Gem-diamines as highly active organocatalysts for carbon-carbon bond formation

Article abstract of DOI:10.1016/j.jcat.2006.11.029

Diamines with neighbour nitrogen atoms have been used as base catalysts in the Knoevenagel condensation reaction between benzaldehyde and ethyl cyanoacetoacetate. The catalytic results show that a good basic catalyst requires a combination of two factors: high proton affinity and the ability to return the proton to the oxoanion intermediate. Computational chemistry calculations show this by characterizing the reactants, products, and transition states and by calculating the activation energies of the different reaction steps. A diamine, di(3-methylpiperidine)methane (B), has been found with a higher catalytic activity than DMAN despite its lower proton affinity, demonstrating that not only the proton affinity, but also the steric ability to abstract the protons, are important in explaining the catalytic results.

Full text of DOI:10.1016/j.jcat.2006.11.029

Journal of Catalysis 246 (2007) 136–146
www.elsevier.com/locate/jcat
Gem-diamines as highly active organocatalysts for carbon–carbon bond
formation
∗
María J. Climent, Avelino Corma , Irene Domínguez, Sara Iborra, María J. Sabater, Germán Sastre
Instituto de Tecnología Química UPV-CSIC, Av. Los Naranjos, s/n 46022, Valencia, Spain
Received 28 September 2006; revised 28 November 2006; accepted 28 November 2006
Available online 4 January 2007
Abstract
Diamines with neighbour nitrogen atoms have been used as base catalysts in the Knoevenagel condensation reaction between benzaldehyde and
ethyl cyanoacetoacetate. The catalytic results show that a good basic catalyst requires a combination of two factors: high proton affinity and the
ability to return the proton to the oxoanion intermediate. Computational chemistry calculations show this by characterizing the reactants, products,
and transition states and by calculating the activation energies of the different reaction steps. A diamine, di(3-methylpiperidine)methane (B), has
been found with a higher catalytic activity than DMAN despite its lower proton affinity, demonstrating that not only the proton affinity, but also
the steric ability to abstract the protons, are important in explaining the catalytic results.
©
2006 Elsevier Inc. All rights reserved.
Keywords: Gem-diamines; Basic catalysis; Computational chemistry; Knoevanagel condensation
1
. Introduction
stance, the proton sponge 1,8-bis(dimethyl amino) naphthalene
DMAN; see Scheme 1) has a pKa of 12.1. It has been proposed
(
that the steric strain is the main reason for the high basicity
constant of these compounds [8]. Thus, it has been reported
that DMAN or DMAN supported on MCM-41 is an excellent
base catalyst for Knoevenagel and Claisen–Schmidt condensa-
tions [5,9].
Knoevenagel condensation is widely used in organic syn-
thesis to produce important intermediates and end products for
perfumes, pharmaceuticals, and polymers [10] (Scheme 1). It
can be catalyzed by a wide variety of bases from weak to strong
basicities and has proven to be a very adequate test reaction for
base catalysis. In fact, reacting benzaldehyde with compounds
containing activated methylenic groups with different pKa val-
ues, such as ethyl cyanoacetate (pKa ꢀ 9), ethyl acetoacetate
The development of clean processes for production of fine
chemicals has driven research in the areas of solid base- and
organic base-supported catalysts [1]. Micelle template silicas
(
MTS), such as MCM-41, containing a large number of silanol
groups that allows the grafting of amines, provide an excellent
inorganic support for producing heterogenized base catalysts
with tailored pore sizes and high surface areas, in which the
basicity can be controlled by changing the pK of the grafted
a
amine. Primary, secondary, and tertiary amines [2] ammonium
hydroxide [3], guanidines [4], diamines, and proton sponges [5]
have been grafted on MTS and used as heterogeneous basic
catalysts for various basic catalyzed organic reactions, such as
aldol and Knoevenagel condensations, alkylation and elimina-
tion reactions, Michael-type additions, transesterifications, and
so on [6]. Among these, diamines with neighbouring nitrogen
atoms at short distances and aromatic frames, such as naphtha-
lene, fluorine, and phenantrene, which are denominated proton
sponges, show a well-defined and strong basicity [7]; for in-
(
pKa ꢀ 10.7), and diethyl malonate (pKa ꢀ 13.3), makes it
possible to evaluate the basic strength of a catalyst [11]. The
kinetics of the Knoevenagel reaction have been widely studied,
and it is generally agreed [12] that the reaction is first order with
respect to each reactant and the catalyst.
In the present work, different diamines with vicinal nitrogen
atoms that exhibit a strong basic character have been prepared.
The behaviour of these stable diamines as homogeneous base
catalysts for the Knoevenagel condensation reaction of ben-
*
Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.11.029

M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
137
Scheme 1. Schematic representation of dipiperidinemethane (amine A), di(3-methylpiperidine)methane (amine B), di(3,5-dimethylpiperidine)methane (amine C),
-methylpiperidine (amine D), 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) and 1,8-bis(dimethyl amino) naphthalene (DMAN).
3
zaldehyde with reactants with different pKa values has been
studied. To explain the catalytic results, the proton affinities of
the different amines have been calculated and the interactions
between the transition state complexes and the base catalyst
established. This work demonstrates that these gem-diamines
behave as strong homogeneous base catalysts and are excellent
candidates for heterogenization.
2
. Experimental
Scheme 2. Schematic representation of the Knoevenagel condensation between
benzaldehyde and carboxylates 1a–1c.
2
.1. Synthesis of the amines
Amines A and B are known compounds and were prepared
according to a procedure described in the patent literature [13].
2
5.9, 31.4, 33.9, 53.0, 60.9 and 82.7 ppm; IE MS (m/z): 209
+
(M ), 112, 98.
2
.1.1. Synthesis of dipiperidinemethane (A)
A 50-mL round-bottomed flask was charged with 3.74 g
2
.1.3. Synthesis of di(3,5-dimethylpiperidine)methane (C)
The same procedure was applied for the synthesis of amine
(
44 mmol) of piperidine, 0.66 g (22 mmol) of paraformalde-
C. A 50-mL round-bottomed flask was charged with 4.97 g
44 mmol) of 3,5-dimethylpiperidine, 0.66 g (22 mmol) of
hyde, and 17 mL of acetonitrile. The mixture was heated to
reflux for 20 h. The solvent was evaporated to dryness, and
the residue was distilled under vacuum to give 3.1 g of the di-
(
paraformaldehyde, and 17 mL of acetonitrile. The mixture was
heated to reflux for 20 h. The solvent was evaporated to dry-
ness, and the residue was distilled under vacuum to give 4.6 g
of the diamine (yield: 89%). Spectroscopic data were as fol-
1
amine A (yield: 78%). Spectroscopic data were as follows: H
NMR (300 MHz, CDCl3): δ = 1.42–1.44 (m), 1.51–1.56 (m),
.38–2.40 (m) and 2.82 (s) ppm; 13C NMR (75 MHz, CDCl3):
2
1
+
lows: H NMR (300 MHz, CDCl ): δ = 0.9 (d), 1 (d), 1.4 (t)
3
δ = 25.3, 26.5, 53.4 and 83.1 ppm; IE-MS (m/z): 182 (M ),
and 2.9 (s) ppm; 13C NMR (300 MHz, CDCl3): δ = 19.7, 31.0,
1
10, 97, 84.
+
4
1
1
1
2.9, 59.9 and 82 ppm; IE-MS (m/z): 238 (M ), 237, 126,
12, 110, 98, 97, 83, 84. Elemental analysis: found—75.3% C,
2.7% H, 11.7% N; C15H30N2; requires—75.6% C, 12.6% H,
1.8% N.
2
.1.2. Synthesis of di(3-methylpiperidine)methane (B)
A 50-mL round-bottomed flask was charged with 4.35 g
(
44 mmol) of 3-methylpiperidine, 0.66 g (22 mmol) of para-
formaldehyde, and 17 mL of acetonitrile. The mixture was
heated to reflux for 20 h. The solvent was evaporated to dry-
ness, and the residue was distilled under vacuum to give 4.1 g
of the diamine B (yield: 90%). Spectroscopic data were as fol-
2.1.4. Synthesis of 1,2-di(3-methylpiperidine)ethane (E)
A 100-mL round-bottomed flask was charged with 3.1 g
(31.3 mmol) of 3-methylpiperidine, 13.3 g (96.2 mmol) of
K2CO3, and 40 mL of chloroform. The mixture was stirred
at room temperature for 15 min, and 1.4 mL (16.4 mmol)
1
lows: H NMR (300 MHz, CDCl3): δ = 0.9 (d), 1.4–1.8 (m),
13
1
.8 (t), 2.8 (s) ppm; C NMR (75 MHz, CDCl3): δ = 20.1,

138
M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
Scheme 3. Mechanism of the Knoevenagel condensation reaction. (A) Formation of the carbanion on the methylenic group upon the action of the basic catalyst;
B) attack of this carbanion intermediate to the carbonyl group and (C) elimination of the hydroxyl group to form a C=C bond and water.
(
(
1)
(
2)
Scheme 4. Reaction steps that take place in the condensation between the ethyl cyanoacetate and benzaldehyde using amine B (reaction (1)) and using DMAN
reaction (2)) as catalyst. Step 1 is the proton transfer from the ethyl cyanoacetate to the catalyst; step 2 is the formation of the oxoanion intermediate; and step 3 is
the release of the proton from the catalyst to the oxoanion to give the Knoevenagel condensation product.
(
of 1,2-dibromoethane was added. The suspension was stirred
at room temperature while being monitored by chromatog-
raphy. The solid was filtered off and washed with chloro-
form. The organic extracts were dried with Na2SO4, fil-
12.7% H, 12.3% N; C14H28N2; requires—75% C, 12.5% H,
12.5% N.
2.1.5. General procedure for the Knoevenagel reactions
In a typical experiment, 0.14 mmol of the basic catalysts
were added to a solvent-free solution of ethyl cyanoacetate
(3.1 g, 28 mmol) while being stirred under inert atmosphere.
After temperature adjustment, 0.85 g of benzaldehyde (3.4 g,
32 mmol) was added, with the reaction periodically monitored
by gas chromatography using a Hewlett Packard GC 5988 A
tered, and concentrated to dryness to give a white solid
1
(
(
(
3
2
yield: 78%). Spectroscopic data were as follows: H NMR
300 MHz, CDCl3): δ = 0.8 (d), 1.5–1.7 (m), 1.8–1.9 (m), 2.4
s) and 2.8 (t) ppm; 13C NMR (300 MHz, CDCl3): δ = 25.6,
1.4, 32.8, 33.0, 54.5, 56.6 and 62.6 ppm; IE-MS (m/z):
24 (M ), 126, 112. Elemental analysis: found—74.6% C,
+

M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
139
with an FID detector and an EQUITYTM-5 Fused silica capil-
lary column of 5% phenylmethylsilicone. Dodecane was used
as an internal standard in all of the experiments.
2
.1.6. Quantum chemical calculations
Quantum chemical calculations were performed at the ab
initio Hartree–Fock (HF) level of theory using the 3-21G ba-
sis sets [14]. We chose this method in an attempt to reach a
compromise between accuracy and computational cost. This
methodology was used to characterise all of the reaction in-
termediates in the condensation reaction of benzaldehyde and
ethyl cyanoacetate, using amine B (see Scheme 1) as the cat-
alyst. The global minima of the reactant and product species
were optimised using Schlegel’s algorithm [15], which includes
the rational function optimisation (RFO) approach [16]. The
transition states were fully characterised by the TS search algo-
rithm [17], and the vibrational frequencies were subsequently
calculated by determining the second derivatives of the en-
ergy with respect to the Cartesian nuclear coordinates and then
transforming to mass-weighted coordinates. This allowed us
to verify that only one eigenvalue was negative, correspond-
ing to a transition state species. We did not include zero point
energies and other more sophisticated techniques for charac-
terising the energy of the reactions in our energy calculations,
because we were interested only in comparing the different
reaction steps as shown in Scheme 4. For the calculation of
proton affinities, a density functional methodology with the
B3LYP [18] functional and the 6-31G** basis set [19] was
used. All quantum chemical calculations were performed us-
ing GAUSSIAN98 [20].
Fig. 1. Knoevenagel condensation of benzaldehyde (32 mmol) and ethyl
cyanoacetate (28 mmol) without solvent at room temperature with diamine A
(P) (0.14 mmol), diamine B (!), and diamine C (×).
3
. Results and discussion
Fig. 2. Knoevenagel condensation of benzaldehyde (32 mmol) and ethyl cya-
noacetate (28 mmol) without solvent at room temperature with diamine B (!),
DMAN (1) and TBD (×) (0.14 mmol).
To check the basicity of the different synthesised diamines
A, B, and C; Scheme 1), we used the Knoevenagel reaction
(
as the reaction test. Thus, condensation between benzaldehyde
and ethyl cyanoacetate was carried out in the absence of sol-
vent at room temperature using the different diamines as base
catalysts (Scheme 2). In all cases, the E and Z isomers of α-eth-
yl-2-cyanocinnamate in an E/Z ratio >160 were obtained with
a selectivity of 100%.
The accepted mechanism for the base-catalyzed Knoeve-
nagel condensation involves a first step in which the formation
of the carbanion on the methylenic group occurs by abstraction
of the proton by the basic catalyst. This is followed by attack of
the carbanion intermediate to the carbonyl group. Finally, elim-
ination of the hydroxyl group occurs to form a C=C bond and
water, while the basic site is restored (Scheme 3).
The kinetic curves obtained with catalysts A, B, and C are
given in Fig. 1. As shown, the conversion of ethyl cyanoac-
etate was 55% in the presence of diamine A after 30 min of
reaction time. In contrast, when two methyl groups were intro-
duced in the cyclohexane ring (diamine B) in A, the activity
increased, providing 80% conversion at the same reaction time.
It should be noted that the reaction is thermodynamically con-
trolled. Nevertheless, when zeolite 4A was introduced to adsorb
the water formed during the reaction with diamine B, the equi-
librium was shifted, and 100% conversion with 100% selectiv-
ity was achieved after 3 h reaction time. The higher basicity of
diamine B found in the catalytic test is likely due to the do-
nating effect of methyl groups in position 3 of the cyclohexane
ring. If this hypothesis were correct, then the presence of two
additional methyl groups (diamine C) should give an even more
active catalyst for the condensation reaction. However, contrary
to the hypothesis, a lower catalytic activity was observed in this
case (Fig. 1).
For comparison purposes, the condensation reaction was
also performed in the presence of 1,5,7-triazabicyclo[4.4.0]dec-
5
-ene (TBD; Scheme 1) and the proton sponge DMAN. Re-
sults, shown in Fig. 2, demonstrate the following order of ac-
tivity: B > DMAN > TBD. The fact that amine B exhibited
higher activity than DMAN appears to indicate that the for-
mer behaved as a stronger base. The apparent activation ener-
gies (Eact) were estimated by performing the reaction at 273,
298, and 313 K and fitting the initial rate values to the Ar-
rhenius equation, with a correlation coefficient of 0.99. The
apparent Eact for the reaction catalyzed by diamine B was

140
M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
Table 1
Results of TOF and yield of 2a (see Scheme 2) obtained in the Knoevenagel
condensation of benzaldehyde with ethyl cyanoacetate using different amines
as base catalysts
−
1
a
Catalysts (Scheme 1)
TOF (min
)
Yield of 2a
A
B
C
D
8.1
26.0
9.6
0.6
1.0
63
80
70
14
31
71
45
E
DMAN
TBD
11.2
2.2
a
Benzaldehyde (32 mmol), ethyl cyanoacetate (28 mmol) catalyst (0.14
mmol) without solvent at room temperature at 1 h reaction time.
Fig. 3. Knoevenagel condensation of benzaldehyde (32 mmol) and ethyl cya-
noacetate (28 mmol) without solvent at room temperature with diamine B (!),
diamine E (1) and diamine D (×) (0.14 mmol).
Table 2
Proton affinities (kcal/mol) calculated as differences in energy between neutral
and protonated species. The quantum chemical calculations were performed
by fully optimising the geometry of both species at the DFT/B3LYP 6-31G**
level. Species are shown in Scheme 1
5
± 0.1 kcal/mol, lower than that for DMAN (7.3 kcal/mol),
consistent with the hypothesis of the stronger basicity of di-
amine B. In contrast, TBD, which has been claimed to be a
strong basic catalyst able to perform various base-catalysed re-
actions, including carbonylation of amines [21], Michael-type
additions [22] and transesterification reactions [23], exhibited
much lower activity than diamine B and DMAN.
Because catalyst B has the greater catalytic activity, we ex-
plored another variable related to the molecular structure that
may influence the basicity of B: increasing the distance between
the two nitrogen atoms by means of two carbon bridges (di-
amine E; Scheme 1) instead of one carbon bridge between the
two nitrogen atoms. Condensation of benzaldehyde with ethyl
cyanoacetate (see Fig. 3) showed that the initial rate was lower
for catalyst E than for catalyst B. In addition, the 3-methyl
piperidine (D), which has only one amine group, is included for
comparison. These results indicate that the molecular geome-
try of the catalyst plays an important role in catalytic activity,
probably through the different stabilities of the reaction transi-
tion states.
Amines
Proton affinities (kcal/mol)
DMAN
258.8
258.2
251.5
250.9
250.1
241.8
232.2
233.3
Amine E
Amine C
Amine B
Amine A
Amine D
TBD (H1)
TBD (H2)
fer to and from the basic site [24]. It also has been reported that
a usual drawback of, for instance, proton sponges, which are
excellent from the standpoint of low nucleophilicity, can be the
very low rates of proton transfer [9,25].
In an attempt to correlate the catalytic activity with the basic-
ity of the different catalysts, we calculated the proton affinities
of the different amines as the energy difference between the
neutral and the protonated forms of DMAN and amines A, B,
C, D, and E (Scheme 1). The results, given in Table 2, show that
DMAN was the most basic molecule, with a protonation energy
of 258.8 kcal/mol, followed by (in order) the amines E, C, B,
A, and D. The proton affinity of TBD also was calculated and
found to be 232.2 or 233.3 kcal/mol, depending on whether the
protonation occurs at N1 or N2 (Scheme 1). The order of the di-
amines A, B, and C indicates that the electron-donating methyl
groups has an additive effect on increasing the electronic charge
at the nitrogen atom, which can then stabilise the protonated
form and increase the proton affinity. These amines contain two
neighbour nitrogen atoms with relative orientation dictated by
the rigidity of the piperidine cycle and the low flexibility of the
Table 1 summarizes TOF values (initial reaction rate per
mmol of catalyst) and yields at 1 h reaction time of 2a (see
Scheme 2) obtained for the different amines evaluated. From
the results presented, we can establish the following order of
catalytic activity (initial rates) for the various catalysts tested:
B > DMAN > C > A > TBD > E > D.
3
.1. Theoretical calculations
To explain the catalytic results, we have considered that the
proton affinity of the basic catalyst, as well as the geometry of
the transition state (induced by the geometry of the basic mole-
cule), can determine the energy of the proton transfer to and
from the basic site. Indeed, it should be taken into account that
although the proton may be easy to abstract from the reactant
by the basic catalyst, the step of releasing the proton back to the
catalyst may not be that easy or at least, when comparing dif-
ferent catalysts, may not maintain the correlation with proton
affinity. If this is indeed the case, then the determinant property
of a basic catalyst will necessarily be not the basic strength, but
rather the kinetic activity, that is, adequate rates of proton trans-
–CH – bridge between the two piperidine cycles. The optimisa-
2
tions show that in the protonated form, the lone pair of nitrogen
+
cannot overlap with the proton as in DMANH [9], resulting in
poorer stabilisation of the protonated form and thus lower pro-
ton affinity with respect to DMAN. A different effect is seen
in diamine E, in which the flexibility of the central –CH2–
CH – group allows the correct reorientation of the two rings,
2
enabling overlap of the nitrogen lone pair with the proton. The
final geometry of the protonated form of diamine E gives a

M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
141
bonding NH distance of 1.06 Å and a nonbonding N···H dis-
tance of 1.93 Å. In comparison, when the protonation of di-
amines A and B occurs, the bonding NH distance is 1.02 Å,
whereas the nonbonding N···H distance is 2.52 Å. This indi-
cates a lower proton affinity and a greater ability to release the
proton back in diamines A and B with respect to diamine E.
Up to now, we have shown that the proton affinities calcu-
lated for TBD and diamines D, A, and B correlate with the
catalytic activity observed, with amine B (which exhibits the
higher proton affinity) exhibiting the greatest catalytic activity.
However, there are exceptions to the direct correlation between
proton affinity and catalytic activity. In our case, we have seen
that amines C, E, and DMAN, despite their higher proton affin-
ity compared with amine B, exhibit lower catalytic activity. For
amine C, the lower activity with respect to B can be attributed
to the presence of the four methyl groups in the piperidine
rings, which can cause steric hindrance of proton abstraction.
However, for amines E and DMAN, the lower catalytic activity
exhibited with respect to the amine B may be related to a lower
rate of proton transfer from the basic site to the oxoanion inter-
mediate, although another possibility, involving the difficulty of
protonating the base due to steric problems, also may explain
the catalytic results. To clarify this point, following the reaction
mechanisms presented in Scheme 4, we studied the different re-
action steps by means of computational chemistry techniques.
the reactant, product, and transition state. The activation en-
ergy is 22.4 kcal/mol for this reaction; this high value is due
to the fact that the incoming acid molecule, when approaching
the proton-attracting (N···N) location in the DMAN molecule,
finds a steric hindrance due to the bulky methyl groups, pre-
cluding an otherwise easy approach between the acid and basic
molecules. Further along in the reaction, this forces a nonopti-
mised angular behaviour of the lone pairs of the N atoms in the
DMAN molecule that, together with the problems in the further
approach of the acid molecule, yields a transition state of high
energy. Despite the clear proton-attracting power of the DMAN
molecule, as reflected in its high proton affinity (Table 2), the
overall result of the reaction analysis indicates a series of prob-
lems resulting in a high activation energy. An analysis of the
calculated geometries shown in Fig. 5 indicates the follow-
ing relevant bond distances in the transition state geometry:
r(NH) = 1.222 Å and r(CH) = 1.562 Å. In the equivalent tran-
sition state of the same reaction with amine B (Fig. 4), we have
r(NH) = 1.214 Å and r(CH) = 1.510 Å, indicating that the
proton in the transition state is less transferred to the amine (and
less detached from the ethyl cyanoacetate) when using DMAN,
at a higher activation energy (22.4 kcal/mol, compared with
8.8 kcal/mol in Table 3).
The second step (step 2 in reaction (2) of Scheme 4) pro-
ceeds exactly the same as with the previous catalyst, because
this is a condensation reaction between benzaldehyde and ethyl
cyanoacetate anion that forms an adduct, with the catalyst play-
ing no role whatsoever. Therefore, the species involved, as
shown in Fig. 5, are the same as in step 2 of reaction (1), and
the same activation energy of 8.4 kcal/mol holds for this case
as well. Although initially the counterion may play a role in this
reaction, this is considered negligible, because the anionic com-
plex between the carbanion and the benzaldehyde is so bulky
that the counterion is at a considerable separation. Also, taking
into account that the counterion (the corresponding protonated
amine, either amine B or DMAN) is itself also a bulky mole-
cule adds to the considerable separation between the cation and
the anion, which all together results in a less important ionic in-
teraction with the cation and thus a negligible role of the cation
in step 2 of the reaction.
3
.2. Knoevenagel condensation using amine B
The first step (step 1 in reaction (1) of Scheme 4) is protona-
tion of amine B, in which the hydrogen of the methylene group
is transferred to the amine [12]. This is shown in Fig. 4, which
characterises the reactant, product, and transition state. An acti-
vation energy of 8.8 kcal/mol has been found for this reaction.
The second step (step 2 in reaction (1) of Scheme 4) pro-
ceeds by approaching the benzaldehyde to the resulting ethyl
cyanoacetate anion to form an adduct with a negative charge
localised in the carbonyl oxygen of benzaldehyde, which be-
comes simply bonded to the carbon atom. This reaction, in
which all of the species involved have been characterised, is
shown in Fig. 4; an activation energy of 8.4 kcal/mol has been
calculated.
The third step (step 3 in reaction (1) of Scheme 4) of the
reaction [12] occurs between the adduct and the basic catalyst
Table 3
Energies of the molecular species corresponding to the reaction between ethyl
cyanoacetate, amine B, and benzaldehyde (reaction (1) in Scheme 4); and to the
reaction between ethyl cyanoacetate, DMAN, and benzaldehyde (reaction (2) in
Scheme 4). The energies were obtained by ab initio Hartree–Fock calculations
with 3-21G basis set. Details of the geometries are shown in Figs. 4 and 5. The
energies of the reactants, transition states and products are in units of Hartrees
(
amine B), which, by giving back the proton, forms the final
Knoevenagel condensation product. The proton is given back
very easily because the reaction proceeds without activation en-
ergy, as can be seen from the species characterised in Fig. 4.
We have not considered the final elimination step, because
the dehydration of the alcohol intermediate giving the cinna-
mate is very rapid, and it is rarely found to be the controlling
step [26].
(1 Hartree/particle is 627.50959 kcal/mol)
b
Reactant Transition state Product
Eact
Step 1 (amine B) −1008.29629 −1008.28233
−1008.28426
−736.31457
−1349.84594
8.8
8.4
–
a
Step 2
−736.32293
−736.30956
Step 3 (amine B) −1349.82254
–
Step 1 (DMAN) −1041.26760 −1041.23186
−1041.24206 22.4
3
.3. Knoevenagel condensation using DMAN
Step 3 (DMAN) −1382.78056 −1382.77410
−1382.81084
4.1
a
Step 2 is the same regardless the base used (either amine B or DMAN), as
The first step (step 1 in reaction (2) of Scheme 4) is the
protonation of DMAN. Fig. 5 shows the characterisation of
can be seen from Scheme 4.
b
Activation energies (Eact) are in kcal/mol.

142
M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
Fig. 4. Molecular species corresponding to the reaction between ethyl cyanoacetate, amine B, and benzaldehyde (this corresponds to reaction (1) in Scheme 4).
The drawings correspond to the geometries obtained by ab initio Hartree–Fock calculations with 3-21G basis set, and the corresponding energies are in Table 3.
The relevant bond distances for step 1 are: (a) reactants: r(NH) = 1.848 Å, r(CH) = 1.127 Å; (b) transition state: r(NH) = 1.214 Å, r(CH) = 1.510 Å; (c) prod-
ucts: r(NH) = 1.002, r(CH) = 1.954 Å. The relevant bond distances for step 2 are: (a) reactants: r(CO) = 1.219 Å; (b) transition state: r(CC) = 1.979 Å,
r(CO) = 1.274 Å; (c) products: r(CC) = 1.627, r(CO) = 1.352 Å. The relevant bond distances for step 3 are: (a) reactants: r(NH) = 1.008 Å, r(OH) = 1.465 Å;
(b) products: r(NH) = 1.709, r(OH) = 1.002 Å.

M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
143
Fig. 5. Molecular species corresponding to the reaction between ethyl cyanoacetate, DMAN, and benzaldehyde (this corresponds to reaction (2) in Scheme 4).
The drawings correspond to the geometries obtained by ab initio Hartree–Fock calculations with 3-21G basis set, and the corresponding energies are in Table 3.
The relevant bond distances for step 1 are: (a) reactants: r(NH) = 2.557 Å, r(CH) = 1.090 Å; (b) transition state: r(NH) = 1.222 Å, r(CH) = 1.562 Å; (c) prod-
ucts: r(NH) = 1.032 Å, r(CH) = 2.464 Å. The relevant bond distances for step 2 are: (a) reactants: r(CO) = 1.219 Å; (b) transition state: r(CC) = 1.979 Å,
r(CO) = 1.274 Å; (c) products: r(CC) = 1.627, r(CO) = 1.352 Å. The relevant bond distances for step 3 are: (a) reactants: r(NH) = 1.069 Å, r(OH) = 2.491 Å;
(b) transition state: r(NH) = 1.054 Å, r(OH) = 1.806 Å; (c) products: r(NH) = 2.155, r(OH) = 0.979 Å.

144
M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
The third step (step 3 in reaction (2) of Scheme 4) of the re-
action occurs between the adduct and the protonated DMAN,
which, by giving back the proton, forms the final Knoevenagel
condensation product. Instead of a reaction without activation
energy (as occurs in step 2 of reaction (1)), this reaction gives
an activation energy of 4.1 kcal/mol, as shown for the species
characterised in Fig. 5 and Table 3. Again, the approach be-
tween the reactants is hindered by the bulky oxoanion and its
+
interaction with the methyl groups of the DMANH molecule.
This results in an activation energy higher than that of the cor-
responding step 3 of the previous reaction (reaction (1) with
amine B).
Fig. 6. Condensation of benzaldehyde (32 mmol) with: (!) ethyl acetoacetate
◦
(
28 mmol), amine B (0.14 mmol) at 60 C; (P) diethyl malonate (28 mmol),
◦
amine B (0.28 mmol) at 80 C; (") acetophenone (28 mmol), amine B
3
.4. Comparison between the results of the Knoevenagel
◦
(0.28 mmol) at 130 C.
condensation using amine B and DMAN
3
.5. Carbon–carbon bond formation with different reactants
After rationalising why the activation energies (shown in Ta-
ble 3) are higher or lower in each case, a simple analysis of the
overall results shows that in either case (reaction (1) or reac-
tion (2)), the rate-determining step is step 1. This points to the
fact that the capability to accept the proton is controlling the
overall reaction rate, and this is easier when amine B instead
of DMAN is used as catalyst, despite the fact that DMAN has
a higher proton affinity than amine B. Consequently, from our
reaction study, we have established a difference between the
ideal capacity of accepting protons (proton affinity, as shown in
Table 2) and the actual capacity of accepting protons, which de-
pends on other factors besides proton affinity, which in this case
are dominated by steric hindrance, as explained in the previous
paragraphs.
To estimate the capacity of amine B for abstracting pro-
tons with pKa > 9, we performed Knoevenagel condensation of
benzaldehyde with more demanding basic strength methylenic
active compounds, ethyl acetoacetate (pKa = 10.9) and di-
ethyl malonate (pKa = 13.3). Moreover, we tested the Claisen–
Schmidt condensation between acetophenone (pKa = 15.8) and
benzaldehyde. Fig. 6 shows the results obtained for the three
substrates. In all cases, the selectivities for the condensation
products are 100%. Fig. 6 shows that the order of activity cor-
relates with the pKa of the different acidic substrates: 1a >
1b > 1c > acetophenone. On the other hand, we have previ-
ously found [9] that when carrying out the condensation of ethyl
acetoacetate with benzaldehyde in the presence of DMAN and
the absence of solvent, no condensation occurred. This behav-
iour was attributed to the fact that the protonated DMAN is so
stable that the proton is not returned toward the oxoanion in-
termediate, and thus the catalyst becomes inactive. However,
when the condensation is carried out in the presence of a sol-
The above calculations also indicate that basicity is not the
only characteristic that can explain the relative order of catalytic
activity when using different catalysts. Along with the ability
to stabilise the proton (proton affinity; step 1), proton release
(
step 3) also must be considered. The steric conformation be-
+
tween amine B and the negatively charged adduct (see Fig. 4)
allows proton elongation from amine B and its transfer to the
oxoanion without activation energy much more easily than can
occur in DMAN, in which the proton is more firmly attached
and the release requires activation energy (4.1 kcal/mol). Al-
though the protonic transfer to an anion should be considered
an easy reaction with no activation energy, geometry and initial
conformations (Fig. 5) show that the proton is shielded and pro-
vent such as DMSO, which polarizes the N–H –N bond, the
intramolecular hydrogen bond of the protonated DMAN loses
strength, facilitating release of the proton, after which the cat-
alyst becomes active. We conclude that diamine B has a strong
basic character, with the ability to abstract protons with pKa up
to 15.8 (see Fig. 6). This can be attributed to the combination
of two factors: the high proton affinity and the ability to return
the proton to the oxoanion intermediate.
+
tected in the DMANH cation, and thus the strong tendency to
transfer this to a negatively charged molecule does not occur
until the proton is separated from its initial geometry, which
requires an activation energy. Although step 3 is not the control-
ling step in this case, nevertheless it also should be considered.
These results explain the experimental observations in the
sense that the most basic DMAN catalyst does not give the
largest conversion, which is found for amine B. Recall that this
is explained by the higher activation energy of step 1 when us-
ing DMAN compared with using amine B, with the latter giving
a lower activation energy and thus greater conversion, as shown
from the experimental results in Fig. 2.
3.6. Influence of the solvent on catalytic activity
It is known that when charged species are involved in a reac-
tion, changes in the polarity of the media can strongly affect the
reaction rate. We investigated the influence of the solvent on
the rate of reaction between benzaldehyde and ethyl acetoac-
etate, using amine B as the catalyst in the presence of solvents
with different dielectric constants, ε: DMSO (ε = 48.9), DMF
(ε = 36.7), EtOH (ε = 24.3), t-butanol (ε = 12.47), chloroben-
zene (ε = 5.6), and toluene (ε = 2.4). Fig. 7 shows the yields
of the Knoevenagel adduct versus reaction time. As can be

M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
145
Acknowledgments
Support was provided by the Ministerio de Educacion y
Cultura of Spain (project MAT2003-07769-C02-01) and the
Generalidad Valenciana (GV04B-270). I.D. thanks the Consejo
Superior de Investigaciones Científicas for an I3P fellowship.
G.S. thanks Centro de Investigaciones Energeticas, Medioambi-
entales y Tecnologicas and Centro de Proceso de Datos, Univer-
sidad Politecnica de Valencia for the use of their computational
facilities. The contribution of Mr. Pablo Ramos to the experi-
mental work reported herein is also gratefully acknowledged.
References
[1] M.J. Climent, A. Corma, S. Iborra, A. Velty, J. Catal. 221 (2004) 472.
[2] A. Cauvel, G. Renard, D. Brunel, J. Org. Chem. 62 (1997) 749.
[3] (a) M. Laspéras, T. Llorett, L. Chaves, I. Rodriguez, A. Cauvel, D. Brunel,
Stud. Surf. Sci. Catal. 108 (1997) 75;
(
b) F. Bigi, S. Carloni, R. Maggi, A. Mazzacani, G. Sartori, Stud. Surf.
Sci. Catal. 130 (2000) 3501;
c) I. Rodriguez, S. Iborra, A. Corma, Appl. Catal. 194–195 (2000) 241.
Fig. 7. Condensation of benzaldehyde (32 mmol) and ethyl acetoacetate
◦
(
28 mmol) at 60 C with diamine B (0.14 mmol) in different solvents
5 mL): DMSO (!); DMF (×); EtOH ("); t-butanol (Q); chlorobenzene (P);
(
(
[
4] (a) Y.V. Subba Rao, D.E. De Vos, P.A. Jacobs, Angew. Chem. Int. Ed.
Engl. 36 (1997) 2661;
toluene (+).
(
b) R. Sercheli, R.M. Vargas, R. Sheldon, U. Schuchardt, J. Mol. Catal.
A 148 (1999) 173;
c) S. Jaenicke, G.K. Chuah, X.H. Lin, X.C. Hu, Microporous Mesoporous
Mater. 35–36 (2000) 143;
d) A.C. Blanc, D.J. Macquarrie, S. Valle, G. Renard, C.R. Quinn, D. Bru-
nel, Green Chem. 2 (2000) 383.
[5] A. Corma, S. Iborra, I. Rodriguez, F. Sanchez, J. Catal. 211 (2002) 208.
seen, the greatest activity is obtained with the most polar sol-
vents [9], (see also the relative rates of ethanol and tert-butanol
in Fig. 7), and the order of reactivity correlates very well with
the dielectric constant: DMSO > DMF > EtOH > t-butanol >
chlorobenzene > toluene. Furthermore, when the condensation
reaction is carried out with nonpolar solvents, the reaction rate
decreases significantly, as occurs when the reaction involves a
charged transition state or a neutral but polarizable transition
state formed through a single concerted step.
(
(
[6] J.-P. Besse, D. Brunel, P. Massiani, D. Tichit, Actual. Chimique 5–6 (2002)
111.
[7] R.W. Alder, Chem. Rev. 89 (1989) 1215.
[8] (a) H.A. Staab, T. Saupe, Angew. Chem Int. Ed. Engl. 27 (1988) 865;
(b) A.L. Llamas-Saiz, C. Foces-Foces, J. Elguero, J. Mol. Struct. 328
(1994) 297.
[9] I. Rodriguez, G. Sastre, A. Corma, S. Iborra, J. Catal. 183 (1999) 14.
4
. Conclusion
[10] (a) B. Siebenhaar, WO 9721659 (1997);
(b) A.J. Kesel, W. Oberthür, WO 9820013 (1998);
(c) J.P. Ferraris, T.L. Lambert, S. Rodriguez, WO 9305077 (1993).
[
11] (a) A. Corma, V. Fornes, R.M. Martin-Aranda, F. Rey, J. Catal. 134 (1992)
8;
b) M.J. Climent, A. Corma, V. Fornés, A. Frau, R. Guil-Lopez, S. Iborra,
J. Primo, J. Catal. 163 (1996) 392.
In this work, diamines with neighbour nitrogen atoms were
5
(
used as base catalysts in the Knoevenagel condensation reac-
tion between benzaldehyde and ethyl cyanoacetoacetate. The
diamines used as catalysts show the feature of stabilising the
proton abstracted in the first step of the reaction due to the ef-
[12] (a) F.S. Prout, U.D. Beaucaire, G.R. Dyrkarcz, W.M. Koppes, R.E. Kuz-
nicki, T.A. Marlewski, J.A. Pienkowski, J.M. Puda, J. Org. Chem. 38
(1973) 1512;
+
fect of a N–H –N interaction between the neighboring atoms
(b) G. Jones, Org. React. 15 (1967) 204;
of the diamine. Nevertheless, the catalytic results do not show
a correlation between proton affinity of the diamine and basic
strength, suggesting that other effects related to the mecha-
nism should be studied. Computational chemistry techniques
have been used to study the three steps of the Knoevenagel
condensation using amine A and DMAN as catalysts and the
results show that the rate-determining step is step 1, which in-
volves release of the proton to the basic catalyst. The ability
to accept the proton by the basic molecules does not corre-
late with the calculated proton affinities due to the influence
of steric factors, making it more difficult to accept the pro-
ton when the catalyst is DMAN. Therefore, the reaction rate
is higher when using amine B, which explains the greater con-
versions found experimentally when using amine B compared
with DMAN.
(
c) J. Guyot, A. Kergomard, Tetrahedron 39 (1983) 1166.
[13] V.K. Valerij, G.I. Rutman, USSR SU1482910 (1989).
[14] (a) J.S. Binkley, J.A. Pople, W.J. Hehre, J. Am. Chem. Soc. 102 (1980)
9
(
39;
b) M.S. Gordon, J.S. Binkley, J.A. Pople, W.J. Pietro, W.J. Hehre, J. Am.
Chem. Soc. 104 (1982) 2797;
c) W.J. Pietro, M.M. Francl, W.J. Hehre, D.J. Defrees, J.A. Pople, J.S.
(
Binkley, J. Am. Chem. Soc. 104 (1982) 5039.
[
[
15] H.B. Schlegel, J. Comput. Chem. 3 (1982) 214.
16] (a) J. Simons, P. Jorgensen, H. Taylor, J. Ozment, J. Phys. Chem. 87 (1983)
2
745;
(b) A. Bannerjee, N. Adams, J. Simons, R. Shepard, J. Phys. Chem. 89
(1985) 52.
[
17] (a) J.T. Golab, D.L. Yeager, P. Jorgensen, Chem. Phys. 78 (1983) 175;
(b) C.J. Cerjan, W.H. Miller, J. Chem. Phys. 75 (1981) 2800.
[
[
18] A.D. Becke, J. Chem. Phys. 10 (1993) 5648.
19] (a) R. Ditchfield, W.J. Hehre, J.A. Pople, J. Chem. Phys. 54 (1971) 724;
(b) W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (1972) 2257;

146
M.J. Climent et al. / Journal of Catalysis 246 (2007) 136–146
(
(
(
c) P.C. Hariharan, J.A. Pople, Mol. Phys. 27 (1974) 209;
d) M.S. Gordon, Chem. Phys. Lett. 76 (1980) 163;
e) P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213.
[22] (a) E. Van Aken, H. Wynberg, F. Van Bolhuis, Acta Chem. Scand. 47
(1993) 122;
(b) D. Bensa, J. Rodriguez, Synth. Commun. 34 (2004) 1515;
(c) A. Horvath, Tetrahedron Lett. 37 (1996) 4423.
[23] (a) F. Jerome, G. Kharchafi, I. Adam, J. Barrault, Green Chem. 6 (2004)
7;
[
20] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.
Cheeseman, V.G. Zakrzewski, J.A. Montgomery Jr., R.E. Stratmann, J.C.
Buran, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain,
O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala,
Q. Cui, K. Morokuma, N. Rega, P. Salvador, J.J. Dannenberg, K.D. Mal-
ick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V.
Ortiz, A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Ko-
maromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham,
C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson,
W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S.
Replogle, J.A. Pople, Gaussian 98, Revision A.11.2, Gaussian, Inc., Pitts-
burgh, PA, 1998.
(b) A.L.M. Guimaraes, R.M. Vargas, J. Am. Oil Chem. Soc. 75 (1998)
755;
(c) U. Schuchardt, R.M. Vargas, G. Gelbard, J. Mol. Catal. A 99 (1995)
65.
[24] R.W. Alder, J. Am. Chem. Soc. 127 (21) (2005) 7924.
[25] (a) R.W. Alder, N.C. Goode, N. Miller, F. Hibbert, K.P.P. Hunte, H.J. Rob-
bins, J. Chem. Soc. Chem. Commun. (1978) 89;
(b) F. Hibbert, G.R. Simpson, J. Chem. Soc. Perkin Trans. 2 (1987)
243.
[26] (a) S. Patai, Y. Israeli, J. Chem. Soc. (1960) 2025;
(b) R.L. Reeves, S. Patai (Eds.), The Chemistry of Carbonyl Group, Inter-
science, New York, 1966, p. 567.
[
21] R. Ballini, D. Fiorini, R. Maggi, P. Righi, G. Sartori, R. Sartorio, Green
Chem. 5 (2003) 396.