Interplay of orbital symmetry and nonstatistical dynamics in the thermal rearrangements of bicyclo[n.1.0]polyenes

Article abstract of DOI:10.1021/ja017083j

CASSCF and CASPT2 calculations have been carried out on some of the thermal rearrangements of bicyclo[2.1.0]pentene (BCP), bicyclo[4.1.0]hepta-2,4-diene (BCH), bicyclo[6.1.0]nona-2,4,6-triene (BCN), and 9,9-dicyanobicyclo[6.1.0]nona-2,4,6-triene (DCBCN). In addition, experiments have been conducted to determine the stereoselectivity and temperature dependence of the nondegenerate rearrangement of 9,9-dicyanobicyclo[6.1.0]nona-2,4,6-triene-exo-¹⁵N. The calculations and experiments allow a consistent picture to be drawn for these reactions. The principal conclusions are as follows. (1) The ring-walk rearrangements of BCP, BCN, and DCBCN are pericyclic reactions occurring with a strong preference for inversion of configuration at the migrating carbon. However, the ring-walk rearrangement of BCH is a nonpericyclic reaction. (2) The rearrangement of DCBCN to 9,9-dicyanobicyclo[4.2.1]nona-2,4,7-triene occurs with a preferred stereochemistry corresponding to a 1,3 migration with retention. However, this reaction is not a pericyclic process; the stereoselectivity is probably of dynamic origin. (3) Cyano substituents can significantly reduce the activation energy for a reaction occurring via a singlet biradical, but they do'not necessarily cause the intermediate to sit in a deeper local minimum on the potential energy surface.

Full text of DOI:10.1021/ja017083j

Published on Web 01/03/2002
Interplay of Orbital Symmetry and Nonstatistical Dynamics in
the Thermal Rearrangements of Bicyclo[n.1.0]polyenes
Mayra B. Reyes, Emil B. Lobkovsky, and Barry K. Carpenter*
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853-1301
Received September 14, 2001
Abstract: CASSCF and CASPT2 calculations have been carried out on some of the thermal rearrangements
of bicyclo[2.1.0]pentene (BCP), bicyclo[4.1.0]hepta-2,4-diene (BCH), bicyclo[6.1.0]nona-2,4,6-triene (BCN),
and 9,9-dicyanobicyclo[6.1.0]nona-2,4,6-triene (DCBCN). In addition, experiments have been conducted
to determine the stereoselectivity and temperature dependence of the nondegenerate rearrangement of
9,9-dicyanobicyclo[6.1.0]nona-2,4,6-triene-exo-15N. The calculations and experiments allow a consistent
picture to be drawn for these reactions. The principal conclusions are as follows. (1) The ring-walk
rearrangements of BCP, BCN, and DCBCN are pericyclic reactions occurring with a strong preference for
inversion of configuration at the migrating carbon. However, the ring-walk rearrangement of BCH is a
nonpericyclic reaction. (2) The rearrangement of DCBCN to 9,9-dicyanobicyclo[4.2.1]nona-2,4,7-triene occurs
with a preferred stereochemistry corresponding to a 1,3 migration with retention. However, this reaction is
not a pericyclic process; the stereoselectivity is probably of dynamic origin. (3) Cyano substituents can
significantly reduce the activation energy for a reaction occurring via a singlet biradical, but they do not
necessarily cause the intermediate to sit in a deeper local minimum on the potential energy surface.
Introduction
However, one could certainly conceive of constrained transition-
state geometries where good pericyclic overlap for a [1,n]
Theoretical and experimental work from this laboratory has
led to the conclusion that the formal [1,n] sigmatropic migrations
of carbon will usually not be pericyclic processes, and that the
stereochemistry of such reactions will consequently not be
subject to the rules of orbital symmetry conservation.1 Instead,
these reactions are expected to occur with the intervention of
transient singlet-state biradicals, whose behavior is frequently
not well described by statistical kinetic models such as transition
state theory (TST).2 Calculations on the nominal [1,3] rear-
rangement of bicyclo[3.2.0]hept-2-ene3 and on the vinylcyclo-
propane rearrangement4 provide supporting examples for the
general proposal.
migration should be possible. The number of such exceptions
is not expected to be large; in fact, only two are readily apparent.
The first has been previously discussed:3a the [1,5] sigmatropic
migration in 1,3-cyclopentadienes.5 Because this reaction is
really only a 1,2 rearrangement, and because the thermally
allowed reaction should occur with retention of configuration
at the migrating atom, it is easy to attain a pericyclic transition
structure with strong overlap among the orbitals of the making
and breaking bonds. In accord with this picture is the known
very high stereoselectivity of the reaction5 and the fact that the
experimental activation enthalpy6 plus the estimated heat of
formation for the reactant in the spiro[4.4]nona-1,3-diene
rearrangement place the transition structure roughly 20 kcal/
mol below that calculated for a mechanism involving a singlet
biradical.3a
The simple physical picture that leads to this generalization
also allows ready identification of situations where exceptions
may occur. The thesis is merely that the overlap of the orbitals
in the transition structure for a [1,n] migration of carbon is
usually too weak to permit the significant “aromatic” delocal-
ization associated with a thermally allowed pericyclic process.
The second potential exception could occur for the “ring-
walk” rearrangements7 of those bicyclo[n.1.0]polyenes for which
the thermally allowed process should occur with inVersion at
the migrating carbon. The first two members of the series are
the [1,3] shift in bicyclo[2.1.0]pentene8 and the [1,7] shift in
bicyclo[6.1.0]nona-2,4,6-triene.9 In this class of reaction, ready
(1) Newman-Evans, R. H.; Simon, R. J.; Carpenter, B. K. J. Org. Chem. 1990,
55, 695-711.
(2) (a) Carpenter, B. K. Acc. Chem. Res. 1992, 25, 520-528. (b) Carpenter,
B. K. Angew. Chem., Int. Ed. 1998, 37, 3340-3350.
(3) (a) Carpenter, B. K. J. Am. Chem. Soc. 1995, 117, 6336-6344. (b)
Carpenter, B. K. J. Am. Chem. Soc. 1996, 118, 10329-10330. (c) Beno,
B. R.; Wilsey, S.; Houk, K. N. J. Am. Chem. Soc. 1999, 121, 4816-4826.
(d) Wilsey, S.; Houk, K. N.; Zewail, A. H. J. Am. Chem. Soc. 1999, 121,
5772-5786.
(4) (a) Davidson, E. R.; Gajewski, J. J. J. Am. Chem. Soc. 1997, 119, 10543-
10544. (b) Houk, K. N.; Nendel, M.; Wiest, O.; Storer, J. W. J. Am. Chem.
Soc. 1997, 119, 10545-10546. (c) Doubleday, C.; Nendel, M.; Houk, K.
N.; Thweatt, D.; Page, M. J. Am. Chem. Soc. 1999, 121, 4720-4721. (d)
Doubleday, C. J. Phys. Chem. A 2001, 105, 6333-6341.
(5) Boersma, M. A. M.; De Haan, J. W.; Kloosterziel, H.; Van de Ven, L. J.
M. J. Chem. Soc. D 1970, 1168-1169.
(6) (a) Dane, L. M.; De Haan, J. W.; Kloosterziel, H. Tetrahedron Lett. 1970,
2755-2758. (b) Semmelhack, M. F.; Weller, H. N.; Foos, J. S. J. Am.
Chem. Soc. 1977, 99, 292-294. (c) Replogle, K. S.; Carpenter, B. K. J.
Am. Chem. Soc. 1984, 106, 5751-5753.
(7) (a) Woodward, R. B.; Hoffmann, R. The ConserVation of Orbital Symmetry;
Verlag Chemie: Weinheim, 1970. (b) Kla¨rner, F.-G. Top. Stereochem. 1984,
15, 1-42.
9
10.1021/ja017083j CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 641

A R T I C L E S
Reyes et al.
Figure 1. Schematic representation of a “ring-walk” reaction occurring with inversion of configuration at the migrating carbon.
Scheme 1. Summary of Possible Rearrangements of
Bicyclo[6.1.0]nonatriene, with the Products Selected Being Those
Corresponding to the Observed Stereochemical Preference
access to a pericyclic transition state seems possible because
the cyclopropane ring opening is expected to be accompanied
by a rehybridization at the migrating carbon and a more-or-
less “automatic” completion of the pericyclic array. The situation
is depicted schematically in Figure 1.
One might anticipate that a pericyclic transition structure
would be especially favored if the angle θ were small, bringing
the pπ orbitals at the ends of the conjugated chain into better
overlap with the p-type orbital on the migrating carbon. A small
value of θ implies a small ring and hence a short conjugated
chain of π orbitals.
A second factor favoring a short conjugated chain can be
discerned if one views the pericyclic transition structure from
the perspective of a perturbation theory model. Interaction of
the p-type orbital on the migrating carbon with the nonbonding
π molecular orbital (NBMO) of the chain should be most
favorable when the coefficients of the NBMO at the termini of
the chain are largest. This will necessarily be the case for the
shortest chain, since the NBMO has to be “shared” by more
carbons as the chain lengthens. It seems reasonable to anticipate,
therefore, that the ring-walk rearrangement in bicyclo[2.1.0]-
pentene would be the most likely candidate for a pericyclic
mechanism in this class of [1,n] carbon migrations.
Scheme 2. Hypothetical Biradical Mechanism for the
Rearrangements of Bicyclo[6.1.0]nonatriene
However, the second member of the seriessthe 1,7 ring-walk
in bicyclo[6.1.0]nonatrienessis in some ways more interesting
because it is accompanied by another nominal sigmatropic
rearrangement that is either a 1,3 or a 1,5 shift (or both).10
Studies on substituted molecules have shown that the ring-walk
occurs with very high stereochemical fidelity for the “allowed”
inversion of configuration.9 The competing 1,3/1,5 rearrange-
ment occurs with lower stereoselectivity and favors the stereo-
chemistry that would be retention for a 1,3 migration or
inversion for a 1,5 migration (see Scheme 1)si.e., the “forbid-
den” stereochemistry from an orbital symmetry conservation
standpoint. The 1,3/1,5 shift also appears to have a higher barrier
than the ring-walk since in similarly substituted molecules it
occurs at a reasonable rate only at about 180 °C, whereas the
ring-walk occurs at about 100 °C.9,10 The computational and
experimental results reported in the present paper are designed
to distinguish among several mechanisms that could be invoked
to explain these facts. The mechanisms under consideration are
as follows:
(A) The ring-walk and the skeletal rearrangements are both
pericyclic sigmatropic reactions, with the minor product of the
[1,3]/[1,5] migration coming from a nominally allowed [1,3]-
inversion and/or [1,5]-retention pathway.
(B) The ring-walk is a purely pericyclic reaction, but the
competing rearrangement occurs by parallel pericyclic and
biradical mechanisms, with the biradical giving both stereo-
isomers of the bicyclo[4.2.1]nonatriene product.
(C) The ring-walk is a purely pericyclic reaction, but the
competing rearrangement is a purely biradical process.
(D) Both the ring-walk and the competing rearrangement
occur by biradical mechanisms. To explain the markedly
different rates of the two rearrangements, this mechanism would
appear to require two ancillary hypotheses: that the exo and
endo conformers of the reactant be separated by a barrier higher
than that for C-C bond scission, and that the putative biradicals
obtained from the exo and endo conformers of the reactant be
distinct and separated from interconversion by a barrier higher
than any of those for product formation (see Scheme 2).
Mechanisms C and D can be considered to have subcatogories
in which the reactions occurring via singlet biradicals have
stereochemical outcomes determined either by conformational
(8) (a) Schoeller, W. W. J. Am. Chem. Soc. 1975, 97, 1978-1980. (b) Kla¨rner,
F. G.; Adamsky, F. Angew. Chem. 1979, 91, 738-739. (c) Kla¨rner, F. G.;
Adamsky, F. Chem. Ber. 1983, 116, 299-322. (d) Kla¨rner, F. G.; Glock,
V.; Figge, H. Chem. Ber. 1986, 119, 794-812. (e) Skancke, P. N.;
Yamashita, K.; Morokuma, K. J. Am. Chem. Soc. 1987, 109, 4157-4162.
(f) Jensen, F. J. Am. Chem. Soc. 1989, 111, 4643-4647.
(9) (a) Kla¨rner, F. G. Angew. Chem., Int. Ed. Engl. 1972, 11, 832-833. (b)
Kla¨rner, F. G.; Wette, M. Chem. Ber. 1978, 111, 282-298.
(10) Kla¨rner, F. G. Tetrahedron Lett. 1971, 3611-3614.
9
642 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Thermal Rearrangements of Bicyclo[n.1.0]polyenes
A R T I C L E S
Table 1. Summary of CASPT2//CASSCF Results on the
barriers on the potential energy surface or by nonstatistical
reaction dynamics.2
Ring-Walk Reactions
na
∆H q − ∆H q (kcal/mol)b
∆WBIc
ret
inv
The relationship of these issues to the general proposal
described in the opening paragraph is that mechanisms A and
B invoke pericyclic [1,3] and/or [1,5] shifts of carbon that would
constitute clear counter examples to the general thesis, since
they belong to neither of the classes for which exceptions are
anticipated.
4
6
8
11.1
0.09
6.3
0.1435
-0.0173
0.0910
a Index defining the size of the larger ring in the reactant (n), the number
of active electrons and orbitals in the MCSCF calculations (n,n), and the
order of the sigmatropic rearrangement constituting the ring-walk ([1,n -
1]). b Difference in CASPT2(n,n)//CASSCF(n,n) activation enthalpies for
the ring-walk with retention and inversion of configuration. All calculations
used the 6-31G(d) basis set and unscaled CASSCF vibrational frequencies.
c Difference in the Wiberg bond index between the inversion and retention
transition structures, where the bond in question is the one made (or
equivalently the one broken) in the ring-walk reaction.
We have approached the resolution of the mechanistic
questions by joint application of ab initio electronic structure
theory and experiment. Calculations have been performed on
the ring-walk reactions of bicyclo[2.1.0]pentene, bicyclo[4.1.0]-
hexadiene, and bicyclo[6.1.0]nonatriene, as well as on the 9,9-
dicyano derivative of the last molecule. In addition, the
rearrangements of the parent and 9,9-dicyanobicyclo[6.1.0]-
nonatrienes to the corresponding bicyclo[4.2.1]nonatrienes have
been explored computationally. Experiments were carried out
on a 9,9-dicyanobicyclo[6.1.0]nonatriene that was stereoselec-
tively 15N-labeled in one of the nitrile substituents.
calculated in the natural atomic orbital basis.12 The results are
summarized in Table 1, but their analysis is deferred to the
Discussion section of this paper.
CASPT2 and CASSCF Calculations on the Bicyclo[6.1.0]-
nonatriene-to-Bicyclo[4.2.1]nonatriene Rearrangement. CASS-
CF(8,8)/6-31G(d) calculations revealed the existence of several
biradical-like stationary points on the bicyclo[6.1.0]/bicyclo-
[4.2.1]nonatriene potential energy surface (PES). Their con-
nections to each other and to the closed-shell minima were
revealed by intrinsic reaction coordinate (IRC) analyses. This
connectivity pattern is summarized in Schemes 3 and 4. Energies
of the stationary points were refined at the CASPT2(8,8)/6-
31G(d) level. The presence of at least one valley ridge inflection
(VRI) point13 on the PES was inferred, for reasons presented
in the Discussion section.
Because the experiments described below were carried out
on 9,9-dicyanobicyclo[6.1.0]nonatriene, and because computa-
tional studies on related systems have suggested significant
influence from such substituents,14 it seemed desirable to
undertake additional calculations on this substituted compound.
However, the substitution not only increases the number of
heavy atoms by four, but it also increases the size of the minimal
active space from (8,8), corresponding to 1764 singlet-coupled
configuration state functions (CSFs), to (12,12), corresponding
to 226 512 singlet CSFs. It was consequently not feasible to
carry out as detailed a survey of the substituted PES as had
been attempted for the parent hydrocarbon. Neither CASSCF
vibrational frequency nor single-point CASPT2 calculations
proved to be computationally tractable.
Results
CASPT2 Calculations on the Ring-Walk Reaction in the
Parent Hydrocarbons. CASPT2/6-31G(d)//CASSCF/6-31G-
(d) calculations were carried out on the transition states and
any intermediates for the ring-walk reactions of bicylo[2.1.0]-
pentene, bicyclo[4.1.0]hepta-2,4-diene, and bicylo[6.1.0]octa-
2,4,6-triene. In each case, the Cs-symmetry stationary points
corresponding to the inversion and retention stereochemistries
were located at the CASSCF level, using (4,4), (6,6), and (8,8)
active spaces respectively for the three different reactants. The
nature of the stationary point was determined by harmonic
normal-mode analysis at the same level. Single-point CASPT
calculations from the CASSCF references were then performed
to allow for dynamic electron correlation corrections to the
relative energies.
If all three molecules underwent allowed pericyclic ring-walk
reactions, one would expect the allowed-stereochemistry station-
ary points (inversion for the first and third molecules, but
retention for the second) all to be transition structures and all
to be lower in energy than their forbidden-stereochemistry
counterparts. However, this relative-energy criterion is not
sufficient to show that the reactions are pericyclic in nature.
The reason is that molecular symmetries permitting the correct
pericyclic topolgy for “aromatic” stabilization are also those
for which biradical and zwitterion configurations would be
permitted to mix in a nonpericyclic intermediate or transition
structure. Specifically, in the Cs point group the closed-shell
(zwitterionic) configurations are obviously of A′ symmetry for
all three molecular systems. The open-shell (biradical) configu-
rations are also of A′ symmetry in the “allowed” geometry but
are of A′′ symmetry in the “forbidden” geometry. Thus, a lower
energy for the “allowed” structure might be due only to the
fact that more electronic configurations can mix in that
geometry. To determine whether pericyclic overlap plays a
significant role, one needs to analyze some measure of bonding
of the migrating carbon to the migration origin (or, equivalently
by symmetry, to the migration terminus) in the transition
structure. We chose for that purpose the Wiberg bond index,11
Several studies on reactions involving singlet biradicals have
used unrestricted density functional theory (DFT) or hybrid
Hartree-Fock/DFT models.3c,15 Such approaches would be
computationally feasible for the present study, but we were
aware that they were controversial.16 We consequently elected
(12) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83,
735-746.
(13) (a) Valtazanos, P.; Ruedenberg, K. Theor. Chim. Acta 1986, 69, 281-
307. (b) Taketsugu, T.; Tajima, N.; Hirao, K. J. Chem. Phys. 1996, 105,
1933-1939. (c) Ramquet, M.-N.; Dive, G.; Dehareng, D. J. Chem. Phys.
2000, 112, 4923-4934.
(14) Kless, A.; Nendel, M.; Wilsey, S.; Houk, K. N. J. Am. Chem. Soc. 1999,
121, 4524-4525.
(15) See, for example: (a) Goldstein, E.; Beno, B.; Houk, K. N. J. Am. Chem.
Soc. 1996, 118, 6036-6043. (b) Houk, K. N.; Beno, B. R.; Nendel, M.;
Black, K.; Yoo, H. Y.; Wilsey, S.; Lee, J. THEOCHEM 1997, 398-399,
169-179. (c) Cramer, C. J. J. Am. Chem. Soc. 1998, 120, 6261-6269.
(16) (a) Yamaguchi, K.; Jensen, F.; Dorigo, A.; Houk, K. N. Chem. Phys. Lett.
1988, 149, 537-542. (b) Wittbrodt, J. M.; Schlegel, H. B. J. Chem. Phys.
1996, 105, 6574-6577. (c) Staroverov, V. N.; Davidson, E. R. J. Am. Chem.
Soc. 2000, 122, 186-187. (d) Orlova, G.; Goddard, J. D. J. Chem. Phys.
2000, 112, 10085-10094. (e) Grafenstein, J.; Cremer, D. Mol. Phys. 2001,
99, 981-989.
(11) Wiberg, K. B. Tetrahedron 1968, 24, 1083-1096.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 643

A R T I C L E S
Reyes et al.
Scheme 3. CASPT2(8,8)//CASSCF(8,8)/6-31G(d) Structures and Relative Enthalpies (kcal/mol) for the Rearrangement of
cis-Bicyclo[6.1.0]nona-2,4,6-trienea
a VRI is a valley ridge inflection point.
to restrict our attention to CASSCF(12,12) calculations on key
stationary points. The results are summarized in Scheme 5.
Since vibrational frequency calculations were not feasible for
the cyano-substituted species, the identity of each stationary
point and its connections to the others on the PES were identified
by analogy with the results from the calculations on the parent
hydrocarbon. In most cases the structures were very similar,
and so this comparison was straightforward. The single excep-
tion was the Cs-symmetry structure XIX. In the unsubstituted
hydrocarbon the analogous structure, VIII, was found to be a
transition state at the CASSCF level (before inclusion of ZPE
differences). The calculations on XIX suggest that it is probably
an intermediate, since geometry optimization without symmetry
restriction did not cause it to relax to a lower-energy C1 structure.
However, this difference is probably inconsequential since the
ZPE and CASPT2 corrections removed the shallow local minima
on the PES for the parent system and would very probably do
the same for the PES of the cyano-substituted molecule.
9
644 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Thermal Rearrangements of Bicyclo[n.1.0]polyenes
A R T I C L E S
Scheme 5. CASSCF(12,12)/6-31G(d) Structures and Relative
Scheme 4. CASPT2(8,8)//CASSCF(8,8)/6-31G(d) Structures and
Relative Enthalpies for the Rearrangement of
trans-Bicyclo[6.1.0]nona-2,4,6-triene
Potential Energies for the Ring-Walk and Skeletal Rearrangements
of 9,9-Dicyanobicyclo[6.1.0]nona-2,4,6-trienea
For reasons described in the Discussion section, we also
carried out CASPT2(6,6) calculations on singlet 1,1-dicyano-
trimethylene biradical and its conrotatory transition state for
formation from 1,1-dicyanocyclopropane. These results are
summarized in Table 2.
Synthesis and Rearrangement of 9,9-Dicyanobicyclo[6.1.0]-
nonatriene-exo-15N. It has been recognized for some time that
for the parent cis-bicyclo[6.1.0]nona-2,4,6-triene the most facile
thermal rearrangement is an electrocyclic reaction that involves
breaking the C1-C8 bond.17 Scission of the C1-C9 bond,
required for both the ring-walk rearrangement and the skeletal
rearrangement to bicyclo[4.2.1]nona-2,4,7-triene, is apparently
unable to compete to any significant extent. However, the
situation changes if appropriate substituents are attached to C9.
Groups that can weaken the C1-C9 bond while strengthening
the C1-C8 bond are known to be capable of reversing the
relative favorability of cleavage of these two bonds.9,10
We sought to take advantage of this substituent effect in the
design of a compound that could provide some mechanistic
insight into the formation of the bicyclo[4.2.1]nonatriene isomer.
In particular, we were interested in applying the criterion of
temperature dependence of stereoselectivity, which has proven
useful in elucidating the mechanisms of other rearrangements.18
The idea behind this approach is that reactions occurring by
parallel mechanisms are expected to exhibit temperature-
dependent kinetic product ratios, since the heats of formation
of the competitive transition structures will be different except
by improbable coincidence. However, reactions whose stereo-
chemical outcome is determined by nonstatistical dynamics have
thus far been found to exhibit little or no temperature depen-
dence of their product ratios, provided that the products differ
only in absolute configuration or in the position of a kinetically
a By analogy with the calculations on the unsubstituted hydrocarbon,
structure XVII is expected to be the ring-walk transition state. The asterisk
shows the location of the 15N label in the experiments conducted on this
system. The structure XXI corresponds to the preferred product found in
these experiments.
Table 2. Summary of Calculations on Trimethylene and
1,1-Dicyanotrimethylene
b
c
structure
CASSCF energya
CASPT2 energya
ZPE
E
rel
cyclopropane
conrotatory TS
trimethylene
1,1-dicyano
cyclopropane
conrotatory TS
(dicyano)
-117.0754726110 -117.4628191820 54.443 -59.1
-116.9903067925 -117.3610644412 49.099
-116.9903428141 -117.3606419232 49.474
-0.69
[0]
-300.5731417157 -301.4806041218 54.025 -44.1
-300.5177969213 -301.4021522091 48.376
-300.5184540962 -301.4021193831 48.871
-0.52
1,1-dicyano-
trimethylene
[0]
a Absolute energies in hartree. b Unscaled harmonic zero-point energies
in kcal/mol. c Relative enthalpies at 0 K. The first three entries use
trimethylene as the reference; the last three use 1,1-dicyanotrimethylene as
the reference. All species are in their lowest singlet electronic states.
innocent isotopic label.4d,18b,19 This last condition is necessary
because products that are diastereomeric to the extent of more
than an isotope difference are likely to be formed in temperature-
dependent ratios by almost any mechanism.
Bearing in mind these criteria, we elected to prepare a 9,9-
dicyanobicyclo[6.1.0]nonatriene that was 15N-labeled in just one
(17) Vogel, E. Angew. Chem., Int. Ed. Engl. 1963, 2, 1-11.
(18) (a) Newman-Evans, R. H.; Carpenter, B. K. J. Am. Chem. Soc. 1984, 106,
7994-7995. (b) Lyons, B. A.; Pfeifer, J.; Carpenter, B. K. J. Am. Chem.
Soc. 1991, 113, 9006-9007.
(19) Lyons, B. A.; Pfeifer, J.; Peterson, T. H.; Carpenter, B. K. J. Am. Chem.
Soc. 1993, 115, 2427-2437.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 645

A R T I C L E S
Reyes et al.
Scheme 6. Synthetic Procedure for the Preparation of
9,9-Dicyanobicyclo[6.1.0]nona-2,4,6-triene-exo-15N a
Figure 3. Structure of compound 6 determined by X-ray crystallography.
Scheme 7. Summary of the Thermal Rearrangement and the
Procedure Used To Deduce Its Stereochemical Preference
a The reagents and conditions for each step were as follows: (a) ClO2S-
NCO, Et2O, -70 °C; (b) H2O, EtOH, room temperature; (c) Cl3CCOCl,
Et3N, CH2Cl2, 0-5 °C; (d) cyclooctatetraene, [Rh(OAc)2]2 (catalyst), CCl4,
room temperature; (e) 15NH3, H2O, EtOH, room temperature.
mixture with aqueous ammonia caused only one of the
components to be converted to an amide, while the other
remained unaffected. The amide was isolated and recrystallized,
and an X-ray diffraction structure was determined. It is shown
in Figure 3. Replacement of the normal ammonia in this last
reaction by 15NH3, followed by dehydration of the resulting
amide, permitted selective generation of compound 4x, thereby
allowing the preferred stereochemistry of the rearrangement to
be deduced. In accord with earlier studies on unsymmetrically
substituted bicyclo[6.1.0]nonatrienes,10 the rearrangement to the
bicyclo[4.2.1]nonatriene skeleton was found to occur with a
preference for the stereochemistry that would correspond to
retention in a 1,3 migration and/or inversion in a 1,5 migration.
The stereoselectivity of the rearrangement of compound 3
was determined as a function of temperature and reaction
medium, with results that are summarized in Tables 3 and 4.
Figure 2. Structure of compound 2 determined by X-ray crystallography.
of the nitriles. The synthetic sequence that permitted this is
shown in Scheme 6. The stereochemistry of the cyclopropana-
tion reaction was established by obtaining single-crystal X-ray
diffraction data on the unlabeled version of amide 2. The
molecular structure deduced from this experiment is shown in
Figure 2.
Discussion
Calculations on the Ring-Walk Rearrangements. The
results summarized in Table 1 show evidence at the CASPT2//
CASSCF level of theory for pericyclic transition structures of
the ring-walk reactions with inversion of configuration for both
bicyclo[2.1.0]pentene and bicyclo[6.1.0]nonatriene. The sig-
nificant enthalpy difference between inversion and retention
transition structures, in combination with the substantially
greater Wiberg bond indices for the former stereochemistry, lead
us to this conclusion. As anticipated, the magnitude of pericyclic
Thermal rearrangement of compound 3 to bicyclo[4.2.1]-
nonatrienes 4x and 4n occurred cleanly at 120-170 °C. 15N
NMR revealed that the products were formed in a 2.7:1 ratio
but did not reveal which was which. To determine that
information, compound 1 was heated to 200 °C, causing it to
rearrange to a mixture of stereoisomeric bicyclo[4.2.1]nonatrienes
5x and 5n (Scheme 7). Surprisingly, treatment of this product
9
646 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Thermal Rearrangements of Bicyclo[n.1.0]polyenes
A R T I C L E S
Table 3. Temperature Dependence of the Product Ratio in
Benzene Solution
temperature (±0.2 °C)
product ratio (4x/4n)
120.0
130.0
140.0
150.0
160.0
170.0
2.64
2.61
2.62
2.65
2.72
2.70
Table 4. Solvent Dependence of the Product Ratio, 4x/4n, at 160
°C
benzene-d0
benzene-d6
cis-decalin
trans-decalin
2.82
2.78
3.29
3.49
3.44
3.39
3.50
2.79
2.79
2.80
2.79
2.79
2.81
2.84
2.79
3.22
3.35
3.15
2.80 ( 0.02
2.80 ( 0.03
3.25 ( 0.16
3.46 ( 0.09
Figure 4. CASSCF(8,8)/6-31G(d) potential energy profile along the IRC
from the ring-walk transition structure of bicyclo[6.1.0]nona-2,4,6-triene.
overlap is greater in the transition structure for the rearrangement
of the smaller bicyclic molecule. In fact, at the CASSCF(4,4)/
6-31G(d) level there is even a shallow minimum in the PES
for the Cs-symmetry geometry corresponding to the ring-walk
“transition structure” with inversion. If this reflected the true
PES, the reaction might be considered pericylic but not
concerted! However, ZPE and CASPT2 corrections remove the
minimum and restore a more traditional description of the
reaction as a concerted, pericyclic process.
inversion transition structure, is strongly suggestive of the
presence of a VRI.
Calculations on the Bicyclo[6.1.0]nonatriene-to-Bicyclo-
[4.2.1]nonatriene Rearrangement. We have chosen to analyze
the computational results for this reaction in terms of the
mechanistic hypotheses labeled A-D in the Introduction. We
begin with the CASPT2//CASSCF calculations on the parent
hydrocarbon.
For the calculations to be consistent with mechanism A or
B, at least some component of the bicyclo[6.1.0]nonatriene-to-
bicyclo[4.2.1]nonatriene rearrangement would have to occur by
a pericyclic mechanism. However, the calculations offered no
support for such a pathway. The only pericyclic transition
structure that could be found on the entire PES was that for the
ring-walk reaction, IV. Three other ring-opening transition
structures were located (V, VI, and XII), but in each case they
connected a stereoisomer of bicyclo[6.1.0]nonatriene to a singlet
biradical.
Mechanism D was found to be inconsistent with the com-
putational results on several grounds. First, as described above,
was the fact that the ring-walk transition structure IV showed
clear indications of being pericyclic. Actually, two additional
nonpericyclic mechanisms (II f VI f VII f VIII f VII′
fVI′ f II′ and I f V f XI f V′ f I′) for the ring-walk
reaction were found, but they were calculated to be substantially
higher in activation enthalpy than the pericyclic process. Second
was the fact that, while discrete exo and endo conformers of
the biradical were located, as mechanism D had posited,
formation of the exo biradical was calculated to have a higher
barrier by nearly 7 kcal/mol than formation of the endo one.
Third was the computational result that the barrier to inter-
conversion of conformers I and II was much lower than the
barrier to any ring-opening reaction.
Mechanism C was found to be the most consistent with the
calculations. Within this mechanism, the original analysis of
the problem considered two possible explanations for the
stereochemical preference observed experimentally. One was
that there were conformational barriers on the biradical PES,
and the other was that the singlet biradical may be subject to
nonstatistical dynamical effects of the kind proposed in several
other reactions. At least in the case of the parent hydrocarbon,
The story is very different for the ring-walk reaction in
bicyclo[4.1.0]hexadiene (norcaradiene). In this case the allowed
pericyclic process should occur with retention of configuration.
However, we find virtually no difference between the energies
or Wiberg bond indices of the retention and inversion structures.
It appears that the retention of stereochemistry mandated by
orbital symmetry conservation leads to a transition structure in
which overlap is too poor to permit the completion of a
pericyclic array. This ring-walk reaction is thus calculated to
occur by a nonpericyclic biradical mechanism. Previous studies
have come to the same conclusion and have discussed its
rationalization with the experimental stereoselectivity.14,20
Before leaving the analysis of the ring-walk reactions, it is
perhaps worth noting that the bicyclo[6.1.0]nonatriene version
probably involves a valley ridge inflection point on the PES.
The calculated geometry for the transition structure reveals an
almost perfectly flat eight-membered ring, suggesting that both
exo and endo conformations of the reactant undergo the
degenerate rearrangement via the same transition state. This
would be possible only if there existed a VRI between the
stereoisomeric reactants and the transition structure. IRC analysis
in mass-weighted coordinates showed the transition state to be
connected only to the endo reactant, but it is known that IRC
calculations cannot reveal the existence of a VRI, since all
common IRC algorithms force selection of one of the two paths
when they encounter such a bifurcation point.21 In support of
the existence of a VRI is the unusual energy profile along the
IRC, shown in Figure 4. The “elbow” in the IRC profile,
occurring at an energy just below that found for the ring-
(20) Jarzecki, A. A.; Gajewski, J. J.; Davidson, E. R. J. Am. Chem. Soc. 1999,
121, 6928-6935.
(21) Yanai, T.; Taketsugu, T.; Hirao, K. J. Chem. Phys. 1997, 107, 1137-
1146.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 647

A R T I C L E S
Reyes et al.
Scheme 8. Top View of the Pathway for the Skeletal
Rearrangement That Is Experimentally Found To Be Preferred
the calculations offered no support at all for the former
explanation. As has been found in several other reactions, the
PES in the vicinity of the singlet biradical structures was
calculated to be extremely flat. At the CASSCF level, structure
VII was found to be a very shallow minimum, with potential
energy barriers of 0.08 kcal/mol for reclosure to II and 0.27
kcal/mol for closure to the product. The barrier to interconver-
sion of enantiomers VII and VII′ via Cs-symmetry TS VIII
was found to be less than 0.01 kcal/mol. With ZPE, thermal,
and CASPT2 corrections, the minimum corresponding to
structure VII disappeared completely. The reaction is best
described at this level of theory as involving ring opening onto
an energetic plateau from which barrierless exits lead to both
bicyclo[6.1.0]- and bicyclo[4.2.1]nonatrienes. For such reactions,
we have argued previously that the selection of exits may be
determined largely by the phenomenon of dynamic matching.3a,22
In effect, this hypothesis says that the biradical has the highest
probability of leaving the plateau by the exit that it encounters
first. The order of encounter of exits is determined by the
detailed trajectories that biradicals follow across the PES, and
those, in turn, owe their behavior in large measure to the
dynamics of formation of the biradical. A complete computa-
tional test of such a hypothesis requires a full molecular
dynamics simulation, but the results of the electronic structure
calculations reported here at least lend a measure of plausibility
to the idea.
The first thing to notice about the stereochemical issue is
that the calculations show a strong preference for retention of
configuration at the migrating carbon. The lack of complete
stereoselectivity in the formation of bicyclo[4.2.1]nonatriene is
due to a competition between 1,3 and 1,5 migrations with
retention, not to a competition between rearrangements with
retention and inversion of configuration. All attempts to find a
transition structure for internal rotation in biradical VII resulted
only in collapse to the pericyclic transition structure VI. Thus,
according to the calculations, the formation of bicyclo[4.2.1]-
nonatriene occurs with retention at the migrating carbon not
because there is a large barrier to internal rotation in the biradical
intermediate, but instead because those biradicals undergoing
internal rotation find themselves simply returning to the bicyclo-
[6.1.0]nonatriene minimum.
The sequence of stationary points II f VI f VII f IX f
X defines a path that corresponds to the experimentally preferred
stereochemistry of the rearrangement (at least in substituted
analogues.) However, it is important to recognize that there is
calculated to be no additional barrier preventing a molecule from
following the alternative path II f VI f VII f VIII f VII′
f IX′ f X′, which would give the minor stereoisomer of the
product if the groups on the migrating carbon were made
distinguishable. We believe that the preference for the former
path derives from the fact that it is dynamically coherent. As
shown in the top view (Scheme 8), the conversion of II to the
ring-opening transition structure VI involves substantial clock-
wise rotation about the C8-C9 bond. This motion orients the
migrating methylene in such a fashion that the first product-
forming transition state to be encountered is the one corre-
sponding to a 1,3 migration with retention. This is precisely
the phenomenon that we have referred to as dynamic matching.3a
However, this explanation requires further scrutiny. We have
previously suggested that the generally observed preference for
inVersion at the migrating carbon in a 1,3 migration is explicable
in terms of dynamic matching.3a How, then, could the preference
for retention observed in this case have the same origin? We
believe that the answer is an interesting one having to do with
the competitive ring-walk process. The issue is really the sense
of rotation that accompanies the initial bond cleavage. The
situation is summarized in Scheme 9, which compares the
transition structures for the skeletal rearrangement and the
pericyclic ring-walk reaction, using the same top view employed
for Scheme 8. While the former reaction, as noted, involves
clockwise rotation about the C8-C9 bond, the latter clearly
entails counterclockwise rotation in order to achieve the overlap
that closes the pericyclic array. In the absence of the ring-walk
reaction, previously presented arguments3a suggest that the
preferred mode of bond cleavage for the 1,3 shift would have
involved counterclockwise rotation, leading eventually to prod-
uct formation with predominant inversion of configuration.
However, the presence of the ring-walk pathway causes
trajectories that involve breaking the C1-C9 bond with
counterclockwise torsion about C8-C9 to be diverted from the
skeletal rearrangement into the lower-energy ring-walk manifold.
To avoid this “trap”, the initial bond scission has to occur with
the less favorable clockwise internal rotation, which then leads
to the 1,3 shift occurring with preferential retention.
Although not directly relevant to the questions addressed in
this study, we note in passing that trans-bicyclo[6.1.0]nonatriene,
XIII, is calculated to be capable of undergoing a degenerate
(22) Hrovat, D. A.; Fang, S.; Borden, W. T.; Carpenter, B. K. J. Am. Chem.
Soc. 1998, 120, 5603-5604.
9
648 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Thermal Rearrangements of Bicyclo[n.1.0]polyenes
A R T I C L E S
Scheme 9. Comparison of the Sense of Internal Rotation about
the C1-C9 Bond That Accompanies the Ring-Walk or Skeletal
Rearrangement
the conrotatory transition structure was investigated, since this
is found to be the one of lowest energy for the stereomutation
of unsubstituted cyclopropane.25
A simple interpretation of the computational results on the
substituent effect is the following. The formation of a singlet
biradical from a closed-shell reactant is generally a reaction that
is highly endothermic and hence, by the Hammond postulate,
should have a very “late” or biradical-like transition state.
Introduction of radical-stabilizing substituents does lead to a
reduction in the overall endothermicity of the reaction, but the
effect of the substituents is quantitatively similar on the biradical
and the transition state. The result is reduction in the barrier to
bond cleavage without a qualitative change in the topography
of the PES.
The qualitative similarity of the PES for the substituted and
unsubstituted bicyclo[6.1.0]nonatrienes leads us to conclude that
the mechanisms of the rearrangements in both systems should
also be similar. In particular, the bicyclo[6.1.0]nonatriene-to-
bicyclo[4.2.1]nonatriene rearrangements for both systems are
expected to be nonpericyclic and to occur with a dynamically
determined preference for the 1,3-migration with retention of
configuration.
Experimental Studies on the Rearrangement of 9,9-
Dicyanobicyclo[6.1.0]nonatriene-exo-15N. The experimentally
determined stereochemical preference for the rearrangement of
9,9-dicyanobicyclo[6.1.0]nonatrieneto9,9-dicyanobicyclo[4.2.1]-
nonatriene is in accord with both the calculations and earlier
experimental studies on unsymmetrically substituted analogues.
There is no doubt that the observed preference for the product
corresponding to a 1,3 shift with retention (or, indistinguishably,
a 1,5 shift with inversion) is kinetic rather than thermodynamic
in origin, since the products differ only in the position of
nitrogen isotopes, and so would have an equilibrium constant
very close to unity. If the kinetic preference arose from the
existence of parallel mechanisms occurring over potential energy
barriers of different heights, one would expect to see a
temperature-dependent product ratio. However, within the
experimental error, the product ratio is found to be temperature
independent over a range of 50 °C.26 The lack of temperature
dependence, and the dependence on reaction medium, are in
accord with experimental results that we have found for other
reactions in which the product ratio is believed to be determined
by the dynamic behavior of a transient intermediate.18 That is
consequently the explanation that we favor for the present
results, too.
ring-walk reaction that would be nonpericyclic, i.e., unlike the
ring-walk in the cis stereoisomer (see Scheme 4). This com-
putational result seems to be consistent with conclusions derived
from experiments on the trans isomer.23
In the case of the norcaradiene ring-walk reaction, it has been
suggested that substituents on the migrating carbon can signifi-
cantly change the shape of the PES.14 Since the experimental
studies in the bicyclo[6.1.0]nonatriene series required substit-
uents on the migrating carbon, we felt it necessary to address
the possibility that the analysis presented above for the parent
hydrocarbon might not be relevant for the system that was
studied experimentally. Of particular concern was the known
radical-stabilizing ability of nitriles. Experiments suggest that
a pair of geminal nitriles should provide 12.4 ( 0.9 kcal/mol
of radical stabilization.24 If these substituents similarly stabilized
biradicals, one might anticipate that the very flat region of the
PES found for the bicyclo[6.1.0]nonatriene-to-bicyclo[4.2.1]-
nonatriene rearrangement in the parent hydrocarbon would be
transformed in the dicyano derivative to a surface with distinct
local minima and barriers to reaction that might control the
observed stereochemistry. However, the actual results were quite
different from this simple picture. As summarized in Scheme
5, the addition of the nitriles did substantially reduce the barriers
for the ring-walk and skeletal rearrangements of bicyclo[6.1.0]-
nonatriene but did not cause the appearance of substantial PES
minima corresponding to singlet biradical intermediates.
To find out whether this result was peculiar to the bicyclo-
nonatriene system, we also carried out CASPT2(6,6) calculations
on the singlet biradical generated from C1-C2 scission in 1,1-
dicyanocyclopropane. Here, too, the calculations showed that
the substituents should cause a 15 kcal/mol reduction in the
barrier to ring opening without generating a significant minimum
on the PES for the resulting singlet biradical (see Table 2). Only
The quantitative stereoselectivity observed in this rearrange-
ment (73-77%, depending on reaction medium) is similar to
that found for the corresponding reaction of the 9-cyano-9-
methyl analogues (85% for the endo isomer and 70% for the
exo).9,10 However, without a detailed theoretical study of the
9-cyano-9-methyl compounds, it is unclear whether this similar-
ity is mechanistically significant.
(25) (a) Getty, S. J.; Davidson, E. R.; Borden, W. T. J. Am. Chem. Soc. 1992,
114, 2085-2093. (b) Baldwin, J. E.; Yamaguchi, Y.; Schaefer, H. F., III.
J. Phys. Chem. 1994, 98, 7513-7522. (c) Doubleday, C., Jr. J. Phys. Chem.
1996, 100, 3520-3526. (d) Baldwin, J. E.; Freedman, T. B.; Yamaguchi,
Y.; Schaefer, H. F., III. J. Am. Chem. Soc. 1996, 118, 10934-10935.
(26) The discrepancy between the uncertainty deduced from the replicate
measurements at constant temperature (Table 3) and the single-point
measurements at varying temperature (Table 2) suggests that the experi-
ments must have been subject to a small systematic error. Even when this
is included, there is still no signficant temperature dependence detectable.
(23) Glock, V.; Wette, M.; Kla¨rner, F. G. Tetrahedron Lett. 1985, 26, 1441-
1444.
(24) Pakusch, J.; Beckhaus, H. D.; Ru¨chardt, C. Chem. Ber. 1991, 124, 1191-
1198.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 649

A R T I C L E S
Reyes et al.
room temperature while ethyl R-cyanodiazoacetate (1.02 g, 9.18 mmol)
was added dropwise. The solution became warm but required no
external cooling if the addition was done slowly. After the addition,
the reaction mixture was stirred until no starting material could be
detected by TLC. The reaction mixture was then filtered, and the solvent
was removed under vacuum. The crude product was purified by
chromatography on silica gel utilizing 30% ethyl acetate in hexane as
eluent. Pure product (0.43 g, 22%) was obtained as a very thick
yellowish oil. 1H NMR (400 MHz, CDCl3): δ 1.35 (t, J ) 7 Hz, 3H),
2.64 (s, 2H), 4.28 (q, J ) 7 Hz, 2H), 5.94 (d, J ) 11 Hz, 4H), 6.17(d,
J ) 11 Hz, 2H). 13C NMR(100 MHz, CDCl3): δ 14.04, 26.86, 35.16,
62.91, 115.65, 122.39, 125.98, 132.30, 167.20.
Experimental Section
Computational Details. GAMESS27 was used for most of the
CASSCF calculations, although analytical vibrational frequency and
Wiberg bond index calculations were carried out with Gaussian 98.28
CASPT2 calculations were carried out with MOLCAS.29 All calcula-
tions used the 6-31G(d) basis set with six Cartesian d functions.
Calculations were carried out on a 500 MHz dual-processor Compaq
Alpha DS-20E workstation, or a Dell Precision 330 workstation
employing a 1.7 GHz Pentium 4 processor and running the Linux
operating system.
Ethyl r-(Chlorosulfonylaminocarbonyl)diazoacetate. To a solution
of ethyl diazoacetate (5.0 g, 43.8 mmol) in ether (50 mL) that was
stirred under nitrogen at -70 °C (dry ice/acetone bath) was added a
solution of chlorosulfonyl isocyanate (6.2 g, 43.8 mmol) in ether (20
mL) over a 30 min period. After the addition was complete, the reaction
was stirred at the same temperature for an additional 3 h. The reaction
mixture was warmed to room temperature, and the ether was removed
using a rotary evaporator. No further purification was required, and
9-Cyanobicyclo[6.1.0]nona-2,4,6-triene-9-carboxamide, 2. To a
solution of 9-cyanobicyclo[2.2.l]nona-2,4,6-triene-9-carboxylic acid
ethyl ester (0.33 g, 1.53 mmol) in ethanol (3 mL) was added aqueous
ammonia (5 mL, 3.3 N solution). For preparation of the 15N-labeled
amide, 15N-labeled aqueous ammonia (99 at. %, 3.3 N solution) was
used. The reaction mixture was stirred for several days until no starting
material was observed by GC in a quenched aliquot. When the reaction
was complete, the solvent was removed and the resulting solid washed
with ether several times. The ether was decanted, and the solid was
dried under vacuum. No further purification was needed. 9-Cyanobicyclo-
[2.2.l]nona-2,4,6-triene-9-carboxamide was obtained in 53% yield. 1H
NMR (200 MHz, DMSO): δ 2.97 (s, 2H), 4.5 1 (s, 1H), 6.39 (d, J )
10 Hz, 2H), 6.42 (s, 2H), 6.64 (d, J ) 10 Hz, 2H). 13C NMR (75 MHz,
DMSO): δ 28.19, 32.40, 117.04, 123.51, 125.68, 131.10, 166.13.
1
the product was obtained in 95% yield (11.1 g). H NMR (400 MHz,
CDCl3): δ 1.39 (t, J ) 7 Hz, 3H), 4.39 (q, J ) 7 Hz, 2H), 11.32 (bs,
1H). 13C NMR (100 MHz, CDCl3): δ 63.51, 157.39, 163.33.
Ethyl r-(Aminocabonyl)diazoacetate. A solution of ethyl R-
(chlorosulfonylaminocabonyl)diazoacetate (9.23 g 36.25 mmol) in
ethanol (50 mL) and water (5 mL) was allowed to stand overnight at
room temperature. The solvent was decanted from the precipitate that
formed, and the solid was washed with ether. The ether was decanted,
and traces were removed under vacuum. This gave the product (4.33
g, 76%) as yellow crystals. 1H NMR (400 MHz, CDCl3): δ 1.33 (t, J
The stereochemistry of the product was determined by X-ray
crystallography. Details of the diffraction experiment and the structure
are provided in the Supporting Information. The diffraction data were
obtained on a colorless monoclinic single crystal of dimensions 0.6 ×
0.4 × 0.1 mm3. The crystal was mounted on a Bruker diffractometer
equipped with a CCD detector, using Mo KR radiation at 0.71073 Å
and operating at 293 K. The structure was solved in space group P21/
c, with a ) 15.675(3) Å, b ) 8.651(2) Å, c ) 7.1630(14) Å, â )
99.38(3)°, V ) 958.3(3) Å3, Z ) 4, D ) 1.291 g cm-3, and µ ) 0.085
mm-1. A total of 4080 reflections were collected in the range 1.32 e
2θ e 23.28°, leading to a set of 1364 independent reflections. The
structure was solved by direct methods using full-matrix least-squares
on F2. All non-hydrogen atoms were refined with anisotropic thermal
parameters. The refinement, using 168 parameters, converged to R(F)
) 0.0588, Rw(F) ) 0.1357, and GOF ) 1.032.
9,9-Dicyanobicyclo[6.1.0]nona-2,4,6-triene, 3. To a stirred mixture
of 9-cyanobicyclo[2.2.l]nona-2,4,6-triene-9-carboxamide (0.10, 0.54
mmol) and dry triethylamine (0.108 g, 1.08 mmol) in CH2Cl2 (5 mL)
was added a solution of trichloroacetyl chloride (0.108 g, 0.54 mmol)
in dry CH2Cl2 (5 mL) dropwise between 0 and 5 °C. After the addition
was finished, the mixture was treated sequentially with water and
aqueous sodium bicarbonate. The separated organic solution was dried
over magnesium sulfate, and the solvent was removed using a rotary
evaporator. No further purification was needed (40 mg, 44%). 1H NMR
(200 MHz, CDCl3): δ 2.73 (s, 2H), 5.98 (d, J ) 10.0 Hz, 2H), 6.00 (s,
2H), 6.26 (d, J ) 10.0 Hz, 2H). 13C NMR(100 MHz, CDCl3): δ 11.20,
34.95, 112.89, 115.89, 121.45, 126.06, 132.91 ppm.
) 7 Hz, 3 H), 4.31 (q, J ) 7 Hz, 2H), 5.76 (bs, 1H), 7.58 (bs, 1H). 13
NMR (100 MHz, CDCl3): δ 14.25, 61.78, 163.36, 163.67.
C
Ethyl r-Cyanodiazoacetate. A mixture of ethyl R-(aminocabonyl)-
diazoacetate (1.47 g, 9.35 mmol) and triethylamine (3.12 g, 28.05 mmol)
in dichloromethane (20 mL) was cooled in an ice bath and stirred while
a solution of trichloroacetyl chloride (5.61 g, 30.86 mmol) in dry CH2-
Cl2 (20 mL) was added dropwise at a rate to maintain the temperature
between 0 and 5 °C. After the addition was finished, the solution was
treated sequentially with water and aqueous sodium bicarbonate. The
separated organic layer was dried over magnesium sulfate, and the
solvent was removed using a rotary evaporator. The crude product was
purified by chromatography on silca gel, utilizing 30% ethyl acetate
in hexane as eluent. Pure ethyl R-cyanodiazoacetate (0.72 g, 69%) was
obtained as a yellow oil. IR (neat): 2229, 2131, 1714 cm-1. 1H NMR
(400 MHz, CDCl3): δ 1.34 (t, J ) 7 Hz, 3H), 4.35 (q, J ) 7 Hz, 2H).
13C NMR (100 MHz, CDCl3): δ 14.21, 63.42, 107.30, 161.17.
9-Cyanobicyclo[6.1.0]nona-2,4,6-triene-9-carboxylic Acid Ethyl
Ester, 1. A 50 mL round-bottomed flask was charged with rhodium
acetate dimer (2 mol %) in carbon tetrachloride (15 mL). Cycloocta-
tetraene (0.96 g, 9.18 mmol) was added, and the mixture was stirred at
(27) GAMESS (Version 26 OCT 2000 R4 for UNIX): Schmidt, M. W.;
Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.;
Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L. J.
Comput. Chem. 1993, 14, 1347-1363.
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz J.
V.; Baboul, A. G.; Stefanov, B. B.; Liu G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.;
Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98, Rev. A.9; Gaussian, Inc.: Pittsburgh, PA, 1998.
9,9-Dicyanobicyclo[4.2.1]nona-2,4,7-triene, 4x and 4n. A solution
of 9,9-dicyanobicyclo[6.1.0]nona-2,4,6-triene (20 mg) in benzene-d6
was placed in a 16.5 cm × 5 mm i.d. Pyrex tube. The solution was
degassed by three freeze-pump-thaw cycles and sealed under vacuum.
The tube was heated to a constant temperature in the range 120-170
°C in a thermostated oil bath. The tube was cooled quickly to room
temperature, opened, and the contents were transferred to a NMR tube.
The products were analyzed by 1H NMR. In the case of the 15N-labeled
reactant, the ratio of product stereoisomers was determined by their
15N NMR shifts at 4.86 and 8.93 ppm downfield from 9,9-dicyanobicyclo-
(29) Andersson, K.; Barysz, M.; Bernhardsson, A.; Blomberg, M. R. A.; Cooper,
D. L.; Fleig, T.; Fu¨lscher, M. P.; de Graaf, C.; Hess, B. A.; Karlstro¨m, G.;
Lindh, R.; Malmqvist, P.-A° .; Neogra´dy, P.; Olsen, J.; Roos, B. O.; Sadlej,
A. J.; Schu¨tz, M.; Schimmelpfennig, B.; Seijo, L.; Serrano-Andre´s, L.;
Siegbahn, P. E. M.; Stålring, J.; Thorsteinsson, T.; Veryazov; V.; Widmark,
P.-O. MOLCAS, Version 5; Lund University, Sweden, 2000.
1
[6.l.0]nonatriene-exo-15N used as reference at 0 ppm. H NMR (400
MHz, CDCl3): δ 2.85 (d, J ) 3.8 Hz, 2H), 4.52 (s, 2H), 5.28 (m, 2H),
9
650 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Thermal Rearrangements of Bicyclo[n.1.0]polyenes
A R T I C L E S
5.72 (m, 2H). 13C NMR (100 MHz, CDCl3): δ 35.50, 52.61, 114.02,
117.67, 120.54, 127.65, 130.77.
allowed stereochemistry is inversion at the migrating carbon
but nonpericyclic when the orbital-symmetry-allowed stereo-
chemistry is retention.
9-Cyanobicyclo[4.2.1]nona-2,4,7-triene-9-carboxylic Acid Ethyl
Ester, 5x and 5n. A solution of 9-cyanobicyclo[6.1.0]nona-2,4,6-triene-
9-carboxylic acid ethyl ester (20 mg) in C6D6 was placed in a 16.5 cm
× 5 mm i.d. Pyrex tube. The solution was degassed by three freeze-
pump-thaw cycles and sealed under vacuum. The tube was heated to
200 °C in a thermostated oil bath for 24 h. The tube was cooled to
room temperature, opened, and the contents were used in the next
reaction.
9-Cyanobicyclo[4.2.1]nona-2,4,7-triene-9-carboxylic acid amide,
6. To a solution of 9-cyanobicyclo[4.2.l]nona-2,4,7-triene-9-carboxylic
acid ethyl ester (0.100 g, 0.47 mmol) in ethanol (3 mL) was added
aqueous ammonia (3 mL, 3 N solution in water). For preparation of
the 15N-labeled amide, 15N-labeled aqueous ammonia (99 at. %, 3.3 N
solution) was used. The reaction mixture was then stirred until no
starting material in a quench aliquot was observed by GC (only one
isomer reacted, and the reaction took several days). When the reaction
was complete, the solvent was removed, and the solid residue was
washed with ether several times to remove the ester that had not reacted.
The ether was decanted, and the solid was dried under vacuum. The
crude product was recrystallized from cold THF/pentane. 1H NMR (400
MHz, CDCl3): δ 3.77 (d, J ) 4.0 Hz, 2H), 5.26 (d, J ) 2.0 Hz, 2H),
5.51 (br s, 2H), 6.21 (m, 4H).
The stereochemistry of the product was determined by X-ray
crystallography. Details of the diffraction experiment and the structure
are provided in the Supporting Information. Single crystals of the
compound were obtained as a mono-THF solvate. The diffraction data
were obtained on a colorless triclinic single crystal of dimensions 0.4
× 0.4 × 0.05 mm3. The crystal was mounted on a Bruker diffractometer
equipped with a CCD detector, using Mo KR radiation at 0.71073 Å
and operating at 293 K. The structure was solved in space group P1h,
with a ) 6.3516(6) Å, b ) 10.3942(10) Å, c ) 10.7314(10) Å, R )
94.2070(17)°, â ) 96.9300(13)°, γ ) 102.4230(16)°,V ) 683.23(11)
Å3, Z ) 2, D ) 1.256 g cm-3, and µ ) 0.084 mm-1. A total of 3417
reflections were collected in the range 1.92 e 2θ e 24.71°, leading to
a set of 2267 independent reflections. The structure was solved by direct
methods using full-matrix least-squares on F2. All non-hydrogen atoms
were refined with anisotropic thermal parameters. The refinement, using
213 parameters, converged to R(F) ) 0.0697, Rw(F) ) 0.1449, and
GOF ) 0.617.
(2) The degree of pericyclic overlap in ring-walk reactions
occurring with inversion decreases as the size of the ring
containing the conjugate π system increases.
(3) The bicyclo[6.1.0]nonatriene-to-bicyclo[4.2.1]nonatriene
rearrangement is a nonpericyclic process both for the hypotheti-
cal reaction of the unsubstituted hydrocarbon and for the actual
reaction of the 9,9-dicyano derivative. This is in accord with
the general principle concerning [1,n] sigmatropic rearrange-
ments of carbon outlined in the Introduction to this paper.
(4) There is a preference for retention of configuration at the
migrating carbon in the bicyclo[6.1.0]nonatriene-to-bicyclo-
[4.2.1]nonatriene rearrangement. This derives not from a
significant barrier to internal rotation but from the fact that the
internal rotation that would be expected to lead to inversion
instead returns the system to the lower-energy ring-walk
manifold on the PES. The principal stereoisomer of the bicyclo-
[4.2.1]nonatriene product arises from a 1,3 migration with
retention. The minor stereoisomer comes not from a 1,3
migration with inversion but rather from a 1,5 migration with
retention.
(5) The preference for the 1,3 migration with retention in
the bicyclo[6.1.0]nonatriene-to-bicyclo[4.2.1]nonatriene rear-
rangement is probably of dynamic origin. The stereochemistry
is initiated during the initial bond-breaking event, which has to
occur with the particular sense of internal rotation that will avoid
the ring-walk reaction. This torsion defines the first product-
forming transition structure to be encountered as that corre-
sponding to a 1,3 migration with retention. The dependence of
the product ratio on the sequence of encounter of the product-
forming transition states is incompatible with a statistical kinetic
model for the reaction.
(6) Stabilization of singlet biradicals by nitrile substituents
may well facilitate the formation of such species, but the
stabilization does not necessarily mean that the biradical will
sit in a deeper minimum on the PES than its unsubstituted
analogue would. In the two cases examined here, the nitrile
substitution causes a reduction in the endothermicity of forma-
tion of the biradical by several kilocalories per mole but reduces
the kinetic barrier by an essentially identical amount, thereby
leaving the intermediate in a potential energy well just about
as shallow as that for the parent hydrocarbon.
9,9-Dicyanobicyclo[4.2.1]nona-2,4,7-triene-exo-15N, 4x. To a stirred
mixture of compound 6 (50 mg, 0.27 mmol) and dry triethylamine (0.54
g, 0.54 mmol) in CH2Cl2 at 0-5 °C was added a solution of
trichloroacetyl chloride (0.05 1 g, 0.27 mmol) in dry CH2Cl2 dropwise.
After the addition was finished, the mixture was treated with water
and a saturated solution of sodium bicarbonate (5 × 15 mL). The
separated organic solution was dried with magnesium sulfate, and the
solvent was removed using a rotary evaporator. The crude mixture was
placed in an NMR tube, and a 15N NMR was recorded. 1H NMR (200
MHz, C6D6): δ 2.85 (d, J ) 3.8 Hz, 2H), 4.52 (s, 2H), 5.28 (m 2H),
5.72 (m 2H). 15N NMR (40.5 MHz, C6D6): 4.86 ppm downfield from
9,9-dicyanobicyclo[6.l.0]nonatriene-exo-15N used as reference at 0.
Acknowledgment. We thank the National Science Foundation
(Grant No. CHE-9876387) for support of this research.
Supporting Information Available: Cartesian coordinates and
energies of all stationary points from the calculations (PDF);
X-ray crystal structure data for compounds 2 and 6 (CIF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
Conclusion
The calculations and experiments reported in this paper appear
to provide a consistent picture. The conclusions from the study
can be summarized as follows.
(1) The ring-walk reaction in the bicyclo[n.1.0]polyenes
studied is a pericyclic reaction when the orbital-symmetry-
JA017083J
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 651