Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

Palladium supported catalysts for the selective

hydrogenation of sunflower oil
Bendaoud Nohaira , Catherine Especela,∗ , Gwendoline Lafayea , Patrice Marécota ,
Lê Chiên Hoangb , Jacques Barbiera,b
a LACCO, UMR 6503, Université de Poitiers, 40 Avenue du Recteur Pineau, 86022 Poitiers Cedex, France
bValagro, Valorisation des Agroressources, 40 Avenue du Recteur Pineau, 86022 Poitiers Cedex, France

Received 29 October 2004; received in revised form 15 November 2004; accepted 20 November 2004
Available online 25 December 2004

Abstract

Selective hydrogenation of ethyl esters of traditional sunflower oil (SOEE) was carried out at low temperature (40 ◦ C) in ethanol as solvent
in the presence of supported palladium catalysts. In the range of studied dispersions (12–55%), the SOEE hydrogenation reaction is insensitive
to the size of the palladium particles deposited on silica, but the largest metallic particles enhance the C18:1 cis–trans isomerization. The use of
various oxide supports (␣-Al2 O3 , ␥c -Al2 O3 , TiO2 , MgO, ZnO, CeO2 , CeZrO2 ) to deposit palladium does not allow to improve the selectivity
of the reaction toward the cis C18:1 compared to Pd/SiO2 catalysts. On the other hand, the modification of palladium by lead, introduced by
surface redox reaction (catalytic reduction), promotes the selectivity in cis C18:1. In addition, this technique of preparation involves a moderate
decrease of the palladium activity compared to a traditional method of successive impregnations. The introduction of amines in the reaction
medium modifies the hydrogenating properties of the Pd/SiO2 catalyst. According to the quantity and the nature of the added amine (aliphatic
with linear or ramified chain and cyclic compounds), the catalytic activity can either be unchanged or inhibited. These evolutions result from a
promoter electronic effect generated by the presence of the amine and from a poison geometric effect related to the adsorption of this nitrogen-
containing compound on the palladium surface. Whatever the nature of the amine, it induces an increase of the selectivity toward the cis C18:1.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Vegetable oils; Selective hydrogenation; Palladium; Lead; Amines

1. Introduction acid, C18:3), linoleic acid (cis-9, cis-12-octadecadienoic

acid, C18:2) and oleic acid (cis-9-octadecenoic, C18:1).
The vegetable edible oils hydrogenation is a very impor- Their oxygen absorption rate is 800:100:1, respectively.
tant operation in the chemical industry. The main qualities Therefore, partial hydrogenation of rapeseed oil to the C18:1
of natural substances are their environment compatibility acid should result in a strong stabilization over against
(low ecotoxicity, great biodegradability). Moreover, these oxygen. Moreover, the oleic acid C18:1 in the cis isomer
resources are renewable. The industrial applications are as form, contrary to the trans C18:1 and the totally saturated
varied as pharmacy, cosmetics, plastics, detergents, lubri- (stearic acid, C18:0) compounds, remains liquid at ambient
cants, etc. The aim of this study is to produce lubricants. Oils temperature. However, a traditional sunflower oil can contain
used as lubricants have to remain liquid at ambient tempera- only 25–30% of C18:1 cis isomer, the main constituent being
ture and to be stable under oxygen atmosphere to avoid their the cis-9, cis-12 C18:2. The selective hydrogenation of such
polymerization [1]. Main unsaturated fatty acids in vegetable vegetable oil must lead to a minimum of 80% toward cis
oils are linolenic acid (cis-9, cis-12, cis-15-octadecatrienoic oleic acid to correspond to the industrial needs. Currently,
the catalyst used in the industrial hydrogenation process is a
∗ Corresponding author. Tel.: +335 49 45 39 94; fax: +335 49 45 34 99. nickel catalyst deposited on a silica support [2–5]. Studies on
E-mail address: catherine.especel@univ-poitiers.fr (C. Especel). the hydrogenation of different vegetable oils have been also

1381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.11.017
118 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

undertaken on catalysts containing copper [6] or copper hydrogen activated on palladium particles and the lead pre-
chromite [7]. Generally, these different catalysts require a cursor (Pb(CH3 COO)2 ) dissolved in methanol (“catalytic re-
reaction temperature above 150 ◦ C and a high hydrogen pres- duction” method [26]). This preparation method of bimetallic
sure. Recently, we have studied the selective hydrogenation catalysts is well known to induce a strong interaction be-
of ethyl esters of a traditional sunflower oil (SOEE) in the tween the two metals whatever the prepared bimetallic sys-
presence of monometallic catalysts containing noble metals tem [27–29]. A known amount of the prereduced palladium
(Pd, Pt, Ru) supported over silica, which allow to work under catalyst was introduced into a reactor under nitrogen and was
gentle conditions [8]. The activity decreases in the order activated at 300 ◦ C for 1 h under hydrogen. Then the solution
Pd > Pt > Ru at 40 ◦ C under 10 bar of hydrogen. However, the of the lead precursor in methanol, previously degassed un-
three metals are not selective in cis oleic isomer since trans der nitrogen flow, was introduced onto the catalyst at room
oleic and stearic derivatives were produced in large amounts. temperature. After a 1 h reaction time under hydrogen bub-
In order to promote the selectivity toward the cis oleic isomer, bling, the solution was filtered out and the catalyst was dried
noble metal can be modified by metallic additives or by addi- overnight at 100 ◦ C. Finally, the bimetallic catalysts were re-
tion of amines in the reaction medium. The effect of amines duced under hydrogen flow at 300 ◦ C for 2 h (2 ◦ C/min heat-
during the selective hydrogenation of unsaturated hydro- ing rate). Using the same approach, a blank Pd catalyst over
carbon compounds was the subject of many studies [9–16]. SiO2 , named “Pd1 treated in CH3 OH/H2 ” further in the text,
For example, it was shown that the addition of pyridine or was also prepared in a methanol solution without addition of
quinoline allows to selectively hydrogenate the acetylenic the lead precursor.
compounds in ethylenic ones on palladium catalysts [17–23].
In the same way, the presence of amines allows to selectively
2.2. Electron microscopy
obtain monoolefins starting from diolefins on catalysts con-
taining palladium, nickel or cobalt [24,25]. These results are
Transmission electron microscopy (TEM) studies were
generally explained by a competition of adsorption between
performed on a Philips CM 120 instrument operating at
the amine and the hydrocarbon compounds and by a modifica-
120 kV. The Pd/SiO2 samples were ultrasonically dispersed
tion of the metal electronic properties resulting from an elec-
in ethanol and the suspension was brought onto a copper
tron transfer from the free nitrogen doublet toward the metal
grid with a carbon support film. The preparation of the grids
phase.
was different for the catalysts supported onto TiO2 , MgO,
Preliminary experiments concerning the addition of cop-
CeO2 and CeZrO2 because of some contrast problems. These
per or lead to palladium, and the addition of amines in
samples were embedded in a polymeric resin (spurr) and
the reaction medium were performed [8]. The present work
cut into a section as small as 40 nm with an ultramicrotome
completes the previous paper on the selective hydrogena-
equipped with a diamond knife. Cuts were then deposited
tion of ethyl esters of traditional sunflower oil on sup-
on an Al grid previously covered with a thin layer of car-
ported palladium catalysts. We study the influence of the
bon.
following parameters: the palladium particle size, the na-
EDX microanalysis of palladium and lead was carried
ture of the support, the addition of lead on palladium
out in the nanoprobe mode. The diameter of the probe is
catalyst and the addition of various amines in the reac-
5.0–7.0 nm.
tion medium during the SOEE hydrogenation on Pd/SiO2
catalyst.
2.3. Metal accessibility

2. Experimental Metal accessibilities of catalysts were determined by

chemisorption of hydrogen at 70 ◦ C. The catalysts were first
2.1. Catalyst preparation reduced under hydrogen at 400 ◦ C (2 h), then evacuated at the
same temperature under argon for 3 h. Hydrogen chemisorbs
The various supports were ground and then sieved to re- on palladium according to the stoichiometry H/Pds = 1. The
tain particles with sizes between 0.10 and 0.04 mm. The particle size (d) was determined from the metal accessibility
monometallic Pd based catalysts were prepared either via value (%D) according to Hugues et al. [30], which assim-
cationic exchange of the support (SiO2 ) by Pd(NH3 )4 (NO3 )2 ilated particles to cubes with one face in contact with the
in basic medium (pH = 11) or via simple impregnation support. The equation is the following:
(Al2 O3 , TiO2 , MgO, ZnO, CeO2 , CeZrO2 ) using the same Pd
precursor. The monometallic catalysts were dried overnight 5 × 106 M
d(Å) = (1)
at 120 ◦ C, then calcined in flowing air at 300 ◦ C for 4 h, and DSρ
finally reduced under hydrogen flow at 400 ◦ C for 4 h. The
Pd/TiO2 catalyst was reduced at 300 ◦ C instead of 400 ◦ C with M is the atomic weight (g/mol), S the surface by mol of
in order to avoid the SMSI effect. Bimetallic Pd–Pb cata- Pd metal (47,776 m2 /mol) and ρ is the metal volumic weight
lysts were prepared by the surface redox reaction between (g/cm3 ).
 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126 119

2.4. Hydrogenation of sunflower oil

Classical refined sunflower oil (from VALAGRO, Poitiers)

was transformed in presence of ethanol and sodium
ethanolate, at 60 ◦ C, into glycerol and ethyl esters. After the
separation of the reaction products by decantation, the sun-
flower oil ethyl esters (SOEEs) were washed and dried at
110 ◦ C for 3 h under vacuum. They were then purified on
a molecular distillation apparatus at 120 ◦ C (P < 10−3 mbar)
until a high purity near 98%. The hydrogenation of SOEE was
performed in liquid phase in a stirred autoclave, at constant
pressure of 10 bar and at low temperature (40 ◦ C) accord-
Fig. 2. SOEE hydrogenation vs. time at 40 ◦ C over 1 wt.% Pd/SiO2 catalyst
ing to an experimental procedure previously implemented (mcatalyst = 122 mg): (䊉) % C18:2; (
) % C18:1 (tot); (△) % cis C18:1; ()
at the laboratory [8]. A known amount of prereduced cata- % trans C18:1; () % C18:0 (mol%).
lyst was immersed into 50 mL of solvent (ethanol) without
exposure to air and introduced into the autoclave. Then a It must be recall that the aim of the study is to selectively
mixture of 75 mL of SOEE in 25 mL of solvent was poured hydrogenate C18:2 toward cis C18:1, avoiding the cis–trans
into the autoclave and flushed with nitrogen. When the de- isomerization and the complete hydrogenation to C18:0. It
sired temperature was raised, 10 bar hydrogen pressure was should be noticed that the C18:1 compounds include the var-
introduced and the stirring was switched on (zero time for ious isomers of position of the C C double bond (C C in
the reaction). In some experiments, amines were introduced positions 9, 10, 11 or 12, Fig. 1). An example of the temporal
directly in the reaction medium. Liquid samples were ana- percentage profiles of the various compounds is represented
lyzed by VPC on a Varian 3400CX chromatograph equipped in Fig. 2. The reagent to be hydrogenated consists of a mix-
with a flame ionization detector and a capillary column ture of C18:0, C18:1 and C18:2 in the following initial pro-
DB-23 (J&W, 30 m, 0.25 mm i.d.) using nitrogen as carrier portions: C18:0, 3.72%; cis-9 C18:1, 27.49%; cis-9, cis-12
gas. C18:2, 60.11%.

3.1. Particle size effect

3. Results and discussion
In order to achieve a series of catalysts displaying vari-
Fig. 1 presents the reaction pathway of SOEE hydrogena- ous Pd particle sizes, two different strategies were followed:
tion, with the various intermediates which can be observed. first, the Pd loading was increased from 0.3 to 1.5 wt.%.

Fig. 1. Reaction pathway of SOEE hydrogenation.

120 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

Table 1
Main characteristics and TOF (C18:2 hydrogenation) of silica supported monometallic and bimetallic catalysts containing palladium
Pdwt.% catalyst/SiO2 Tcalcination , Treduction Metala (wt.%) db (nm) Dc (%) MCd (×103 mol/L) TOF C18:2 hydrogenation
(◦ C) (×1019 mol/h)
Pd0.3 300, 400 0.39 1.7** 55 0.030 2.08
Pd1 300, 400 0.91 1.8** 53 0.070 1.92
Pd1.5 300, 400 1.45 2.0** 47 0.037 2.25
Pd1 treated in CH3 OH/H2 –, 300 0.91 3.5** 28 0.070 1.99
Pd1 treated in H2 O/H2 –, 500 0.91 7.8** 12 0.070 2.12
Pd1 + 0.05 wt.% Pb –, 300 0.91 (Pd), 0.04 (Pb) 3.8* – 0.070 –
Pd1 + 0.10 wt.% Pb –, 300 0.91 (Pd), 0.07 (Pb) 4.0* – 0.070 –
a Determined by ICP.
b Mean particle size deduced from electronic microscopy (*) or hydrogen chemisorption (**).
c Pd accessibility.
d MC: metallic concentration in the reaction medium during the SOEE hydrogenation.

The principal characteristics of the resulting catalysts are C18:2 hydrogenation were expressed as turnover frequencies
summarized in Table 1. Pd1.5 /SiO2 presents slightly larger (Table 1). Whatever the metallic particle size, the accessible
particle size (2.0 nm) than Pd0.3 and Pd1 /SiO2 , which are atoms of the various Pd/SiO2 catalysts seem to work at a
similar in size (mean diameter about 1.7–1.8 nm). The sec- comparable initial rate (about (2.10 ± 0.20) × 10−19 mol/h).
ond strategy was a treatment of the Pd1 /SiO2 catalyst un- This observation suggests that the particle size in the range
der hydrogen atmosphere in a methanol or water solution. 1.7–7.8 nm does not have an effect on the intrinsic rate of
This treatment was suggested from previous results from the SOEE hydrogenation. Likewise, Fig. 3b shows no no-
our group, which show a sintering of the metallic particles table evolution concerning the C18:0 product. As for the
under similar conditions on various monometallic catalysts trans C18:1 (Fig. 3c), it tends to increase for the Pd1 /SiO2
containing Pd, Pt or Rh [28,29,31]. In our case, this treat- catalyst treated in H2 O/H2 medium, which presents the
ment induces an increase of the particle size from 1.8 nm largest mean particle size of the Pd catalysts series (Table 1,
(Pd1 ) to 3.5 nm (Pd1 in methanol) and 7.8 nm (Pd1 in water) d = 7.8 nm).
(Table 1). Consequently, the rate of hydrogenation of the SOEE
Fig. 3a gives the evolution of the C18:2 conversion as a is insensitive to the palladium particle size in the range
function of time obtained on the five different monometal- 1.7–7.8 nm. However, larger particles should be avoided be-
lic catalysts. From these various curves, initial rates of cause they promote the cis–trans isomerization.

Fig. 3. SOEE hydrogenation on the various monometallic palladium/SiO2 catalysts: (a) C18:2 conversion (%) vs. time, (b) C18:0 (%) and (c) trans C18:1
(%) vs. C18:2 conversion (mol%); (䊉) Pd0.3 catalyst; () Pd1 catalyst; () Pd1.5 catalyst; () Pd1 treated in CH3 OH/H2 catalyst; (
) Pd1 treated in H2 O/H2
catalyst (for each sample, mcatalyst = 122 mg except for Pd1.5 catalyst, mcatalyst = 41 mg).
 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126 121

Table 2
Main characteristics and TOF (C18:2 hydrogenation) of Pd catalysts supported on various oxides
Pd0.3 catalyst/support Support surface Tcalcination , Treduction Metala db (nm) Dc (%) Accessible atom TOF C18:2 hydrogenation
area (m2 /g) (◦ C) (wt.%) (×10−19 gcata ) (×1019 mol/h)
SiO2 200 300, 400 0.39 1.7** 55 1.21 2.08
␣-Al2 O3 , SCS 9 9 300, 400 0.32 1.6** 59 1.07 3.52
␥c -Al2 O3 GFS 215 300, 400 0.30 1.3** 70 1.18 3.70
TiO2 48 300, 300 0.38 2.0* 47 1.01 3.29
MgO 297 300, 400 0.30 2.4* 38 0.65 2.62
ZnO 40 300, 400 0.28 3.5** 26 0.41 3.80
CeO2 127 300, 400 0.30 3.6* 26 0.44 3.69
CeZrO2 87 300, 400 0.31 3.2* 29 0.51 3.35
a Determined by ICP.
b Mean particle size deduced from electronic microscopy (*) or hydrogen chemisorption (**).
c Pd accessibility.

3.2. Support effect sicity of the various oxides established by Martin and Duprez
[32]: MgO ≫ CeO2 > ZrO2 > ␥c -Al2 O3 > SiO2 , we observe
The aim of the study was to examine the influence of the that the catalysts producing higher amounts of C18:0 and
nature of the support on the catalytic properties of supported trans C18:1 compounds (series 4d, 4f) are more basic than
palladium for the SOEE hydrogenation. The amount of Pd the alumina, silica and titanium oxide supports (series 4c,
was fixed at 0.3 wt.%. The main characteristics of the differ- 4e). Nevertheless, the results cannot be explained only by
ent catalysts are summarized in Table 2. this characteristic of the oxides because the more basic sup-
Fig. 4a and b represent the C18:2 conversion versus time port (MgO) does not lead to higher percentages of C18:0
during the SOEE hydrogenation on the various supported and trans C18:1 compared to a less basic support such as the
Pd0.3 catalysts. According to Fig. 4a and b, respectively, two cerium oxide for example.
tendencies become apparent concerning the activity in SOEE In conclusion, this study shows that the use of various ox-
hydrogenation: ides to prepare supported palladium catalysts did not improve
the selectivity of the reaction of C18:2 hydrogenation toward
• The Pd0.3 catalysts supported on the two aluminas, as well cis C18:1.
as titanium oxide, present a conversion slightly higher than
the Pd0.3 /SiO2 catalyst.
• The Pd0.3 catalysts supported on magnesium oxide, zinc 3.3. Modification of the palladium by lead addition
oxide, cerium oxide as well as cerium–zirconium, tend to
be less active than the Pd0.3 /SiO2 catalyst. Two bimetallic Pd–Pb catalysts were prepared from
the monometallic Pd1 /SiO2 catalyst by catalytic reduction
The reaction of SOEE hydrogenation being insensitive (Table 1). The blank catalyst “Pd1 treated in CH3 OH/H2 ”
to the particle size in the 12–55% dispersion range, which prepared under the same operating conditions, but in ab-
includes the whole of studied catalysts, it is then appropriate sence of the lead precursor, allows to estimate the influ-
to express the activities as turnover frequencies (Table 2). The ence of the preparation method of bimetallic catalysts by
results indicate that the support effect is really not significant catalytic reduction (CR), on the hydrogenating properties of
despite the extremely variable specific surface areas, from the Pd1 /SiO2 catalyst. EDX microanalysis of Pd–Pb/SiO2
9 m2 /g (␣-Al2 O3 ) to 297 m2 /g (MgO), and the difference in catalysts reduced at 300 ◦ C indicates that both metals are de-
acido-basic properties of the supports [32]. posited into closed contact. The mean particle size of the
Fig. 4c–f represent the evolution of the C18:0 and trans Pd–Pb catalysts is approximately 50% larger than obtained
C18:1 molar percentages, respectively, as a function of the for the parent catalyst (Table 1), indicating that some sintering
C18:2 conversion during the SOEE hydrogenation on the var- is taking place during the preparation of bimetallic catalysts.
ious catalysts separated into two series according to their be- Indeed, this phenomenon of sintering has been previously
havior compared to the Pd0.3 /SiO2 system. The first series reported in the course of bimetallic catalyst preparation by
of catalysts (Fig. 4c and e) includes the Al2 O3 supports and catalytic reduction [28,33].
TiO2 . These samples present similar selectivities in C18:0 sFor comparison with the CR bimetallic catalysts, a
and trans C18:1. The second series of catalysts MgO, ZnO, 1 wt.% Pd–0.05 wt.% Pb/SiO2 catalyst was also prepared
CeO2 and CeZrO2 (Fig. 4d and f), shows more important by a traditional technique of successive impregnations (SI)
amounts of C18:0 and C18:1 trans isomers, compared with using the same Pb(CH3 COO)2 precursor. The blank and
the percentages obtained on the Pd0.3 /SiO2 sample at iso- the bimetallic samples were reduced at 300 ◦ C for 2 h. The
conversion of C18:2. These differences cannot be related to comparison of the blank “Pd1 treated in CH3 OH/H2 ” and
the structural properties of the supports nor to the size of the the CR Pd–Pb bimetallic catalysts indicates that the addition
metallic particles. If we examine the classification of the ba- of lead involves a small decrease of the C18:2 conversion
122 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

Fig. 4. SOEE hydrogenation on the monometallic Pd0.3 catalysts (mcatalyst = 122 mg) supported on various oxides: (a and b) C18:2 conversion (%) vs. time,
(c and d) C18:0 (%) and (e and f) trans C18:1 (%) vs. C18:2 conversion (mol%); () SiO2 ; () ␣-Al2 O3 ; (䊉) ␥c -Al2 O3 ; (×) TiO2 ; (+) MgO; (♦) ZnO; ()
CeO2 ; (△) CeZrO2 .

(Fig. 5a). The result can be explained by the deposit of 3.4. Addition of amines in the reaction medium
inactive lead atoms on the surface of the palladium particles.
This decrease is all the more significant as the content The effect of amine addition was studied on the Pd1 /SiO2
of the deposited lead is high (Fig. 5a). The deposit of a catalyst. The different amines are listed in Table 3 with the
comparable lead content (0.05 wt.%) by the SI method led introduced quantities and the pKa value of their conjugated
to a more severe poisoning of the catalyst activity. This acid. The amine/Pdtotal molar ratio is also specified in Table 3
phenomenon also means a strong interaction between the for each experiment. The various amines, introduced directly
two metals introduced by successive impregnations. The in the reaction medium, are likely to be found in the hy-
different behaviors between the two modes of preparation drogenated oil after reaction. These compounds can be then
can be explained by a selective deposit of lead on particular eliminated by acidic hydrolysis. However, it should be spec-
sites of palladium when this one is introduced by catalytic ified again that the considered application is located in the
reduction. Similar results have been observed during the nonfood field.
selective hydrogenation of isoprene on bimetallic catalysts
PdFe/Al2 O3 prepared by successive impregnations and by 3.4.1. Addition of primary aliphatic amines
a surface controlled reaction [34]. Moreover, the addition of The hydrogenation tests were performed in the presence
lead by both method of preparation does not influence the of methylamine and butylamine (Table 3, entries 1 and 2) in
proportion in C18:0 (Fig. 5b) but decreases the formation of order to examine the influence of the length of the alkyl chain
trans C18:1 isomers (Fig. 5c). Then, the addition of lead is of the amine on the activity and selectivity of the Pd1 /SiO2
beneficial to limit the C18:1 cis–trans isomerization during catalyst. In addition, different amounts of these amines were
the SOEE hydrogenation on palladium. Furthermore, the introduced to observe the effect of their concentration on the
hydrogenation rate of palladium is slightly inhibited when reaction. Fig. 6a represents the C18:2 conversion in presence
this inactive additive is introduced by catalytic reduction. of methylamine as a function of time. From this figure, it can
 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126 123

Fig. 5. SOEE hydrogenation on bimetallic Pd–Pb/SiO2 catalysts (mcatalyst = 122 mg): (a) C18:2 conversion (%) vs. time, (b) C18:0 (%) and (c) trans C18:1 (%)
vs. C18:2 conversion (mol%); () Pd1 /SiO2 ; () Pd1 treated in CH3 OH/H2 ; () Pd1 –Pb0.04 /SiO2 (CR); (△) Pd1 –Pb0.07 /SiO2 (CR); (+) Pd1 –Pb0.05 /SiO2 (SI).

be seen that the addition of methylamine strongly inhibits the catalyst (Fig. 6b). Moreover, in the case of the addition of
activity of palladium, this effect being all the more significant small quantity of butylamine (0.15 mL), no deactivation is
as the quantity of methylamine is high. On the other hand, observed. Concerning the C18:0 (Fig. 6d) and trans C18:1
the introduction of this amine does not have any effect on the (Fig. 6f), these products present the same evolution in pres-
percentage of C18:0 (Fig. 6c). Moreover, Fig. 6e indicates ence of methylamine or butylamine: no inhibition of the
that the presence of methylamine allows to decrease the for- formation of C18:0 but a significant decrease of the trans
mation of trans C18:1. However, the increase of the quantity C18:1.
of this amine does not induce a more significant drop in the In conclusion, addition of primary aliphatic amines in the
formation of trans C18:1. reaction medium induces a more or less important deactiva-
The same kind of experiments conducted in the presence tion according to the length of the chain of the amine and the
of butylamine shows that even the addition of 0.40 mL of introduced quantity. For a similar amine/Pdtotal molar ratio,
this amine (leading to an amine/Pdtotal molar ratio compa- the methylamine deactivates more strongly the catalyst than
rable to the addition of 0.15 mL of methylamine, Table 3) the butylamine. The pKa values of the conjugated acid of these
does not induce such a strong deactivation of the Pd1 /SiO2 two amines being comparable, this result can be explained by

Table 3
Quantities of introduced amines in the reaction medium during the SOEE hydrogenation on the Pd1 /SiO2 catalyst, amine/Pdtotal molar ratios and values of the
pKa of the conjugated acid of each amine
Entry Amine pKa Formula Introduced volume (mL) Amine/Pdtotal molar ratio
1 Methylamine 10.65 H3 C NH2 0.15, 0.30 377, 755
2 Butylamine 10.77 H3 C (H2 C)3 NH2 0.15, 0.25, 0.40 131, 217, 351

3 Triethylamine 10.70 0.10, 0.20, 0.25, 0.60 62, 124, 155, 347

4 Piperidine 11.12 0.30, 0.60 264, 528

5 Pyridine 5.21 0.20, 0.70 215, 752

124 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

Fig. 6. SOEE hydrogenation on the Pd1 /SiO2 catalyst in presence of primary amines: (a and b) C18:2 conversion (%) vs. time, (c and d) C18:0 (%) and (e and
f) trans C18:1% vs. C18:2 conversion (mol%); () 0.15 mL methylamine; (△) 0.30 mL methylamine; (䊉) 0.15 mL butylamine; () 0.25 mL butylamine; (+)
0.40 mL butylamine; () without amine.

a higher coverage of the metallic phase by the methylamine 3.4.2. Addition of tertiary aliphatic amines
because of its shorter carbonaceous chain. The more com- In order to examine the assumption of the steric hindrance
pact fitting of the methylamine on the surface of palladium of the amines on the catalyst surface, the study was continued
particles would thus inhibit more strongly the adsorption of in the presence of a tertiary aliphatic amine, the triethylamine
the C18:2 unsaturated ethyl ester. Concerning the selectivity (Table 3, entry 3). The conversion of C18:2 in presence of
of the reaction, the addition of these two amines has no effect different amounts of triethylamine is given in Fig. 7a. From
on the formation of the C18:0 saturated ester, but allows to this figure, it can be seen that contrary to the addition of
decrease the formation of the trans C18:1 compound. methylamine or butylamine, the presence of triethylamine

Fig. 7. SOEE hydrogenation on the Pd1 /SiO2 catalyst in presence of different amounts of triethylamine: (a) C18:2 conversion (%) vs. time, (b) trans C18:1
(%) vs. C18:2 conversion (mol%); () 0.10 mL; () 0.20 mL; (䊉) 0.25 mL; (△) 0.60 mL; () without amine.
 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126 125

Fig. 8. SOEE hydrogenation on the Pd1 /SiO2 catalyst in presence of cyclic amines: (a and b) C18:2 conversion (%) vs. time and (c and d) trans C18:1 (%) vs.
C18:2 conversion (mol%); (䊉) 0.30 mL piperidine; () 0.60 mL piperidine; () 0.20 mL pyridine; (△) 0.70 mL pyridine; () without amine.

does not inhibit the C18:2 hydrogenation. Concerning the 3.4.3. Addition of cyclic amines
selectivity of the reaction, no evolution of the C18:0 was The cyclic amines were often used as additives in the reac-
observed while the trans C18:1 decreases after addition of tion medium during selective hydrogenations [13,14,17–23].
the triethylamine (Fig. 7b). Two various cyclic amines are presented in this work, the
These results confirm that the effects observed on piperidine and the pyridine. The introduced quantities are
the hydrogenating activity of a Pd/SiO2 catalyst depend listed in Table 3, entries 4 and 5, respectively. The C18:2 con-
on the nature of the aliphatic amine. For a compara- versions versus time on the Pd1 /SiO2 catalyst are reported in
ble amine/Pdtotal molar ratio, the inhibiting effect of the Fig. 8a (piperidine) and b (pyridine). We can notice that the
amine on the C18:2 conversion is all the more high as presence of cyclic amine involves a more or less important de-
the hydrocarbon fragment of the amine is small: methy- crease of the catalyst activity. In particular, the pyridine does
lamine > butylamine > triethylamine. In addition, the pKa not deactivate the Pd1 /SiO2 catalyst for a amine/Pdtotal molar
value of the conjugated acid of these three amines being sim- ratio = 215 (Fig. 6b, 0.20 mL introduced) whereas piperidine
ilar, these different behaviors can be explained by a lower involves a drop of conversion for a comparable amine/Pdtotal
coverage of palladium surface by the amine (amine/Pdsurface molar ratio (Fig. 6a, 0.30 mL introduced, molar ratio = 264).
ratio) having the highest steric hindrance. The absence of in- These differences could be explained by the weaker pKa value
hibiting effect on the conversion of the C18:2 in the case of of the conjugated acid of pyridine compared to piperidine and
triethylamine probably results from a donor effect of nitro- thus by a weaker adsorption of the former on the palladium
gen toward the metal (positive electronic effect) which com- surface.
pensates the loss of active surface related to the amine ad- Concerning the reaction products, the presence of the
sorption (negative geometric effect). So, the metallic surface cyclic amines does not involve modifications on the forma-
which is accessible to the reagents decreases in the presence tion of the C18:0 on the Pd1 /SiO2 catalyst. On the other hand,
of amines but it would be more active because of a weaker the two cyclic amines induce a diminution of the trans C18:1
adsorption of C18:2 resulting from the donor effect of the (Fig. 8c and d). These evolutions are similar with those pre-
amine. Indeed, kinetic studies have shown a zero order for viously observed in the case of the aliphatic amines.
the hydrogenation of this ester meaning a strong adsorption
on the metal [35]. This modification of the electronic proper-
ties of the metal also would explain the limited formation of 4. Conclusion
trans C18:1. Beneficial effects of nitrogen containing com-
pounds addition were already observed during the selective The objective of this work was to selectively hydrogenate
liquid phase hydrogenation of alkynes, explained by a ligand C18:2 ethyl esters of the sunflower oil (SOEE) toward
effect, which would decrease the complexation strength of C18:1 cis ethyl oleate on supported palladium catalysts, for
the highly unsaturated hydrocarbons [9–11]. a nonfood use such as lubricants. The influence of various
126 B. Nohair et al. / Journal of Molecular Catalysis A: Chemical 229 (2005) 117–126

parameters related to the catalyst structure was evaluated in [5] D. Jovanovic, R. Radovic, L. Mares, M. Stankovic, B. Markovic,
order to improve the selectivity of the reaction. In the range Catal. Today 43 (1998) 21.
[6] N. Ravasio, F. Zaccheria, M. Gargano, S. Recchia, A. Fusi, N. Poli,
of studied dispersions (12–55%), the reaction of SOEE
R. Psaro, Appl. Catal. A 233 (2002) 1.
hydrogenation is insensitive to the size of the palladium [7] C. Fragale, M. Gargano, N. Ravasio, M. Rossi, I. Santo, Inorg. Chim.
particles deposited on silica. On the other hand, the largest Acta 82 (1984) 157.
metallic particles enhance the C18:1 cis–trans isomerization. [8] B. Nohair, C. Especel, P. Marécot, C. Montassier, L.C. Hoang, J.
In addition, the use of various oxides (␣-Al2 O3 , ␥c -Al2 O3 , Barbier, C.R. Chim. 7 (2004) 113.
[9] J.P. Boitiaux, J. Cosyns, S. Vasudevan, Appl. Catal. 15 (1985) 317.
TiO2 , MgO, ZnO, CeO2 , CeZrO2 ) to deposit palladium did
[10] J.P. Boitiaux, J. Cosyns, E. Robert, Appl. Catal. 49 (1989) 219.
not improve the selectivity of the reaction obtained in the [11] J.P. Boitiaux, J. Cosyns, E. Robert, Appl. Catal. 49 (1989) 235.
presence of the Pd/SiO2 catalyst. By opposite, some supports [12] T.A. Nijhuis, G. van Koten, J.A. Moulijn, Appl. Catal. 238 (2003)
led to higher selectivities in trans C18:1. The modification of 259.
palladium by lead addition showed that this inactive deposit [13] W. Oroshnik, A.D. Mebane, J. Am. Chem. Soc. 76 (1954) 5719.
improves the selectivity in cis C18:1, the conversion being [14] W. Oroshnik, G. Karmas, A.D. Mebane, J. Am. Chem. Soc. 74
(1952) 3807.
only slightly decreased when this additive is introduced [15] U.O.P. US Patent 3,433,842, 1969.
by surface redox reaction (catalytic reduction). Finally, [16] J.P. Boitiaux, J. Cosyns, G. Martino, Stud. Surf. Sci. Catal. 11 (1982)
the introduction of amines modifies the hydrogenating 355.
properties of supported palladium catalyst. According to the [17] O. Isler, W. Huber, A. Ronco, M. Kefler, Helv. Chim. Acta 30 (1974)
quantity and the nature of the added amine (aliphatic with 1911.
[18] E.N. Marvell, T. Li, Synthesis 8 (1973) 457.
linear or ramified chain, cyclic compounds), the activity of [19] E.N. Marvell, T. Gosink, P. Churchley, J. Org. Chem. 37 (1972)
the Pd/SiO2 catalyst can either be unchanged or inhibited. 2989.
These evolutions would result from a promoter electronic [20] Sokol, Skaya, Kinet. Katal. 12 (1971) 525.
effect generated by the presence of the amine and from a [21] D.V. Sokol’Skii, P. Gadenov, Jzv. Akad. Nauk. Ser. Khim. 20 (1970)
poison geometric effect related to adsorption of this nitrogen- 62.
[22] H. Lindlar, Helv. Chim. Acta 35 (1952) 446.
containing compound on the palladium surface. Whatever the [23] H. Lindlar, Org. Synth. 46 (1966) 89.
nature of the amine, the selectivity in cis C18:1 is improved [24] Phillips Petroleum, US Patent 3,637,877, 1972.
compared to the nonmodified monometallic Pd catalyst. [25] Phillips Petroleum, US Patent 3,636,175, 1972.
In conclusion, this study relating to the selective hydro- [26] J. Barbier, Advances in Catalyst Preparation, Study Number 4191CP,
genation of ethyl ester of sunflower oil on palladium catalysts Catalytica Studies Division, Mountain View, 1992.
[27] C.L. Pieck, P. Marécot, J. Barbier, Appl. Catal. A 145 (1996) 323.
shows that it is difficult to be freed from the secondary re- [28] G. Lafaye, C. Micheaud-Especel, C. Montassier, P. Marécot, Appl.
action of cis–trans isomerization of the required compound Catal. A 230 (2002) 19.
(C18:1). All the results obtained in this work show that the [29] F. Epron, F. Gauthard, J. Barbier, Appl. Catal. A 237 (2002) 253.
best control of this selectivity was obtained by introduction [30] T.R. Hugues, R.H. Houston, R.P. Sieg, in: Proceedings of the 135th
of amines into the reaction medium. Natural Meeting, Am. Chem. Soc. Div. Petrol Chem., Prep. 4 (1959)
33.
[31] A. Douidah, P. Marécot, J. Barbier, Appl. Catal. A 225 (2002) 11.
[32] D. Martin, D. Duprez, J. Mol. Catal. 118 (1997) 113.
References [33] A. Douidah, P. Marécot, S. Labruquere, J. Barbier, Appl. Catal. A
210 (2001) 111.
[1] P.C. Naegely, Seed Oils Future (1992) 14. [34] R. Bachir, P. Marécot, B. Didillon, J. Barbier, Appl. Catal. A 164
[2] J.T. Richardson, R.J. Dubus, J. Catal. 54 (1978) 207. (1997) 313.
[3] Y. Nitta, T. Imanaka, S. Teranishi, J. Catal. 96 (1985) 429. [35] G. Cecchi, R.P. Dennis, P.J. Player, E. Ucciani, Rev. Fr. Corps Gras
[4] S.M. Echeverria, Appl. Catal. 66 (1990) 73. 31 (6) (1984) 243.