquids, independent of the
cation, act very well as selectivity modifiers for the hetero-
geneously catalysed citral hydrogenation on Pd supported
catalysts. Regardless of the nature and origin of the prepared Pd
catalysts, excellent selectivities towards citronellal were observed
by using these ILs as catalyst additive or coating, whereby con-
secutive and side reactions are strongly inhibited. This concept
offers a greener opportunity to a solvent-less hydrogenation of
citral, as experiments on neat citral at high citronellal selectivity
in considerable space-time-yield have shown.
3.3.2. Dicyanamide containing ionic liquids. By applying
dicyanamide based ionic liquids as catalyst coating, citral con-
version and citronellal selectivity could be improved indepen-
+
+
+
dent of which cation (either [BMIM] , [BMPL] or [B3MPYR] )
was used (Table 6, entries 5–9). Moreover, nearly quantitative
citronellal yield was obtained if the catalyst was loaded with
2
9wt% [B3MPYR][DCA]. In comparison to previous findings,
the citronellal selectivity is moderately lowered due to a de-
creased amount of ionic liquid. With decreased IL loading, the
citronellal selectivity at 80% conversion was diminished either
with [BMIM][DCA] (S = 81%) or [B3MPYR][DCA] (S = 82%)
as catalyst coating (entries 5 and 6); dihydrocitronellal was then
observed as the consecutive product (S ~ 10%).
In comparison with [BMIM][DCA] and [B3MPYR][DCA]
at equal IL loading, experiments with 10wt% [BMPL][DCA]
gave the highest selectivity towards citronellal, supporting the
Acknowledgements
The authors want to thank Dipl.-Ing. K. Lehnert, Dipl.-Ing.
F. Klasovsky and Dipl.-Ing. A. Pachulski for catalyst character-
ization, and Dipl.-Ing. N. W o¨ rz for catalyst preparation. P. C.
thanks the Fonds der Chemischen Industrie for financial support.
findings with [NTf ]-based ILs, whereas the [BMPL]-cation
showed increased selectivities compared to the [BMIM]-cation.
Quantitative citronellal yield was obtained with a 24wt%
2
References
¨
1
P. Claus and Y. Onal, Regioselective Hydrogenations, in Handbook
[
BMPL][DCA] coated catalyst (entry 9).
of Heterogeneous Catalysis, ed. G. Ertl, H. Kn o¨ zinger, F. Sch u¨ th and
J. Weitkamp, 2nd Edition, Wiley-VCH, Weinheim, 2008, pp. 3308–
The latter supports the findings with [BMIM][DCA] as
3
329.
30
catalyst additive, where an increasing amount of IL leads
to increased citronellal selectivities. Obviously, it is necessary
to introduce a threshold amount of DCA-IL to perform the
reaction at full citronellal selectivity, facilitating a full coverage
of the active sites with DCA-containing ionic liquid.
2
P. Claus, Top. Catal., 1998, 5, 51–62.
3 P. Claus, Appl. Catal. A General, 2005, 291, 222–229.
4 P. Gallezot and D. Richard, Catal. Rev. Sci. Eng., 1998, 40, 81–126.
5
6
D. Murzin, Appl. Catal. A General, 2005, 292, 1–49.
M. A. Aramendia, V. Borau, C. Jimenez, J. M. Marinas, A. Porras
and F. J. Urbano, J. Catal., 1997, 172, 46–54.
The results show that generally dicyanamide based ILs are
able to act as a modifier in the citral hydrogenation with a
selectivity enhancement towards citronellal. So the electronic
and chemical structure of the anion plays a crucial role,
affecting the mechanism of hydrogenation which may also be
7 V. Satagopan and S. B. Chandalia, J. Chem. Tech. Biotechnol., 1994,
5
9, 257–263.
8
G. Neri, L. Mercadante, A. Donato, A. M. Visco and S. Galvagno,
Catal. Lett., 1994, 29, 379–386.
9 U. K. Singh and M. A. Vannice, J. Catal., 2001, 199, 73–84.
10 S. Mukherjee and M. A. Vannice, J. Catal., 2006, 243, 108–130.
11 C. Milone, M. L. Tropeano, G. Gulino, G. Neri, R. Ingoglia and S.
Galvagno, Chem. Commun., 2002, 868–869.
-
1
concluded by IR spectroscopy, whereas the blue shift (~15 cm )
of the nitrile vibrations indicates a partial formation of DCA
complexed palladium species.
1
2 M. Lucas and P. Claus, Chem. Ing. Tech., 2001, 73, 252–257.
13 M. Steffan, M. Lucas, A. Brandner, M. Wollny, N. Oldenburg and P.
Claus, Chem. Eng. Technol., 2007, 30, 481–486.
Furthermore, the DCA based ionic liquids may also modify
the adsorption, reaction and mass transport processes, as is
known for the liquid phase hydrogenation of benzene, where
suspending of the supported ruthenium catalyst in water leads
to a selectivity enhancement towards the preferred cyclohexene
by lowering the desorption of cyclohexene from the catalyst
1
1
1
1
4 M. Steffan, F. Klasovsky, J. Arras, C. Roth, J. Radnik, H. Hofmeister
and P. Claus, Adv. Synth. Catal., 2008, 350, 1337–1348.
5 F. Klasovsky, M. Steffan, J. Arras, J. Radnik and P. Claus, Open Phys.
Chem. J., 2007, 1, 1–4.
6 H. G. G o¨ bbel, G. Wegner, H. Fuchs, S. Unverricht, and A. Salden,
WO2004007414A1, 2004.
7 F. J. Br o¨ cker, G. Kaibel, W. Aquila, H. Fuchs, G. Wegner, and M.
Stroezel, EP0947493B1, 1999.
40
surface. In our study, extraction experiments with ILs and
solutions of citral and citronellal were performed at room
temperature in order to determine the concentration of citral
and citronellal in the ionic liquid. The product mixture of the
organic phase was analyzed by GC and a “conversion” level
18 V. I. Parvulescu and C. Hardacre, Chem. Rev., 2007, 107, 2615–2665.
1
9 C. C. Cassol, A. P. Umpierre, G. Machado, S. I. Wolke and J. Dupont,
J. Am. Chem. Soc., 2005, 127, 3298–3299.
0 A. Riisager, P. Wasserscheid, R. van Hal and R. Fehrmann, J. Catal.,
2003, 219, 452–455.
2
7
22 | Green Chem., 2009, 11, 716–723
This journal is © The Royal Society of Chemistry 2009

View Online
2
2
2
2
2
2
2
2
2
3
1 C. P. Mehnert, E. J. Mozeleski and R. A. Cook, Chem. Commun.,
002, 3010–3011.
2 M. Haumann and A. Riisager, Chem. Rev., 2008, 108, 1474–
497.
31 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc., 1938,
2
60, 309–319.
32 E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem. Soc.,
1951, 73, 373–380.
1
3 A. Riisager, B. Jorgensen, P. Wasserscheid and R. Fehrmann, Chem.
33 G. Bergeret and P. Gallezot, in Handbook of Heterogeneous Catalysis,
st
Commun., 2006, 994–996.
eds. G. Ertl, H. Kn o¨ zinger, J. Weitkamp, (Eds.), 1 edition, Wiley-
4 U. Kernchen, B. Etzold, W. Korth and A. Jess, Chem. Eng. Technol.,
VCH, Weinheim, 1999, pp. 439–441.
2
007, 30, 985–994.
34 C. A. Ohlin, P. J. Dyson and G. Laurenczy, Chem. Commun., 2004,
5 K. Anderson, P. Goodrich, C. Hardacre and D. W. Rooney, Green
Chem., 2003, 5, 448–453.
1070–1071.
35 D. R. MacFarlane, J. Golding, S. Forsyth, M. Forsyth and G. B.
Deacon, Chem. Commun., 2001, 1430–1431.
36 H. K o¨ hler, M. Jeschke and H. Wusterhausen, Z. anorg. Allg. Chem.,
1987, 549, 199–203.
37 F. Delbecq and P. Sautet, J. Catal., 1995, 152, 217–236.
38 J. Arras, M. Steffan, and P. Claus, 1st SynTOP Smart Synthesis and
Technologies for Organic Processes, VDI-Berichte 2039, VDI-Verlag
GmbH, D u¨ sseldorf, 2008, pp. 109–114.
6 J. L. Anthony, E. J. Maginn and J. F. Brennecke, J. Phys. Chem. B,
2
002, 106, 7315–7320.
7 P. J. Dyson, G. Laurenczy, C. A. Ohlin, J. Vallance and T. Welton,
Chem. Commun., 2003, 2418–2419.
8 J. P. Mikkola, P. Virtanen, H. Karhu, T. Salmi and D. Y. Murzin,
Green Chem., 2006, 8, 197–205.
9 J. P. Mikkola, P. P. Virtanen, K. Kordas, H. Karhu and T. O. Salmi,
Appl. Catal. A General, 2007, 328, 68–76.
39 P. J. Dyson, G. Laurenczy, A. Ohlin, J. Vallance and T. Welton, Chem.
0 J. Arras, M. Steffan, Y. Shayeghi and P. Claus, Chem. Commun.,
Commun., 2003, 2418–2419.
2
008, 4058–4060.
40 E. Dietzsch, PhD thesis, TU Chemnitz, 2004; and references therein.
This journal is © The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 716–723 | 723