Strained allenes as dienophiles in the Diels-Alder reaction: An experimental and computational study

Article abstract of DOI:10.1021/jo982091c

Strained allenes, such as 1,2-cyclohexadiene, undergo facile Diels- Alder reactions with otherwise unreactive dienes. Substituted cyclohexa-1,2- dienes add to furan via a Diels-Alder reaction, forming only two of four possible regioisomers and stereoisomers. Alkyl cyclohexa-1,2-diene carboxylates yield the nonconjugated endo adduct as the major product. However, chiral cyclohexa-1,2-dienecarboxylates, such as l-menthyl and l- bornyl cyclohexa-1,2-dienecarboxylate, show no diastereoselectivity in [4 + 2] cycloadditions. DFT (B3LYP/6-31G*) calculations were performed in order to explain these results. A comparison to the calculated B3LYP/6-31G* transition structures and intermediates along the reaction paths of 1,2- cyclohexadiene with 1,3-butadiene and with furan (as well as propadiene with butadiene) show that the diradical stepwise pathways are preferred over the concerted paths. At the same time, the concerted transition structures are extremely asynchronous. A QM/MM study indicates a minimal influence of the chiral auxiliary even in the concerted scenario.

Full text of DOI:10.1021/jo982091c

976
J . Org. Chem. 1999, 64, 976-983
Str a in ed Allen es a s Dien op h iles in th e Diels-Ald er Rea ction : An
Exp er im en ta l a n d Com p u ta tion a l Stu d y
Maja Nendel,†,‡ Laren M. Tolbert,*,† Laura E. Herring,† Md. Nurul Islam,§ and K. N. Houk*,‡
School of Chemistry and Biochemistry, Georgia Institute of Technology,
Atlanta, Georgia 30332-0400, and the Department of Chemistry & Biochemistry,
University of California, Los Angeles, California 90095-1569
Received October 19, 1998
Strained allenes, such as 1,2-cyclohexadiene, undergo facile Diels-Alder reactions with otherwise
unreactive dienes. Substituted cyclohexa-1,2-dienes add to furan via a Diels-Alder reaction, forming
only two of four possible regioisomers and stereoisomers. Alkyl cyclohexa-1,2-diene carboxylates
yield the nonconjugated endo adduct as the major product. However, chiral cyclohexa-1,2-
dienecarboxylates, such as l-menthyl and l-bornyl cyclohexa-1,2-dienecarboxylate, show no dias-
tereoselectivity in [4 + 2] cycloadditions. DFT (B3LYP/6-31G*) calculations were performed in order
to explain these results. A comparison to the calculated B3LYP/6-31G* transition structures and
intermediates along the reaction paths of 1,2-cyclohexadiene with 1,3-butadiene and with furan
(as well as propadiene with butadiene) show that the diradical stepwise pathways are preferred
over the concerted paths. At the same time, the concerted transition structures are extremely
asynchronous. A QM/MM study indicates a minimal influence of the chiral auxiliary even in the
concerted scenario.
In tr od u ction
metric concerted transition structures or even to stepwise
pathways. In contrast, fumarates represent the sym-
metric extreme. We explored the possibility of chiral
induction by using l-bornyl and l-menthyl 1,2-cyclohexa-
diene carboxylates. If the reaction of 1,2-cyclohexadiene
is concerted, its chirality is transferred to the reaction
products, and the incorporation of a chiral auxiliary
might influence the formation of the chiral center. The
additional driving force provided by the relief of strain
can prompt otherwise unreactive dienophiles to react.
In contrast to extensive studies on simple dienophiles,
allenic systems are only now being explored by compu-
tational methods.7 We have examined the reaction of 1,2-
cyclohexadiene and of the parent allene, 1,2-propadiene,
with dienes, such as 1,3-butadiene and furan using
density functional theory (DFT) in order to explain the
experimentally observed results. DFT methods provided
excellent results for the Diels-Alder reaction between
ethylene and 1,3-butadiene,8 as well as for other pericy-
clic reactions.9
Strained allenes have provided fruitful ground for a
number of theoretical and experimental investigations
into the effect of strain on reactivity.1 They are nonpla-
nar, chiral allenes rather than planar zwitterionic or
carbene-like species, even in the case of the highly
strained 1,2-cyclohexadiene.2,3 1,2-Cyclohexadiene is stable
in a frozen matrix below 170 K. It has been studied
spectroscopically3a and by trapping experiments.3b,c Above
190 K, 1,2-cyclohexadiene dimerizes via a [2 + 2] cy-
cloaddition involving diradical intermediates in the ab-
sence of trapping agents.3c
In earlier studies, we developed a method for the
generation of strained allenes through thermal and
photochemical chloride expulsion from 2-chloroallyl an-
ions.4 We have also been interested in the determination
of the symmetry of Diels-Alder transition states through
the use of “cooperativity” by chiral auxiliaries5 and of
high-level calculations.6 Because of their inherent asym-
metry, allenes are expected to lead to extremely asym-
The Diels-Alder reaction can proceed by two mecha-
nisms. The first is a concerted, allowed [4 + 2] cycload-
dition, while the second is a two-step mechanism involv-
† Georgia Institute of Technology.
‡ University of California at Los Angeles.
§ Present address: Department of Chemistry, Bangladesh Institute
of Technology, Dhaka, Gazipur-1700, Bangladesh.
(6) (a) Borden, W. T.; Loncharich, R. J .; Houk, K. N. Annu. Rev.
Phys. Chem. 1988, 39, 213. (b) Houk, K. N.; Lin, Y.-T.; Brown, F. K. J .
Am. Chem. Soc. 1986, 108, 554. (c) Li, Y.; Houk, K. N. J . Am. Chem.
Soc. 1993, 115, 7478.
(1) (a) Christl, M.; Braun, M. Chem. Ber. 1989, 122, 1939. (b) Christl,
M.; Schreck, M. Chem. Ber. 1987, 120, 915. (c) Christl, M.; Schreck,
M. Angew. Chem. 1987, 99, 474.
(2) Schmidt, M. W.; Angus, R. O., J r.; J ohnson, R. P. J . Am. Chem.
Soc. 1982, 104, 6838.
(7) (a) The use of allene and fluoroallenes as dienophiles in Diels-
Alder reactions has been studied by semiempirical methods: Manohan,
M.; Venuvanalingam, P. J . Chem. Soc., Perkin Trans. 2 1996, 1423.
(b) The [2 + 2] and [2 + 4] reactivity of heterocumulenes and ketenes
was studied: Fabian, W. M. F.; J anoschek, R. J . Am. Chem. Soc. 1997,
119, 4253. (c) Salzner, U.; Bachrach, S. M. J . Org. Chem. 1996, 61,
237. (e) The reactivity of strained allenes in [2 + 2] cycloadditions was
also studied: Burrell, R. C.; Daoust, K. J .; Bradley, A. Z.; DiRico, K.
J .; J ohnson, R. P. J . Am. Chem. Soc. 1996, 118, 4218.
(8) Goldstein, E.; Beno, B.; Houk, K. N. J . Am. Chem. Soc. 1996,
117, 6036.
(9) (a) E.g., the vinylcyclopropane rearrangement: Houk, K. N.;
Nendel, M.; Wiest, O.; Storer, J . W. J . Am. Chem. Soc. 1997, 119,
10545. (b) As well as the Cope and Claisen rearrangements: Wiest,
O.; Black, K. A.; Houk, K. N. J . Am. Chem. Soc. 1994, 116, 10336.
(3) (a) Wentrup, C.; Gross, G.; Maquestiau, A.; Flammang, R. Angew.
Chem. 1983, 95, 551; Angew. Chem., Int. Ed. Engl. 1983, 22, 542. (b)
Balci, M.; J ones, W. M. J . Am. Chem. Soc. 1980, 102, 7607. (c) Moore,
W. R.; Moser, W. R. J . Am. Chem. Soc. 1970, 92, 5469.
(4) (a) Tolbert, L. M.; Islam, Md. N.; J ohnson, R. P.; Loiselle, P. M.;
Shakespeare, W. C. J . Am. Chem. Soc. 1990, 112, 6416. (b) Tolbert, L.
M.; Siddiqui, S. J . Am. Chem. Soc. 1984, 106, 5538. In this case, the
distinction between a planar isoindenylidene and a nonplanar 4,5-
benzo-1,2,4-cyclopentatriene was not resolved. However, recent com-
putational evidence by J ohnson supports the nonplanar structure:
J ohnson, R. P. Private communication.
(5) (a) Tolbert, L. M.; Ali, M. B. J . Am. Chem. Soc. 1981, 103, 2104.
(b) Tolbert, L. M.; Ali, M. B. J . Am. Chem. Soc. 1984, 106, 3806.
10.1021/jo982091c CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/15/1999

Allenes as Dienophiles in the Diels-Alder Reaction
J . Org. Chem., Vol. 64, No. 3, 1999 977
Sch em e 1. P ossible Mech a n ism s of th e Rea ction
betw een 1,2-Cycloh exa d ien e a n d 1,3-Bu ta d ien e
Sch em e 3. Tw o Ster eoch em ica lly Distin ct
In ter m ed ia tes a n d th e Cor r esp on d in g P r od u cts of
th e Step w ise P a th 2 of th e Rea ction betw een
1,2-Cycloh exa d ien e a n d F u r a n a
Sch em e 2. Differ en t Con cer ted Tr a n sition Sta tes
a n d Cor r esp on d in g P r od u cts for th e Rea ction
betw een 1,2-Cycloh exa d ien e a n d F u r a n
a
The stereochemistry of the chiral center opposite to the
bridgehead carbon is indicated as well.
Sch em e 4. F ou r Ster eoch em ica lly Distin ct
In ter m ed ia tes a n d th e Cor r esp on d in g P r od u cts of
th e Step w ise P a th 1 of th e Rea ction betw een
1,2-Cycloh exa d ien e a n d F u r a n
ing a diradical intermediate. In the case of the parent
Diels-Alder reaction between 1,3-butadiene and ethyl-
ene, the concerted path is computed to be favored by 2-7
kcal/mol, and B3LYP/6-31G* predicts this preference to
be 3.4 kcal/mol.8
to the products. The influence of the chiral substituents
on the orientation of the reactants in the initial attack
will be reflected in the product ratio.
The introduction of a cumulated double bond is ex-
pected to exert a significant perturbation on the syn-
chronicity of the concerted transition structures. The
release of strain associated with the formation of diradi-
cal intermediates should exert an additional driving force
for the stepwise path: one of the diradical intermediates
of the stepwise path contains two allyl radicals (Scheme
1). An estimate of the heats of formation by Benson’s
group additivity rules10 predicts that the diallylic diradi-
cal intermediate of the stepwise path 2 is 1 kcal/mol
above the reactants. The heat of formation of the reac-
tants (in particular that of 1,2-cyclohexadiene) is under-
estimated compared to the heat of formation of the
diradical intermediate by this method. There are no
experimental data about the strain exerted by the
introduction of a normally linear allene into a six-
membered ring. While it can be assumed that this strain
is significant, an allyl radical will not experience much
strain in such a system. Therefore, the heat of formation
of the reactants is underestimated compared to the
diradical intermediate, leading to an overestimation of
the heat of reaction by Benson’s group additivity rules.
There are three distinct stereochemical scenarios in the
case of the reaction between 1,2-cyclohexadiene and a
dienophile, one for each possible mechanism. Four ste-
reochemically different transition structures can be
formed in the concerted case (Scheme 2). Two result in
endo products, while the other two provide exo products.11
The stereochemistry of the initial attack is transferred
In the stepwise path 2, the stereochemistry of the
bridgehead carbon is fully determined only in both steps.
Two stereochemically distinct intermediates can be formed,
again leading to the four different products (Scheme 3),
but either intermediate leads only to two products. The
stereochemistry is thus predetermined in the first step.
A chiral auxiliary would have to influence the R/ S
stereochemistry of the chiral center on the diene moiety
in the first step and influence the exo/endo selectivity in
the second step to lead only to one product.
If the other stepwise path (path 1) is taken, the chiral
center is formed in the first step, and the geometry of
the initial approach is transferred to the products. There
are four stereochemically distinct intermediates (Scheme
4).
A substituted 1,2-cyclohexadiene can also lead to
regioisomeric products, which differ in the position of the
substituent relative to the double bond. The R/ S stere-
ochemistry described by Schemes 2-4 can be applied.
Only the position of the substituent shifts, increasing the
possible products from four to eight.
Exp er im en ta l Resu lts
P r ep a r a tion a n d Tr a p p in g Exp er im en ts w ith
Alkyl 1,2-Cycloh exadien ecar boxylates. Methyl 2-chlo-
rocyclohex-2-enecarboxylate (1a -Me) and its conjugated
isomer methyl 2-chlorocyclohex-1-enecarboxylate (1b-
Me), as well as the ethyl esters 1a -Et and 1b-Et, were
prepared as a 72:28 (77:23 for 1-Et) mixture of isomers
by chlorodeoxygenation of the corresponding alkyl cyclo-
hexanone-2-carboxylate with triphenylphosphine in car-
(10) Benson, S. W. Thermochemical Kinetics. Methods for Estimation
of Thermochemical Data and Rate Parameters, 2nd ed.; Wiley: New
York, 1976.
(11) Endo and exo refer to the position of the diene relative to the
former cyclohexadiene, rather than to the position of the substituent
on the ring.

978 J . Org. Chem., Vol. 64, No. 3, 1999
Nendel et al.
Sch em e 5. P r ep a r a tion a n d Tr a p p in g of
2-Alk ylcycloh exa -1,2-d ien e Ca r boxyla tea
Discu ssion
The reaction is regioselective (the formation of the
bridgehead-substituted product is preferred) and endo
stereoselective11 with alkyl 1,2-cyclohexadienecarboxy-
lates, 2. By contrast, there is a complete lack of R/ S
diastereoselectivity with the chiral dienophiles 2-Bn and
2-Mn . The selectivity for the endo product 3 as well as
for the conjugated product 4 is within expectations for a
concerted reaction, but the absence of diastereoselectivity
is surprising, given at least minimal diastereoselectivity
observed with simple chiral acrylates.14 An explanation
can be found by studying the transition structures
obtained by computations.
Com p u ta tion a l Meth od ology
GAUSSIAN94 was used for the density functional theory
calculations.15 The reported energies are corrected by unscaled
zero point energies. The B3LYP functional was used for
calculating the ground states and concerted transition struc-
tures, while the open-shell UB3LYP was used for structures
on the stepwise pathways. This functional provided structures
and energies in good agreement with experimental results in
the case of the parent reaction and other pericyclic reactions.8,9
The 6-31G* basis set was used in all calculations. We also
computed the spin projected energies of the open-shell struc-
tures using the procedure developed by Yamaguchi et al.16
However, the results did not change the relative order of the
pathways, and we choose to include these energies only in the
Supporting Information. MACROMODEL and the MM2* force
field were used for the QM/MM molecular modeling.17
a
Key: (a) PPh3, CCl4, reflux; (b) 1-hydroxy-3-isothiocyanatotet-
rabutyldistannoxane (0.1 equiv), R′′OH, toluene, reflux; (c) KO-t-
But, furan, THF.
bon tetrachloride (Scheme 5).12 The corresponding l-men-
thyl and l-bornyl cyclohexenecarboxylates 1a -Mn , 1b-
Mn , 1a -Bn , and 1b-Bn were prepared by catalytic
transesterification of 1-Et with l-menthol or l-borneol and
1-hydroxy-3-isothiocyanatotetrabutyldistannoxane.13 The
chlorocyclohexene carboxylates 1 were subjected to a
potassium tert-butoxide elimination in a 50:50 solution
of furan/THF. Of the four possible regio- and stereoiso-
meric trapping products from methyl 1,2-cyclohexadien-
ecarboxylate 2-Me, two were formed in 66% yield and in
a ratio of 3.4:1. The major regioisomer (3-Me) is substi-
tuted at the bridgehead position. The substituent of the
minor product (4-Me) was in conjugation with the
remaining double bond. The ethyl, l-menthyl, and l-
bornyl esters produced the analogous products 3-Et, 4-Et,
3-Mn , 4-Mn , 3-Bn , and 4-Bn in a ratio of 2:1, 2.1:1, and
2.9:1, respectively. In each case, the isomers were sepa-
rated by preparative thick-layer silica gel chromatogra-
phy.
Resu lts a n d Discu ssion of th e Com p u ta tion s
Th e Diels-Ald er Rea ction betw een P r op a d ien e
a n d 1,3-Bu ta d ien e. Structures along the concerted and
both stepwise pathways were located for the reaction of
propadiene and 1,3-butadiene using (U)B3LYP/6-31G*
computations.
The concerted transition structure is asynchronous
(Figure 1) and has a computed activation energy of 29.1
kcal/mol, about 5 kcal/mol higher than in the case of 1,3-
butadiene and ethylene.8 This result is surprising insofar
as the reaction of 1,3-butadiene and the parent allene is
exothermic by 44.3 kcal/mol, 8 kcal/mol more so than the
1H NMR experiments and comparison to the Diels-
Alder reaction product of 1-phenylcyclohexa-1,2-diene
indicated that both regioisomers, 3 and 4, were the endo
adducts with the cyclohexyl endo relative to furan and
the bridgehead substituent exo relative to the cyclohexyl
(Scheme 2).5a The fact that only the endo adducts were
observed was also predicted by computations, which
indicate that the endo pathway is favored. Thus, the
reaction is stereoselective.
(14) E.g., the reaction of cyclopentadiene with (-)-menthyl acry-
late: (a) Sauer, J .; Kredel, J . Tetrahedron Lett. 1966, 6359. (b) Sauer,
J .; Kredel, J . Angew. Chem. 1965, 77, 1037; Angew. Chem., Int. Ed.
Engl. 1965, 4, 989. (c) Farmer, R. F.; Hamer, J . J . Org. Chem. 1966,
31, 2418. (d) Sauer, J .; Kredel, J . Tetrahedron Lett. 1966, 731. Or the
reaction of butadiene with di-(-)-menthyl fumerate: (e) Korolev, A.;
Mur, V. Dokl. Akad. Nauk USSR 1948, 59, 251. (f) Korolev, A.; Mur,
V. Chem. Abstr. 1948, 42, 6776. (g) Walborsky, H. M.; Barash, L.;
Davis, T. C. Tetrahedron 1963, 19, 2333. (h) Walborsky, H. M.; Barash,
L.; Davis, T. C. J . Org. Chem. 1961, 26, 4778.
(15) Gaussian 94 (Revision B.1): Frisch, M. J .; Trucks, G. W.;
Schlegel, H. B.; Gill, P. M. W.; J ohnson, B. G.; Robb, M. A.; Cheeseman,
J . R.; Keith, T. A.; Peterson, G. A.; Montgomery, J . A.; Raghavachari,
K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.;
Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.;
Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J . L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.;
Defrees, D. J .; Baker, J .; Stewart, J . P.; Head-Gordon, M.; Gonzalez,
C.; Pople, J . A. Gaussian, Inc., Pittsburgh, PA, 1995.
Use of the chiral shift reagent tris[3-heptafluoropro-
pylhydroxymethylene-d-camphorato]europium(III) al-
lowed separation of the δ 6.4 olefinic absorption into two
absorptions. In the case of the ethyl esters, 3-Et, as well
as the l-bornyl and l-menthyl esters 3-Bn and 3-Mn , the
integrated areas for the two absorptions were 1.0:1.0
within experimental error.
(16) (a) Yamaguchi, K.; J ensen, F.; Dorigo, A.; Houk, K. N. Chem.
Phys. Lett. 1988, 149, 537. (b) Yamakami, S.; Kawakami, T.; Nagao,
H.; Yamaguchi, K. Chem. Phys. Lett. 1994, 231, 25.
(17) (a) MM2* is a modified version of Allinger’s MM2 force field:
Allinger, N. L. J . Am. Chem. Soc. 1977, 99, 8127. (b) MACROMODEL
V5: Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J . Comput. Chem.
1990, 11, 440.
(12) (a) Gruber, L.; To¨mo¨sko¨zi, I.; Radics, L. Synthesis 1975, 708.
(b) Isaacs, N. S.; Kirkpatrick, D. J . Chem. Soc., Chem. Commun. 1972,
443.
(13) Otera, J .; Ioka, S.; Nozaki, H. J . Org. Chem. 1989, 54, 4013.

Allenes as Dienophiles in the Diels-Alder Reaction
J . Org. Chem., Vol. 64, No. 3, 1999 979
F igu r e 1. Two views of the transition structure of the
concerted pathway of the reaction between butadiene and
propadiene (selected bond lengths in Å).
F igu r e 3. Two views of the diradical intermediate of the
stepwise path 2 of the reaction between butadiene and
propadiene (selected bond lengths in Å).
Sch em e 6. Sch em a tic En er gy Dia gr a m of th e
Rea ction betw een s-tr a n s-1,3-Bu ta d ien e a n d
P r op a d ien e
F igu r e 2. Structures along the stepwise path 1 of the reaction
between butadiene and propadiene (selected bond lengths in
Å).
butadiene-ethylene reaction.18 The forming bond be-
tween butadiene and C-2 of allene is almost 0.1 Å shorter
than that between butadiene and C-1 (2.261 Å compared
to 2.357 Å). In comparison, the corresponding forming
bond lengths are 2.273 Å in the transition structure for
the reaction between ethylene and butadiene.8 Despite
the considerable asynchronicity of the transition struc-
ture, the intrinsic reaction coordinate (IRC) of the transi-
tion structure leads directly to reactants and products.
Two intermediates and one transition structure were
found on the stepwise path 1 of Scheme 1, in which the
first bond is formed to C-1 of allene (Figure 2). The 1,3-
butadiene moiety is anti to the allene moiety in one
intermediate, while it is gauche-in in the other.19 Both
have essentially the same relative energy: the anti
intermediate is 30.2 kcal/mol above the reactants, com-
pared to 30.1 kcal/mol in the case of the gauche-in
intermediate.20 The transition structure leads to the anti
intermediate. It is 33.5 kcal/mol above the reactants. This
barrier is 4.4 kcal/mol higher in energy than that of the
concerted path (Scheme 6). The forming bond length is
similar to that in the corresponding transition structure
of the ethylene-butadiene reaction (1.879 Å compared
to 1.884 Å).8 The transition structure for the ring closure
could not be located, as it is the case of the reaction
between butadiene and ethylene. This second step is
therefore expected to have no significant barrier.18b
Only one structure was found on the stepwise path 2.
It is the gauche-in diradical intermediate (Figure 3). This
intermediate is 7.9 kcal/mol above the reactants, 26.0
kcal/mol below the concerted transition structure, and
21.2 kcal/mol below the transition structure leading to
the other diradical intermediate.21 An extensive search
did not produce any transition structure on this path.
The region of the hypersurface immediately surrounding
the intermediate is flat, and the spin contamination
increases quickly with an increase in the forming bond
length. It will be necessary to use CASSCF calculations
in a future study to map accurately this region of the
potential energy surface.
Con cer ted ver su s Step w ise Mech a n ism . The rela-
tive energies are summarized in Scheme 6. The less
favored stepwise path is only 4.4 kcal/mol higher in
energy than the concerted pathway. Unfortunately, no
transition structures were located on the more favored
stepwise path, but the hypersurface surrounding the
intermediate was flat. If the energy difference between
transition structure and intermediate is similar to the
first stepwise mechanism, the barrier of this second
stepwise mechanism will still be about 11 kcal/mol lower
in energy than the concerted transition state. This
reaction therefore proceeds via a stepwise mechanism.
Th e Diels-Ald er Rea ction betw een 1,2-Cyclo-
h exa d ien e a n d 1,3-Bu ta d ien e. In this case, not only
one concerted and two stepwise pathways but also endo
and exo concerted transition structures have to be
considered. While the endo and exo concerted transition
(18) (a) The reaction of ethylene and butadiene is exothermic by 36.6
kcal/mol (ref 8). (b) The experimental barrier of the reaction of 1,3-
butadiene and ethylene is 25.1 ( 2 kcal/mol: Houk, K. N.; Li, Y.;
Evanseck, J . D. Angew. Chem. 1991, 104, 711; Angew. Chem., Int. Ed.
Engl. 1991, 31, 682. (c) The experimental heat of formation is 38.5
kcal/mol: Uchiyama, M.; Tomioka, T.; Amano, A. J . Phys. Chem. 1964,
68, 1878.
(19) Since the propadiene moiety is at the “inside” of the former
butadiene, it is referred to as “gauche-in”.
(21) Geometry optimizations starting with either an anti or
gauche-out conformation about the newly formed bond provide the
structure of Figure 3.
a
(20) The structures on the stepwise pathways are diradicals with
S2 ≈ 1.

980 J . Org. Chem., Vol. 64, No. 3, 1999
Nendel et al.
F igu r e 6. Diradical intermediate and transition structure of
stepwise path 1 of the reaction between s-trans-1,3-butadiene
and 1,2-cyclohexadiene (selected bond lengths in Å).
Sch em e 7. Sch em a tic En er gy Dia gr a m of th e
Rea ction betw een 1,2-Cycloh exa d ien e a n d
s-tr a n s-1,3-Bu ta d ien e
F igu r e 4. Concerted transition structures for the reaction of
1,2-cyclohexadiene and 1,3-butadiene (selected bond lengths
in Å).
F igu r e 5. Diradical intermediate of stepwise path 2 of the
reaction between butadiene and 1,2-cyclohexadiene (selected
bond lengths in Å).
structures are geometrically distinct, the corresponding
products cannot be distinguished.
Both concerted transition structures are extremely
asynchronous (Figure 4). In the endo case, the forming
bond between C-2 and C-7 is 2.046 Å, about 0.5 Å shorter
than that between C-1 and C-10. In the exo case, the
difference is even larger. C-2 and C-7 are 2.134 Å apart,
almost 0.9 Å closer than C-1 and C-10. In comparison,
the corresponding value is 0.1 Å in the reaction of
propadiene and 1,3-butadiene. However, the transition
vectors (motion along the directions of negative curva-
ture) of the transition states indicate that both structures
belong to the concerted pathways. Exo and endo transi-
tion structures are close in energy. The exo path is
favored by 0.2 kcal/mol over the endo. It is 7.2 kcal/mol
above the separate reactants, and the reaction is exo-
thermic by 75.5 kcal/mol.
The extremely low activation energy and the high heat
of reaction are due to the strain of 1,2-cyclohexadiene.
The geometry of 1,2-cyclohexadiene undergoes only small
changes to reach the transition structure and the product,
so that bond formation to C-2 is easy. At the same time,
the bond formation releases the inherent strain of the
allene.
The relative ease of formation of the bond to C-2 is even
more pronounced in the formation of the diradical
intermediate of the stepwise path 2. This intermediate
has two allyl radical moieties, the bond has been formed
to C-2. The gauche-in conformer is 27.6 kcal/mol more
stable than the reactants (Figure 5). The cyclohexadiene
ring is in an envelope conformation, the atoms of the allyl
radical site and those connected to it are located in the
same plane. The R/ S stereochemistry of the bridgehead
carbon in the product is determined in the second step
of path 2. This is also the case for the endo/exo stereo-
chemistry, since the former butadiene part is perpen-
dicular to the six-membered ring. The influence of the
ring conformation on the relative energy is negligible. A
structure in which the ring was flipped “out” relative to
the butadiene moiety, rather than being flipped “in” as
in Figure 6, is 0.1 kcal/mol higher in energy. The ring
itself is expected to flip easily between the two conforma-
tions.22
In the reaction between allene and butadiene, the
transition structure of the stepwise path 1 was close in
energy to the intermediate; however, this is not the case
here. The anti intermediate of stepwise path 1, in which
a vinyl radical is formed, is only 1.5 kcal/mol above the
reactants (Figure 6),23 but the transition structure lead-
ing to this intermediate is located 23.8 kcal/mol above
the separate reactants (Scheme 7). The transition struc-
ture is early. The forming bond length is 2.117 Å, 0.242
Å longer than in the reaction between ethylene and 1,3-
butadiene.8 Therefore, not much strain of the allene is
released at this point. The approach also forces the proton
on C-1 out of planarity and crowds the transition
structure relative to the concerted transition state where
the forming bond to C-1 is significantly longer (2.66 and
2.92 Å).
(22) In the case of cyclopentene, a typical example for a molecule
with an envelope conformation, the ring flip requires only 0.8 kcal/
mol: Cavagnat, D.; Roberts, M. P.; Cavagnat, R. M.; Vahedi-Banisaeid,
S. J . Phys. Chem. 1991, 95, 134.
(23) This is the lowest energy conformer. The other conformers are
located within 1.7 kcal/mol above it.

Allenes as Dienophiles in the Diels-Alder Reaction
J . Org. Chem., Vol. 64, No. 3, 1999 981
F igu r e 8. Intermediate and transition structure on the
stepwise path 1 of the reaction between furan and 1,2-
cyclohexadiene (selected bond lengths in Å).
F igu r e 7. Concerted endo and exo transition structures of
the reaction of 1,2-cyclohexadiene and furan (selected bond
lengths in Å).
The energies are summarized in Scheme 7. Although
no transition structure was located on the second, lower
energy, stepwise path, it can be assumed that the
transition structures of this path will be lower in energy
than the concerted transition states. The energy differ-
ence between the intermediate and the transition struc-
ture in the nonfavored stepwise case is 22.3 kcal/mol.
Even if the difference is in the same order of magnitude
in the most favored case, the energy of the transition
structure would be 5.3 kcal/mol below the reactants. The
transition structure is thus expected to be very early and
to have a similar energy to the reactants.
F igu r e 9. Intermediate of the more favored stepwise path
(path 2) of the reaction between furan and 1,2-cyclohexadiene
(selected bond lengths in Å).
Sch em e 8. Sch em a tic En er gy Dia gr a m of th e
Rea ction of 1,2-Cycloh exa d ien e a n d F u r a n
Th e Diels-Ald er Rea ction betw een 1,2-Cyclo-
h exa d ien e a n d F u r a n . In contrast to the reaction of
1,2-cyclohexadiene and 1,3-butadiene, the endo and exo
products can be distinguished in the reaction of 1,2-
cyclohexadiene and furan. The formation of the endo
adduct is exothermic by 38.9 kcal/mol, and its formation
is favored by 0.8 kcal/mol over that of the exo product.
The concerted transition structures are extremely asyn-
chronous (Figure 7). In the endo structure, the two
forming bonds differ by almost 0.8 Å. In the exo case,
the difference is slightly more than 1.0 Å, the largest
difference of the systems we studied in this account. The
barrier of the concerted endo pathway is 7.0 kcal/mol,
0.9 kcal/mol lower than the concerted exo path. The endo
product is thus kinetically favored, if the reaction pro-
ceeds via a concerted mechanism.
In analogy to the reaction between 1,2-cyclohexadiene
and 1,3-butadiene, an anti transition structure and
intermediate were located on the stepwise path 1. The
intermediate is 12.8 kcal/mol above the reactants, 5.8
kcal/mol above the concerted endo transition state (Fig-
ure 8, Scheme 8). The activation barrier for its formation
is 28.2 kcal/mol. The energy difference between the
transition structure and the intermediate is again large
compared to the reaction between allene and butadiene.
The transition structure in Figure 8 leads to the diaste-
reomer of the intermediate, but because this path re-
quires 21.2 kcal/mol more energy than the concerted endo
path, no attempt was undertaken to explore the confor-
mational space of the intermediate or transition struc-
ture.
formation of this intermediate is considered barrierless
for practical purposes.
The energy gap between the stepwise path 2 and the
concerted reaction is not as large as in the case of the
reaction between 1,2-cyclohexadiene and 1,3-butadiene,
but it is still large enough to make the concerted paths
improbable (Scheme 8). Two regioisomers are observed
experimentally, while there is no chiral induction. The
experimental results are consistent with two scenarios.
(1) The reaction is concerted, but the influence of the
chiral auxiliary is very small due to the long forming bond
to C-1. (2) Path 2 is taken, but the chiral auxiliary has
no influence on the formation of the chiral center opposite
to C-1. The formation of the endo product is controlled
by the second step. The energetics depicted in Scheme 8
clearly show that the second scenario is more likely.
The intermediate of the more favored stepwise path is
located 16.2 kcal/molbelow the reactants, 23.2 kcal/mol
below the concerted path (Figure 9). Again, no transition
structure could be located on this pathway, and the

982 J . Org. Chem., Vol. 64, No. 3, 1999
Nendel et al.
Another experimental observation is in favor of this
scenario: two regioisomers are formed with a maximal
selectivity of 3.9:1. If the reaction followed a concerted
pathway, electronic factors should govern the regiochem-
istry significantly and the product ratios should be more
weighted toward the product resulting from reaction at
the more electron-deficient bond.
No transition structure could be located for the ring
closure, and the endo/exo stereochemistry is determined
in the second step of the reaction. The exclusive formation
of the endo product can be explained by secondary orbital
interactions. In the formation of the exo product, the two-
carbon unit of furan will have approximately the same
distance to the cyclohexadiene ester substituent as the
oxygen atom in the transition state leading to the endo
product. The distance between the oxygen atom and the
bridgehead hydrogen atom of the endo product is 2.54
Å, and the distance to the two-carbon unit in the exo
product is 2.54 Å as well. The steric requirements are
expected to be similar in the formation of both products,
but stabilizing interactions between the empty orbitals
of the oxygen and the ester substituent on C-1 are
possible, when the endo product is formed. Therefore
secondary orbital interaction should lower the barrier to
the formation of the endo product in the second half of
the stepwise path 2.
Because no transition structures were located on the
stepwise path 2, we modeled the concerted pathway for
the l-bornyl reaction using a QM/MM approach. The
B3LYP/6-31G* concerted transition structures of 1,2-
cyclohexadiene and furan which lead to the major regio-
isomer were substituted with the bornyl ester, and the
ester substituent was optimized with MM2*. The inter-
actions of furan and the chiral substituent would be
similar to, if weaker than, those observed in the reactions
of 1,3-diphenylisobenzofuran and l-bornyl fumarates
described earlier.5a Figure 10 shows the endo transition
structures. The exo transition structures were considered
as well, and the relative energies are listed in Table 1.
Because the ester can be in either a E or a Z conformation
(relative to 1,2-cyclohexadiene), there are eight conforma-
tions to be considered for each regiochemical scenario.
As can be seen in Table 1, the R/ S stereochemical
induction by the chiral substituent is minimal. The four
transition structures are located within 0.4 kcal/mol of
each other in the endo case and within 0.7 kcal/mol in
the exo transition structures. The energy difference
between the lowest R and S structures is only 0.2 kcal/
mol, well within the expected error of this method. In
similar QM/MM studies, the induction by a chiral sub-
stituent is generally overestimated; i.e., the differences
in relative energies in Table 1 are even smaller, so that
they are no longer chemically significant.24 Therefore, the
QM/MM method predicts no stereoselectivity for the
l-bornyl cyclohexadiene carboxylate cycloadditions.
Our DFT results show that the reaction proceeds via
stepwise path 2. The stereochemistry is determined in
the first and the second step. Two diastereomeric diradi-
cal intermediates can be formed, and either one can only
lead to two of four products (Scheme 4). The endo
products are favored and formed exclusively in the
experiments (within the limits of detection); the stereo-
chemistry of the bridgehead carbon is determined in the
step forming the diradical intermediate. If a chiral
F igu r e 10. Endo transition structures of the reaction of
l-bornylcyclohexa-1,2-diene carboxylate and furan leading to
the nonconjugated product (the hydrogen atoms of the bornyl
ester are removed for clarity).
Ta ble 1. Rela tive MM2* En er gies of th e Differ en t
Con cer ted Tr a n sition Str u ctu r es of th e Rea ction
betw een F u r a n a n d
1-(-)-Bor n ylcycloh exa -1,2-d ien eca r boxyla te
CdC-CdO
dihedral (deg)
Erel (kcal/mol)
Z endo R
Z endo S
E endo R
E endo S
Z exo R
Z exo S
2
-4
171
179
-22
24
0.0
0.5
0.5
0.2
6.4
6.1
6.3
6.8
E exo R
E exo S
167
-171
auxiliary influences the formation of either diastereo-
meric intermediate, this influence will be reflected in the
product ratio. However, the influence on the formation
of the intermediate will be even less than on the
concerted transition structure. The bond to C-2 is formed
first and it is reasonable to assume that the distance of
the dienophile to C-1 will be even larger than in the
concerted case, thus allowing for less steric and electronic
interactions. Therefore, modeling the concerted path
provides an upper limit to the diastereomeric ratio. Since
no transition structure was found on this path, this
hypothesis could not be verified by molecular modeling.
Con clu sion s
The incorporation of a cumulated double bond into an
otherwise unreactive cyclohexene carboxylate provides
(24) Eksterowicz, J . E.; Houk, K. N. Chem. Rev. 1993, 93, 2439.

Allenes as Dienophiles in the Diels-Alder Reaction
J . Org. Chem., Vol. 64, No. 3, 1999 983
a mixture of 84% menthyl 2-chlorocyclohexenecarboxylate 1b-
Mn and 16% menthyl 2-chlorocyclohex-2-enecarboxylate 1a -
Mn was obtained. Anal. Calcd for C17H27O2Cl: C, 68.32; H,
a useful method for generating a reactive dienophile as
well as other similarly activated dienophiles. These are
open to trapping by Diels-Alder reactions with kineti-
cally controlled regiochemistry.
The reaction of propadiene with 1,3-butadiene is step-
wise. Further perturbation of the dienophile, such as its
integration into a six-membered ring as in the case for
1,2-cyclohexadiene, increases the energetic gap between
the stepwise and concerted pathways even further. No
detectable diastereoselectivity was observed in the l-
bornyl and l-menthyl cyclohexa-1,2-diene carboxylate
cycloadditions, consistent with the computational prefer-
ence of the stepwise mechanism
1
9.11; Cl, 11.86. Found: C, 68.32; H, 9.27; Cl, 11.90. H NMR
(250 MHz, CDCl3): δ 5.98-6.06 (16 H, m), 4.63-4.90 (l H, dt,
J ) 4.3, 10.9), 3.21-3.32 (0.16 H, m), 2.30-2.70 (3 H, m), 1.85-
2.25 (H, m), 1.60-1.85 (5.68 H, m), 1.30-1.60 (2 H, m), 0.80-
1.20 (9 H, m), 0.70-0.80 (3 H, d, J ) 7). FTIR (thin film): 1725,
1665, 1627, 1456, 1341, 1277, 1252, 846, 757, 667 cm-1. MS:
(EI); 283 °C, m/z (relative intensity): 302.1 (M + 2, 8.9), 301.1
(M + 1, 30.6), 300.2 (M, 19.8), 143.0 (64.3), 109.1 (8.4), 81.1
(25.1).
Bor n yl 2-ch lor ocycloh exen eca r boxyla te (1-Bn ) was
prepared similarly to 1-Mn (see the Supporting Information).
Cycloa d d ition of Allen es to F u r a n . A. Meth yl Cyclo-
h exa -1,2-d ien e-1-ca r boxyla te (2-Me). A flask was charged
with 0.69 g of a mixture of 66% methyl chlorocyclohexenecar-
boxylate 1 and 34% of methyl cyclohexanone-2-carboxylate in
an argon atmosphere. Furan (150 mL, Lancaster, dried over
MgSO4) and THF (75 mL, freshly distilled over sodium
benzophenone, Baker) were added. Potassium tert-butoxide
(0.60 g, 1.4 mol equiv) dissolved in THF (75 mL) was added to
the stirred mixture over 30 min. Stirring was continued for
24 h, and the reaction was quenched with water (3 × 50 mL).
The combined aqueous phases were saturated with sodium
chloride and extracted three times with diethyl ether (50 mL).
The combined organic layers were dried over MgSO4, filtered,
and concentrated to yield 0.94 g of a yellow oil. The product
mixture was separated by flash silica gel column chromatog-
raphy (9:1 hexane/ethyl acetate). The Diels-Alder adducts
were the last two fractions. The nonconjugated 3-Me was the
major product (0.25 g, 1.2 mmol, 44%), while the conjugated
4-Me was the minor one (0.07 g, 0.4 mmol, 13%). Major
product 3-Me. HRMS: calcd for C12H14O3 206.0943, found
206.0943. MS (EI+, 51 °C): 207.1 (M + 1), 206.1 (M), 174.1,
Exp er im en ta l Section
An a lyses. 1H NMR spectra were recorded at 250, 300, and
400 MHz using tetramethylsilane as internal reference; cou-
plings are expressed in hertz. Thin-layer chromatography
(TLC) was performed with Analtech precoated silica gel GF
(250-micron) plates and with basic alumina IB-F sheets (J . T.
Baker), both 2.5 cm × 7.5 cm. Preparative TLC was performed
with Analtech precoated silica gel GF (1000 µm, 20 cm × 20
cm) plates.
Meth yl 2-Ch lor ocycloh exen eca r boxyla te (1-Me). Triph-
enylphosphine (22.30 g, 84.9 mmol, dried under vacuum over
P2O5 for 8 h) was dissolved in CCl4 (40 mL, Aldrich, freshly
distilled from P2O5 onto molecular sieves), and methyl cyclo-
hexanone-2-carboxylate (4.50 mL, 5.00 g, 32.0 mmol, Fluka)
was added. The reaction was heated to reflux and kept under
argon. After 4 days, the mixture was added to hexane (250
mL). Stirring was maintained for an additional 12 h, and the
mixture was cooled in an ice bath and filtered, yielding 16.40
g of brown crystals that were washed by an additional hexane
(250 mL). The solvent was evaporated under reduced pressure,
yielding yellow-white crystalline triphenylphosphine oxide and
a brown oil (2.14 g). Distillation at 50 °C (0.75 Torr) gave a
pale yellow oil (1.35 g), containing the starting material (34%)
and methyl 2-chloro-2-cyclohexenecarboxylate (1a -Me) and
methyl 2-chloro-1-cyclohexenecarboxylate (1b-Me, 66% of 1a -
Me and 1b-Me, 8% purified yield) in a ratio of 72:28. HRMS:
1
147.1, 77.1, 65.0. H NMR (400 MHz, CDCl3): δ 6.489 (dd, 1
H, J ) 5.6, 1.7), 6.122 (dd, 1 H, J ) 1.6), 5.722 (t, 1 H, J )
3.5), 5.248 (broad s, 1 H), 5.159 (broad s, 1 H), 3.721 (s, 3 H),
2.225 (m, 1 H), 1.950 (m, 2 H, J ) 8.0), 1.710 (m, 1 H), 1.425
(m, 1 H, J ) 20.0, 16.0, 12.0, 4.0), 0.569 (m, 1 H, J ) 14.1,
12.1, 4.0). Minor product 4-Me. HRMS: calcd for C12H14O3
206.0943, found 206.0940. MS (EI+, 46 °C): 207.1 (M + 1),
206.1 (M), 174.1, 146.1, 77.1, 65.0. 1H NMR (CDCl3, TMS, 300
MHz): 6.415 (dd, 1 H, J ) 5.7, 2.0), 6.186 (dd, 1 H, J )
1.7),5.949 (broad s, 1 H), 5.037 (broad d, 1 H, J ) 4.5), 3.760
(s, 3 H), 2.420 (m, 1 H), 2.160 (m, 1H), 2.020 (m, 1 H), 1.900
(m, 1 H), 1.560 (m, 2 H), 0.380 (m, 1 H).
B. Eth yl 1,2-Cycloh exadien ecar boxylate (2-Et). C. Men -
th yl 1,2-Cycloh exa d ien eca r boxyla te (2-Mn ). D. Bor n yl
1,2-Cycloh exa d ien eca r boxyla te (2-Bn ). 2-Et, 2-Mn , and
2-Bn were treated similarly to 2-Me (see the Supporting
Information).
1
calcd for C8H11O2Cl 174.0448, found 174.0450. H NMR (300
MHz, CDCl3): δ 6.05 (t, 1 H), 3.78 (s, 3 H), 3.75 (s, 3 H), 3.30
(t, 1 H), 2.45 (m, 2 H), 2.38 (m, 2 H), 2.20 (m, 2 H, 2.00 (m, 2
H), 1.70 (m, 6 H).
Eth yl 2-ch lor ocycloh exen eca r boxyla te (1-Et) was pre-
pared similarly (see the Supporting Information).
l-Men th yl 2-Ch lor ocycloh exen eca r boxyla te (1-Mn ). A
flask equipped with a coldfinger condenser and trap was
charged with 2.00 g (10.6 mmol) of ethyl 2-chlorocyclohexene
carboxylate 1-Et, 1.66 g (10.6 mmol) of (1R,2S,5R)-(-)-menthol
(Aldrich), 0.59 g (1.06 mmol) of 1-hydroxy-3-isothiocyanatotet-
rabutyldistannoxane, and 8 mL of toluene (Fisher). The
solution was heated at reflux for 2.5 days. The condensate (5.5
mL) was found by infrared spectroscopy to contain 11%, or
0.61 mL (98% yield), of ethanol. The dark brown solution was
evaporated to give 3.90 g of residue. Silica gel column chro-
matography (29 cm × 2.2 cm, 32 g, 85:15 hexane/ethyl acetate
yielded 2.45 g of a high Rf fraction that was distilled to give
1.93 g (66%) of colorless liquid, bp 111-115 °C (0.18 Torr).
Further basic alumina column chromatography of a 1.63 g
sample (hexane, then 95:5 hexane/ethyl acetate) allowed
isolation of menthyl 2-chlorocyclohexenecarboxylate from ethyl
2-chlorocyclohexenecarboxylate. The yield of 1-Mn was 1.12
g (66%) of colorless liquid, bp 105-110 °C (0.085 Torr).
Integration of the 1H NMR signal at 5.98-6.06 ppm relative
to the bornyl methine proton at 4.63-4.90 ppm indicated that
Ack n ow led gm en t. We are grateful to the National
Science Foundation for financial support of this research
and to the San Diego Supercomputer Center and UC-
LA’s Office of Academic Computing for computational
support.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and Cartesian coordinates of the computed structures
including energies and zero point energies as well as S2
values and triplet single point energies of the UB3LYP
structures. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O982091C