You are on page 1of 9

Tetrahedron 60 (2004) 6443–6451

A DFT rationalization for the observed regiochemistry in the


nitrile oxide cycloaddition with anthracene and acridineq
Antonino Corsaro,a Venerando Pistarà,a Antonio Rescifina,a,* Anna Piperno,b
Maria A. Chiacchioa and Giovanni Romeob
a
Dipartimento di Scienze Chimiche, Università di Catania, Viale Andrea Doria 6, Catania 95125, Italy
b
Dipartimento Farmaco-Chimico, Università di Messina, Viale SS. Annunziata, Messina 98168, Italy
Received 30 March 2004; revised 21 May 2004; accepted 10 June 2004

Available online 26 June 2004

Abstract—A series of computational approaches, based on the global and local indexes defined in the context of the DFT, at the B3LYP/
6-31G(d) computational level, were investigated to elucidate the regiochemistry and the energetics of the mesitonitrile oxide 1,3-dipolar
cycloaddition with anthracene and the aza-analogue acridine. The results are in agreement with the observed regioselectivity and are in
contrast with the ones predicted in terms of FMO theory.
q 2004 Elsevier Ltd. All rights reserved.

1. Introduction

Nitrile oxides constitute a powerful class of 1,3-dipoles,


able to react with the poorest dipolarophiles such as
polycyclic and heterocyclic aromatics.1 The synthetic utility
of this kind of 1,3-dipolar cycloadditions (1,3-DCs),
although limited by the low yields obtained, is noteworthy
because 1,3-DCs open a new path to the selective
functionalization of this type of dipolarophile.2

Recently, within our studies aimed at examining the


dipolarophilic reactivity of polycyclic aromatic hydro-
carbons (PAHs) and their aza-analogues (N-PAHs), we
reported the reactions of some of them (anthracene,
phenanthrene, pyrene, perylene, acridine, benzo[h]quino-
line, 1,10-, 1,7-, 4,7-phenanthroline) with 2,4,6-trimethyl-
phenyl- and 3,5-dichloro-2,4,6-trimethylphenylnitrile
oxide, which gave mainly monocycloadducts. 3 As Scheme 1.
examples, the reactions of nitrile oxides with anthracene
1a and acridine 1b and those with phenanthrene 2a and HOMO and LUMO energies; the results are consistent with
benzo[h]quinoline 2b are illustrated in Scheme 1. the dominant interaction HOMO (dipolarophile)—LUMO
(dipole) and with energy gaps.3a Thus, with regard to the
The dipolarophilic reactivity of PAHs towards nitrile oxides site-selectivity of the process, the formation of the unique
has been rationalized in terms of Frontier molecular orbital isolated monocycloadduct for PHAs3a is controlled by the
(FMO) theory using PM3 methods4 by calculating the adjacent pair with higher orbital coefficients.5
q
Supplementary data associated with this article can be found in the online The outcome of aza-analogues of anthracene and phenan-
version, at doi: 10.1016/j.tet.2004.06.052 threne reactions is clearly influenced by the presence of the
nitrogen atoms which lead to the initial formation of a labile
Keywords: 1,3-Dipolar cycloadditions; Density functional theory;
Anthracene; Acridine; Mesitonitrile oxide.
1,3-adduct at their nitrogen atom in equilibrium with
* Corresponding author. Tel.: þ39-095-7385014; fax: þ39-06-233208980; reactants.6 The normal pericyclic pathway is then changed
e-mail address: arescifina@unict.it into a pseudo pericyclic reaction7 which involves the initial

0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.06.052
6444 A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451

attack of nitrile oxide to nitrogen atom and then the later TS, i.e. TS is product-like, and then the frontier orbital
subsequent electrocyclic closure.6 effects will be less noticeable.

Because of the reversibility of the first addition step, in the Thus, although the FMO theory works well for a wide range
case of N-PAHs, it was not possible to deduce such a fair of reactants, there are some exceptions11e,12 which, on the
rationalization of the relative reaction rates as for the nitrile contrary, are well rationalizable working with a more
oxide cycloaddition of pyridine6 and diazines.8 The powerful level of theory, such as density functional theory
different behaviour observed arises from the different (DFT).
positions of the nitrogen atoms in the molecule which
affords 1,3-adducts with different thermodynamic For this reason, we have investigated a more powerful
stabilities. theoretical approach which is based on the formulation of
the interaction energy of 1a,b and mesitonitrile oxide 7 in
In the case of the reaction of acridine in refluxing toluene, a terms of DFT,13 which proved to be in agreement with the
monocycloadduct with the same structure of that deriving experimental results.
from anthracene was isolated, while at room temperature an
addition product of the 1,3-dipole to the nitrogen atom was
afforded which lead to the uncommon dioxadiazine dimer.9 2. Computational methods

The FMO arguments, however, do not provide a satisfactory DFT calculations were carried out with the G98 system of
explanation of the experimentally observed regioselectivity programs.14 Critical points (reactants, transition structures
in the nitrile oxide cycloadditions to 1a,b. In fact, the and products) were fully characterized as minima or first-
interaction of the atoms with the largest coefficients10 order saddle points by diagonalizing the Hessian matrices of
should lead to the formation of cycloadducts with an inverse the optimized structures at the B3LYP/6-31G(d) level.15 All
regiochemistry. the critical points were further characterized by analytic
computation of harmonic frequencies at the B3LYP/
The failure of FMO approach in the regioselectivity 6-31G(d) level. Transition structures were found to have
prediction for the examined reactions can be rationalized only one negative eigenvalue with the corresponding
on the basis of the subsequent considerations: (a) the FMO eigenvector involving the formation of the newly created
theory is based only on the properties of the Frontier C – C and C – O bonds. Vibrational frequencies were
molecular orbitals of the reactants and just for its simplicity calculated (1 atm, 298.15 K) for all B3LYP/6-31G(d)
it does not account for all other orbital interactions;11 (b) optimized structures and used, unscaled, to compute both
moreover, FMO theory is normally applicable to exothermic ZPVE and activation energies. The electronic structures of
reactions where, according to the Hammond postulate, an critical points were studied by the natural bond orbital
early transition state is involved. Conversely, in the case of (NBO) method.16 Atomic electron populations were
endothermic reactions, the Hammond postulate claims a evaluated following the Merz-Kollman (MK) scheme17

Scheme 2.
A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451 6445

formulas.20 The intrinsic reaction coordinates21 (IRC


analysis) were also calculated to analyze the mechanism
in detail for all the transition structures obtained.

We studied the regioselectivity of the 1,3-DC reaction of


1a,b with 7. We have considered two reaction channels,
meta and ortho, corresponding to the formation of D2-
isoxazolines P1, P3, P5 and P2, P4, P6, respectively.
Consequently, the two transition states leading to the two
cycloadducts have been located for the dipolarophile 1a,
whereas the four transition states leading to the four
cycloadducts have been located for the dipolarophile 1b.
The nomenclature used for defining stationary points is
given in Scheme 2.

3. Results and discussion

3.1. Reactants

The parameters of lowest energy structures for 1a22 and


1b23 corresponds to that reported in literature; to our best
knowledge, no DFT calculations have been already
reported for 3 (geometries and energies are reported as
Supplementary data).

The optimized geometries, with the adopted numeration, of


compounds 1a,b and 3 are illustrated in Figure 1.

From the hardness values,24 reported in Table 1, it emerges


that 1b is slightly more aromatic than 1a, in agreement with
experimental data.25 Moreover, the global electrophilicity
value of 3 is lower than that reported for benzonitrile
oxide26 and therefore 3 is a more moderate electrophile.

3.2. Prediction of regiochemistry

A simple way to predict the regiochemistry of 1,3-DC


reactions and the trend of regioselectivity consists of using
the charges obtained by the MK molecular electrostatic
potential method,17 which has been recently considered as
an appropriate local descriptor of charge.27 In the case at
hand 1a has a negative charge on both C1 (20.20) and C2
(20.12) atoms, so as 1b on C1 (20.12), C2 (20.17), C3
(20.07) and C4 (20.30) atoms, whereas 3 has a negative
charge on O1 (20.33) and a positive one on C3 (þ0.09). In
the reaction of 1a with 3, the more nucleophilic oxygen of
Figure 1. B3LYP/6-31G(d) optimized structures of the reactants. the dipole is prone to attach the C2 atom of 1a, because it is
less negative than the C1 one (Dq¼0.08), leading to a
that, other than already being proved to be reliable,18 has prevalence of the P1 regioisomer. In the reaction of 1b with
been used in most DFT calculations of regiochemistry of 3, the difference of charge at the two possible sites of
1,3-DCs.19 The enthalpy and entropy changes were attachment C 1vC 2 and C 3vC 4, is 0.04 and 0.23,
calculated from standard statistical thermodynamic respectively, with the C1 and C3 atoms being less negative.

Table 1. Global properties (electronic chemical potential m, chemical hardness h and chemical softness S values are in a.u.; electrophilicity power v values are
in eV) and local softness (s þ for nucleophilic and s 2 for electrophilic attack) of anthracene 1a, acridine 1b and mesitonitrile oxide 3

m h s v s2 sþ s2 sþ

C1 C2 C3 C4 C1 C2 C3 C4 O1 C3 O1 C3

1a 20.12601 0.13213 3.78 1.63 0.288 0.223 0.159 0.292


1b 20.14082 0.13606 3.67 1.98 0.355 0.264 0.147 0.369 0.124 0.351 0.198 0.252
3 20.13107 0.17910 2.79 1.31 0.646 0.278 0.510 0.574
6446 A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451

So it is expected that the O1 atom of 3 predominantly attacks the other on the base of local softness s.33b,34 This quantity is
the C3 atom of 1b leading to the P3 arising from the 1,3-DC so defined:
on C3vC4 double bond. The 1,3-DC on C1vC2 double
bond is less favoured and, in any case, leads to a near Dkl 2
ij ¼ ðsi 2 sk Þ þ ðsj 2 sl Þ
2
ð2Þ
equimolar mixture of P5 and P6 regioisomers.
where i and j are the atoms of a molecule A involved in the
Moreover, these 1,3-DC reactions have been analyzed using formation of a cycloadduct with atoms k and l of a molecule
the global and local indexes, as defined in the context of B, and si’s are the appropriate type of atomic softnesses (if si
DFT,28 which are useful tools to understand the reactivity of and sj are electrophilic then sk and sl are obviously
molecules in their ground states. The electronic chemical nucleophilic).
potential m, that is the negative of electronegativity x, is
usually associated with the charge transfer ability of the The idea is based on the simultaneous fulfillment of the
system in its ground state geometry and it can be local HSAB concept at both termini. This is because, in
approximated, using Koopmans’ theorem, to the half the case of a multicenter addition reaction, it is not the
value of the sum of the one-electron energies of the FMO similarities of softness at one center that are important. D
HOMO and LUMO.28a,29 The chemical hardness h is can be considered a measure of how extensively the HSAB
considered to be a measure of the stability of a system: the principle is satisfied. The reaction associated with a lower D
system having the maximum hardness being the most value will be the preferred one.
stable.30 Essentially the hardness is approximated to be
the difference between LUMO and HOMO energies. The The local softness s was calculated as a product fS, where f is
chemical softness parameter S is strictly related to the the condensed form of Fukui functions calculated as
chemical hardness and it is due to the inverse of 2h.28 reported elsewhere,35 utilizing the MK derived charges.

Besides these indexes it is also possible to define the global The values obtained for si’s and the corresponding Ds
electrophilicity power v, which measures the stabilization values, referred at both meta and ortho channels, for the
in energy when the system acquires an additional electronic cycloadditions of 3 with 1a and 1b are reported in Tables 1
charge DN from the environment. The approximate and 2, respectively.
expression for v, in the ground-state parabola model, is31
In the case of the 1,3-DC of 1a with 3, we should consider
m2 only the D values for a HOMO1a –LUMO3 approach, in
v¼ ð1Þ
2h agreement to global parameter analysis. The D values for
both meta and ortho channels are similar with the meta
and it may be classified within a unique relative scale in channel slightly predominating and both regioisomers are
order to evaluate the polarity of transition state structures. expected to be formed. On the contrary, in the 1,3-DC
Both Diels – Alder reactions32 and 1,3-DCs26 can be between 1b and 3, we should consider only the D values for
evaluated using such an absolute scale. a LUMO1b – HOMO3 approach. In the case of the C1vC2
approach, the D value of the meta channel is lower than that
The values of m, h, S and v for compounds 1a,b and 3, of the ortho one, clearly indicating a net predominance of
calculated with the reported formulas, are listed in Table 1. the P5 regioisomer. On the contrary, in the case of the
The electronic chemical potential of 1a is lower than that of C3vC4 approach, the D value for the ortho channel slightly
3, thereby indicating that a net charge transfer will take predominates and the P4 regioisomer is favored. Never-
place from 1a to 3. Instead, the chemical potential of 1b is theless, the meta D value for the C1vC2 approach is the
higher than that of 3, thereby indicating that a net charge lowest one and then P5 is the only product expected. Thus,
transfer will take place from 3 to 1b. Indeed, the the charge approach predicts the experimental results for the
electrophilicity differences between 1a and 3 1,3-DC reaction of 1a well, whereas the local softness
(Dv¼0.32 eV) indicates a lower polar character for this approach reproduces the experimentally observed results for
cycloaddition than that for the reaction between 1b and 3 the reaction of 1b.
(Dv¼0.67 eV). Both Dv values are characteristic of non-
polar (pericyclic) reactions26,32 as is also indicated by the 3.3. Cycloaddition with anthracene
low charge transfer found in all cases. Whereas the global
parameters help to understand the behaviour of a system, in 3.3.1. Energies of the transition structures. The absolute
a more local approach, the same parameters, also emerge as and relative free and electronic energies with respect to
a useful tool for rationalizing, interpreting and predicting reactants for the two transition structures located for the
diverse aspects of chemical bonding and reaction mechan- reaction between 1a and 3 are collected in Table 3.
ism. Recently, local softness s has been successfully applied
for explaining the regiochemistry in more complex The predicted activation free energy for the meta channel is
pericyclic reactions.31 However, in order to explain the lower than that corresponding to the ortho channel.
regiochemistry of four center reactions, like these cyclo- Accordingly, it can be predicted that the P1 regioisomer
additions, the HSAB principle, in a local sense, needs to also will be formed preferentially, in agreement with
be considered apart from the local softness. Chandra and experimental observations.
Nguyen have correlated the idea of the local HSAB concept
and regioselectivity defining a quantity, said ‘delta’ (D), that For TS1 and TS2, which correspond to the transition
suggests a measure of predominance of one reaction over structures for the approach in the meta and ortho channels,
A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451 6447

Dortho
Table 3. B3LYP/6-31G(d) relative free energies (DG) (kcal/mol) and

0.16
relative electronic energies (DE) (kcal/mol) for the reaction of 1a with 3

Direct DEa Inverse DEb Direct DGa Inverse DGb

LUMO1b –HOMO3
C3vC4
TS1 23.81 30.65 36.99 29.95
TS2 24.50 32.03 37.67 30.46
P1 26.84 7.04
P2 27.53 7.21
Dmeta
0.20
a
Referred to 1aþ3.
b
Referred to the corresponding products.
Dortho

respectively, Mulliken population analysis (MPA) provides


0.20

some evidence for secondary orbital interactions (SOIs)


between the two reactants.36 In fact the TS1 shows a
positive overlap density of 0.003, 0.004, 0.003, 0.004 and
C3vC4a

HOMO1b –LUMO3

0.006 for C6H28, C7H28, C10C18, C10C19 and H27H29


respectively; similar interactions can be seen in TS2
(C 8H 30¼0.005, C9 H 30¼0.006, H 27 H 29¼0.005 and
C18H28¼0.003), but these latter are, globally, less intense.
Dmeta
0.17

Moreover, if we take into account that this 1,3-DC is mainly


1bþ3

conducted in toluene at reflux temperature, i.e. at 110 8C, on


the basis of the reverse barrier energies reported in Table 3,
Dortho
0.28

we can assume that the reactions at hand are reversible and


then under thermodynamic control. Nevertheless, in this
case the kinetic and thermodynamic products are coincident
LUMO1b –HOMO3
C1vC2

and then only the formation of P1 is noticeable, in good


agreement with the experimental.3a

3.3.2. Geometries of the transition structures. The


Dmeta

optimized geometries of two transition structures corre-


0.11

sponding to the reaction of 1a with 3, and leading to the two


possible products P1 and P2, are illustrated in Figure 2.
Some relevant structural features of the transition structures
Dortho
0.12

are given in Table 4.


Table 2. D values, for the 1,3-DC reactions of 1a and 1b with 3, calculated as described in the text

For both transition structures TS1 and TS2, the lengths of


HOMO1b –LUMO3
C1vC2

the C – C forming bonds (2.08 and 2.09 Å) are markedly


shorter than the lengths of the C – O ones (2.25 and 2.24 Å).
Taking into consideration that C – O sigma bonds are shorter
than C –C sigma bonds, the difference of the lengths of the
Dmeta

two forming bonds are borderline for a concerted process


0.11

with a significant asynchronicity. It is well-known that when


a 1,3-DC cycloaddition presents asynchronous TSs,
diradical structures could in principle be involved.37 This
Dortho
0.24

as been ruled out repeating the TSs calculations using the


keyword STABLE, present in Gaussian 98, at both
Referred to the attach of 3 on the double bond of 1a,b.

RB3LYP/6-31G(d) and UB3LYP/6-31G(d) levels. In all


LUMO1a –HOMO3
C1vC2

cases the wavefunction was stable under the perturbations


considered.

3.3.3. Bond order and charge analysis. The concept of


Dmeta
0.14

bond order (BO) can be utilized to obtain a deeper analysis


of the extent of bond formation or bond breaking along a
reaction pathway. This theoretical tool has been used to
Dortho

study the molecular mechanism of chemical reactions. To


0.17

follow the nature of the formation process for C1 –C2 and


C3 – O4 bonds, the Wiberg bond indexes38 have been
computed by using the NBO population analysis as
HOMO1a –LUMO3

implemented in Gaussian 98. The results are included in


Table 4.
C1vC2a

The general analysis of the bond order values for the two
1aþ3

Dmeta
0.16

TSs structures showed that the cycloaddition process is


a
6448 A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451

Figure 2. Optimized geometries at B3LYP/6-31G(d) level for transition structures leading to P1 and P2. Some hydrogen atoms have been omitted for clarity.
Distances of forming bonds are given in angstroms. SOI are represented by coloured dotted lines.

Table 4. Delocalization energies (DE in kcal/mol) from reactants to TSs, weak, their sum correspond to a delocalization energy of
bond orders (Wiberg indexes) for the two forming bonds in the TSs and 1.37 kcal/mol. On the other hand, for the TS2, a p!sp
charge transfer (a.u.) in terms of the residual charge of the nitrile oxide (0.57 kcal/mol) stabilization, due to the C8vC9 double
fragment in the transition state
bond of 1a with C26 – H30 single bond of a methyl on 3, and a
DEa C1 –C2 C3 –O4 NPA qCT (e) s!sp one (0.43 kcal/mol) due to the C25 – H29 with C2 –
H27, for a total of 1.00 kcal/mol, are evident. This is
TS1 319.12 0.394 0.247 20.036 completely in agreement with the greater stability of TS1
TS2 293.87 0.384 0.255 20.040
vs. TS2 due to the strongest SOI achieved in the approach.
a
Calculated using a secondary-order perturbation theory (SOPT) analysis
of the Fock matrix. 3.4. Cycloaddition with acridine

asynchronous with a greater value for the forming C1 – C2 3.4.1. Energies of the transition structures. The absolute
bonds in both cases, showing that the extent of and relative free energies for the four transition structures
asynchronicity is the same for the two channels. located for the reaction between 1b and 3 are indicated in
Table 5.
The natural population analysis39 allows the evaluation of
the charge transferred between the two reactants at the TS Table 5. B3LYP/6-31G(d) relative free energies (DG) (kcal/mol) and
geometry. The charge transfer in terms of the residual relative electronic energies (DE) (kcal/mol) for the reaction of 1b with 3
charge on the nitrile oxide, for all the optimized TSs, is Direct DEa Inverse DEb Direct DGa Inverse DGb
shown in Table 4. The negative values are indicative of an
electron flow from the HOMO of the 1a to the LUMO of 3, TS3 22.62 31.28 36.65 28.62
in agreement to their electronic chemical potential values, TS4 25.02 31.82 37.94 30.03
but their magnitudes reveal an almost neutral reaction. TS5 23.70 30.44 36.83 29.51
TS6 24.12 31.60 37.36 30.09
P3 28.65 8.03
From the results of the secondary-order perturbation theory P4 26.80 7.91
(SOPT) analysis shown in Table 4, an important increment P5 26.74 7.32
in the delocalization energy appears on passing from P6 27.48 7.93
reactants to TSs. Even though there is not a strict correlation a
Referred to 1bþ3.
between the energies of the TSs and the total energies of b
Referred to the corresponding products.
delocalization, an examination of the contributions due to
inter-reactant delocalizations, accounts for the SOI pointed As in the case of 1a, the activation free energies for the meta
out by the MPA. Thus, TS1 is stabilized by a series of channel are lower in energy than those corresponding to the
delocalizations that are: p!sp (0.13 kcal/mol) of the ortho channel and they point to the prediction of a near
C6vC10 aromatic double bond of the anthracene moiety equimolar mixture of P3 and P5 adducts. However, if we
with the antibonding orbital of the C25 – H28 bond of a look to the inverse activation free energies, the 1,3-DC at
methyl on nitrile oxide moiety and the analogous p!sp hand is, once again, reversible and then under thermo-
(0.19 kcal/mol), due to the C7vC13 with C25 – H28. More- dynamic control and therefore the P5 product is preferen-
over, two other stabilizations due to a p!pp of the C6vC10 tially formed, in good agreement with experimental
with C18vC23 and C19vC20 double bonds of the mesito observations.9
ring (0.24 and 0.16 kcal/mol, respectively), are evident.
Finally, the two mutually corresponding delocalizations Even in the case of TS3 and TS5, MPA provides some
s!sp of C2 – H27 with C26 –H29 and C26 – H29 with C2 –H27 evidence for their greater stability with respect to TS4 and
(0.22 and 0.43 kcal/mol, respectively) complete this frame- TS6, due to SOI effects. In particular, the TS5 shows a
work. Although these singularly taken interactions are series of positive overlap density of 0.002, 0.004, 0.004,
A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451 6449

Figure 3. Optimized geometries at B3LYP/6-31G(d) level for transition structures leading to P3–P6. Some hydrogen atoms have been omitted for clarity.
Distances of forming bonds are given in angstroms. SOI are represented by coloured dotted lines.

0.003 and 0.006 for C6H28, C7H28, C10C18, C10C19 and TSs.40 What has been said, is also valid for the comparison
H27H29, respectively, that stabilizes it over the others. of reactions of 3 with 1a and 1b: in the first case, the lower
imaginary frequency value corresponds to the more
3.4.2. Geometries of the transition structures. The asynchronous TS.
optimized geometries of four transition structures corre-
sponding to the reaction of 1b with 3, and leading to the four The charge transfer, evaluated by the natural population
possible products P3, P4, P5, and P6, are illustrated in analysis, in terms of the residual charge on the nitrile oxide,
Figure 3. for all the optimized TSs is shown in Table 6. The negative
values are indicative of an electron flow from the HOMO of
For the cycloaddition of 1b with 3, the lengths of the 1b to the LUMO of 3, in contrast to the lower value of the
forming bonds are closer than those for the reaction with 1a. electronic chemical potential of 3 (m¼20.13107) with
The lengths of the C –C forming bonds (2.08 – 2.09 Å) are respect to that of 1b (m¼20.14082), revealing an almost
shorter than the C – O forming bonds (2.21 – 2.24 Å) in both neutral reaction.
meta and ortho channels. Since C –O bonds are longer than
Table 6. Delocalization energies (DE in kcal/mol) from reactants to TSs,
C – C bonds, once again, this 1,3-DC is borderline of a bond orders (Wiberg indexes) for the two forming bonds in the TSs and
concerted process with a significant asynchronicity. charge transfer (a.u.) in terms of the residual charge of the nitrile oxide
fragment in the transition state
3.4.3. Bond order and charge analysis. The general
DEa C1 –C2 C3 –O4 NPA qCT (e)
analysis of the bond order values for all the TSs structures
showed that the cycloaddition process is particularly TS3 476.21 0.387 0.256 20.014
asynchronous with an interval, in the Wiberg bond indexes, TS4 349.73 0.381 0.264 20.013
of 0.381 –0.394 for the forming C1 – C2 bonds and 0.250 – TS5 378.98 0.394 0.250 20.022
0.264 for the forming C4 –O3 (Table 6). This analysis is also TS6 381.84 0.385 0.260 20.029
in agreement with the corresponding imaginary frequency a
Calculated using a secondary-order perturbation theory (SOPT) analysis
values that are generally lower for the more asynchronous of the Fock matrix.
6450 A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451

From the results of the SOPT analysis showed in Table 6 p 291. (b) Corsaro, A.; Pistarà, V. Trends Heterocycl. Chem.
there is an important increment in the delocalization energy 1995, 4, 169– 215.
on passing from reactants to TSs. Even though there is not a 2. Corsaro, A.; Chiacchio, U.; Librando, V.; Pistarà, V.;
strict correlation between the energies of the TSs and the Rescifina, A. Synthesis 2000, 1469– 1473.
total energies of delocalization, an examination of the 3. (a) Corsaro, A.; Librando, V.; Chiacchio, U.; Pistarà, V.
contributions due to inter-reactant delocalizations, accounts Tetrahedron 1996, 52, 13027 – 13034. (b) Librando, V.;
for the SOI pointed out by the MPA. Thus, e.g. the TS5 is Chiacchio, U.; Corsaro, A.; Gumina, G. Polycyclic Aromat.
stabilized by a series of delocalizations that are: the p!sp Compd. 1996, 11, 313– 316. (c) Corsaro, A.; Chiacchio, U.;
(0.11 kcal/mol) of the C6vC10 aromatic double bond of the Librando, V.; Fisichella, S.; Pistarà, V. Heterocycles 1997, 45,
acridine moiety with the antibonding orbital of the C25 –H28 1567 –1572. (d) Corsaro, A.; Librando, V.; Chiacchio, U.;
bond of a methyl on nitrile oxide moiety and the analogous Pistarà, V.; Rescifina, A. Tetrahedron 1998, 54, 9187– 9194.
p!sp (0.11 kcal/mol), due to the C7vN13 with C25 –H28, 4. Stewart, J. J. P. J. Comp. Chem. 1989, 209. see also p 221.
and the two s!pp (0.03 and 0.05 kcal/mol) of C25 –H28 5. Fleming, I. In Frontier Orbitals and Organic Chemical
with C6vC10 and C7vN13, respectively. Moreover, another Reactions; Wiley: New York, 1976; p 171.
two stabilizations due to a p!pp of the C6vC10 with 6. Corsaro, A.; Perrini, G.; Caramella, P.; Marinone Albini, F.;
C18vC23 and C19vC20 double bonds of the mesito ring Bandiera, T. Tetrahedron Lett. 1986, 27, 1517– 1520.
(0.25 and 0.15 kcal/mol, respectively), are found evident. 7. Ross, J. A.; Seiders, R. P.; Lemal, D. L. J. Am. Chem. Soc.
Finally, the four mutually corresponding delocalizations 1976, 98, 4325– 4327.
s!sp of C2 – H27 with C26 –H29 and C26 – H29 with C2 –H27 8. Corsaro, A.; Perrini, G.; Pistarà, V.; Quadrelli, P.; Gamba
(0.21 and 0.43 kcal/mol, respectively) and p!pp of the Invernizzi, A.; Caramella, P. Tetrahedron 1996, 52,
C18vC23 and C19vC20 with C6vC10 (0.09 and 0.06 kcal/ 6421 –6436.
mol, respectively) complete this framework. Although these 9. Corsaro, A.; Pistarà, V.; Rescifina, A.; Romeo, G.; Romeo, R.;
singularly taken interactions are weak, their sum corre- Chiacchio, U. Arkivoc 2002, 8, 5 – 15.
sponds to a delocalization energy of 1.49 kcal/mol; thus, the 10. (a) Houk, K. N.; Sims, J.; Watts, C. R.; Luskus, L. J. J. Am.
TS5 results, as a rough average, to be 0.27 kcal/mol more Chem. Soc. 1973, 95, 7301– 7315. (b) Houk, K. N.; Sims, J.;
stabilized with respect to the other analogous TSs, by SOIs. Duke, R. E.; Strozier, R. W.; George, J. K. J. Am. Chem. Soc.
1973, 95, 7287– 7301.
11. (a) Sustman, R.; Sicking, W. Chem. Ber. 1987, 120,
4. Conclusions 1471– 1480. (b) Sustman, R.; Sicking, W. Chem. Ber. 1987,
120, 1653– 1658. (c) Craig, S.; Stone, A. J. J. Chem. Soc.,
In conclusion, it may be generalized that in endothermic Faraday Trans. 1994, 18, 1663. (d) Méndez, F.; Tamariz, J.;
1,3-DC reactions, such as the reactions at hands, the FMO Geerlings, P. J. Phys. Chem. A 1998, 102, 6292 –6296.
theory does not provide a satisfactory explanation of the (e) Herrera, R.; Nagarajan, A.; Morales, M. A.; Méndez, F.;
experimental results, while these are rationalized in terms of Jiménez-Vázquez, H. A.; Zepeda, L. G.; Tamariz, J. J. Org.
DFT calculations coupled to HSAB principle; this method- Chem. 2001, 66, 1252– 1263.
ology has already demonstrated a more powerful level of 12. (a) Grunanger, P.; Vita-Finzi, P. In Isoxazoles. The Chemistry
reliability.11d,e,12j of Heterocyclic Compounds; Taylor, E. C., Ed.; Wiley: New
York, 1991; Vol. 49, part 1. (b) Padwa, A.; Burgess, E. M.;
In the 1,3-DCs of 1a,b with 3, the predicted preference for Gingrich, H. L.; Roush, D. M. J. Org. Chem. 1982, 47,
the meta channel transition structure is substantial, leading 786– 791. (c) Yebdri, O.; Henri-Rousseau, O.; Texier, F.
to a theoretical preference for P1, P3 and P5 regioisomers, Tetrahedron Lett. 1983, 24, 369– 372. (d) Padwa, A.; Kline,
respectively. The predominant production of P1 and P5 D. N.; Koehler, K. F.; Matzinger, M.; Venkatramanan, M. K.
adducts is correctly predicted on further considerations of J. Org. Chem. 1987, 52, 3909– 3917. (e) Dalla Croce, P.; La
MPA and SOI interactions and thermodynamic control; Rosa, C. Heterocycles 1988, 27, 2825– 2832. (f) Jiménez, R.;
moreover, their low yields are justifiable in consideration of Pérez, L.; Tamariz, J.; Salgado, H. Heterocycles 1993, 35,
the elevated cycloreversion process of the reactions, due to a 591– 598. (g) Coppola, B. P.; Noe, M. C.; Schwartz, D. J.;
partial loss of aromaticity. The results obtained agree with Abdon, R. L.; Trost, B. M. Tetrahedron 1994, 50, 93 –116.
the experimental findings for both 1a,b. Moreover, (h) Fisera, L.; Konopikova, M.; Ertl, P.; Pronayova, N.
considerations based on the ground state lead to successful Monatsh. Chem. 1994, 125, 301– 312. (i) Pelkey, E. T.; Chang,
rationalization of the selectivity. L.; Gribble, G. W. J. Chem. Soc., Chem. Commun. 1996,
1909 –1910. (j) Avalos, M.; Babiano, R.; Cabanillas, A.;
Cintas, P.; Jimenez, J. L.; Palacios, J. C.; Aguilar, M. A.;
Corchado, J. C.; Espinosa-Garcı́a, J. J. Org. Chem. 1996, 61,
Acknowledgements
7291 – 7297. (k) Nagarajan, A.; Zepeda, G.; Tamariz, J.
Tetrahedron Lett. 1996, 37, 6835 –6838. (l) Coppola, B. P.;
This work was supported by M.I.U.R. (project P.R.I.N. 2002
Noe, M. C.; Hong, S. S.-K. Tetrahedron Lett. 1997, 38,
and F.I.R.B.).
7159 –7162. (m) Shanmugasundaram, M.; Raghunathan, R.;
Padma Malar, E. J. Heteroatom Chem. 1998, 9, 517–522.
(n) Pelkey, E. T.; Gribble, G. W. Synthesis 1999, 1117 –1122.
References and notes (o) Krajsovszky, G.; Gaál, A.; Haider, N.; Mátyus, P. J. Mol.
Struct. (THEOCHEM) 2000, 528, 13 – 18. (p) Gribble, G. W.;
1. (a) Caramella, P.; Grünanger, P. 1,3-Dipolar Cycloaddition Pelkey, E. T.; Simon, W. M.; Trujillo, H. A. Tetrahedron
Chemistry; Padwa, A., Ed.; Wiley: New York, 1984; Vol. 1, 2000, 56, 10133– 10140.
A. Corsaro et al. / Tetrahedron 60 (2004) 6443–6451 6451

13. (a) Kohn, W.; Sham, L. J. Phys. Rev. A 1965, 140, 1133– 1138. 27. (a) Green, D. F.; Tidor, B. J. Phys. Chem. B 2003, 107,
(b) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 10261– 10273. (b) Chattaraj, P. K. J. Phys. Chem. A 2001, 105,
785– 789. 511– 513.
14. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; 28. (a) Parr, R. G.; Pearson, R. G. J. Am. Chem. Soc. 1983, 105,
Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; 7512– 7516. (b) Parr, R. G.; Yang, W. Density Functional
Montgomery J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Theory of Atoms and Molecules; Oxford University Press:
Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; New York, 1989.
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; 29. Parr, R. G.; Donnelly, R. A.; Levy, M.; Palke, W. E. J. Chem.
Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, Phys. 1978, 68, 3801– 3807.
S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; 30. (a) Pearson, R. G. J. Chem. Ed. 1987, 64, 561– 567. (b) Parr,
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, R. G.; Chattaraj, P. K. J. Am. Chem. Soc. 1991, 113,
K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, 1854– 1855.
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; 31. Parr, R. G.; von Szentpály, L.; Liu, S. J. Am. Chem. Soc. 1999,
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, 121, 1922– 1924.
M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; 32. Domingo, L. R.; Aurell, M. J.; Perez, P.; Contreras, R.
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Tetrahedron 2002, 58, 4417– 4423.
Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; 33. (a) Chandra, A. K.; Nguyen, M. T. J. Phys. Chem. A 1998, 102,
Replogle, E. S., Pople, J. A., Gaussian 98 Revision A.9. 6181– 6185. (b) Chandra, A. K.; Nguyen, M. T. J. Comput.
Gaussian, Inc., Pittsburgh PA, 1999. Chem. 1998, 19, 195– 202. (c) Nguyen, T. L.; De Proft, F.;
15. (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648– 5652.
Chandra, A. K.; Uchimaru, T.; Nguyen, M. T.; Geerlings, P.
(b) Becke, A. D. Phys. Rev. A 1988, 38, 3098– 3100. (c) Lee,
J. Org. Chem. 2001, 66, 6096– 6103. (d) Sengupta, D.;
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785– 789.
Chandra, A. K.; Nguyen, M. T. J. Org. Chem. 1997, 62,
16. Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys.
6404– 6406. (e) Damoun, S.; Woude, Van de; Mendez, F.;
1985, 83, 735– 746.
Geerlings, P. J. Phys. Chem. A 1997, 101, 886– 893.
17. (a) Singh, U. C.; Kollman, P. A. J. Comput. Chem. 1984, 5,
34. Chandra, A. K.; Nguyen, M. T. Int. J. Mol. Sci. 2002, 3,
129– 145. (b) Besler, B. H.; Merz, K. M., Jr.; Kollman, P. A.
310– 323.
J. Comput. Chem. 1990, 11, 431–439.
35. (a) Mendez, F.; Gazquez, J. L. J. Am. Chem. Soc. 1994, 116,
18. De Proft, F.; Martin, J. M. L.; Geerlings, P. Chem. Phys. Lett.
1996, 250, 393– 401. 9298– 9301. (b) Yang, W.; Mortier, W. J. J. Am. Chem. Soc.
19. (a) Ponti, A.; Molteni, G. J. Org. Chem. 2001, 66, 5252– 5255. 1986, 108, 5708– 5711. (c) Yang, W.; Parr, R. G. Proc. Natl.
(b) Ponti, A.; Molteni, G. New J. Chem. 2002, 26, 1346– 1351. Acad. Sci. U. S. A. 1983, 82, 6723.
(c) Ponti, A.; Molteni, G. Chem. Eur. J. 2003, 9, 2770– 2774. 36. (a) Loncharich, R. J.; Brown, F. K.; Houk, K. N. J. Org. Chem.
20. Hehre, W. J.; Radom, L.; Schleyer, P. V. R.; Pople, J. A. Ab 1989, 54, 1129– 1134. (b) Birney, D. M.; Houk, K. N. J. Am.
initio Molecular Orbital Theory; Wiley: New York, 1986. Chem. Soc. 1990, 112, 4127– 4133. (c) Apeloig, Y.; Matzer, E.
21. (a) Fukui, K. Acc. Chem. Res. 1981, 14, 363– 368. (b) Head- J. Am. Chem. Soc. 1995, 117, 5375– 5376.
Gordon, M.; Pople, J. A. J. Chem. Phys. 1988, 89, 5777– 5786. 37. Tietze, L. F.; Fennen, J.; Geibler, H.; Schulz, G.; Anders, E.
22. Ikuta, S. J. Mol. Struct. (THEOCHEM) 2000, 530, 201– 207. Liebigs Ann. 1995, 1681– 1687.
23. Fu, A.; Du, D.; Zhou, Z. Spectrochim. Acta Part A 2003, 59, 38. Wiberg, K. B. Tetrahedron 1968, 24, 1083 –1096.
245– 253. 39. (a) Carpenter, J. E.; Weinhold, F. J. Mol. Struct.
24. De Proft, F.; Geerlings, P. Chem. Rev. 2001, 101, 1451– 1464. (THEOCHEM) 1988, 169, 41 – 62. (b) Reed, A. E.; Weinstock,
25. (a) Jackman, L. M.; Packham, D. I. Proc. Chem. Soc., London R. B.; Weinhold, F. J. Chem. Phys. 1985, 83, 735– 746.
1957, 349 – 350. (b) Sabbah, R.; Tabet, D.; Bélaadi, S. 40. (a) Domingo, L. R. J. Org. Chem. 1999, 64, 3922– 3929.
Thermochim. Acta 1994, 247, 201– 207. (c) Helal, M. R. (b) Domingo, L. R.; Arnó, M.; Andrés, J. J. Org. Chem. 1999,
J. Mol. Struct.: Theochem 2000, 528, 255– 261. 64, 5867– 5875. (c) Domingo, L. R.; Asensio, A. J. Org.
26. Pérez, P.; Domingo, L. R.; Aurell, M. J.; Contreras, R. Chem. 2000, 65, 1076– 1083.
Tetrahedron 2003, 59, 3117– 3125.

You might also like