The aldol condensation of acetaldehyde and heptanal on hydrotalcite-type catalysts

Article abstract of DOI:10.1016/S0021-9517(03)00192-1

The aldol condensation of acetaldehyde and heptanal has been carried out in the liquid phase between 353 and 413 K using different types of solid base catalysts: MgO with strong Lewis basic sites, Mg(Al)O mixed oxides with acid-base pairs of the Lewis type obtained from hydrotalcite precursor, and rehydrated Mg(Al)O mixed oxides with Bronsted basic sites. The influence of several reaction parameters, temperature, acetaldehyde to heptanal molar ratio, nature of solvent (hexane, toluene, ethanol), has been investigated. A comparative study of the catalysts has been performed in the such defined optimal reaction conditions, i.e., 373 K, acetaldehyde/heptanal molar ratio, 2/1; and ethanol/reactants molar ratio, 5/1. Mg(Al)O mixed oxides calcined below 673 K are the most selective catalyst to 2-nonenal, the cross-condensation product formed when in the first step proton abstraction occurs from acetaldehyde. Acid-base pairs of moderate basic strength are suitable when this cross-condensation is the desired reaction. Stronger Lewis basic sites of MgO or Bronsted-type basic sites of the rehydrated mixed oxide tend to favor the formation of carbanion from heptanal. This latter leads to the formation of 2-pentyl-2-butenal and 2-pentyl-2-nonenal by cross-condensation with acetaldehyde and self-condensation, respectively.

Full text of DOI:10.1016/S0021-9517(03)00192-1

Journal of Catalysis 219 (2003) 167–175
www.elsevier.com/locate/jcat
The aldol condensation of acetaldehyde and heptanal
on hydrotalcite-type catalysts
Didier Tichit,a,∗ Doina Lutic,a,1 Bernard Coq,a Robert Durand,a and Rémy Teissier b
a
Laboratoire de Matériaux Catalytiques et Catalyse en Chimie Organique, UMR 5618 CNRS-ENSCM, 8, Rue Ecole Normale,
34296 Montpellier cedex 5, France
b
ATOFina, CRRA, Rue Henri Moissan BP 63, 69310 Pierre-Bénite cedex, France
Received 13 December 2002; revised 18 March 2003; accepted 8 April 2003
Abstract
The aldol condensation of acetaldehyde and heptanal has been carried out in the liquid phase between 353 and 413 K using different
types of solid base catalysts: MgO with strong Lewis basic sites, Mg(Al)O mixed oxides with acid–base pairs of the Lewis type obtained
from hydrotalcite precursor, and rehydrated Mg(Al)O mixed oxides with Brønsted basic sites. The influence of several reaction parameters,
temperature, acetaldehyde to heptanal molar ratio, nature of solvent (hexane, toluene, ethanol), has been investigated. A comparative study
of the catalysts has been performed in the such defined optimal reaction conditions, i.e., 373 K, acetaldehyde/heptanal molar ratio, 2/1;
and ethanol/reactants molar ratio, 5/1. Mg(Al)O mixed oxides calcined below 673 K are the most selective catalyst to 2-nonenal, the cross-
condensation product formed when in the first step proton abstraction occurs from acetaldehyde. Acid–base pairs of moderate basic strength
are suitable when this cross-condensation is the desired reaction. Stronger Lewis basic sites of MgO or Brønsted-type basic sites of the
rehydrated mixed oxide tend to favor the formation of carbanion from heptanal. This latter leads to the formation of 2-pentyl-2-butenal and
2-pentyl-2-nonenal by cross-condensation with acetaldehyde and self-condensation, respectively.
2003 Elsevier Inc. All rights reserved.
Keywords: Acetaldehyde; Heptanal; 2-Nonenal; Hydrotalcite; Aldol condensation
1. Introduction
nominated layered double hydroxides or anionic clays, con-
stitute a class of compounds with positively charged layers
and exchangeable anions in the interlayer space [2–4]. The
Reactions involving C–C bond formation are of utmost
importance for obtaining many fine chemicals of commer-
cial interest. The base-catalyzed aldol condensation, which
belong to this type of reaction is indeed very useful for the
preparation of higher molecular weight aldehydes and/or ke-
tones from lower easily available homologs. The exciting
possibility of carrying out these reactions through a het-
erogeneous catalytic process can now be contemplated by
taking advantage of the increasing number of available ma-
terials with finely tunable basicities. This is particularly the
case of the hydrotalcite-type catalysts which have recently
attracted much attention for various base-catalyzed reactions
in fine chemistry [1]. Hydrotalcite-like compounds of gen-
eral formula [MI1I−xMIxII(OH)2]x+[Axn−/n] · mH2O, also de-
structure is similar to that of brucite Mg(OH)2, where Mg2+
-
centered octahedra are linked by the edges to form infinite
sheets. In the natural mineral hydrotalcite, whose name has
been extended to this family of materials, some Mg2+ are
isomorphously substituted for Al3+ and the formal positive
charge thus appearing in the hydroxyl layers is usually com-
pensated by carbonates linked by hydrogen bonds to water
molecules. Dehydroxylation takes place and volatile anions
are decomposed upon calcination, leading to mixed oxides.
These materials obtained from hydrotalcite are the catalysts
most largely used until now, though interest in activated
hydrotalcites obtained by rehydration of the mixed oxides
is growing. Using hydrotalcite-like catalysts, attractive re-
sults have been thus reported in: (i) the self-condensations
of acetone [5–7], acetaldehyde [8], or butyraldehyde [9];
(ii) the condensation between aromatic aldehydes like ben-
zaldehyde or substituted benzaldehydes and acetone [10],
*
Corresponding author.
E-mail address: tichit@cit.enscm.fr (D. Tichit).
On leave from Al.I. Cuza University of Iasi, Faculty of Chemistry,
1
11 Bd. Carol I, Iasi, Romania.
0021-9517/$ – see front matter
2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00192-1

168
D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
substituted acetophenones [11,12], and heptanal [13]; and
(iii) the condensations of acetone and citral [14,15].
2. Experimental
2.1. Material preparation
However, the field of cross-condensation between lin-
ear aliphatic aldehydes to yield longer linear aldehydes was
not explored till now despite the great interest of high-
molecular-weight aldehydes as valuable monomers in some
polymerization processes. It is worth noting the very chal-
lenging aspect of this proposal because various self- and
cross-condensations can occur simultaneously. It was there-
fore the aim of this work to perform the aldol condensa-
tion of acetaldehyde and heptanal toward 2-nonenal, with
catalysts obtained from hydrotalcite materials, as a model
reaction of chain lengthening between aldehydes of close re-
activities.
The Mg–Al hydrotalcite was synthesized by coprecipita-
tion of a gel at constant pH under air. With this aim 250-
cm3 aqueous solution of Mg(NO3)2 · 6H2O, and Al(NO3)3 ·
6H2O (Mg/Al =3) was delivered into a polypropylene re-
actor by a chromatography-type pump at a constant flow of
1 cm3 min−1. A second aqueous solution of 2 M NaOH +
0.5 M Na2CO3 was simultaneously fed by a pH-stat appara-
tus (718 Stat Titrino, Metrohm). The careful control of the
flow maintained the pH during precipitation in the reactor at
a constant value of 10.0 0.2. After completion of the pre-
cipitation, the resulting suspension was aged at 353 5 K
for 15 h with stirring. The solid thus obtained was then iso-
lated by centrifugation, washed thoroughly with deionized
water, and dried overnight in oven at 353 K.
The catalysts were Mg(Al)O mixed oxides obtained from
calcination at various temperatures of an Mg/Al hydrotal-
cite precursor, and the rehydrated form of these calcined
materials. The main advantage of the hydrotalcites as cat-
alysts comes from the opportunity to control to a certain
degree their acido-basic properties [13,16]. The thermal de-
composition in the range 623–823 K results in the formation
of a Mg(Al)O mixed oxide with both strong Lewis basic
and mild Lewis acid sites. The former are due to the pres-
ence of oxygen atoms of low coordination, and the latter to
Al3+ cations. Rehydration of these calcined hydrotalcites al-
lows the reconstruction of the lamellar structure owing to
its memory effect. The counteranions in the interlayer space
then become OH−; the basic Lewis sites are thus moving
to hydroxyls species which are Brønsted sites [5]. At the
same time, the mild acidity of the mixed oxides disappears.
This material is called meixnerite. The different nature of
the active sites in the mixed oxides, on one hand, and the re-
hydrated samples, on the other hand, has been illustrated by
the activities and selectivities reached in different reactions
demanding specific types of sites. The rehydrated materi-
als are the most active in the aldol condensations alone.
They are thus very selective toward diacetone alcohol in
the self-condensation of acetone [5], or toward 4-hydroxy-4-
phenylbutan-2-one and pseudoionone in the condensations
of benzaldehyde and acetone [17], and of citral and ace-
tone [14], respectively. Lower selectivities are obtained with
the mixed oxides because dehydration of the aldol, self-
condensations reactions of the reagents, and oligomeriza-
tions may occur. Mixed oxides are in contrast highly active
and selective in the Meerwein–Pondorf–Verley (MPV) re-
duction of aldehydes and ketones requiring Lewis-type sites
only [18].
The exchange of nitrate anions was performed by dis-
persing 2 g of the as-synthesized hydrotalcite in a 1.5 ×
10−3 M Na2CO3 solution under magnetic stirring at 353 K
for 2 h. After filtration the solid was washed and dried
in an oven at 353 K. This material will be hereafter re-
ferred as HT. The chemical analysis gives the formula
[Mg0.75Al0.25(OH)2]0.25+[CO3
]0.125 ·mH2O, well consis-
2−
tent with the composition of the synthesis solution.
The Mg–Al mixed oxides were obtained by calcination
of HT at the desired temperature (from 643 to 823 K) for 4 h
in a dry synthetic air flow (mass, 2 g; flow, 100 cm3 min−1
;
ramp, 1 K min−1). These samples were hereafter labeled as
HTc(673) for HT calcined at 673 K, for instance.
The rehydration of the mixed oxides to yield the meixner-
ite was carried out in situ after calcination. The sample was
cooled to room temperature in synthetic air flow (CO2 free),
and then contacted with a flow (100 cm3 min−1) of N2 satu-
rated with the vapor pressure of water at the same tempera-
ture (16 h for 1 g of sample).
A commercial Mg(Al)O mixed oxide, labeled KW 2200,
was purchased from Kyowa. This sample has a molar ratio
Mg/Al = 2.12 and a specific surface area of 160 m2 g−1 af-
ter activation at 723 K.
MgO was obtained by thermal treatment of Mg(OH)2
(Strem Chemicals) at 673 K following the same procedure
as described above. The rehydration process from MgO back
to Mg(OH)2 was conducted as well.
2.2. Characterizations
Chemical analyses of the as-prepared solids were per-
formed at the Service Central d’Analyse du CNRS (Solaize,
France) by ICP-MS.
It was therefore interesting to compare the catalytic per-
formances of mixed oxides obtained by calcination at dif-
ferent temperatures of an hydrotalcite precursor and of the
rehydrated forms in the condensation of acetaldehyde and
heptanal. The reaction will be also performed with MgO
and rehydrated MgO whose efficiencies in aldolization have
been put in evidence previously [19].
XRD powder patterns were collected on a CGR Theta 60
instrument using monochromatized Cu-Kα radiation (λ =
1
0.15401 nm, 40 kV, and 50 mA).
BET specific surface areas were determined by N2 ad-
sorption at 77 K with a Micromeritics ASAP 2000 apparatus

D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
169
on samples calcined at 723 K and outgassed at 523 K and
10−4 Pa.
The basicity of the calcined samples was studied by CO2
adsorption followed by calorimetry and gravimetry using a
Setaram TG-DSC-111 apparatus. The samples were previ-
ously outgassed at 723 K, cooled to 373 K, and contacted
with flowing CO2.
The acidity of the calcined samples was estimated from
temperature-programmed desorption of NH3 (NH3-TPD)
using a conductivity cell for the detection of the effluent
gases. The samples were previously outgassed at 723 or
873 K in some cases, cooled to 373 K, and contacted with
NH3 vapor. After the sample was purged, the temperature
was increased to 723 K (ramp: 10 K min−1), and evolved
ammonia was trapped in a HCl solution and finally titrated.
2.3. Catalytic tests
The reaction between acetaldehyde (+95%, Aldrich) and
heptanal (+98%, Aldrich) was performed in liquid phase in
a sealed stainless-steel autoclave (100 mL), equipped with
mechanical stirring and sampling device. Usually, about
70% of the solvent (ethanol, +95%, Carlo Erba), the inter-
nal standard (decane, +98%, Fluka), and 2 g of the catalyst
previously activated and kept under dry synthetic air flow
at room temperature were poured in the autoclave before
closing. The reactor was purged for 1 min with N2 flow, pres-
surized at 2×105 Pa, and then heated at the reaction temper-
ature. Typically the solution containing 0.07 mol of heptanal
and 0.14 mol of acetaldehyde in a volume of solvent neces-
sary to reach a total volume of 80 mL was introduced in the
autoclave with a pump, as soon as the reaction temperature
was reached. The reaction mixture was analyzed periodically
by gas chromatography on a Hewlett Packard HP 4890 ma-
chine, equipped with a capillary HP-5 column (30 m length,
0.32 mm i.d.), using decane as internal standard. The prod-
ucts were first identified by a gas chromatograph HP 5890
coupled with a mass spectrometer. Expression of results:
Fig. 1. XRD powder patterns of HT (a), HTc(673) (b), and HTc(673)R (c).
XRD patterns show that the lamellar structure is present in
HTc(643)R and HTc(673)R (Fig. 1).
In order to further explain their catalytic behavior the
acido-basic properties of the catalysts have been character-
ized by CO2 and NH3 adsorptions and TPD experiments.
The heat of CO2 adsorption is considered as a measure of
the strength of basic sites. The strength of the acid sites
is evaluated from the temperature of maximum desorption
rate of NH3 (Tmax). The number of basic (Nbase) and acid
sites (Nacid) are determined from the amounts of CO2 in
TG-DSC and NH3-TPD experiments, respectively. The re-
sults are reported in Table 1. MgO which exhibit sites of the
higher basic strength and a low amount of acid sites behave
therefore as a strong basic solid. Regarding the mixed oxides
calcined in the range 643–823 K, they are characterized by
Reactant conversion (mol%) = 100 × (reactantin
reactantout)/reactantin;
−
Selectivity to product i = (mole of reactant converted
to i)/(total number of mole of reactant converted).
3. Results and discussion
Table 1
Some characteristics of the texture, acid, and basic properties of the cata-
lysts
3.1. Characterization of the materials
Catalysts
N
ꢀH(CO )
N
T
in
N /
base
Surface
area
(m g
max
base
−1
2
acid
The calcination at temperature T (643–823 K) of the hy-
drotalcite precursor HT leads to the materials HTc(T ). All
present typical XRD patterns of the Mg(Al)O phase (Fig. 1).
These mesoporous materials exhibit very similar specific
surface areas (220–240 m2 g−1) and pores of 60 nm mean
size, whatever the calcination temperature (Table 1). After
rehydration of HTc(643) and HTc(673) mixed oxides, the
−1
−1
(meq g ) (kJ mol ) (meq g ) NH -TPD N
3
(K)
acid
2
−1
)
HTc(643)
HTc(673)
HTc(723)
HTc(823)
MgO
0.15
0.27
0.46
0.47
1.03
−65
0.32
0.29
0.15
0.10
0.06
–
0.47
0.93
3.07
4.70
217
225
238
224
160
−61.5
−73.5
−78
543
548
548
548
−83
17

170
D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
Scheme 1. Simplified scheme of the reaction of acetaldehyde and heptanal on a solid basic and acid catalyst.
the presence of acid–base pairs, as expected. These data call
for two comments:
1. There is a net increase of basic strength with the tem-
perature of calcination. In contrast, the acid strength
remains almost constant in all the temperature ranges
investigated.
2. There is a threefold increase of the number of basic
sites while that of acid sites decreases to the same ex-
tent when the calcination temperature goes from 643 to
823 K.
Two distinct behaviors could actually be identified for the
mixed oxides. Upon calcination below 673 K, a medium ba-
sic strength with −ꢀH(CO2) not exceeding 65 kJ mol−1
and ratios Nbase/Nacid < 1 show that they behave as mod-
erately basic solids. The basic character is magnified for
calcination at temperature above 673 K.
3.2. The reaction scheme
The reactions between acetaldehyde and heptanal in-
volve self- and cross-condensations (Scheme 1). The se-
lectivities can thus be expressed with respect to products
coming from either heptanal or acetaldehyde conversion.
This is illustrated in Fig. 2 for the reaction carried out on
HTc(673) at 393 K. With respect to heptanal conversion
(Fig. 2a), 2-nonenal (1) and 2-pentyl-2-butenal (3, mixture
of the E and Z isomers) are the main products from cross-
condensations. The self-condensation of heptanal leads to 2-
pentyl-2-nonenal (4, mixture of the E and Z isomers) which
amounted to 20% after 3 h reaction. The formations of hep-
tanol (5) and diacetal (6) from secondary reactions also take
place to a lower extent. Heptanol comes from the MPV re-
duction of heptanal by hydrogen transfer from ethanol. This
reaction has already been reported as catalyzed by Mg(Al)O
Fig. 2. Conversion and selectivity to the different products as a function
of time in the reaction of acetaldehyde and heptanal on HTc(673) catalyst;
results given on an heptanal conversion basis (a), and acetaldehyde con-
version basis (b); (2) heptanal or acetaldehyde conversion, (F) 1, (e) 2,
(Q) 3, (P) 4, (") 5, (1) 6, and (E) oligomers from self-condensation of ac-
etaldehyde. Conditions: solvent, ethanol; ethanol/reactants molar ratio, 5;
T
= 393 K; acetaldehyde 0.14 mol and heptanal 0.07 mol, 2 g cata-
reaction
lyst.

D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
171
mixed oxides [18]. The diacetal comes from the condensa-
tion of heptanal and ethanol on acid sites. Selectivities to-
ward both compounds were in the range 5–10%. It is worth
noting the changes of selectivity when the conversion goes
on. The selectivity to 1 decreases whereas those toward 3
and 4 increase. In Fig. 2b the results were reported with
respect to acetaldehyde conversion. Compared to the data
determined on an heptanal conversion basis, one should note
(i) the larger conversion of acetaldehyde as compared to hep-
tanal, (ii) the lower selectivities to 1 and 3 (almost 10%), and
(iii) the decrease of selectivity when the conversion goes on
of 2-propenal (crotonaldehyde, 2), the product of acetalde-
hyde self-condensation.
For the sake of comparison, the reaction was carried out
under the same conditions and using 10−4 mol NaOH as
homogeneous catalyst. The reaction rate was faster but the
selectivity to 1 very low (< 10%); the main products were
crotonaldehyde and oligomers from the self-condensation of
acetaldehyde.
The aldol condensation on basic sites occurs in the first
step by the abstraction of an hydrogen atom at the α po-
sition of the carbonyl group of the aldehyde leading to a
carbanion species. In a second step the aldol is formed by
reaction of the carbanion on the carbonyl group of an other
aldehyde molecule. Generally the aldol is dehydrated on the
weak acid sites of the catalyst giving the α, β-unsaturated
aldehyde. When two different aldehydes are concerned, four
products resulting from self- and cross-condensation reac-
tions could be obtained.
Scheme 3. Cross- and self-condensation from the carbanion of heptanal.
carbanion by a second acetaldehyde molecule leads to the C4
aldol which is then dehydrated to 2 on acid sites. In contrast
when the carbanion was formed from heptanal, its cross-
condensation with acetaldehyde leads to 3, whereas 4 results
from the self-condensation with an other heptanal molecule
(Scheme 3).
Assuming an equal probability for the formation of car-
banions from acetaldehyde or heptanal, and a close reactiv-
ity of those with aldehydes, the selectivities to 1, 3, and 4
should be 25, 25, and 50%, respectively, with respect to hep-
tanal conversion. One should emphasize in the present case
the close reactivity of the two aldehydes which makes the
achievement of a high selectivity in one product much more
challenging. It is worth noting that on HTc(673) (Fig. 2a),
the selectivity to 1 amounted up to ca 43% at low conver-
sion, and even more regarding the products from the aldol
condensation alone (ca 51%). The selectivity to 1 decreases
when the reaction goes on due to the faster consumption of
acetaldehyde than heptanal. The relative coverage by hep-
tanal of the catalyst surface then increases which favors the
self-condensation to 4 (Fig. 2a). An improvement of the se-
lectivity to 1 may result from the adjustment of the reaction
conditions, but also from tuning the acido-basic properties
of the catalyst surface [15,20].
The cross-condensation of acetaldehyde and heptanal
leading to 1 results from the attack of the acetaldehyde car-
banion by heptanal (Scheme 2). However, the attack of this
We have studied these two aspects by first carrying out a
parametric study (reaction temperature, reactants molar ra-
tio, solvent) on a commercial catalyst (KW 2000). Under
the optimal reaction conditions thus defined, we then in-
vestigated the influence on the reaction of the acido-basic
properties of the catalyst.
3.3. Parametric study
The influence of the reaction temperature on the selec-
tivity to 1 is shown in Fig. 3 as a function of the heptanal
Scheme 2. Cross-condensation between the carbanion of acetaldehyde and
heptanal.

172
D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
Fig. 3. Selectivity to 1 as a function of heptanal conversion in the reaction
of acetaldehyde and heptanal at different reaction temperatures; (F) 373,
(2) 393, and (P) 413 K. Conditions: acetaldehyde/heptanal molar ratio, 2;
solvent, ethanol; ethanol/reactants molar ratio, 5; catalyst, KW 2200.
conversion. No reaction proceeds below 373 K and the con-
version remains below 13% with a selectivity around 20% at
this temperature. The conversion and selectivity to 1 increase
markedly with reaction temperature, and the latter reaches
ca 40% at 413 K. However, at higher temperatures, several
consecutive condensation reactions involving acetaldehyde
occur leading to oligomers. Selectivity to 1 levels off in all
cases above 10% conversion. An optimum reaction temper-
ature of 393 K was then chosen to obtain good yields to 1
and to avoid problems arising from oligomer formation.
The selectivity toward cross- and self-condensation re-
actions was expected to strongly depend on the molar ratio
between the two reactants in the reaction medium. Reactions
with acetaldehyde/heptanal molar ratios (AA/HA) of 1, 2,
and 4 were carried out at 393 K, with an ethanol/reactant
molar ratio of 5. The selectivity to 1, as a function of hep-
tanal or acetaldehyde conversion, is reported in Fig. 4. A se-
lectivity in the range of 30–65%, on an heptanal conversion
basis (Fig. 4A), is obtained with a molar ratio AA/HA = 4.
However, this selectivity becomes lower than 5% when ex-
pressed on an acetaldehyde conversion basis (Fig. 4B), due
to the oligomerization of this reactant. The selectivity to 1
on a acetaldehyde conversion basis is of 8–14% for molar
ratios AA/HA = 1 and 2; this is due to a lower extent of
acetaldehyde oligomerization at low AA/HA ratio. Unfor-
tunately, the selectivity to 1 then becomes lower than 20%
on an heptanal conversion basis, for AA/HA = 1. We have
thus considered a molar ratio AA/HA = 2 as a compromise
to get good selectivity to 1 in the whole range of heptanal
conversion.
Fig. 4. Selectivity to 1 as a function of heptanal (A) and acetaldehyde
(B) conversions in the reaction of acetaldehyde and heptanal at different
acetaldehyde/heptanal molar ratios: (F) AA/HA = 1, (2) AA/HA = 2,
and (P) AA/HA = 4. Conditions: solvent, ethanol; ethanol/reactants mo-
lar ratio, 5; T
= 393 K; catalyst, KW 2200.
reaction
ence of reactivity was observed. After 8 h reaction, the hep-
tanal conversion is in the range 20–25% with toluene or
hexane, whereas it reaches 45% with ethanol. Lowering of
the rates when Michael and Knoevenagel reactions were per-
formed in hexane and toluene, respectively, has also been
reported [10,21]. We may anticipate that ethanol, a highly
polar solvent, improves the desorption of products from the
surface of the catalyst. On one side it minimizes the self-
condensations and other side reactions, and on the other side
it increases the activity. The influence of the solvent/reactant
ratio was then examined with ethanol. The selectivity to
1 increased and the conversion simultaneously decreased
when this ratio goes from 2 to 10 (Fig. 5). These results
accounted for the expected lowering of conversion and sec-
ondary self-condensation reactions when the coverage of the
reactants at the surface decreased. The yield to 1 (with re-
Experiments were carried out at 393 K with AA/HA = 2
and toluene, hexane, or ethanol as solvent (solvent to reac-
tants ratio of 5), previously used for condensation reactions
with hydrotalcite-like catalysts [10,12,21]. A great differ-

D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
173
Table 2
Activities and selectivities (on an heptanal conversion basis) obtained in
the reaction of acetaldehyde and heptanal on different catalysts (acetalde-
hyde/heptanal molar ratio, 2; T
= 393 K; solvent, ethanol)
reaction
a
Catalysts
Initial rate
Selectivity (%)
6
−1
−1
(10 mol g min )
1
3
4
5
6
HTc(643)
HTc(643)R
HTc(673)
HTc(673)R
HTc(723)
HTc(823)
MgO
78
8.8
86
32
24
41
39
39
35
17
38
22
39
29
36
33
41
40
37
14
33
14
22
22
23
39
23
7
1
9
3
4
4
2
1
25
3
8
1
2
0
1
0
70
103
130
320
66
Mg(OH)
2
a
At conversion of 30%.
calcination temperature and 3 becomes the most abundant
product with HTc(823). This is also the case with MgO for
which the selectivities to 3 and 4 both reach 40%, clearly
higher than that to 1. Significantly, as selectivities to 3 and 4
increase, that to 5 decreases. The MPV reduction of hep-
tanal, accounting for the formation of 5, occurs also to a
lower extent in the case of the rehydrated mixed oxides The
inhibition of the MPV reaction after rehydration is in line
with the results previously reported by Kumbhar et al. [18].
They show that the synergetic effect between strong Lewis
basic and mild acid sites of the mixed oxides is more suitable
than the Brønsted-type sites of the rehydrated hydrotalcite
for catalyzing this reaction.
Finally the formation of 6, catalyzed by acid sites, de-
creases steadily with the calcination temperature of the
mixed oxides, and this compound is almost absent on re-
hydrated materials, as well as on MgO and Mg(OH)2.
It must be recalled that the formation of 1 on the one side,
and of 3 and 4 on the other side, results in the first step from
an α hydrogen abstraction from acetaldehyde and heptanal,
respectively, to yield the carbanion. This hydrogen abstrac-
tion is expected to be easier as the length of the alkyl chain
becomes shorter, i.e., for acetaldehyde in comparison to hep-
tanal. Consistently, the mixed oxides calcined at or below
673 K, of moderate basicity, lead to 1 predominantly rather
than 3 and 4. In this case, the competitive formation of car-
banions from acetaldehyde or heptanal is clearly in favor of
the former. When the basic strength increases, i.e., for cal-
cination above 673 K, the discrimination between acetalde-
hyde and heptanal for the formation of carbanions is less
efficient. This behavior is put in evidence in Fig. 6 where the
ratio 1/(1+3+4) has been reported as a function of the heat
of CO2 adsorption. Accordingly, MgO which is strongly ba-
sic (Table 1) gives the largest amounts of 3 and 4, coming
from the heptanal carbanion. This behavior is quite similar
to that observed in the synthesis of jasminaldehyde involving
also heptanal [13], for which higher selectivity was obtained
with the less basic catalyst (CsX); stronger basic catalysts
(MgO and hydrotalcites calcined at high temperature) lead to
high amounts of the self-condensation product of heptanal.
Fig. 5. Selectivity to 1 at different ethanol to reactants molar ratios and for
an heptanal conversion of 30%. Conditions: acetaldehyde/heptanal molar
ratio, 2; T
= 393 K; catalysts, KW 2200.
reaction
spect to heptanal) of ca 12% was actually constant above a
solvent/reactants ratio of 5.
The heterogeneously base-catalyzed aldol condensation
of aldehydes and ketones obeys a Langmuir–Hinshelwood-
type kinetic model [12,17,22]. The selectivity to 1 is thus fa-
vored by a high interdispersion in the adsorbed layer of hep-
tanal between activated acetaldehyde carbanions. From the
above results this is achieved by: (i) a higher reaction tem-
perature and a larger ethanol/reactants ratio which decrease
the heptanal coverage, and (ii) a large AA/HA reactant ratio
which favors the interdispersion of adsorbed heptanal in sur-
rounding acetaldehyde. However, we have to pay attention
to the yield and an excessive self-condensation of acetalde-
hyde. For these reasons we have chosen a temperature of
393 K and AA/HA and ethanol/reactants ratios of 2 and 5,
respectively, as a good compromise for the reaction condi-
tions.
Under these conditions we have then evaluated the oppor-
tunities which could be brought by the various hydrotalcite-
based materials to modify the acido-basic properties of the
surface, and in fine the respective activation of heptanal and
acetaldehyde as carbanions.
3.4. Influence of catalytic materials
Initial reaction rates and selectivities obtained with the
different catalysts are reported in Table 2. The initial rate
increases slightly with the calcination temperature of the
mixed oxide. The rehydrated materials are less active.
Interesting features call for several comments regarding
the evolution of selectivity patterns. First of all, the presence
of acid–base pairs in the HTc(T ) catalysts well accounted
for the products obtained. Compound 1 is obtained predom-
inantly with HTc(T ) calcined below 823 K. However, it is
worthy to note that selectivities for 3 and 4 increase with the

174
D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
Fig. 7. Selectivity to diacetal (6) as a function of N
alysts.
in the different cat-
Fig. 6. Selectivity ratio 1/(1 + 3 + 4) as a function of the enthalpy of CO
acid
2
adsorption on the different catalysts.
acetaldehyde, ranged between 16 and 40%. It is worth not-
ing that such yields were achieved with a large excess of
the less reactive acetone (acetone/acetaldehyde, 4/1). This
is obviously inapplicable in the present study due to the very
close reactivity of the two aldehydes.
The rehydrated hydrotalcites possess Brønsted basic
sites. In most studies these catalysts are more active and se-
lective to the cross-condensation products than the mixed
oxides [10,13–15,17,23]. Several explanations have been
proposed for this behavior. From calorimetric experiments
Rao et al. [17] reported a lower basicity of the surface after
rehydration. They concluded that the improvement in activ-
ity accounted for a better specificity of OH− for aldolization,
Lewis basic sites O2− working worse even if their basic
strength is higher. Roelofs et al. [15,24] demonstrated that
less than 5% of the total number of OH− in the rehydrated
samples is available for condensation reactions. However,
they are highly active due to their localization at the edge
of the reconstructed platelets of the layered structure. Other-
wise, Climent et al. [13] suggested that the OH− sites in the
rehydrated samples are of medium basic strength and thus
well adapted to the aldol condensation.
Our results in the condensation of acetaldehyde and hep-
tanal show that the rehydrated hydrotalcites are less active
but more selective for the self-condensation of heptanal than
the mixed oxides. Brønsted OH− sites thus appear less spe-
cific for this type of cross-condensation involving two linear
aldehydes. It comes out that Lewis-type basic sites are more
efficient, provided that they were of moderate strength in
order to maintain the proton abstraction from heptanal, in
competition with acetaldehyde, to a lower extent.
4. Conclusions
The reaction of acetaldehyde and heptanal on catalysts
elaborated from an hydrotalcite precursor (Mg/Al = 3)
yields a wide variety of products. Most of them (usually
more than 80%) originate from cross- and self-condensation
of the aldehydes on basic and acid-pair sites. The most ef-
ficient material for the cross-condensation to 2-nonenal is
Mg(Al)O calcined at 873 K. A maximum yield of 21%
was reached at 393 K in ethanol as solvent and a moderate
excess of acetaldehyde (acetaldehyde/heptanal, 2/1). This
excess is necessary to achieve a good balance between both
reactants in the adsorbed state and then to favor the cross-
condensation.
From a practical point of view, one could conclude that
the initial goal, the selective cross-condensation to produce
2-nonenal, was not achieved. This is due to the very close re-
activity of acetaldehyde and heptanal to form the carbanion
upon interaction with the basic surface.
Finally, the selectivity to 6 increases with the number of
acid sites in the catalysts, as expected (Fig. 7).
However, from a fundamental point of view, a very inter-
esting aspect deals with the high sensitivity of the aldehyde
activation with the fine tune of surface basicity. Other things
being equal, the acid–base pairs are the most efficient sites
for this reaction, and the relative selectivity to 1, in com-
parison to 3 and 4, can suffer a fivefold increase when the
strength of the basic sites are decreased. This is due to a
preferential activation of acetaldehyde. In contrast, on solids
of stronger basicity, there is no discrimination between the
The maximum achieved yield to 1, on an heptanal ba-
sis, was almost 21% (60% conversion and 35% selectiv-
ity) on HTc(673). To date, there is no report on the cross-
condensation between aliphatic aldehydes to be compared.
One can only note that in the cross-condensation between
acetone and acetaldehyde to pent-3-en-2-one on nonzeolitic
molecular sieves [25], the maximum yield, with respect to

D. Tichit et al. / Journal of Catalysis 219 (2003) 167–175
175
abstraction of protons from acetaldehyde or heptanal by the
basic sites; the latter leading to the formation of 2-pentyl-2-
butenal.
[10] M.L. Kantam, B.M. Choudary, C.V. Reddy, K.K. Rao, F. Figuéras,
Chem. Commun. 1033 (1998).
[11] M.J. Climent, A. Corma, S. Iborra, J. Primo, J. Catal. 151 (1995) 60.
[12] A. Guida, M.H. Lhouty, D. Tichit, F. Figuéras, P. Geneste, Appl. Catal.
A 164 (1997) 251.
[13] M.J. Climent, A. Corma, S. Iborra, A. Velty, J. Mol. Catal. 182–183
(2002) 327.
[14] M.J. Climent, A. Corma, S. Iborra, A. Velty, Catal. Lett. 79 (2002)
157.
[15] J.C.A.A. Roelofs, A.J. van Dillen, K.P. de Jong, Catal. Today 60
(2000) 297.
Acknowledgment
D. Lutic gratefully thanks Atofina for the financial sup-
port as a postdoctoral fellowship.
[16] D. Tichit, F. Fajula, Stud. Surf. Sci. Catal. 125 (1999) 329.
[17] K.K. Rao, M. Gravelle, J. Sanchez Valente, F. Figuèras, J. Catal. 173
(1998) 115.
References
[18] P.S. Kumbhar, J. Sanchez-Valente, J. Lopez, F. Figuèras, Chem. Com-
mun. 535 (1998).
[19] G. Zhang, H. Hattori, K. Tanabe, Appl. Catal. 36 (1988) 189.
[20] E. Dumitriu, V. Hulea, C. Chelaru, C. Catrinescu, D. Tichit, R. Durand,
Appl. Catal. A 178 (1999) 145.
[1] B.F. Sels, D.E. De Vos, P.A. Jacobs, Catal. Rev. 43 (2001) 443.
[2] W.T. Reichle, Chemtech 16 (1986) 58.
[3] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 11 (1991) 173.
[4] A. Vaccari, Appl. Clay Sci. 14 (1999) 161.
[5] D. Tichit, M.N. Bennani, F. Figueras, R. Teissier, J. Kervennal, Appl.
Clay Sci. 13 (1998) 401.
[21] B.M. Choudary, M.L. Kantam, C.R.V. Reddy, K.K. Rao, F. Figuèras,
J. Mol. Catal. A 146 (1999) 279.
[6] W.T. Reichle, J. Catal. 63 (1980) 295.
[7] J.I. Di Cosimo, V.K. Diez, C.R. Apesteguia, Appl. Clay Sci. 13 (1998)
433.
[8] J.I. Di Cosimo, V.K. Diez, M. Xu, E. Iglesia, C.R. Apesteguia, J. Ca-
tal. 178 (1998) 499.
[22] A. Aguilera, A.R. Alcantara, J.M. Marinas, J.V. Sinisterra, Can. J.
Chem. 65 (1987) 1165.
[23] D. Tichit, B. Coq, S. Cernaux, R. Durand, Catal. Today 75 (2002) 197.
[24] J.C.A.A. Roelofs, D.J. Lensveld, A.J. van Dillen, K.P. de Jong, J. Ca-
tal. 203 (2001) 184.
[9] W.T. Reichle, US patent 4,086,188, 1978.
[25] K.D. Olson, US patent 4,701,562, 1987.