Experimental and theoretical studies of Diels-Alder dimerization of 1,2,3,4,5-pentachlorocyclopentadiene and of Diels-Alder cycloaddition of polychlorinated cyclopentadienes to norbornadiene

Article abstract of DOI:10.1016/S0040-4020(03)00120-0

Diels-Alder cyclodimerization of 1,2,3,4,5-pentachlorocyclopentadiene (1) affords 2a as the exclusive reaction product. Diels-Alder cycloaddition of 1 to norbornadiene also proceeds stereoselectively to produce a single [4+2] cycloadduct, 4c. The structures of 2a and 4c were established unequivocally via application of single crystal X-ray crystallographic techniques. The origins of the observed diastereofacial selectivity in each of these cycloaddition processes have been investigated by application of semiempirical (AM1 Hamiltonian) and ab initio (Hartree-Fock 3-21G*) calculations. The computational results thereby obtained, which are based upon consideration of the kinetically favored transition state for each of the two cycloaddition reactions studied, are consistent with experiment. These semiempirical and ab initio methods also have been used to investigate the mechanisms of the Diels-Alder reactions that have been used to prepare aldrin and isodrin (7 and 8, respectively). The results thereby obtained suggest that isodrin formation via Diels-Alder cycloaddition of cyclopentadiene to 1,2,3,4,7,7-hexachloronorbornadiene proceeds with kinetic control of product stereochemistry.

Full text of DOI:10.1016/S0040-4020(03)00120-0

Tetrahedron 59 (2003) 1763–1771
Experimental and theoretical studies of Diels–Alder
dimerization of 1,2,3,4,5-pentachlorocyclopentadiene
and of Diels–Alder cycloaddition of polychlorinated
cyclopentadienes to norbornadiene
a
a
a
Alan P. Marchand,a Bishwajit Ganguly, Carolina I. Malagon, Huiguo Lai
,
*
´
and William H. Watsonb
,
*
aDepartment of Chemistry, University of North Texas, Denton, TX 76203-5070, USA
bDepartment of Chemistry, Texas Christan University, Fort Worth, TX 76203-8860, USA
Received 19 November 2002; revised 14 January 2003; accepted 15 January 2003
Abstract—Diels–Alder cyclodimerization of 1,2,3,4,5-pentachlorocyclopentadiene (1) affords 2a as the exclusive reaction product. Diels–
Alder cycloaddition of 1 to norbornadiene also proceeds stereoselectively to produce a single [4þ2] cycloadduct, 4c. The structures of 2a and
4c were established unequivocally via application of single crystal X-ray crystallographic techniques. The origins of the observed
diastereofacial selectivity in each of these cycloaddition processes have been investigated by application of semiempirical (AM1
Hamiltonian) and ab initio (Hartree–Fock 3-21G*) calculations. The computational results thereby obtained, which are based upon
consideration of the kinetically favored transition state for each of the two cycloaddition reactions studied, are consistent with experiment.
These semiempirical and ab initio methods also have been used to investigate the mechanisms of the Diels–Alder reactions that have been
used to prepare aldrin and isodrin (7 and 8, respectively). The results thereby obtained suggest that isodrin formation via Diels–Alder
cycloaddition of cyclopentadiene to 1,2,3,4,7,7-hexachloronorbornadiene proceeds with kinetic control of product stereochemistry. q 2003
Published by Elsevier Science Ltd.
1. Introduction
Compound 1 was considered to be an excellent candidate
substrate for a study of diastereofacial selectivity in Diels–
Alder reactions. Indeed, 1 functions as both diene and
dienophile during thermal [4þ2] cyclodimerization. Thus,
investigation of the course of its kinetically controlled
cyclodimerization is expected to provide important infor-
mation concerning the characteristics of the transition state
for a Diels–Alder reaction in which both the diene and
dienophile components possess diastereotopically non-
equivalent p-faces. In addition, it was of interest to
determine the influence of extensive substitution of C–H
by C–Cl in the diene and dienophile upon the course of this
reaction. Importantly, the presence of five Cl substituents in
1 assures that this molecule will be configurationally stable
by excluding the possibility of [1,5]-sigmatropic shifts. If
operative, thermal, suprafacial sigmatropic shifts would
result in effective scrambling of a 5-substituent throughout
the cyclopentadiene ring.6
Thermal [4þ2] cycloadditions between dienes and dieno-
philes have provided a mainstay of organic synthesis ever
since this important reaction was first discovered by Diels
and Alder in 1928.1 Reactions of this type have been
employed extensively for regio-, stereo-, and enantio-
selective synthesis of carbocyclic2 and heterocyclic3 six-
membered rings.
Diastereofacial selectivity that accompanies product for-
mation has been reported for thermal [4þ2] cycloadditions
between cage-annulated dienes and dienophiles in which
one or both of the reactants contains stereoelectronically
non-equivalent p-faces.4,5 In an effort to gain additional
insight into the origin of this phenomenon, we now report
the results of an investigation of Diels–Alder cycloaddition
reactions that involve 1,2,3,4,5-pentachlorocyclopentadiene
(1) as substrate.
In 1955, McBee and Smith7 reported that thermal [4þ2]
cyclodimerization of 1 occurs spontaneously at ambient
temperature to afford a cycloadduct, 2, that displays mp
220–2218C. Subsequently, Williamson and co-workers8
confirmed that only one product was formed via thermal
cyclodimerization of 1. However, these investigators
Keywords: Diels–Alder reaction; X-ray crystal structures; theoretical
studies.
*
Corresponding authors. Tel.: þ1-940-369-7226; fax: þ1-940-565-4318;
e-mail: marchand@unt.edu
0040–4020/03/$ - see front matter q 2003 Published by Elsevier Science Ltd.
doi:10.1016/S0040-4020(03)00120-0

1764
A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
Scheme 1.
recognized that as many as four [4þ2] cycloadducts, i.e.
2a–2d (Scheme 1), might be formed, if it is assumed that
the reaction proceeds in accordance with the Alder–Stein
rule of ‘maximum accumulation of unsaturation’.9
pentadiene by following a previously reported procedure.10
After purification of 1 had been completed, the pure
material was allowed to stand at ambient temperature,
whereupon spontaneous [4þ2] cyclodimerization ensued to
afford a single product.7,8 A suitable single crystal of this
material, mp 210–2118C, was obtained by fractional
recrystallization from hexane. The structure of this
compound was shown to be 2a via single crystal X-ray
structural analysis (see Section 5). An X-ray structure
drawing of 2a appears in Figure 1.
Williamson and co-workers8 attempted to assign the
configuration of the cyclodimerization product, 2. They
found that irradiation of an acetone solution of this material
affords the corresponding decachloropentacyclic cage
compound, 3, thereby confirming the fact that 2 possess
1
the endo configuration. The H NMR spectrum of 3 was
found to consist of a sharp singlet at d 4.56; this result is
consistent only with structures 3a (W¼Z¼Cl, X¼Y¼H)
and 3c (W¼Z¼H, X¼Y¼Cl), each of which possesses the
required molecular symmetry. This result, in turn, requires
that the cyclodimerization product, 2, from which 3 is
derived must possess either structure 2a or 2c. However,
Williamson and co-workers were not able to distinguish
between these two alternative structures for 2, either by
chemical means or via analysis of the 1H NMR spectrum of
the product formed via cyclodimerization of 1.8
2.2. Synthesis and X-ray structure of the product formed
via thermal [412] reaction of 1 with norbornadiene
In general, cycloadditions of dienes to norbornadiene
(dienophile) occur via approach of the diene toward the
exo face of the dienophile.11 However, by allowing for the
possibility that endo approach of the diene upon a
We now report the results of single crystal X-ray structural
analysis of 2, which provides unequivocal confirmation of
structure 2a for this cyclodimerization product. Correspond-
ing semiempirical and ab initio transition state calculations
also have been performed in an effort to provide additional
insight into the kinetically controlled cyclodimerization
reaction.
In addition, Diels–Alder cycloaddition of 1 to norborna-
diene, another dienophile that contains diastereotopically
non-equivalent p-faces, has been studied, and the structure
of the cycloadduct has been established unequivocally via
application of X-ray crystallographic techniques. Semi-
empirical and ab initio transition state calculations also have
been performed for this reaction.
2. Results and discussion
2.1. Synthesis and X-ray structure of the product formed
via thermal [412] cyclodimerization of 1
Compound 1 was prepared via SnCl2 promoted hydro-
genolysis of one of the C(5)–Cl bonds in hexachlorocyclo-
Figure 1. X-Ray structure drawing of 2a.

A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
1765
Scheme 2.
norbornadiene CvC double bond might occur, it is
conceivable that any or all of eight cycloadducts, 4a–4h
(Scheme 2) could arise via this cycloaddition reaction.
thermodynamic (i.e. product stability) considerations. In
addition, the transition state that leads to the formation of 2a
is the least energetic of the four possible endo transition
states for thermal [4þ2] cyclodimerization of 1. Hence, 2a,
whether formed under conditions of kinetic or thermo-
dynamic control of the cycloaddition reaction, is expected to
be the major (if not exclusive) reaction product, a result that
is consistent with experiment.
In our hands, this reaction afforded a single [4þ2]
cycloadduct. A suitable single crystal of this material, mp
82–838C, was obtained via fractional recrystallization from
hexane. The structure of this compound was shown to be 4c
via single crystal X-ray structural analysis (see Section 5).
An X-ray structure drawing of 4c appears in Figure 2.
The AM1-calculated transition state geometries for 2a–2d
Table 1. Calculated transition state and ground state energy differences
(kcal mol21) for thermal [4þ2] cyclodimerization of 1
3. Theoretical results and discussion
Isomer
AM1
HF/3-21G*
3.1. Computational investigation of the thermal [412]
cyclodimerization of 1
Transition
state
Ground
state
Transition
state
Ground
state
2a (syn, exo)
2b (syn, endo)
2c (anti, endo)
2d (anti, exo)
0.0
0.0
1.7
3.8
2.15
0.0
7.3
8.8
0.0
3.3
5.4
8.5
Fully optimized semiempirical calculations (AM1
Hamiltonian)12 and ab initio transition state calculations
performed at the Hartree–Fock HF/3-21G* level of theory13
have been carried out in an effort to obtain additional insight
into the kinetically controlled cyclodimerization reaction.
The results of these calculations appear in Table 1. Therein,
it can be seen that 2a is favored on the basis of
4.0
9.0
4.8
15.8
˚
Table 2. Lengths of forming s-bonds (a) and (b) [i.e. r(a) and r(b), in A] in
the four transition states for thermal [4þ2] cyclodimerization of 1
Isomer
AM1 transition state
˚
r(a) (A)
˚
r(b) (A)
2a (syn, exo)
2b (syn, endo)
2c (anti, endo)
2d (anti, exo)
2.120
2.153
2.162
2.135
2.249
2.229
2.245
2.258
Figure 2. X-Ray structure drawing of 4c.

1766
A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
Table 3. Calculated transition state and ground state energy differences
(kcal mol21) for thermal [4þ2] cycloaddition of 1 to norbornadiene
transition state is not unexpected, given the unsymmetrical
nature of the transition states.
Isomer
AM1
RHF/3-21G*//
RHF/AM1a
3.2. Computational investigation of the thermal [412]
cycloaddition of 1 to norbornadiene
Transition
state
Ground
state
Transition
state
Ground
state
In the case of this [4þ2] cycloaddition process, both the
diene (1) and the dienophile (norbornadiene) possess
diastereotopically non-equivalent p-faces. Thus, as can be
seen in Scheme 2, a total of eight transition states must be
considered explicitly. Semiempirical calculations (AM1
Hamiltonian)12 and single point ab initio transition state
calculations performed at the Hartree–Fock HF/3-21G*
level of theory13 have been carried out for the eight
transition states of interest, i.e. 6a–6h, that lead to the
formation of 4a–4h, respectively (Scheme 2).
4a (syn, exo, exo)
4b (anti, exo, exo)
4c (syn, endo, exo)
4d (anti, endo, exo)
4e (syn, endo, endo)
4f (anti, endo, endo)
4g (syn, exo, endo)
4h (anti, exo, endo)
3.1
11.2
0.0
2.3
3.0
13.9
6.0
0.8
7.5
0.4
0.3
0.0
9.8
3.0
2.4
5.3
19.2
0.0
2.1
6.9
25.0
4.5
1.5
14.4
1.0
0.0
1.5
16.4
4.7
6.3
6.8
3.2
a
Single point calculations performed at the HF/3-21G* level of theory by
using AM1 geometries.
The computational results obtained at the AM1 and HF/3-
21G* levels of theory both predict that the transition state
which leads to syn,endo,exo cycloadduct 4c is preferred
energetically (Table 3), i.e. 4c is predicted to be the product
of kinetic control of Diels–Alder cycloaddition of 1 to
norbornadiene, in agreement with experiment (Fig. 2). By
way of contrast, the results of AM1 calculations suggest that
4c–4e all possess comparable (thermodynamic) stabilities,
whereas the corresponding results obtained at the HF/3-
21G* level of theory suggest that there may be a somewhat
greater preference for formation of 4d if the cycloaddition
were to be performed under conditions of thermodynamic
control.
˚
Table 4. Lengths of forming s-bonds (a) [i.e. r(a), in A] in the eight AM1
calculated transition states for thermal [4þ2] cycloaddition of 1 to
norbornadiene
Isomer
AM1 transition state
The AM1 calculated transition state geometries suggest that
all eight possible transition states are concerted and
synchronous in nature. The newly forming C· · ·C s-bond
lengths in transition states of the type 6 range between 2.16
˚
r(a) (A)
4a (syn, exo, exo)
4b (anti, exo, exo)
4c (syn, endo, exo)
4d (anti, endo, exo)
4e (syn, endo, endo)
4f (anti, endo, endo)
4g (syn, exo, endo)
4h (anti, exo, endo)
2.170
2.186
2.162
2.172
2.168
2.184
2.162
2.172
˚
and 2.19 A, as shown in Table 4.
3.3. Computational investigation of the thermal [412]
cycloadditions halogenated cyclopentadienes and
norbornadienes: aldrin and isodrin
Thermal reaction of 1,2,3,4,5,5-hexachlorocyclopentadiene
with norbornadiene leads to the formation of an exo,endo
[4þ2] cycloadduct (i.e. ‘aldrin’, 7, Scheme 3) as the
exclusive reaction product. By way of contrast, the
corresponding cycloaddition of cyclopentadiene to
1,2,3,4,7,7-hexachloronorbornadiene affords a single
[4þ2] cycloadduct, which possesses the endo,endo con-
figuration (i.e. ‘isodrin’, 8 Scheme 3).14
are consistent with a concerted [4þ2] cycloaddition process
and are slightly asynchronous in nature. Two new s-bonds
are formed in the transition state during the cyclodimeriza-
tion process. The bonds of interest in this connection are
identified as ‘(a)’ and ‘(b)’ in transition state structure 5
(Table 2); the relevant calculated bond lengths are presented
in Table 2.
The data in Table 2 indicate that bonds (a) and (b) are
similar in length but not equal; the calculated length of bond
(b) is consistently greater than that of bond (a). The fact that
these two bonds are formed to differing extents in each
At the time when both aldrin and isodrin were used
extensively as insecticides, these Diels–Alder reactions
were important commercially.15 In the context of the
Scheme 3.

A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
1767
Table 5. Calculated transition state and ground state energy differences (kcal mol21) and lengths of forming s-bonds (a) [i.e. r(a), in A] in the four transition
states for thermal [4þ2] cycloaddition of 1 to norbornadiene
˚
7a (exo,exo)
7b (exo,endo)
7c (endo,exo)
7d (endo,endo)
˚
Bond lengh, s-bond (a) (A)
2.189
2.179
2.187
2.180
Transition state energies (kcal mol21
)
AM1
RHF/3-21G*//RHF/AM1
10.5
19.0
0.0
0.0
12.9
25.3
4.0
5.0
Ground state energies (kcal mol21
AM1
RHF/3-21G*//RHF/AM1
)
10.8
19.3
0.0
0.0
12.2
21.3
2.4
3.6
present study, their significance lies in the fact that the
stereochemical outcome of these two closely related
reactions is so clearly disparate. This observation suggests
that at least one, if not both, of these cycloadditions is
subject to kinetic control. In the present study, we have
investigated these cycloadditions by using only low-level
semiempirical (AM1 Hamiltonian) and ab initio (HF/3-
21G*) theoretical approaches in an attempt to rationalize the
course of each reaction.
The AM1 calculated transition and ground state results
clearly predict that the exo,endo cycloadduct, i.e. that which
results via transition state 7b, should be the preferred
product formed via Diels–Alder cycloaddition of
1,2,3,4,5,5-hexachlorocyclopentadiene to norbornadiene.
The results of single point calculations performed at
HF/3-21G* level of theory predict an even greater
preference for formation of 7b. Importantly, the prediction
that emerges from the computational results obtained for
both the transition state and ground state calculations agree
with experiment, i.e. that aldrin is expected to be the
preferred product, whether formed via kinetic or via
thermodynamic control of the cycloaddition reaction.
First, Diels–Alder cyclization of 1,2,3,4,5,5-hexachloro-
cyclopentadiene and norbornadiene, which leads to the
formation of aldrin (7) via one of four possible transition
states, i.e. 7a–7d, was examined computationally. The
results thereby obtained are shown in Table 5.
Next, Diels–Alder cyclization of cyclopentadiene to
1,2,3,4,7,7-hexachloronorbornadiene, which leads to the
formation of isodrin (8), via one of four possible transition
states, i.e. 8a–8d, was examined computationally. The
results thereby obtained are shown in Table 6.
The results of AM1-optimized geometry calculations
suggest that all four possible transition states are
concerted and synchronous in nature. The newly forming
˚
C· · ·C s-bond lengths range between 2.18 and 2.19 A
(Table 5).
The AM1 calculated transition state geometries suggest that
Table 6. Calculated transition state and ground state energy differences (kcal mol21) and lengths of forming s-bonds (a) [i.e. r(a), in A] in the four transition
states for thermal [4þ2] cycloaddition of cyclopentadiene to 1,2,3,4,7,7-hexachloronorbornadiene
˚
8a (exo,exo)
8b (exo,endo)
8c (endo,exo)
8d (endo,endo)
˚
Bond lengh, s-bond (a) (A)
2.156
2.132
2.158
2.137
Transition state energies (kcal mol21
)
AM1
RHF/3-21G*//RHF/AM1
11.4
22.0
8.2
14.5
1.1
4.7
0.0
0.0
Ground state energies (kcal mol21
AM1
RHF/3-21G*//RHF/AM1
)
10.8
19.3
12.2
21.3
0.0
0.0
2.4
3.6

1768
A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
all four possible transition states are concerted and
synchronous in nature. The newly forming C· · ·C s-bond
˚
lengths range between 2.13 and 2.16 A (Table 6).
predicted via application of semiempirical (AM1) compu-
tational methodology (see Table 6).
3.4. Aldrin vs isodrin: comparison of relevant transition
states
The results obtained via AM1 calculation of transition state
energetics suggest that the endo,endo cycloadduct, i.e. that
which results via transition state 8d, should be the preferred
product formed via kinetic control of Diels–Alder cyclo-
addition of cyclopentadiene to 1,2,3,4,7,7-hexachloro-
norbornadiene, in accord with experimental observations.
Interestingly, the computational results predict aldrin
(formed via transition state 8c) to be the preferred product
of thermodynamic control of this cycloaddition.
Formation of aldrin and isodrin (i.e. 7 and 8, respectively)
via the kinetically controlled Diels–Alder reactions
described herein is governed in each case by a subtle
interplay among often conflicting stereoelectronic effects.
Aldrin is formed via approach of 1,2,3,4,5,5-hexachloro-
cyclopentadiene toward the exo face of one of the 2 equiv.
norbornadiene CvC double bonds. This same mode of
approach is followed in the corresponding Diels–Alder
cycloaddition of cyclopentadiene to norbornadiene.16 Thus,
the presence of the six C–Cl groups in the dienophile does
not appear to have a dramatic effect upon the stereo- and
regiochemical course of the former process.
As was observed for the corresponding calculations
performed for Diels–Alder cycloaddition of 1,2,3,4,5,5-
hexachlorocyclopentadiene to norbornadiene (vide supra),
the results of single point calculations performed at HF/3-
21G* level of theory agree with the corresponding AM1
calculated results. In addition, the HF/3-21G* results show
larger energy differences among 8a–8d transition states and
also among the corresponding ground states than are
However, the situation is quite different in the case of
the kinetically controlled cycloaddition of cyclopentadiene
to 1,2,3,4,7,7-hexachloronorbornadiene, which leads
Scheme 4.

A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
1769
exclusively to the formation of isodrin (8). Buildup of
intramolecular Cl· · ·H non-bonded interactions is likely to
occur in 8a as the cycloaddition reaction proceeds to afford
the corresponding exo,exo cycloadduct (Scheme 4). Elec-
trostatic repulsions between the negative end of a C–Cl
bond dipole and a neighboring CvC double bond are likely
to be significant as 8b proceeds to form the corresponding
exo,endo cycloadduct (Scheme 4). Kinetically favored
transition state 8d proceeds in accordance with the Alder–
Stein rule9 and develops only minimal intramolecular H· · ·H
non-bonded interactions as it proceeds to afford the
observed endo,endo [4þ2] cycloadduct, 8 (Scheme 4).
chloronorbornadiene to norbornadiene, the resultant of
these effects can override the normal preference for exo
approach of the diene upon a norbornadiene CvC double
bond.11 Similarly, the normal regiochemical course of
[4þ2] thermal cycloadditions that typically adhere to the
familiar Alder–Stein rule9 is not followed in the case of
aldrin formation via [4þ2] cycloaddition of 1,2,3,4,5,5-
hexachlorocyclopentadiene to norbornadiene.
Recently, we noted the fact that application of relatively low
levels of theory (i.e. semiempirical and Hartree–Fock
molecular orbital calculations with small basis sets) can
be used to provide chemically meaningful predictions with
regard to the regio- and stereochemical course of Diels–
Alder reactions.5 The corresponding results obtained in the
present study are consistent with this observation and serve
to further demonstrate the utility of low-level, inexpensive
theoretical methods for this purpose.
4. Summary and conclusions
Thermal [4þ2] cyclodimerization of 1,2,3,4,5-pentachloro-
cyclopentadiene (1) and the corresponding Diels–Alder
cycloaddition of 1 to norbornadiene have been carried out.
In each case, only one cycloadduct was isolated. The
reaction products thereby prepared were shown to be 2a and
4c, respectively. The structure of each cycloadduct was
established unequivocally via application of single crystal
X-ray crystallographic techniques.
5. Experimental
High-resolution mass spectral data reported herein were
obtained by Professor Jennifer S. Brodbelt at the Mass
spectrometry Facility at the Department of Chemistry and
Biochemistry, University of Texas at Austin by using a
ZAB-E double sector high resolution mass spectrometer
(Micromass, Manchester, England) that was operated in the
chemical ionization mode. Elemental microanalytical data
was obtained by personnel at M-H-W Laboratories, Inc.,
Phoenix, AZ.
The mechanisms of the Diels–Alder cycloadditions that
have been used historically to prepare aldrin and isodrin (7
and 8, respectively) have been examined computationally.
The results of semempirical (AM1 Hamiltonian) and ab
initio (Hartree–Fock HF/3-21G*) calculations both suggest
that the formation of isodrin via Diels–Alder cycloaddition
of cyclopentadiene to 1,2,3,4,7,7-hexachloronorbornadiene
is a kinetically controlled process. It seems likely that the
same is true for the analogous formation of aldrin via Diels–
Alder cycloaddition of 1,2,3,4,5,5-hexachlorocyclopenta-
diene to norbornadiene.
5.1. Computational methodology
Semiempirical (AM1) calculations have been performed for
all the possible transition state and ground state
geometries.13 The transition states and ground states have
been characterized via complete vibrational frequency
analyses. Single point calculations have been performed at
the HF/3-21G* level of theory by using AM1 calculated
geometries.
In the cycloaddition that leads to the formation of aldrin, the
kinetically preferred transition state involves approach of
the diene upon the exo face of one of the two equiv. CvC
double bonds in the dienophile (norbornadiene). Of the two
pathways by which exo approach of the diene upon the
dienophile might occur, that which leads to the formation of
aldrin minimizes non-bonded steric interactions between a
C(5)–Cl bond in the diene and the anti C–H(7) bond in the
dienophile.
5.1.1. Pentachlorocyclopentadiene. A solution of hexa-
chlorocyclopentadiene 14 (20.4 g, 74.9 mmol) in acetone
(10 mL) was cooled via application of an external ice-water
bath. To this cooled solution was added dropwise with
stirring a solution of SnCl2·H2O (17.2 g, 76.7 mmol) in
acetone (50 mL) in such a manner that a reaction
temperature of 30–358C was maintained throughout the
addition process. After the addition of the reducing agent
had been complexed (ca. 10 min), the external ice-water
bath was removed, and the brown reaction mixture was
allowed to warm gradually to ambient temperature while
stirring during 1 h. The reaction mixture was concentrated
in vacuo without external heating, and CH2Cl2 (50 mL) was
added to the residue. The aqueous (upper) layer was
removed via pipette, and the remaining organic layer was
dried (CaCl2), and filtered, and the filtrate was concentrated
in vacuo without external heating. The residue was distilled
in vacuo via application of an external oil bath that was
heated rapidly to 1008C. Pure 1,2,3,4,5-pentachlorocyclo-
pentadiene (6.0 g, 34%) was thereby obtained as a yellow
oil: bp 105–1078C (5 torr.) [lit.10 bp 73–768C (4 torr.)]; IR
In the cycloaddition that leads to the formation of isodrin,
the kinetically preferred transition state involves approach
of the diene upon the endo face of the non-chlorinated CvC
double bond in the dienophile (1,2,3,4,7,7-hexachloro-
norbornadiene). Of the four possible pathways by which
the diene might approach the dienophile might occur, that
which leads to the formation of isodrin minimizes non-
bonded steric interactions between a C(5)–H bond in the
diene and the dienophile (see transition states 8a–8d, as
depicted in Table 6 and Scheme 4).
The course of each of the Diels–Alder reactions that leads
to the formation of aldrin and isodrin appears to be governed
by a subtle interplay among frequently conflicting stereo-
electronic effects. In the case of isodrin formation via [4þ2]
cycloaddition of cyclopentadiene to 1,2,3,4,7,7,-hexa-

1770
A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
1
(film) 2938 (m), 1603 cm21 (s); H NMR (CDCl3) d 4.75
Table 7. X-Ray data collection and processing parameters for 2a and 4c
(s); 13C NMR (CDCl3) d 60.2 (d), 129.0 (s), 129.6 (s).
Compound
2a
4c
5.1.2. Diels–Alder dimerization of 1,2,3,4,5-pentachloro-
cyclopentadiene. Upon standing at ambient temperature, a
sample of 1,2,3,4,5-pentachlorocyclopentadiene (3.0 g,
6.3 mmol) slowly was converted into a single Diels–Alder
dimer, C10H2Cl10.8 This material was purified via fractional
recrystallization from hexane. Pure 2a (2.4 g, 80%) was
thereby obtained as a colorless microcrystalline solid: mp
210–2118C (lit7,8 mp 220–2218C); IR (KBr) 2996 (w),
Formula
Size (mm)
Space group
C
10H2Cl10
C12H9Cl5
0.61£0.25£0.16
Pbca
8.0602(6)
14.367(1)
23.144(2)
90
90
90
2680.6(4)
8
1.638
1.055
298
1.76–28.31
15,810
3217
0.0643
3217
155
0.51£0.48£0.21
Pca2(1)
15.239(1)
7.4327(6)
13.708(1)
90
90
90
1552.6(2)
4
2.039
1.777
213(2)
2.67–28.24
9201
3310
0.0393
3310
188
˚
a (A)
˚
b (A)
˚
c (A)
a (8)
b (8)
g (8)
3
V (A )
˚
1
1607 (s), 1313 (m), 1232 (s), 1156 (m), 870 cm21 (m); H
Z-value
Dcalc (g cm23
NMR (CDCl3) d 4.92 (s, 1 H), 5.01 (s, 1 H); 13C NMR
(CDCl3) d 64.7 (d), 78.7 (d), 81.7 (s), 82.1 (s), 82.3 (s), 86.2
(s), 129.5 (s), 131.6 (s), 134.4 (s), 134.7 (s). Anal. Calcd for
C10H2Cl10: Cl, 74.38. Found: Cl, 73.85.
)
m (mm21
T (K)
)
u range for data collection (deg.)
Total reflections
Independent reflections
Rint
I^2s(I)
Parameters
5.1.3. Diels – Alder cycloaddition of 1,2,3,4,5-
pentachlorocyclopentadiene to norbornadiene.
A
solution of 1,2,3,4,5-pentachlorocyclopentadiene (2.38 g,
1.0 mmol) and norbornadiene (920 mg, 1.0 mmol) in
toluene (200 mL) was refluxed during 24 h. The reaction
mixture then was concentrated in vacuo, and the residue was
purified via column chromatography on silica gel by eluting
with hexane. Pure 4c (1.4 g, 50%) was thereby obtained as a
colorless microcrystalline solid: mp 82–838C; IR (KBr)
3072 (w), 3035 (w), 1591 (s), 1470 (s), 1315 (s) 1250 (s),
1015 cm21 (s); 1H NMR (CDCl3) d 1.31 (AB, JAB¼10.7 Hz,
1 H), 1.50 (AB, JAB¼10.7 Hz, 1 H), 2.47 (s, 2H), 2.90–2.92
(m, 2H), 4.34 (s, 1H); 6.28 (s, 2H); 13C NMR (CDCl3) d
40.2 (t), 41.2 (d), 55.2 (d), 75.9 (s), 82.4 (d), 129.8 (s), 140.9
(d). Exact mass (CI-HRMS) Calcd for C12H9Cl5 [Mþr þH]þ
m/z 328.9225. Found: Mrþ 328.9216. Anal. Calcd for
C12H9Cl5: C, 43.61; H, 2.75. Found: C, 43.35; H, 3.00.
Goodness of fit
R, Rw
1.030
0.0278; 0.0728
0.594; 20.288
1.027
0.0421; 0.1086
0.368; 20.394
23
˚
max; rmin (e A )
r
kindly obtained the high-resolution chemical ionization
mass spectral data reported herein.
References
1. Diels, O.; Alder, K. Justus Liebigs Ann. Chem. 1928, 460, 98.
2. (a) Lehr, R.; Marchand, A. P. Operational Criteria for
Evaluation of Concertedness in Potential Pericyclic Reactions.
In Pericyclic Reactions; Marchand, A. P., Lehr, R. E., Eds.;
Academic Press: New York, 1977; Vol. 1, pp 1–51. (b) March,
J. Advanced Organic Chemistry; 4th ed. Wiley: New York,
1992; pp 839–852; and references cited therein.
5.2. X-Ray structure determination of 2a and 4c
All X-ray data were collected on a Bruker SMARTe 1000
CCD based diffractometer. The frames were integrated with
the SAINT program package17 by using a narrow-frame
algorithm, and the structures were solved and were refined
by using the SHELXTL program package.18 The structures
were checked by using PLATON.19 The molecules are well
packed with no solvent voids. Thermal ellipsoid plots are
drawn at the 30% probability level. The five and six-
membered rings exhibit envelope and boat conformations,
respectively. Packing interactions lead to several Cl· · ·H and
Cl· · ·Cl interactions that are slightly shorter than the sum of
the van der Waals radii. X-Ray data for 2a and 4c are listed
in Table 7. Complete crystallographic details are available
as supplementary material and have been deposited at the
Cambridge Crystallographic Data Centre [CCDC 196805
(4c) and 196806 (2a)].
3. 1,4-Cycloaddition Reactions. The Diels–Alder Reaction in
Heterocyclic Synthesis; Hamer, J., Ed.; Academic: New York,
1967.
4. Marchand, A. P.; Chong, H.-S.; Ganguly, B.; Coxon, J. M.
Croat. Chem. Acta 2000, 73, 1027.
5. For reviews, see: (a) Coxon, J. M.; Froese, R. D. J.; Ganguly,
B.; Marchand, A. P.; Morokuma, K. Synlett 1999, 1681.
(b) Marchand, A. P.; Coxon, J. M. Acc. Chem. Res. 2002, 35,
271.
6. Spangler, C. W. Chem. Rev. 1976, 76, 187.
7. McBee, E. T.; Smith, D. K. J. Am. Chem. Soc. 1955, 77, 389.
8. Williamson, K. L.; Hsu, Y.-F. L.; Lacko, R.; Youn, C. H.
J. Am. Chem. Soc. 1969, 91, 6129.
9. Alder, K.; Stein, G. Angew. Chem. 1937, 50, 510.
10. Wellman, M. A.; Burry, L. C.; Letourneau, J. E.; Bridson, J. N.;
Miller, D. O.; Burnell, D. J. J. Org. Chem. 1997, 62, 939.
11. See: Rondan, N. D.; Paddon-Row, M. N.; Caramella, P.; Houk,
K. N. J. Am. Chem. Soc. 1981, 103, 2436, and references cited
therein.
Acknowledgements
12. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
J. Am. Chem. Soc. 1985, 107, 3902.
We thank the Robert A. Welch Foundation [Grants B-963
(A. P. M) and P-074 (W. H. W.] and the US Office of Naval
Research [Grant N00014-98-1-0478 (A. P. M.)] for financial
support of this study. In addition, we thank Professor
Jennifer S. Brodbelt, Department of Chemistry and
Biochemistry, University of Texas at Austin, for having
13. (a) Spartan, version 5.0, purchased from Wavefunction, Inc.
18401 Von Karman., Suite 370 Irvine, CA 92612. (b) Frisch,
M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson,
B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson,
G. A.; Montgomery, J. A.; Raghavachari, K.; AlLatham, M. A.;

A. P. Marchand et al. / Tetrahedron 59 (2003) 1763–1771
1771
Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski,
J.; Stefanov, B. B.; Nanayakara, A.; Challacombe, M.; Reng,
C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.;
Binlley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. GAUSSIAN 94;
Gaussian Inc: Pittsburgh, PA, 1995.
16. (a) Soloway, S. B. J. Am. Chem. Soc. 1952, 74, 1027. (b) Stille,
J. K.; Frey, D. A. J. Am. Chem. Soc. 1959, 81, 4273.
17. SAINT, Version 6.02; Bruker Analytical X-ray Systems, Inc.,
Copyright 1997–1999.
18. SHELXTL, Version 5.1; Bruker Analytical X-ray Systems, Inc.,
Copyright 1998.
19. (a) Spek, A. L. Acta Crystallog. 1990, C34, 46. (b) Speck, A. L.
PLATON—A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2001.
14. Lidov, R. E. US Patent 2,635,977; Chem. Abstr. 1954, 48,
2769.
15. (a) Bluestone, H. US Patent 2,676,132; Chem. Abstr. 1954, 48,
8474. (b) Lidov, R. E. US Patent 2,733,248; Chem. Abstr.
1956, 50, 2914.