Hydrogenation of fatty esters over ruthenium-tin catalysts; characterization and identification of active centers

Article abstract of DOI:10.1006/jcat.1998.2044

The selective hydrogenation of methyl oleate into oleyl alcohol was performed over RuSn/alumina catalysts. A maximum yield of unsaturated alcohol was obtained for a bulk atomic ratio to Sn/Ru = 4. The side reaction of the transesterification between methyl oleate and oleyl alcohol leading to oleyl oleate was strongly decreased. The ruthenium-tin catalysts prereduced with NaBH₄ have been characterized by TPR and XPS analysis. The addition of tin increased the content of Ru° on the catalyst surface without modification of the total Ru content. When the catalysts were reduced with NaBH₄, a decrease of the tin and ruthenium contents on the surface was observed. On the catalyst surface were found two separate tin oxide species which differed in the oxidation state of tin. The proportion and the repartition of each species was a function of the total tin content. Over such catalysts, the reaction scheme involved three steps: (1) the hydrogenation of the methyl oleate into the oleyl alcohol, (2) the transesterification reaction between the methyl oleate and the oleyl alcohol with the formation of the heavy oleyl oleate ester, (3) the hydrogenation of this heavy ester into oleyl alcohol. The first and the third steps could involve mixed ruthenium-tin sites while the second step could require tin species without an interaction with ruthenium. If it is assumed that the rate determining step is the oleyl oleate hydrogenation, the results of this study show that active centers corresponding to a stoechiometry Ru/SnOx = 2 could be involved in this reaction. The activation of esters (methyl oleate and oleyl oleate) would occur via a hemiacetal intermediate over these mixed sites where the more reduced tin species could be close to the ruthenium.

Full text of DOI:10.1006/jcat.1998.2044

JOURNAL OF CATALYSIS 176, 215–224 (1998)
ARTICLE NO. CA982044
Hydrogenation of Fatty Esters over Ruthenium–Tin Catalysts;
Characterization and Identification of Active Centers
,1
ꢀ
ꢀ
ꢀ
Y. Pouilloux, F. Autin, C. Guimon,† and J. Barrault
ꢀ
Laboratoire de Catalyse, UMR 6503, ESIP, 40, Avenue du Recteur Pineau, 86022 Poitiers, France; and †Laboratoire
de Physicochimie mole´culaire, UMR 5624, 64000 Pau, France
Received October 23, 1997; revised February 9, 1998; accepted February 9, 1998
==
the activation of the (C O) carbonyl groups without the
==
reduction of the C C bond. On the other hand, the hy-
The selective hydrogenation of methyl oleate into oleyl alcohol
was performed over RuSn/alumina catalysts. A maximum yield
of unsaturated alcohol was obtained for a bulk atomic ratio to
Sn/Ru = 4. The side reaction of the transesterification between
methyl oleate and oleyl alcohol leading to oleyl oleate was strongly
decreased. The ruthenium–tin catalysts prereduced with NaBH4
have been characterized by TPR and XPS analysis. The addition
of tin increased the content of Ruꢁ on the catalyst surface without
modification of the total Ru content. When the catalysts were re-
duced with NaBH4, a decrease of the tin and ruthenium contents
on the surface was observed. On the catalyst surface were found
two separate tin oxide species which differed in the oxidation state
of tin. The proportion and the repartition of each species was a
function of the total tin content. Over such catalysts, the reaction
scheme involved three steps: (1) the hydrogenation of the methyl
oleate into the oleyl alcohol, (2) the transesterification reaction bet-
ween the methyl oleate and the oleyl alcohol with the formation of
the heavy oleyl oleate ester, (3) the hydrogenation of this heavy ester
into oleyl alcohol. The first and the third steps could involve mixed
ruthenium–tin sites while the second step could require tin species
without an interaction with ruthenium. If it is assumed that the
rate determining step is the oleyl oleate hydrogenation, the results
of this study show that active centers corresponding to a stoechiom-
etry Ru/SnOx = 2 could be involved in this reaction. The activation
of esters (methyl oleate and oleyl oleate) would occur via a hemi-
acetal intermediate over these mixed sites where the more reduced
c
drogenation of esters is more difficult to achieve since the
carbonyl group of ester is less active than the hydrogenation
of a ketone and a aldehyde group.
Generally, these works concerned the hydrogenation of
light esters (5, 6). When using similar catalysts (noble metal
with promoter) for the hydrogenation of more complex
unsaturated esters, it was observed that the selectivity to
unsaturated alcohols could also be improved, i.e., oleyl
alcohol (9-octadecen-1-ol) from methyl oleate (methyl-9-
octadecenoate). But we also demonstrated that the re-
action scheme of the methyl oleate hydrogenation over
RuSnB/Al2O3 (7) catalyst is more complex than the one
proposed by Narasimhan et al. (8). We have observed the
formation of heavy esters (i.e., oleyl oleate) and an unsatu-
rated alcohol yield lower than the one obtained in previous
works (9). Indeed, Scheme 1 shows that oleyl alcohol can re-
act with methyl oleate to form oleyl oleate which decreases
the alcohol yield at least at the beginning of the reaction. It
seems that tin species dispersed on alumina favor the side
reaction of transesterification and that the hydrogenation
of heavy esters (giving two moles of unsaturated alcohol)
is the rate-limiting step of the whole process.
Inordertogetfurtherinformationsonthereactionmech-
anism (including the nature of active sites) we investigate
detailed characterization of RuSn catalysts. In the present
paper, the surface and bulk characteristics of samples differ-
ing in their ruthenium or tin content and the prereduction
process are established and discussed in relation with the
catalytic properties.
tin species could be close to the ruthenium.
ꢂ 1998 Academic Press
Key Words: hydrogenation; fatty esters; unsaturated alcohols
synthesis; ruthenium–tin catalysts.
INTRODUCTION
The selective formation of unsaturated alcohols from
the hydrogenation of ꢀ,ꢁ unsaturated carbonyl compounds,
particularly ꢀ,ꢁ unsaturated aldehydes, in the presence of
solid catalysts has been previously reported (1–4).
This reaction requires the control of the chemioselec-
tivity since the formation of unsaturated alcohol involves
EXPERIMENTAL
Catalytic test. The hydrogenation reaction was car-
ried out in a stainless steel batch reactor (300 ml). The
reagent (100 ml of methyl oleate) and the catalyst (2.2 g)
were introduced into the reactor under ambient condi-
tions. The methyl oleate, supplied by Stearinerie Dubois
with a purity of 80%, also contained other methyl esters:
1 Corresponding author. E-mail: joel.barrault@esip.univ-poitiers.fr.
215
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

216
POUILLOUX ET AL.
SCHEME 1. Reaction scheme of the hydrogenation of methyl oleate in the presence of RuSn/Al2O3 catalyst.
C14 = 1.5%, C16 = 15%, C18 = 80%. The reactor was then aqueous solution of sodium borohydride. The solids were
purged (4 times) with nitrogen at 5 MPa and stirred conti- then filtered, washed with water, and dried at 120ꢁC under
nuously. The temperature was slowly increased up to 270ꢁC nitrogen for 4 h.
at constant pressure (5 MPa), nitrogen was substituted for
Before use the catalysts were reduced with hydrogen at
hydrogen and the pressure increased up to (8 MPa) was 400ꢁC and passivated with air at room temperature.
maintained during the reaction.
XPS characterization. (ꢀ) X-ray photoelectron spec-
troscopy (XPS) analyses were performed with an SSI (Sur-
Analysis. Liquid samples were mixed with dodecane
face Science Instruments, Mountain View, CA) model 301
and analyzed by a GPC equipped with FID and a
spectrometer with focused (diameter of the irradiated area
Chrompack Cp Sil-5 column (L, 25 m; ID, 0.25 mm; thick-
(600 ꢂm)) monochromatic Alk ꢀ radiation (10 kV, 10 mA)
ness of film, 0.11 ꢂm). The carrier gas was nitrogen.
and coupled with a glove-box which was used for the trans-
fer of the samples reduced with hydrogen and passivated
Catalyst Preparation
for 10 min under air. The residual pressure inside the ana-
Coimpregnation and reduction with sodium borohydride. lysis chamber was about 5 ꢅ 10ꢄ8 Pa. The calibration of the
The catalysts were prepared by coimpregnation of the sup- spectra was performed with the Al 2p line (74.4 eV) from
port (Al2O3 GFSC from Rhoˆne Poulenc, area 200 m2 ꢃ gꢄ1
)
an Al2O3 support.
The XPS peaks were decomposed into subcomponents
with RuCl3 and/or SnCl2, according to the method de-
scribed by Narasimhan et al. (10). After the impregnation usingaGaussian(80%)–Lorentzian(20%)curvefitting pro-
step (12 h), the metallic precursors were reduced with an gram with a nonlinear background (11). The quantitative

HYDROGENATION OF FATTY ESTERS
217
analyses were performed with the sensitivity factors given
TABLE 1
by Scofield (12).
XPS Analysis of the RuSnB/Al2O3 Catalysts: Effects of the
Tin Content
Temperature-programmed reduction. TPR experiments
were performed in a conventional flow system using a ther-
mal conductivity detector (TCD). The catalytic sample was
swept by Argon at 300ꢁC for2 h. Aftercooling, it was heated
under hydrogen at a heating rate of 4ꢁC/min from 25ꢁC to
450ꢁC.
Ru
Sn
RuSnB/Al2O3
Ru(2)B(0.7)
Ru(1.5)Sn(4.1)B(1.1) 0.60
Ru(1.6)Sn(8.5)B(1.1) 0.48
(at%) (at%) Sn/Rua Ru/Al Sn/Al Sn/Ru
0.53
—
1.8
3.0
3.0
0
0.014
—
—
3.0
6.2
6.9
2.4
4.4
5.5
0.016 0.048
0.014 0.087
0.012 0.082
H2 or CO chemisorption. Chemisorption of H2 or CO
has been performed using a conventional pulse system op-
erating at room temperature. After reduction with H2 at
400ꢁC (12 h), a He stream was passed over the sample at
the same temperature and the catalyst was cooled down to
room temperature. A fixed amount of gas was then peri-
odically injected into the He stream over the sample and
hydrogen or carbon monoxide not adsorbed on the catalyst
was analyzed with a TC cell.
Ru(1.5)Sn(10)B(1.3)
0.43
a Atomic ratio determined by elementary analysis.
erature of the reduction step increases with the Sn/Ru ratio.
The H2 and CO chemisorption values confirm that tin in-
hibits totally the adsorption both of hydrogen and of CO
(7).
According to the TPR results, we can say that there are
three kinds of catalysts, depending on the Sn/Ru ratio:
Nomenclature. All catalysts are referenced as
(i) the RuSn systems (0 ꢆ Sn/Ru ꢆ 1) which are easily
Ru(x)Sn(y)B(z)/Al2O3,
ꢁ
ꢇ
reducible (Tmax 175 C).
=
(ii) the RuSn systems (2 ꢆ Sn/Ru ꢆ 5) for which only
one reduction peak (200ꢁC < T < 250ꢁC) is detected.
(iii) the RuSn system (Sn/Ru > 7) for which a reduction
at higher temperature is observed. From these results a
modification of the accessibility and of the electronic pro-
perties of ruthenium can be expected.
wherex, y, andzaretheweightcontent(%)ofeachelement.
RESULTS
Effect of the Addition of Tin
In order to examine the effect of the tin content on
bimetallic RuSn/Al2O3 catalysts, we have prepared a series
by varying the tin content.
XPS analysis of the RuSn/Al2O3 catalysts. The addition
of tin has no effect on the surface ruthenium content since
theRu/Alismoreorlessconstant(0.014 ꢈ 0.002)(Table1).
On the other hand, a surface tin enrichment can be ob-
served, the surface Sn/Ru ratio varies from 3 to 7 instead
of 2.4 to 5.5 for the bulk Sn/Ru ratio (Table 1). This result
could indicate the presence of isolated tin oxides without
any interaction with ruthenium. The percentage of all the
species detected on the solid surface (Fig. 2) reported in
Table 2 shows that, in the presence of tin, there is only
zerovalent Ruꢁ species varying by binding energy. Under
Temperature-programmed reduction of RuSn/Al2O3 cata-
lysts. In Fig. 1, the effect of the tin content on TPR pro-
files of RuSn–alumina is presented. First, we can see, since
there is only one reduction peak, that ruthenium and tin are
in strong interaction. Furthermore, the addition of tin de-
creases the hydrogen consumption and the maximum temp-
TABLE 2
XPS Analysis of RuSnB/Al2O3: Influence of the Tin Content
Ru 3d5/2
Ruꢁ (%)
Sn 3d5/2
RuSnB/Al2O3
Ru(2)B(0.7)a
SnOx (%)
SnOy (%)
—
100
—
—
20
30
25
35
—
80
70
75
65
Ru(1.5)Sn(4.1)B(1.1)a
Ru(1.6)Sn(8.5)B(1.1)a
Ru(1.5)Sn(10)B(1.3)a
Ru(1.5)Sn(4.1)B(1.1)b
100
100
100
100
—
—
—
BE (eV)
279.3
280.1
484.9
486.7
FIG. 1. TPR profiles of RuSnB/Al2O3 catalysts and effect of the tin
content.
a Reduced under H2 at 400ꢁC and passivated 10 min.
b Reduced under H2 at 400ꢁC.

218
POUILLOUX ET AL.
FIG. 2. XPS spectra of the Ru 3d of the RuSn/Al2O3 catalysts. Influence of the Sn/Ru ratio: (a) Sn/Ru = 0; (b) Sn/Ru = 2.4; (c) Sn/Ru = 4.4;
(d) Sn/Ru = 5.5.
these conditions, the hydrogenating character of metal par- pare the results with that observed over passivated samples
ticles should not be modified. It can also be noticed that (Table 2).
two kinds of tin oxides (SnOx and SnOy with 0 < x < y)
Over reduced RuSn catalyst, the content of SnOx species
are formed and that the relative amount of these oxides is more significant than over passivated materials, showing
varies with the tin content (Fig. 3). Indeed, a maximum of that they could have a role during the passivation.
SnOx (30%) is obtained for a Sn/Ru ratio of 4 (Table 2).
Contrary to Narasimhan’s previous propositions (10), we
Hydrogenation of the methyl oleate. Figure 4 shows that
did not detect any zerovalent tin (Snꢁ) at the surface of the addition of tin leads to a modification of the hydro-
these solids nor any alloys or intermetallics (RuxSny) that genating properties of ruthenium. But the increase of the
Narasimhan (10) or Ferretti (13) assumed to be present. But tin content does not change significantly the catalyst ac-
we can observe that for SnOx the binding energy of Sn 3d is tivity. However, Figs. 5 and 6 show that the selectivity is
not very different of the binding energy of Snꢁ (BE = 484, strongly modified by the addition of tin. Figure 5 presents
5 eV). Some experiments have been done with samples pre- the variation, at different methyl oleate conversions, of
reduced in situ with hydrogen (Fig. 3d); so that we can com- the unsaturated alcohol selectivity with the Sn/Ru atomic
FIG. 3. XPS spectra of the Sn 3d of the RuSn/Al2O3 catalysts reduced with H2 and passivated 10 min under air. Influence of the Sn/Ru ratio:
(a) Sn/Ru = 2.4; (b) Sn/Ru = 4.4; (c) Sn/Ru = 5.5; (d) Sn/Ru = 2.4 reduced in situ and not passivated.

HYDROGENATION OF FATTY ESTERS
219
FIG. 6. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalysts. Variation of the selectivity to heavy esters with
the Sn/Ru ratio at different conversions.
FIG. 4. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalysts. Variation of the activity with Sn/Ru ratio.
ratio. The selectivity to oleyl alcohol is maximum for a
Sn/Ru ratio of 4 and in these conditions a yield of 75% of
unsaturated alcohol is obtained for a conversion of about
90%, the formation of heavy esters being very low (Fig. 6).
These facts clearly show that there is: (i) an increase of
the hydrogenation rate of the oleyl oleate which seems
to be the limiting step; (ii) a limitation of the side reac-
tion of transesterification; (iii) the selectivity to the satu-
rated alcohol (stearyl alcohol) is even low whatever the tin
content.
tial leaching of the catalyst with the removal of ruthenium
during the reduction step and an increase of the Sn/Ru ra-
tio. On the other hand, the boron content increases to about
1% with the addition of sodium borohydride.
Temperature-programmed reduction of RuSnB systems.
Figure 7 presents the TPR profiles of RuSnB solids after
a prereduction with a progressive increase of the amount
of sodium borohydride (NaBH4/Ru). We can see that the
increase of the NaBH4 content modifies the hydrogen con-
sumption.
First, we observed a decrease of the H2 consumption
which could indicate that the species formed during the
reduction with NaBH4 are either less reducible or not reox-
idized during the passivation step. Indeed, once the ruthe-
nium and tin are totally oxidized after the passivation with
air, the H/Ru ratio cor+re4sponding to the reduction of (Ru+4
to Ruꢁ) and two (Sn to Sn+2) is equal to 8 (Table 3).
After a prereduction with a small amount of NaBH4, the
H/Ru ratio first increases and then drops to a value close to
Effect of the NaBH4 Prereduction
In order to study the effect of the reducing agent on the
activity and the selectivity of the RuSnB/Al2O3 catalysts
and the effect of boron, the amount of sodium borohydride
used during the reduction step was varied.
Table 4 shows that the increase of NaBH4 involves a de-
crease of the ruthenium content. This could be due to a par-
FIG. 5. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalysts. Variation of the selectivity to unsaturated alco-
hol with the Sn/Ru ratio at different conversions.
FIG. 7. TPR profiles of RuSnB/Al2O3 systems. Influence of the pre-
reduction with increasing amount of NaBH4.

220
POUILLOUX ET AL.
TABLE 3
Influence of NaBH4 Amount on the H2 Consumption on the Formation Rate of Unsaturated Alcohol
and the Selectivity to Unsaturated Alcohol over RuSnB/Al2O3 Catalysts
runsat. alcohol
Selectivityb to
unsat. alcohol (%)
ꢄ1
RuSnB/Al2O3
Sn/Rua
NaBH4/Ru
H/Ru
(mol/h ꢃ g
)
Ru
Ru(1.9)Sn(4.3)
—
2.2
2.4
3.3
0
4.2
12.7
38.0
9.8
10.4
5.1
0.26
0.32
0.37
0.49
27
28
45
34
Ru(1.6)Sn(4.2)B(0.3)
Ru(1.5)Sn(4.1)B(1.1)
Ru(1.2)Sn(4.6)B(1.1)
3.2
a Atomic ratio.
b Calculated at 70%; PH2 = 8 MPa, temperature = 270ꢁC, mcata = 2.2 g.
3. These results could indicate that after the prereduction cific surface area since the initial activities are comparable
with a large amount of sodium borohydride, new species (Table 5). However, the selectivity to unsaturated alcohol
like alloys (RunSnp) are formed at the surface of the cat- and to heavy esters (Table 5) are significantly changed. In-
alyst. Indeed, tin and ruthenium, which are practically not deed, the oleyl oleate formation is more significant when
miscible, can give different kinds of alloys such as RuSn3, the ruthenium and tin contents are increased. It has been
RuSn2, Ru3Sn7 (14).
suggested that the distribution of active species is modi-
fied when the ruthenium and tin contents are increased.
Tin species without interaction with ruthenium or covering
ruthenium particles are more numerous at the surface thus
leading to an increase of the transesterification reaction and
the formation of heavy esters.
XPS analysis of RuSnB catalysts. Although on the solid
surface the tin content is always larger than the amount
of ruthenium, the reduction with NaBH4 leads to a rela-
tive ruthenium surface enrichment. The Ru/Al ratio is dou-
bled which means that the (Sn/Ru)surf. ratio decreases after
the NaBH4 reduction from 3.7 to 3.0. Moreover, the XPS
analysis also showed that besides the formation of metallic
ruthenium particles (Ruꢁ), two kinds of tin oxides (SnOx
and SnOy) were also formed in the same proportion 20/80
(Table 4).
DISCUSSION
In this work, we have studied the selective hydrogena-
tion of methyl oleate into oleyl alcohol in the presence of
bimetallic RuSn catalysts supported on alumina. The selec-
tive hydrogenation of fatty unsaturated esters (or acids),
via the reaction scheme proposed in this paper, is not fully
understood but apparently both the basic and the hydro-
gen adsorption properties of the catalyst are involved in a
multistep process.
Catalyst activity. The rate of formation of oleyl alcohol
rises with the sodium borohydride content (Table 3) with-
out significant change of the total activity. We can notice
that the unsaturated alcohol selectivity is higher when the
NaBH4/Ru ratio is equal to 12 (Table 3). Then the varia-
tions of catalytic properties could be due to the change of
the Sn/Ru ratio after the NaBH4 reduction.
With regard to the overall activity (Fig. 4), Narasimhan
in previous works (8) shows that the addition of tin leads to
a decrease of activity. Ferretti (13) proposed, on one hand,
that tin acts as a poisoning agent by decreasing the hydro-
genating and the hydrogenolysis activity of rhodium and on
the other that tin participates to the activation of carbonyl
bond (increasing the overall activity). Furthermore, to jus-
Effect of the Ruthenium Content
When the ruthenium content is increased (ꢅ 2), there is
no modification of the activity of the catalyst nor of the spe-
==
tify the change of the C O bond polarization the change of
the electronic properties of the metal particles close to tin
species has been suggested by numerous authors (15, 16).
However, the percentage of oleyl alcohol is maximum for
a Sn/Ru ratio of 4 at which the heavy esters issued from the
transesterification reaction are strongly decreased (Fig. 8).
The other products are the saturated alcohol and the methyl
stearate.
TABLE 4
XPS Analysis of RuSnB/Al2O3: Effect of the Reduction
with NaBH4
Sn/Ru
Ru 3d5/2
Sn 3d5/2
RuSnB/Al2O3
Ru(1.9)Sn(4.3)
Bulk XPS Ru/Al Sn/Al Ruꢁ
SnOx SnOy
1.9 3.7 0.008 0.030 100
Ru(1.5)Sn(4.1)B(1.1) 2.4 3.0 0.016 0.050 100
20
20
80
80
According to these results, one can assume that the tin
species located at the solid surface change (Scheme 2):
BE (eV) 279.3
485 486.7
(i) the rate of the direct formation of oleyl alcohol (k1)

HYDROGENATION OF FATTY ESTERS
221
TABLE 5
Effect of the Ruthenium Content; Characteristics of the RuSnB/Al2O3 Catalysts
Selectivityb (%)
Activitya
BET area
RuSnB/Al2O3
Sn/Ru
(mol ꢃ hꢄ1 ꢃ gꢄR1u
)
(m2 ꢃ gꢄ1
)
Unsat. Al.
Heavy esters
Ru(1.6)Sn(8.5)B(1.1)
Ru(2.9)Sn(14.9)B(1.6)
4.5
4.4
6.10ꢄ2
134
129
73
27
10
72
7.10ꢄ2
a Calculated at isoconversion.
b Calculated at 70% conversion, PH2 = 8 MPa, temperature = 270ꢁC, mcata = 2.2 g.
(ii) the rate of the heavy esters formation (k2 or k02)
step which could be due to a significant coverage of ruthe-
(iii) the rate of the heavy esters hydrogenation (k3 or k03) nium by tin oxides SnOy in agreement with the TPR profiles
(Fig. 1). But as stated by Mizukami et al., it could be that the
carbonyl groups of esters are more strongly adsorbed over
the catalyst, thus leading to an inhibition of the hydrogen
adsorption and of the hydrogenation steps (17).
Identification of Active Sites
At low tin content, Fig. 8 shows that there is mainly (i) for-
mation of saturated ester (methyl stearate) issued from the
Over RuSnB/Al2O3 catalyst (Sn/Ru = 4), the maximum
of selectivity to oleyl alcohol can be explained if the varia-
tion of the unsaturated alcohol selectivity and the propor-
tion of SnOx over the solid surface versus the Sn/Ru ratio
are considered (Fig. 9). Indeed, this figure shows clearly
that (i) there is a synergetic effect between Ruꢁ and tin
species for the formation of SnOx and (ii) the selectiv-
ity to unsaturated alcohol changes like the SnOx content.
According to surface ruthenium and tin content (XPS anal-
ysis), there are six tin atoms in the oxide form for one
ruthenium atom in the RuSnB/Al2O3 (Sn/Ru = 4) catalyst
(Table 6). As the proportion of SnOx is around 30% there
are two SnOx sites which could be very close to Ru parti-
cles (this assumption could explain the binding energy shift
recorded by XPS). This result suggests that reactive centers
such as Ruꢁ—(SnOx)2 are formed in which Ruꢁ could be
associated with two SnOx species.
==
hydrogenationoftheC Cbondand(ii)formationofheavy
esters (SS) due to the transesterification between stearyl al-
cohol and methyl stearate. The hydrogenation of the C C
bond (k4) is much faster than the reduction of the C O
==
==
bond (k1). This means that the hydrogenating properties
of ruthenium are only slightly modified and that the hy-
drogenation rate of heavy ester (k03) is still low. Indeed,
there remains 40% of oleyl oleate at total ester conversion.
Several hypotheses can be proposed: (i) the addition of tin
inhibits the hydrogen adsorption and the tin content is not
==
enough for the C O bond activation; (ii) tin species could
be on the ruthenium particles or/and dispersed on the sup-
port with or without interaction with ruthenium particles.
Depending on the localization of the tin species, the oxi-
dation state of Sn varies and associations between tin and
ruthenium particles could or could not occur.
Over RuSnB/Al2O3 catalyst (Sn/Ru ratio = 7), the trans-
esterification step (k2) is apparently very fast. In this case,
the hydrogenation of heavy esters (k3) may be the limiting
In Fig. 9, we also report the data of oleyl alcohol selec-
tivity obtained over RuSnB/Al2O3 with Sn/Ru ratios of 0.6,
1.1, and 7 (represented by 1). We can see that these values
FIG. 8. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalysts. Selectivity to unsaturated alcohol, to saturated
alcohol, and to heavy esters with the Sn/Ru ratio (compared at isoconver-
sion of methyl oleate, 80%).
SCHEME 2. Kinetic scheme of the hydrogenation of methyl oleate
(OM): OO, oleyl oleate; OA, oleyl alcohol; MS, methyl stearate; SS, stearyl
stearate; SA, stearyl alcohol.

222
POUILLOUX ET AL.
SCHEME 3. Model representating the RuSnB centers with different
tin contents: (1) Sn/Ru < 4; (2) 4 < Sn/Ru < 5.5; (3) Sn/Ru > 5.5.
FIG. 9. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalysts. Relationship between selectivity to unsaturated
alcohol and SnOx species with Sn/Ru ratio.
Reaction Mechanism
Concerning the elementary chemical steps occurring
over these different sites, metallic ruthenium activates hy-
drogen into a “hydride form”. Tin oxides (SnOx) which are
are in good agreement with the other results. The discrep-
ancy observed for the lowest Sn/Ru ratio (i.e., 0.6) proves
that (i) the surface SnOx content could be very small and
(ii) the formation of SnOx is quite dependent of the Sn con-
tent as already suggested. The transesterification is favored
over SnOy sites. The high reducibility of this catalyst (TPR
profile) is in agreement with the presence of ruthenium
species in weak interactions with tin.
==
in interaction with Ru favor the O adsorption of the C O
bond of the ester. The activated hydrogen on ruthenium
attacks the carbon atom of carbonyl groups to obtain an
acetal of tin. This acetal is converted to an aldehyde which
is rapidly hydrogenated into an alcohol form under high
hydrogen pressure (Scheme 4).
In our case, we suggest that the alcohol is directly
formed via a hemiacetal adsorbed at the catalyst surface
(Scheme 5). This mechanism is in accordance with the one
proposed by Rieke et al. (18).
When the tin content is more significant, the rate of
the transesterification is greater than the one of the al-
cohol desorption and the alcoholate adsorbed on the tin
surface reacts with an ester molecule adsorbed on ad-
jacent SnOy species (Scheme 6). Then the heavy esters
can lead to two moles of oleyl alcohol by hydrogenation
(Scheme 7).
According to these results, we propose a schematic rep-
resentation of the catalyst surface (Scheme 3):
(1) when the tin content is low, ruthenium and tin species
are dispersed without interaction, tin being in a high oxida-
tion state.
(2) when the tin content increases, some tin species are
near the ruthenium particles, mixed species (SnO-Ru) be-
ing formed at the surface of ruthenium particles and alu-
mina(typeI)orby“decoration”ofrutheniumparticleswith
tin oxides (type II). As in the first case, tin oxides dispersed
on alumina having the same nature than the one described
in (1) remain.
(3) after increasing the relative amount of tin, there
would be a rather complete covering of alumina and a “dec-
oration”oran“encapsulation”ofrutheniumparticlesbytin
oxides and a decrease of mixed sites (active for the forma-
tion of unsaturated alcohol).
Influence of NaBH4 Reduction
The prereduction of monometallic catalysts with NaBH4
modifies the reducibility of the metal particles as well as
TABLE 6
Relative Composition to Ruthenium, Tin, and SnOx
of the Catalyst RuSnB/Al2O3
Ru
(atom.)
Sn
(atom.)
SnOx
(%)
RuSnB/Al2O3
SnOx/Ru
Ru(1.5)Sn(4.1)B(1.1)
Ru(1.6)Sn(8.5)B(1.1)
Ru(1.5)Sn(10)B(1.3)
1
1
1
3
6
7
20
30
25
0.5
1.9
1.8
SCHEME 4. Mechanism of the hydrogenation of methyl oleate into
oleyl alcohol via the formation of aldehyde.

HYDROGENATION OF FATTY ESTERS
223
SCHEME 5. Mechanism of the direct hydrogenation of methyl oleate into unsaturated alcohol RuSnB/Al2O3 catalyst.
CONCLUSION
the surface composition and similar results are obtained
with bimetallic systems. Indeed, TPR results show that the
prereduction step lead to:
The hydrogenation of methyl oleate over bimetallic
RuSnB/Al2O3 catalysts is selective to unsaturated alcohol
while saturated esters are formed over Ru monometallic
catalysts. Indeed, a yield to oleyl alcohol of 75% for a con-
versionof90%wasobtainedoveraRuSnB/Al2O3 forabulk
atomic ratio Sn/Ru of 4. Owing to the side reaction of the
transesterification between methyl oleate and oleyl alcohol
the tin content strongly affects the selectivity to unsatu-
rated alcohol. This side reaction is catalyzed by tin oxides
without interaction with ruthenium particles. The charac-
terizations of RuSnB/Al2O3 by TPR and XPS analysis have
shown that:
—ruthenium species easily reoxidized
—a similar surface distribution of SnOx and SnOy
—the possible formation of a RuxSny alloy.
From XPS analysis, a surface enrichment of ruthenium
and especially of tin was observed. The (Sn/Ru)surf. ratio is
greaterthanthebulkSn/Ruratiobutthe(Sn/Ru)surf. ratioof
the unreduced and reduced catalysts are comparable. Such
a ruthenium surface enrichment can easily explain the en-
hancement of the RuSnB activity (Table 3). Nethertheless,
there is no significant change of the selectivity (Fig. 10) be-
cause the relative contents of SnOx and SnOy are quite
identical (Table 4).
(i) Ruꢁ particles are formed at the catalyst surface in
the presence of tin species
Furthermore, these experiments prove that the introduc-
tion of boron during the prereduction step with NaBH4
has no effect on the catalytic properties of these catalysts.
These results are in agreement with our previous works
(19), where we showed that the activity of the catalyst pre-
pared via a sol-gel method (without boron) was the same as
that of a solid prepared in accordance with the Narasimhan
method.
(ii) the addition of tin does not modify the surface ruthe-
nium content
(iii) there are two tin oxides species (SnOx and SnOy)
whose surface content varies with the tin content
(iv) theSnOxspecies(x < y)wouldbeininteractionwith
ruthenium (without significant formation of alloy RunSnp).
All these results indicate that there is a correlation be-
tween the selectivity to oleyl alcohol and such SnOx species.
The active center could be constituted by a ruthenium atom
in interaction with two SnOx species.
SCHEME 6. Mechanism of formation of heavy esters over
FIG. 10. Hydrogenation of methyl oleate in the presence of
RuSnB/Al2O3 catalyst.
RuSnB/Al2O3 catalysts. Effect of the prereduction step with NaBH4.

224
POUILLOUX ET AL.
SCHEME 7. Mechanism of the hydrogenation of oleyl oleate into oleyl alcohol over RuSnB/Al2O3 catalyst.
ACKNOWLEDGMENTS
8. Desphande, V. M., Ramnarayan, K., and Narasimhan, C. S., J. Catal.
121, 174–182 (1990).
This study was carried out within the framework of a European Pro-
gram concerning the valorisation of natural feedstocks (AAIR). The au-
thors from the University of Poitiers are very grateful to the European
Communities and to the “Re´gion Poitou-Charentes” for their financial
support.
9. Pouilloux, Y., Piccirilli, A., and Barrault, J., J. Mol. Cat. A Chem. 108,
161–166 (1996).
10. Desphande, V. M., Patterson, N. R., and Narasimhan, C. S., J. Catal.
121, 165–173 (1990).
11. Shirley, D. A., Phys. Rev. B 5, 4709 (1972).
12. Scofield, J. H., J. Elec. Spect. Relat. Phenom. 8, 129 (1976).
13. Ferretti, O. A., thesis, Universite´ Pierre et Marie Curie, 1986.
14. Hansen, P. M., and Anderko, K., “Constitution of Binary Alloys,”
Metallurgy and Metallurgical Engineering Series, Mc Graw–Hill, New
York, 1958.
REFERENCES
1. Ponec, V., Appl. Catal. A General 149, 27–48 (1997).
2. Rylander, P. N., “Hydrogenation Methods,” Academic Press, New
York, 1985.
15. Galvagno, S., Donato, A., Neri, G., Pietropaolo, R., and Capanneli,
G., J. Mol. Cat. 78, 227–236 (1993).
3. Kluson, P., and Cerveny, L., Appl. Catal. A General 128, 13–31 (1995).
4. Candy, J. P., Didillon, B., Smith, E. L., Shay, T. B., and Basset, J. M., 16. Galvagno, S., Milone, C., Donato, A., Neri, G., and Pietropaolo, R.,
J. Mol. Cat. 86, 179–204 (1994). Catal. Lett. 17, 55–61 (1993).
5. Basset, J. M., Candy, J. P., Louessard, P., Ferretti, O. A., and 17. Cheah, K. Y., Tang, T. S., Misukami, F., Niwa, S., Toba, M., and Choo,
Bournonville, J. P., Wiss. Zeitschr. THLM 32, 657–667 (1990). Y. M., J.A.O.C.S. 69(5), 410–416 (1992).
6. Ferretti, O. A., Bournonville, J. P., Mabilon, G., Candy, J. P., and Basset, 18. Rieke, R. D., Thakur, D. S., Roberts, B. D., and White, G. T., JAOCS
J. M., J. Mol. Cat. 67, 283–294 (1991). 74(4), 333–339 (1997).
7. Piccirilli, A., Pouilloux, Y., Pronier, S., and Barrault, J., Bull. Soc. Chim. 19. Piccirilli, A., Pouilloux, Y., and Barrault, J., “Catalysis of Organic Re-
France 132, 1109–1117 (1995). actions” (R. E. Malz, Ed.), pp. 355–360, 1996.