Acid-base bifunctional catalysts for the preparation of fine chemicals: Synthesis of jasminaldehyde

Article abstract of DOI:10.1006/jcat.2000.3086

Jasminaldehyde was prepared by condensation between benzaldehyde and heptanal. Large-pore acid zeolites (HY and Beta), mesoporous aluminosilicate (Al MCM-41), and amorphous aluminophospates (ALPO) were used as catalysts. The results indicated that zeolites showed lower activity and selectivity than mesoporous aluminosilicate (Al MCM-41). These results were attributed to the confinement effects of the reactants and products inside of the voids of the microporous materials, which lead to the preferential formation of the heptanal self-condensation product, as well as to a fast deactivation of the catalyst. However, the amorphous aluminophosphate (ALPO) which possesses weaker acid sites than zeolites and MCM-41, but combines acidic and basic sites, showed the maximum activity and selectivity to jasminaldehyde. This finding was explained on the basis of the acid-base bifunctional character of the ALPO catalyst. The role of the weak acid sites is to activate benzaldehyde by protonation of the carbonyl group, favoring then the attack of the enolate heptanal intermediate generated on the relatively weak basic sites of ALPO.

Full text of DOI:10.1006/jcat.2000.3086

Journal of Catalysis 197, 385–393 (2001)
doi:10.1006/jcat.2000.3086, available online at http://www.idealibrary.com on
Acid–Base Bifunctional Catalysts for the Preparation of Fine
Chemicals: Synthesis of Jasminaldehyde
M. J. Climent, A. Corma,1 H. Garcia, R. Guil-Lopez, S. Iborra, and V. Forne´s
Instituto de Tecnolog´ıa Quimica (UPV-CSIC), Avda. de los Naranjos, s/n, 46022 Valencia, Spain
Received August 10, 2000; revised October 6, 2000; accepted October 9, 2000; published online December 21, 2000
naldehyde with high selectivity requires a low concen-
tration of heptanal relative to benzaldehyde in the reac-
tion mixture. This can be achieved using a high benzalde-
hyde/heptanal molar ratio and/or by adding slowly the
heptanal into the reaction mixture.
Jasminaldehyde was prepared by condensation between ben-
zaldehyde and heptanal. Large-pore acid zeolites (HY and Beta),
mesoporous aluminosilicate (Al MCM-41), and amorphous alu-
minophospates(ALPO)wereused ascatalysts. Theresultsindicated
that zeolites showed lower activity and selectivity than mesoporous
aluminosilicate (Al MCM-41). These results were attributed to the
confinement effects of the reactants and products inside of the voids
of the microporous materials, which lead to the preferential forma-
tion of the heptanal self-condensation product, as well as to a fast
deactivation of the catalyst. However, the amorphous aluminophos-
phate (ALPO) which possesses weaker acid sites than zeolites and
MCM-41, but combines acidic and basic sites, showed the maxi-
mum activity and selectivity to jasminaldehyde. This finding was
explained on the basis of the acid–base bifunctional character of
the ALPO catalyst. The role of the weak acid sites is to activate
benzaldehyde by protonation of the carbonyl group, favoring then
the attack of the enolate heptanal intermediate generated on the
c
Different solid base catalysts have been reported for the
synthesis of jasminaldehyde. Among them, the most inter-
esting are anionic exchange resins (2), potassium carbonate
in the presence of a solid–liquid phase-transfer catalyst (3),
and solid–liquid phase-transfer catalyst in dry media under
microwave irradiation (4). It is known that the condensa-
tion of heptanal and benzaldehyde can also be carried out
in the presence of solid acid catalysts. However, due to the
lower selectivity generally found, processes involving the
use of heterogeneous acid catalysts in the production of
jasminaldehyde are scarce (5). In a previous work we ob-
served that using aluminosilicate molecular sieves as acid
catalysts, the selectivity to jasminaldehyde was strongly de-
pendent on the topology of the catalysts and on the acid
strength of the active sites (6). Therefore, we thought that
by controlling these parameters in a heterogeneous acid
catalyst it should be possible to improve the selectivity to
jasminaldehyde. The aim of this work is to study the synthe-
sis of jasminaldehyde by the acid-catalyzed condensation
relatively weak basic sites of ALPO.
2001 Academic Press
Key Words: synthesis of jasminaldehyde; aldol condensation; zeo-
lites;MCM-41acid catalysts;solid acid–basecatalystsfor finechem-
icals; ALPO; acid–base bifunctional catalyst.
INTRODUCTION
Jasminaldehyde, -n-amylcinnamaldehyde (3) with a vi- of benzaldehyde and heptanal in order to find the opti-
olet scent, is considered a fine chemical of commercial in- mum heterogeneous acid catalyst. With this target, we se-
terest and may be obtained by the condensation of heptanal lected different solid catalysts covering a wide range of acid
with benzaldehyde (Scheme 1). This aldol condensation is strengths, including amorphous and crystalline solid acids.
carried out in the presence of alkali as catalyst (sodium or It will be shown that it is possible to obtain very good solid
potassium hydroxide), heptanal being added slowly to the catalysts for the synthesis of jasminaldehyde when a bifunc-
reaction mixture at moderate temperatures (1).
tional acid–base bifunctional system is prepared.
A limitation of aldol condensation reactions in general,
and more specifically in the synthesis of jasminaldehide,
is the formation of by-products which reduce the yield of
EXPERIMENTAL
amylcinnamic aldehyde (3) to 70% . The most abundant
=
Materials
undesired product comes from the self-condensation of
heptanal to form 2-n-pentyl-2-n-nonenal (4) (Scheme 1).
Therefore, the general methodology for producing jasmi-
Amorphous aluminum phosphate (ALPO) with a P/Al
ratio of 1 was prepared following the method proposed by
Lindblad et al. (7). Na-exchanged ALPO (Na-ALPO) was
prepared by adding to a 1 M aqueous solution of sodium
acetate (100 ml), 1 g of ALPO precursor. The mixture was
1 To whom correspondence should be addressed. Fax: 34 963877809.
E-mail: acorma@itq.upv.es.
385
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

386
CLIMENT ET AL.
SCHEME 1
heated at 333 K for 8 h. The solid was then filtered and the accessory for solid samples. BaSO4 was used as the standard
exchange was repeated three times. Finally the solid was for baseline correction.
filtered, thoroughly washed with water, and dried at 333 K
for 12 h.
The IR spectra were obtained with an FT-IR Nicolet 710
instrument using sealed cells provided with CaF2 windows.
2
An Al MCM-41 sample with a Si/Al ratio of 15 and a pore Wafers (10 mg) of solid catalysts were pressed (1 Tm cm
)
diameter of 3.5 nm, as well as an MCM-41 Al-free sample, and outgassed under vacuum (10 2 Pa) for 1 h at increasing
were prepared following the procedures given in Refs. (8) temperatures.
and (9), respectively.
Thermogravimetric analyses and differential scanning
The HY-100 zeolite (where the number indicates the per- calorimetry of the catalysts after the aldol condensation
centage of Na+ to NH4+ exchange) was prepared by stirring reaction were obtained in a Netszch microbalance under
a commercial NaY sample (SK-40, Union Carbide, Si/Al an air stream using kaolin as the inert standard.
2.5) with different aqueous solutions of ammonium acetate
Acetalization of benzaldehyde with trimethyl orthofor-
followed by calcination at 823 K (10).
mate (TMOF). This was selected as a test reaction for
The Beta sample with a Si/Al ratio of 13 was supplied in
determining the presence of weak acid sites on the bifunc-
the acid form by PQ Corp.
tional catalysts.
A 200-mg portion of ALPO or Na-exchanged ALPO
catalyst was added to a solution of benzaldehyde (5 mmol)
and TMOF (25 mmol) in tetrachloromethane (15 ml). The
resulting suspension was magnetically stirred at reflux tem-
Reaction Procedure
Condensation of benzaldehyde with heptanal. All re-
actions were carried out under a nitrogen atmosphere, in
perature. Samples of the reaction mixture were periodically
a flask that was fitted with a reflux condenser. The flask
withdrawn by a filtering syringe within a time period of 0.1
was immersed in a thermostated silicone oil bath and the
to 5 h and analyzed by GC.
reaction mixture was magnetically stirred. A mixture of
previously distilled benzaldehyde (50 mmol) and heptanal
(10mmol) wasplaced in the flask. Once the mixture reached
Knoevenagel condensation between benzaldehyde and
malononitrile. This experiment was carried out as a test
reaction to show the presence or absence of basic sites on
a temperature of 398 K, a specific calculated quantity of the
the catalysts. Thus, a mixture of benzaldehyde (10 mmol)
catalyst was added such that the final wt/wt ratio of cata-
and malononitrile (10 mmol) under an inert atmosphere
lyst versus the total sum of reagents was 10% . All cata-
was magnetically stirred and heated at 333 K, and 85 mg of
lysts except ALPO and Na-ALPO were activated before
the catalyst (ALPO or Na-exchanged ALPO) was added.
reaction, at 323 K under vacuum (1 Torr) for 3 h. Sam-
Samples of the reaction mixture were periodically with-
ples of the reaction mixture were periodically withdrawn
drawn by a filtering syringe within a time period of 0.1 to
by a filtering syringe and analysed by gas chromatography
5 h and analyzed by GC.
(GC) (Hewlett-Packard 5990A spectrometer provided with
a 25-m capillary column of cross-linked 5% phenylmethyl-
silicone).
Quantitative determinations were based on the mea-
sured response factors of the reactants and reaction prod-
ucts. Each reaction product was identified by H NMR and heptanal (2) was carried out with H-Beta zeolite that
spectroscopy (Varian VXR-400 S, 400 MHz) and GC–mass was taken as a prototype of strong acid catalysts. A molar
spectrometry (Hewlett-Packard 5988A mass spectrometer ratio of benzaldehyde/heptanal of 5 was chosen in order to
connected with a 25-m capillary column of phenylmethyl- minimize heptanal self-condensation. Although the actual
RESULTS AND DISCUSSION
The condensation reaction between benzaldehyde (1)
1
silicone).
concentration of reagents inside the pores may be different
Diffuse reflectance UV–vis spectra were recorded using a from the reagent ratio in solution, this large excess still en-
Cary 5G spectrophotometer adapted with a Praying Mantis sures that benzaldehyde is likely the predominant reagent

SYNTHESIS OF JASMINALDEHYDE
387
SCHEME 2
at the catalytic sites. A reactant/catalysts ratio of 10 and rate of formation of jasminaldehyde. This fact should in-
398 K were chosen as the other reaction parameters. After dicate that even with a benzaldehyde/heptanal ratio of 5
8 h of reaction time, the composition of the final mixture heptanal is diffusing faster than benzaldehyde and that the
was 20% of (E)- -n-amylcinnamaldehyde (jasminalde- self-condensation is favored by the smaller size of the tran-
hyde) (3), 40% of the self-condensation product, (E)-2-n- sition state which better fits in the pores of the zeolite.
pentyl-2-n-nonenal (4), 23% of the trans-1-phenylheptene
In order to improve the selectivity to jasminaldehyde,
(5), and 5% of heptanoic acid (6) (Scheme 1). The forma- we have used as acid catalysts a 12-membered-ring tridi-
tion of 3 and 4 can be explained through a competitive acid- rectional Y zeolite, in which the diffusion of reactants and
catalyzed aldolic condensation which involves the elec- product should be easier, and which has larger internal void
trophilic attack of a protonated carbonyl group, from spaces (internal cavities) than H-Beta zeolite. The reaction
benzaldehyde or heptanal, to the enolic form of heptanal was carried out under the same reaction conditions and
(Scheme 2). Product 5 could be formed by oxidation of Fig. 2 shows the yields of the different products obtained
jasminaldehyde followed by a fast decarboxylation or by versus time. As can be seen there, the zeolite pore topology
direct deformylation following an acid-catalyzed tandem clearly affects the rate as well as the final distribution of
Aldol–Grob sequence (11).
the different products with the Y zeolite. In this case, we
In Fig. 1 are presented the yields of the different products observe that jasminaldehyde (3) and 2-n-pentyl-2-nonenal
versus time. As can be seen there, the rate of the compet- (4) are formed with similar reaction rates, the selectivity to
itive self-condensation (4) is significantly larger than the jasminaldehyde being higher than that obtained with Beta
zeolite. Only 2% of product 5 was detected in the final re-
action mixture. However, despite the higher selectivity to
jasminaldehyde, the final conversion of heptanal was low,
FIG. 1. Yields of the different products 2 ( ), 3 (᭹), 4 (᭺), 5 (5),
and 6 (+) versus reaction time in the condensation of benzaldehyde
and heptanal using HBeta zeolite (10 wt% ), a reactant molar ratio
PhCHO : heptanal = 5 : 1 at 398 K.
FIG. 2. Yields of the different products 2 ( ), 3 (᭹), 4 (᭺), and 5 (5)
versus reaction time in the condensation of benzaldehyde and heptanal
using HY-100 zeolite (10 wt% ), a reactant molar ratio PhCHO : heptanal =
5 : 1 at 398 K.

388
CLIMENT ET AL.
achieving after 16 h of reaction time 57% of the maximum
attainable conversion, and this conversion remained con-
stant with time. The behavior of the reaction with time in-
dicates that a fast deactivation of the catalyst is occurring
during the process. In a series of previous articles we have
shown that one of the major causes of zeolite deactivation
during liquid-phase organic reactions is the formation of
persistent carbocationic species (12).
Deactivation derives from two different effects: (i) for-
mal replacement of catalytically active protons by fairly sta-
ble unactive carbocations, and (ii) blocking of the pores
by these strongly adsorbed, bulky reaction intermediates
(12b). In some cases we have proved that these poisons
have a definite structure that can be unequivocally charac-
terized by spectroscopic techniques (13).
Particularly intriguing has been the case of styrene and
its derivatives, wherein earlier reports claiming the gener-
ation of adsorbed radical cations have been shown later by
us to correspond in reality to diphenyl allylic carbocations
(14). In the present case we noticed that one of the reac-
tion products corresponds to an -alkyl-substituted styrene.
We speculate that this product could also lead to diphenyl
allyl carbenium poisons analogous to those previously re-
ported by us. Thus, we undertook a spectroscopic study of
the deactivated catalyst used in the reaction after being sub-
mitted to exhaustive solid–liquid extraction. The purpose
is to demonstrate whether deactivation of the catalyst in
the present aldolic condensation also occurs by formation
of persistent diphenyl allyl carbocations.
FIG. 4. Aromatic region of FT-IR spectrum of self-supported H-Beta
wafersafter exhaustive solid–liquid extraction. The spectrum wasrecorded
at room temperature after outgassing at 473 K for 1 h. Note the absence of
any C O vibration band at about 1690 cm 1 and the presence of a strong
==
absorption at 1615 cm 1 characteristic of allylic cations.
that max of the parent 1,3-diphenyl allyl cation should red-
shift 30 nm by double alkyl substitution on the allylic moi-
ety, thus giving a value of 530 nm, as is in fact observed for
deactivated HY and H-Beta catalysts.
The IR spectrum of self-supported wafers of deactivated
HY catalyst is shown in Fig. 4. It has been proposed that
diphenyl allyl cation has a characteristic vibration band at
ca. 1550 cm 1, probably due to the the stretching vibration
–
of a C C bond with a single and a half bond order. As it can
be seen in Fig. 4, this specific band is present accompanied
1
by other bands at 1605 and 1595 cm attributable to the
Figure 3showsthe diffuse reflectance UV–visspectrum of
deactivated HY-100 and H-Beta catalysts showing a charac-
teristic absorption band centered at 520 nm with a shoulder
at the short-wavelength side of the peak. These bands have
close similaritywith the reported diffuse reflectance UV–vis
spectrum of the parent 1,3-diphenyl allyl cation embedded
within ZSM-5 (13a). Furthermore, according to the Wood-
ward rules of additivity on the influence of substituents on
skeletal vibration of phenyl rings.
Thus, on the basis of the diffuse reflectance UV–vis and
IR spectra aswell asthe known behavior ofrelated styrenes,
we propose that deactivation of HY catalyst is due, in this
case, to the formation of 2-butyl-1,3-diphenyl-2-octen-1-yl
cation. The same species is observed also using H-Beta as
catalyst. A likely reaction mechanism to account for the
generation of this carbocation is shown in Scheme 3.
From the results presented above we can conclude that
the use of microporous acid materials as catalysts for the
==
max of conjugated C C double bonds (15), it is calculated
FIG. 3. Diffuse reflectance UV–vis spectra (plotted as the inverse of
the reflectance, R) of HY (a), H-Beta (b), and ALPO (c) used as catalysts
for the jasminaldehyde formation after exhaustive solid–liquid extraction.
SCHEME 3

SYNTHESIS OF JASMINALDEHYDE
389
TABLE 1
Relevant Physicochemical Parameters of the Catalysts
Used in This Work
Average
Surface area Pore volumea
poreb
1
1
Catalysts
Beta
(m2 g
)
(cm3 g
)
diameter (nm) Si/Al
606
0.175
0.76 0.64
0.55 0.55
0.74
13
HY-100
709
900
211
0.323
0.500
0.328
2.5
15
—
Al MCM-41
ALPO (Al/P = 1)
3.5
7.2
a Measured by N2.
b Measured by Ar.
FIG. 5. Yields of the different products 2 ( ), 3 (᭹), 4 (᭺), and 5 (5)
versus reaction time in the condensation of benzaldehyde and heptanal
using Al MCM-41 catalyst (10 wt% ), a reactant molar ratio PhCHO :
heptanal = 5 : 1 at 398 K.
sample studied: an Al-free MCM-41 sample and an amor-
phous aluminum phosphate (ALPO). The physicochemical
parameters of the catalysts used are summarized in Table 1.
With the first one the acid character should be attributed ex-
clusively to silanol groups, whereas amorphous aluminium
phosphate, built of tetrahedral units of AlO4 and PO4, is
structurally similar to silica and its surface contains both
Al and P hydroxyl groups (17). When the reaction was
tested with both catalysts we observed that while the all-
silica MCM-41 did not catalyze the condensation at all,
with the ALPO material optimum results were obtained.
Indeed, amorphous ALPO behaves as an active and selec-
tive catalyst for this reaction, achieving practically a total
conversion after 3 h of reaction time with a selectivity to
jasminaldehyde of 85% (Fig. 6). In Table 2 the results ob-
tained in the condensation between 1 and 2 on the different
catalysts used in this work are summarized. Moreover, in
contrast to HY, H-Beta, and Al MCM-41, product 5 was
not detected in the reaction mixture and the spectroscopic
condensation of benzaldehyde and heptanal leads to low
selectivities to jasminaldehyde and fast deactivation of
the catalyst due to consecutive reactions occurring within
the pores, which end with the formation of persistent 2-
butyl-1,3-diphenyl-2-octen-1-yl cation. Because of that, we
thought that mesoporous aluminosilicates materials could
be better acid catalysts for this condensation than the
most constrained microporous molecular sieves. Indeed,
the presence of large regular mesopores can allow the dif-
fusion of reactants and conversely the fast diffusion of the
products out, minimizing unwanted consecutive reactions
and catalyst decay. The condensation between benzalde-
hyde and heptanal was then carried out using as catalyst
Al MCM-41 with a Si/Al ratio of 15 and pore diameter of
3.5 nm. The evolution of the different reaction products
with time is presented in Fig. 5. In contrast with the results
obtained with the microporous catalysts, two facts are now
observed: first, it is possible to achieve a total conversion of
heptanal in relatively short reaction times (6 h), indicating
that in this case the catalyst decay is practically negligible.
Second, at heptanal conversions higher than 75% , the for-
mation of jasminaldehyde is favored as could be expected
from our previous hypothesis on the catalytic behavior of
the microporous materials for that reaction.
Furthermore, the better results obtained with Al MCM-
41 despite the weaker acitity of the sites present in this
material (16) allow us to conclude that, under our reaction
conditions, the presence of strong acid sites is not necessary
to perform the aldolic condensation. Moreover, it is not un-
reasonable to think that the absence of confinement effects
along with the weakness of the acid sites in Al MCM-41 de-
creases the fast conversion of jasminaldehyde to compound
5 that was largely obtained when using zeolites as catalysts.
In order to check these hypotheses, we selected as cata-
FIG. 6. Yields of the different products 2 ( ), 3 (᭹), and 4 (᭺) versus
reaction time in the condensation of benzaldehyde and heptanal using
ALPO catalyst (10 wt% ), a reactant molar ratio PhCHO : heptanal = 5 :1
lysts two solids with weaker acid sites than the Al MCM-41 at 398 K.

390
CLIMENT ET AL.
TABLE 2
itating the attack of the carbanion formed on a basic site.
The reaction rate is thereby increased since the controlling
step of the reaction has been shown to be the bimolecu-
lar condensation step, i.e., the attack of the carbanion to
the carbonyl group. Following this idea, it is obvious that
the weaker the basic site of the catalyst the less negatively
charged will be the carbanionic species, and consequently
a higher density of positive charge will be necessary on the
Results Obtained in the Condensation of Benzaldehyde and
Heptanal with Different Catalysts
Poison
in the
Yields
(% )
Reaction Conversion catalyst Selectivity
Catalysts
time (h)
(% )
(% )a
to 3
3
4
5
H-Beta
HY-100
Al MCM-41
ALPO
Na-ALPO
6
16
10
3
93b
53
99
96
71
5.1
7.4
—
1.5
—
21
52
56
86
67
20 40 23 carbon of the carbonyl group in the benzaldehyde deriva-
28 15
56 35
83 13
48 23
3
8
—
—
tive in order for the condensation reaction to proceed effi-
ciently.
Using this model as an analogy, we can assume that in an
acid–base bifunctional catalyst, such as ALPO, the role of
the weak acid sites is to interact with the carbonyl group
of the benzaldehyde producing a polarization of this group
and increasing the positive charge on the corresponding
carbon atom. If this occurs then the attack of the carbanion
formed on the relatively weak basic sites of ALPO should
be favored and this would explain the good reactivity of
ALPO for the reaction studied. Furthermore, it has been
reported that in the homogeneous base-catalyzed conden-
22
a Determined by TG analysis and based on moles of heptanal retained.
b 5% of heptanoic acid was also detected.
characterization of the ALPO catalyst after the reaction in-
dicated that the 1,3-diphenyl allylic cation was absent (see
Fig. 3). We propose that the lower acidity of the latter solid
and its lack of microporosity can explain the absence of any
persistent carbocationic poison. In this case even if such
positive species were formed, they would be able to react
with nucleophiles, giving a neutral compound that would
desorb from the acid site. Thus, ALPO would be less prone
than zeolites to undergo deactivation.
It has been reported in the literature that ALPO is a
bifunctional catalyst whose surface acid–base sites are able
to participate in catalytic processes (18). The presence of
acid and basic sites has been convincingly demonstrated by
tritation with weak bases or acids using Hammet indicators
(19), or spectrophotometric methods (20), as well as by IR
spectroscopy of absorbed probe molecules, and by means
of test reactions (18, 21).
Concerningthe nature ofthe acidicand basicsites, molec-
ular orbital calculations on acidic active sites on ALPO
catalysts (22, 23) showed that P–OH groups represent the
most stable Brønsted acid sites, while phosphorous atoms
function as Lewis acid sites on ALPO surfaces. In addition,
the proton abstraction energies for Al–OH groups indicate
that these centers cannot be considered as strong Brønsted
sites but, however, might enhance the Brønsted acidity of
P–OH groups through H-bonding. On the other hand, ex-
posed oxygens of P–OH groups are found to act as Lewis
basic sites since they have higher electronic density than
bridged oxygens (22).
sation of alkanals with benzaldehyde the selectivity to
-
alkylcinnamaldehydes can be increased by the presence of
a small amount of an acid (25).
In order to support such a model we have adsorbed ben-
zaldehyde on pure silica MCM-41 and ALPO in a vac-
uum IR cell and it can be seen in Fig. 7 that, while in
MCM-41 benzaldehyde is physically adsorbed (see the car-
bonyl group at 1700 cm 1), on ALPO we observe the ma-
jority presence of a chemically interacting carbonyl group
at 1655 cm 1 (interaction shift of 50 cm 1). This is an indi-
cation that a certain polarization of the carbonyl group has
occurred in accordance with the proposed reaction model.
Finally, the presence of those weak acid sites in ALPO has
also been shown by using a reaction test for weak acid sites
such as the acetalization of benzaldehyde with trimethyl
On the basis of the demonstrated acid–base bifunction-
ality of the ALPO catalysts, we have tried to explain the
excellent catalytic activity and selectivity of ALPO for the
synthesis of jasminaldehyde.
In a previous work (24) it has been demonstrated that
the reactivity of benzaldehyde for Knoevenagel condensa-
tion reactions could be enhanced by introducing substitut-
ing groups in the aromatic ring that increase the density of
FIG. 7. IR spectra of benzaldehyde adsorbed on ALPO and Al-free
positive charge on the carbon of the carbonyl group, facil- MCM-41.

SYNTHESIS OF JASMINALDEHYDE
391
SCHEME 5
we have neutralized the acid sites present in ALPO. Never-
theless, by neutralizing the acid sites by Na+ exchange one
also increases the basicity of the solid. We have seen that
this occurs in our case by using a test reaction that can only
be catalyzed by basic sites, i.e., the Knoevenagel condensa-
tion between benzaldehyde and malononitrile (Scheme 5).
The results presented in Fig. 9 show that both the ALPO
and the Na-exchanged ALPO have basic sites active for
this reaction, but the activity of Na-ALPO is higher as a
consequence of its higher basicity.
At this point, we know that ALPO has acid and base
sites, while Na-exchanged ALPO has only basic sites that
are stronger than those ofALPO. Thus, ifonlythe basicsites
of ALPO were responsible for the catalytic behavior of this
materialin the synthesisofjasminaldehyde, we could expect
Na-ALPO to be an even more active catalyst than ALPO
for this reaction. However, the results presented in Fig. 10
clearly indicate that Na-ALPO gives a lower activity and a
worse selectivity to jasminaldehyde than ALPO. Thus, we
can conclude that a bifunctional acid–base mechanism such
as the one presented in Scheme 4 would be consistent with
the results obtained for the synthesis of jasminaldehyde on
ALPO.
SCHEME 4
orthoformate. Indeed, the results showed that ALPO
catalyst is able to catalyze the formation of acetals achiev-
ing the benzaldehyde dimethyl acetal (formed with an ini-
4
1
tial rate of 1.8 10 mol min
g
1), in good agreement
with the results reported by Bautista et al. (26).
In another experiment that can give information on the
potential role of the acid sites in the condensation reaction,
we prepared an ALPO sample in which the acidity was
neutralized by treating the ALPO with a 1 M solution of
sodium acetate followed by washing and drying. After this
treatment, we showed by IR that the stretching bands of
Influence of the Reaction Conditions
1
OH groups associated with Al and P (3793 and 3678 cm
As was said before, the general methodology to obtain
jasminaldehyde with high selectivity requires that the con-
centration of heptanal be kept low in order to decrease the
reaction rate of the self-condensation. This can be achieved
by adding slowly the heptanal into the reaction mixture
respectively), and which were present in the original ALPO,
have disappeared (see Fig. 8). Meanwhile, the catalytic ac-
tivity for the acetalization of benzaldehyde with TMOF was
dropped to zero, confirming that with the above treatment
FIG. 8. IR spectra in the hydroxyl range of ALPO (a) and Na-
exchanged ALPO (b).
FIG. 9. Knoevenagel condensation between benzaldehyde (10 mmol)
and malononitrile (10 mmol) at 333 K on ALPO (᭹) and Na-ALPO (᭺).

392
CLIMENT ET AL.
FIG. 11. Yield of 3–heptanal conversion plot using benzaldehyde to
FIG. 10. Yields of the different products 2 ( ), 3 (᭹), and 4 (᭺) versus
heptanal molar ratios of 3 ( ), 5 (᭹), 7 (᭺), and 9 (5) in the presence of
reaction time in the condensation of benzaldehyde and heptanal using Na-
ALPO catalyst (10 wt% ), a reactant molar ratio PhCHO : heptanal = 5 :1
at 398 K.
ALPO catalyst at 398 K.
out at 383 K and 413 K, using ALPO (10 wt% ) as catalyst.
The results are presented in Figs. 12a and 12b. As was ex-
pected, a decrease of temperature causes a decrease in the
heptanal disappearance rate, whereas an increase of tem-
perature leads to a total heptanal conversion within 1 h of
which involves long reaction times or by directly using a
high benzaldehyde to heptanal molar ratio, which requires
larger reactors.
In order to study the influence of the molar ratio of the
reagents on the selectivity to jasminaldehyde and then to
optimize this parameter, we have carried out the conden-
sation reaction using molar ratios of benzaldehyde to hep-
tanal of 3, 5, 7, and 9 in the presence of ALPO as catalyst
(10 wt% ) at 398 K. Interestingly, on decreasing the ratio
of benzaldehyde to heptanal from 5 to 3, the initial hep-
tanal disappearance rate undergoes a 3-fold increase. This
result could be explained on the basis that when the con-
centration of heptanal is increased, this reagent competes
with benzaldehyde for adsorption on the acid sites, in such a
way that it would be possible that the protonation and con-
sequently the activation of heptanal carbonyl group would
take place to a considerable extent. If this is so, it should
be expected that the self-condensation would be favored
and the selectivity to jasminaldehyde would decrease. In-
deed, when the yield of jasminaldehyde versus conversion
using different molar ratio of reagents is plotted (Fig. 11) it
can be seen the decrease in jasminaldehyde selectivity for
a benzaldehyde/heptanal ratio of 3.
On the other hand, when the molar ratio of reagents
is increased above 5 appreciable changes in the heptanal
disappearance reaction rate were not observed, indicating
that beyond this molar ratio the majority of acid sites are
already preferentially interacting with benzaldehyde, and
therefore, no changes in selectivity to jasminaldehyde are
observed for benzaldehyde/heptanal molar ratios higher
than 5 (Fig. 11).
Temperature is another important parameter which not
only affects the reaction rate but also can affect the selec-
tivity for a given reaction. In order to study the influence
of the temperature, the condensation reaction was carried
FIG. 12. Yields of the different products 2 ( ), 3 (᭹), 4 (᭺), and 6 (+)
versus reaction time in the condensation of benzaldehyde and heptanal
using ALPO catalyst (10 wt% ), a reactant molar ratio PhCHO : heptanal =
5 : 1 at 383 K (a) and 413 K (b).

SYNTHESIS OF JASMINALDEHYDE
393
ACKNOWLEDGMENTS
The authors acknowledged financial support by the Direccio´n General
de Investigacio´n Cient´ıfica y Te´cnica of Spain (Project MAT2000-1392).
REFERENCES
1. (a) Farbenindustrie I. G., D R Patent 284, 458/1927; (b) Payne, L. S.,
European Patent 0392579 A2, 1990.
2. Mastagli, P., and Durr, G., Bull. Soc. Chim. Fr. 268 (1955).
3. Sarkar, A., Dey, P. K., and Datta, K., Indian J. Chem. B 25, 656 (1986).
4. Abenhaim, D., Ngoc Son, C. P., Loupy, A., and Ba Hiep, N., Synth.
Commun. 24, 1199 (1994).
5. Mastagli, P., and Lagrange, G., Compt. Rend. 244, 207 (1957).
6. Climent, M. J., Corma, A., Guil-Lopez, R., Iborra, S., and Primo, J.,
J. Catal. 175, 70 (1998).
7. Lindblad, T., Rebenstorf, B., Zhi-Guang, Y., Lars, S., and Anderson,
T., Appl. Catal. A 112, 187 (1994).
8. Corma, A., Grande, M. S., Gonzalez-Alfaro, V., and Orchilles, A. V.,
J. Catal. 159, 375 (1996).
FIG. 13. Yield of 3–heptanal conversion plot using a benzaldehyde to
heptanal molar ratio of 5 in the presence of ALPO catalyst at 388 K ( ),
398 K (᭹), and 413 K (᭺).
9. Beck, J. S., Vartulli, J. C., Roth, W. J., Leonowicz, M. E., Kresge, C. T.,
Schmitt, K. D., Chu, C. T. W., Olson, D. H., Sheppard, E. W., McCullen,
S. B., Higgins, J. B., and Schlenker, J. L., J. Am. Chem. Soc. 114, 10834
(1992).
reaction time. However, when the yield of jasminaldehyde
is compared at same level of conversion (Fig. 13), the se-
lectivity is practically the same at temperatures between 10. Corma, A., Climent, M. J., Garc´ıa, H., and Primo, J., Appl. Catal. 51,
113 (1989).
383 K and 398 K, while at 413 K or higher a decrease in the
selectivity should be expected.
11. (a) Kabalka, G. W., Li, N. S., Tejedor, D., Malladi, R. R., and Trotman,
S., J. Org. Chem. 63, 6438 (1998); (b) Kabalka, G. W., Li, N. S., Tejedor,
D., Malladi, R. R., and Trotman, S., J. Org. Chem. 64, 3157 (1999).
12. (a) Adam, W., Casades, I., Forne´s, V., Garc´ıa, H., and Weichold, O.,
J. Org. Chem. 65, 3947 (2000); (b) Corma, A., and Garc´ıa, H., J. Chem.
Soc., Dalton Trans. 9, 1381 (2000).
13. (a) Cano, M. L., Forne´s, V., Garc´ıa, H., Miranda, M. A., and Pe´rez-
Prieto, J., J. Chem. Soc., Chem. Commun. 2477 (1995); (b) Cano,
M. L., Corma, A., Forne´s, V., Garc´ıa, H., Miranda, M. A., Baerlocher,
C., and Lengauer, C., J. Am. Chem. Soc. 118, 11006 (1996).
14. (a) Cozens, F. L., Bogdanova, R., Re´gimbald, M., Garc´ıa, H., Mart´ı,
V., and Scaiano, J. C., J. Phys. Chem. B 101, 6921 (1997); (b) Forne´s, V.,
Garc´ıa, H., Mart´ı, V., and Ferna´ndez, L., Tetrahedron 54, 3837 (1998).
15. Silverstein, R. M., Bassler, G. C., and Morrill, T. C., “Spectroscopic
Identification ofOrganicCompounds,” 4th Ed. Wiley, NewYork, 1981.
16. Corma, A., Forne´s, V., Navarro, M. T., and Perez-Pariente, J., J. Catal.
148, 569 (1994).
CONCLUSIONS
It has been shown that the use of zeolites as catalysts
for the condensation of benzaldehyde and heptanal leads
to low selectivities to jasminaldehyde and fast deactivation
of the catalyst. These results are attributed to the confine-
ment effects of the reactants and products inside the voids
of the microporous materials, which lead to the preferen-
tial formation of the heptanal self-condensation product as
well as to consecutive reactions occurring within the pores,
which end with the formation of persistent diphenyl allyl
cation. Whereas mesoporous aluminosilicate (AlMCM-41)
exhibits better activity and selectivity to jasminaldehyde 17. Rebenstorf, B., Lindblad, T., and Andersson, S. L. T., J. Catal. 128, 293
(1991).
than zeolites, the best results are obtained using amorphous
aluminophosphate as catalyst. This finding was explained
on the basis of the acid–base bifunctional character of the
ALPO, where the role of the weak acid sites is the activa-
18. Moffat, J. B., Catal. Rev. Sci. Eng. 18, 199 (1978).
19. (a) Kearby, K., “Proceedings, 2nd International Congress on Catalysis,
Paris, 1960,” p. 2567. Technip, Paris, 1961; (b) Moffat, J. B., and Chao,
E., J. Catal. 46, 151 (1977).
tion of benzaldehyde by protonation of the carbonyl group 20. Campelo, J. M., Garcia, A., Luna, D., and Marinas, J. M., Can. J. Chem.
62, 638 (1984).
favoring then the attack of the enolate heptanal intermedi-
ate generated on the relatively weak basic sites of ALPO.
This hypothesis is supported by the fact that the IR band
of the carbonyl group of benzaldehyde shifts toward lower
21. Aramendia, M. A., Borau, V., Garcia, I. M., Jimenez, C., Marinas,
J. M., Porras, A., and Urbano, F. J., Appl. Catal. A 184, 115 (1999).
22. Moffat, J. B., Vetrivel, R., and Viswanathan, B., J. Mol. Catal. 30, 171
(1985).
wavelength when the benzaldehyde is adsorbed on ALPO, 23. Zhidomirov, G. M., and Kazansky, V. B., Adv. Catal. 34, 131 (1983).
24. Cabello, J. A., Campelo, J. M., Garc´ıa, A., Luna, D., and Marinas,
J. M., J. Org. Chem. 49, 5195 (1984).
25. Kuiterman, A., Castelijns, A. M. C. F., Dielemans, H. J. A., and Green,
indicating that a certain polarization of the carbonyl group
occur. Moreover, when acid sites on ALPO are neutralized
by treating with sodium acetate it is found that Na-ALPO
R., European Patent 0771780 A1, 1997.
gives lower activity and worse selectivity to jasminaldehyde
than ALPO.
26. Bautista, F. M., Campelo, J. M., Garc´ıa, A., Leon, J., Luna, D., and
Marinas, J. M., J. Prakt. Chem. 336, 620 (1994).