Ketene recognizes 1,3-dienes in their s-Cis forms through [4 ± 2] (Diels-Alder) and [2 ± 2] (Staudinger) reactions. An innovation of ketene chemistry

Article abstract of DOI:10.1021/ja990072u

The mechanism of ketene-diene reactions has been studied both experimentally and theoretically. Careful experiments of the reactions of diphenylketene (1) with cyclic (s-cis) 1,3-dienes [cyclopentadiene (2) and cyclohexa-1,3-diene (3)] lead to the first direct detection of the Diels- Alder cycloadducts (10 and 11) by low-temperature NMR spectroscopy. The initially formed cycloadducts are converted to the final Staudinger products, cyclobutanones (6 and 7), by [3,3] sigmatropic (Claisen) rearrangements. In contrast, ketene 1 reacts with open-chain 1,3-dienes [2,3-dimethyl-1,3- butadiene (4) and 1-methoxy-1,3-butadiene (5)] to afford initially both the Staudinger-type (8, 9) and Diels-Alder-type cycloadducts (12, 13). The Staudinger cycloadducts (8, 9) are converted eventually to Diels-Alder products (12, 13) by the retro-Claisen rearrangement. Thus, ketene recognizes dienes in cycloadditions as ketenophiles different from olefins. [4 + 2] and [2 + 2] cycloadducts are generated and can be intermediates or products flexibly according to diene structures.

Full text of DOI:10.1021/ja990072u

J. Am. Chem. Soc. 1999, 121, 4771-4786
4771
Ketene Recognizes 1,3-Dienes in Their s-Cis Forms through [4 + 2]
(Diels-Alder) and [2 + 2] (Staudinger) Reactions. An Innovation of
Ketene Chemistry
Takahisa Machiguchi,*,† Toshio Hasegawa,† Akihiro Ishiwata,‡ Shiro Terashima,‡
Shinichi Yamabe,*,§ and Tsutomu Minato|
Contribution from the Department of Chemistry, Faculty of Science, Saitama UniVersity, Urawa,
Saitama 338-8570, Japan, Sagami Chemical Research Center, Nishi-Ohnuma, Sagamihara,
Kanagawa 229-0012, Japan, Department of Chemistry, Nara UniVersity of Education,
Takabatake-cho, Nara 630-8528, Japan, and Institute for Natural Science, Nara UniVersity,
Misasagi-cho, Nara 631-8502, Japan
ReceiVed January 5, 1999
Abstract: The mechanism of ketene-diene reactions has been studied both experimentally and theoretically.
Careful experiments of the reactions of diphenylketene (1) with cyclic (s-cis) 1,3-dienes [cyclopentadiene (2)
and cyclohexa-1,3-diene (3)] lead to the first direct detection of the Diels-Alder cycloadducts (10 and 11) by
low-temperature NMR spectroscopy. The initially formed cycloadducts are converted to the final Staudinger
products, cyclobutanones (6 and 7), by [3,3] sigmatropic (Claisen) rearrangements. In contrast, ketene 1 reacts
with open-chain 1,3-dienes [2,3-dimethyl-1,3-butadiene (4) and 1-methoxy-1,3-butadiene (5)] to afford initially
both the Staudinger-type (8, 9) and Diels-Alder-type cycloadducts (12, 13). The Staudinger cycloadducts (8,
9) are converted eventually to Diels-Alder products (12, 13) by the retro-Claisen rearrangement. Thus, ketene
recognizes dienes in cycloadditions as ketenophiles different from olefins. [4 + 2] and [2 + 2] cycloadducts
are generated and can be intermediates or products flexibly according to diene structures.
I. Introduction
showed that ketenes react with olefins across the CdC bond
rather than the CdO one. Ketenes are potentially useful
In the first decades of this century, Staudinger discovered
ketene2-6 and its reactions7 in the cycloaddition of diphenyl-
ketene2 (1) and cyclopentadiene (2).8-11 The pioneering studies
on ketene [2 + 2] cycloadditions (Staudinger reactions) have
been confirmed by a huge number of experiments.12-14 They
(8) The cycloaddition of ketenes was first discovered by Staudinger in
1907 from the reaction of diphenylketene and cyclopentadiene (refs 7a and
7b). He noted that there were two possibilities for the regiochemistry of a
[2 + 2] product (ref 7c). The [2 + 2] nature of the product was proved
both by its chemical reaction with NaOH or MeMgI (ref 7d) and by reducing
the CdC bond (refs 7c and 7d) and heating the product to give cyclopentene
and 1 (ref 7c). Farmer et al. confirmed the structures for cyclobutanones
(see, refs 9-11). The Staudinger reaction was found earlier than the Diels-
Alder reaction (Diels, O.; Alder, K. Ann. Chem. 1928, 460, 98-122).
(9) (a) Farmer, E. H.; Farooq, M. O. Chem., Ind. (London) 1937, 1079-
1080. (b) Farmer, E. H.; Farooq, M. O. J. Chem. Soc. 1938, 1295-1930.
(10) (a) Lewis, J. R.; Ramage, G. R.; Simonsen, J. L.; Wainwright, W.
G. J. Chem. Soc. 1937, 1837-1841. (b) Smith, L. I.; Agre, C. L.; Leekley,
R. M.; Prichard, W. W. J. Am. Chem. Soc. 1939, 61, 7-11.
* To whom correspondence should be addressed. For experiments T.M.
(E-mail: machiguc@chem.saitama-u.ac.jp. Fax: +81 48 858 3618). For
computations S.Y. (E-mail: yamabes@mailsrv.nara-edu.ac.jp. Fax: +81
742 27 9291).
† Saitama University.
‡ Sagami Chemical Research Center.
§ Nara University of Education.
| Nara University.
(1) Huisgen, R. The Adventure Playground of Mechanisms and Novel
Reactions. In Profiles, Pathways, and Dreams: Autobiographies of Eminent
Chemists; Seeman, J. I., Ed.; American Chemical Society: Washington,
DC, 1994.
(2) Staudinger first isolated a ketene, diphenylketene (1), generated from
Ph2CCl-COCl with granulated zinc. Staudinger, H. Chem. Ber. 1905, 38,
1735-1739.
(3) (a) Staudinger, H. Die Ketene; Verlag Enke: Stuttgart, 1912. (b)
Staudinger, H. From Organic Chemistry to Macromolecules; Wiley: New
York, 1970.
(4) Historically, Wedekind was the first chemist who obtained ketenes
from acid chlorides, R1R2CHCOCl (R1 ) R2 ) Me; R1 ) Me, R2 ) Br; R1
) H, R2 ) Ph), with pyridine, Et3N, or Pr3N. He quenched ketenes (R1R2Cd
CdO) with water to give acid anhydrides, (R1R2CHCO)2O. However, he
did not isolate those ketenes. Wedekind, E. Chem. Ber. 1901, 34, 2070-
2077.
(5) For the first synthesis of parent ketene, see: Wilsmore, N. T. M. J.
Chem. Soc. 1907, 91, 1938-1941. See also ref 6.
(6) Staudinger, H.; Klever, H. W. Chem. Ber. 1908, 41, 594-600.
Staudinger, H.; Klever, H. W. Chem. Ber. 1908, 41, 1516-1517.
(7) (a) Staudinger, H. Liebigs Ann. Chem. 1907, 356, 51-123. (b)
Staudinger, H. Chem. Ber. 1907, 40, 1145-1148. (c) Staudinger, H.; Suter,
E. Chem. Ber. 1920, 53B, 1092-1105. (d) Staudinger, H.; Rheiner, A. HelV.
Chim. Acta 1924, 7, 8-18.
(11) (a) Dawson, T. L.; Ramage, G. R. J. Chem. Soc. 1950, 3523-3525.
(b) Dryden, H. L., Jr. J. Am. Chem. Soc. 1954, 76, 2841. (c) Dryden, H. L.,
Jr.; Burgert, B. E. J. Am. Chem. Soc. 1955, 77, 5633-5637. (d) Brady, W.
T.; O’Neal, H. R. J. Org. Chem. 1967, 32, 2704-2707.
(12) For reviews, see: (a) Hanford, W. E.; Sauer, J. C. Org. React. 1946,
3, 108-140. (b) Roberts, J. D.; Sharts, C. M. Org. React. 1962, 12, 1-56.
(c) Ulrich, H. Cycloaddition Reactions of Heterocumulenes; Academic: New
York, 1967. (d) Bormann, D. Methoden der Organischen Chemie: Theime
Verlag: Stuttgart, 1968; Vol. 7, Part 4. (e) Ghosez, L.; O’Donnell, M. J. In
Pericyclic Reactions; Marchand, A. P., Lehr, R. E., Eds.; Academic:
Orlando, FL, 1977; Vol. 2, pp 79-140. (f) Huisgen, R. Pure Appl. Chem.
1980, 52, 2283-2302. (g) Brady, W. T. Tetrahedron 1981, 37, 2949-
2966. (h) Clemens, R. J. Chem. ReV. 1986, 86, 241-318. (i) Boger, D. L.;
Weinreb, S. M. Hetero Diels-Alder Methodology in Organic Synthesis:
Academic: San Diego, CA, 1987. (j) Ghosez, L.; Marchand-Brynaert, J.
In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.; Pergamon: Oxford,
U.K., 1991; Vol. 5, pp 85-122. (k) Schaumann, E.; Scheiblich, S. Methoden
der Organischen Chemie: Theime Verlag: Stuttgart, 1993; Vol. E15, Part
3, Chapters 4, 6, and 8.
(13) For a recent, excellent, and comprehensive review on ketenes and
their reactions, see: Tidwell, T. T. Ketenes; Wiley: New York, 1995. See
also: Tidwell, T. T. Acc. Chem. Res. 1990, 23, 273-279.
(14) For a thorough review on ketene cycloaddition reactions, see: Hyatt,
J. A.; Raynolds, P. W. Org. React. 1994, 45, 159-646.
10.1021/ja990072u CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/11/1999

4772 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Scheme 1. The Anomaly of Ketene Cycloaddition
Machiguchi et al.
Scheme 2. The Traditional Scheme for the Cycloaddition
Reaction of Ketenes toward Conjugated 1,3-Dienes To Give
a [2 + 2] Cycloadduct (B)a-d
dienophiles but they do not appear to owe their reactivity toward
dienes to Diels-Alder reactions. They react preferentially with
1,3-dienes to give four-membered rings, cyclobutanones,13,14
rather than six-membered ones, cyclohexanones.15 The unusual
capability and the periselectivity of ketenes in 1,2-cycloadditions
have been widely employed in organic syntheses13,16 including
those of natural products.12,13 The most representative example
is found in a key initial stage of the total synthesis of the natural
primary prostaglandins.17
Scheme 1 shows that allenes18 (>CdCdC<) and heterocu-
mulenes12 [such as ketenimines19 (>CdCdN-) and keten-
iminium ions20 (>CdCdN+<)] participate in the Diels-Alder
([4 + 2] cycloadditions) reactions with conjugated dienes such
as cyclopentadiene. Exceptionally, most ketenes (>CdCdO)
do not appear to react by [4 + 2] cycloadditions.
a Upper: Two-step mechanism via a zwitterionic intermediate (A).
Lower: Concerted [π2s + π2a] pathway via a transition state C. bTS(C)
shows the charge-transfer model from HOMO of 2 to (lu + 1)mo of
c
ketene. L and S stand for large and small substituents, respectively.
d
See ref 27. For experimental results, see refs 9-11. For mechanistic
investigations, see refs 20 and 21.
diphenylketene and 5,5-dimethylcyclopenta-1,3-diene was
reported.23a The result indicated that the reaction takes a
nonconcerted two-step mechanism.23b Other evidence against
the one-step pathway has been provided by both experimental23-25
and theoretical studies.26
Open-chain (acyclic) 1,3-dienes undergo 1,2-cycloadditions
similarly to cyclic 1,3-dienes with ketenes.12-14 Huisgen28 et
al. reported the first kinetic measurement that diphenylketene
(1) reacts with cyclopentadiene (2) 3.05 × 104 times faster than
cyclopentene at 40 °C.15b The rate preference of 2 to cyclo-
pentene has been explained by the secondary orbital interaction
in TS(C) of Scheme 2.12e,15b But, Staudinger reactions of open-
chain 1,3-dienes were found to be considerably slower than those
of cyclic 1,3-dienes.15 Introduction of an electron-donating
substituent, methoxy group, at the 1-position of open-chain 1,3-
butadiene increases the reactivity of the diene.29 It is curious
that the ethylene bond is more reactive than the vinyl ether
double bond of the diene in this reaction.
It was difficult to rationalize the feature of ketene cycload-
ditions (Staudinger reaction), until the proposal of the Wood-
ward-Hoffmann rule.21 The rule explained the easy and
exclusive formation of a four-membered ring by a nonphoto-
chemical process even with 1,3-dienes. As a conventional result,
it has been thought that ketenes do not add as dienophiles to
the 1,4-positions of 1,3-dienes but rather yield cyclobutanones
exclusively. Scheme 2 illustrates the traditional interpretation
of the Staudinger reaction. Ketenes undergo cycloadditions
toward 1,3-dienes through either (i) a nonconcerted two-step
pathway via a zwitterionic intermediate A to give a [2 + 2]
(22) Thermal [2π + 2π] cycloadditions are allowed generally as [π2s +
π2a] processes according to the Woodward-Hoffmann rule. But, the paths
must involve the orthogonal approach of the reacting species. Such reactions
are precluded for normal olefins due to the strain of the transition state.
However, ketenes were regarded as especially good antarafacial components
in [π2s + π2a] cycloadditions. The next lowest unoccupied molecular orbital,
(lu + 1)mo, of ketenes could overlap efficiently with the highest occupied
molecular orbital (HOMO) of olefins orthogonally in Scheme 2.
(23) (a) Holder, R. W.; Graf, N. A.; Duesler, E.; Moss, J. C. J. Am.
Chem. Soc. 1983, 105, 2929-2931. (b) gem-Dimethyl substitution does
not appreciably alter the course of the reaction relative to that of
cyclopentadiene (2).
(24) Hegedus, L. S.; Montgomery, J.; Narukawa, Y.; Snustad, D. C. J.
Am. Chem. Soc. 1991, 113, 5784-5791 and references therein.
(25) (a) Gompper, R. Angew. Chem., Int. Ed. Engl. 1969, 8, 312-327
(b) Wagner, H. U.; Gompper, R. Tetrahedron Lett. 1970, 2819-2822. (c)
Al-Husaini, A. H.; Moore, H. W. J. Org. Chem. 1985, 50, 2595-2597.
(26) (a) Burke, L. A. J. Org. Chem. 1985, 50, 3149-3155. (b) Bernardi,
F.; Bottoni, A.; Olivucci, M.; Robb, M. A.; Schlegel, H. B.; Tonachini, G.
J. Am. Chem. Soc. 1988, 110, 5993-5995. (c) Valent´ı, E.; Perica`s, M. A.;
Moyano, A. J. Org. Chem. 1990, 55, 3582-3593.
cycloadduct B or (ii) a concerted [π2s + 2a] cycloaddition22
π
via a transition state (TS) C. An excellent study of secondary
deuterium kinetic isotope effect on the reaction between
(15) (a) Huisgen, R.; Otto, P. Tetrahedron Lett. 1968, 43, 4491-4495.
(b) Huisgen, R.; Otto, P. Chem. Ber. 1969, 102, 3475-3485.
(16) Combining C-C π-Bonds; Paquette, L. A., Ed.; Comprehensive
Organic Synthesis, Pergamon: Oxford, U.K., 1991; Vol. 5.
(17) (a) Corey, E. J.; Schaaf, T. K.; Huber, W.; Koelliker, U.; Weinsh-
enker, N. M. J. Am. Chem. Soc. 1970, 92, 397-398. (b) Corey, E. J.; Noyori,
R.; Schaaf, T. K. J. Am. Chem. Soc. 1970, 92, 2586-2587. (c) Corey, E.
J.; Shirahama, H.; Yamamoto, H.; Terashima, S.; Venkateswarlu, A.; Schaaf,
T. K. J. Am. Chem. Soc. 1971, 93, 1490-1491.
(18) The Chemistry of Ketenes, Allenes and Related Compounds; Patai,
S., Ed.; The Chemistry of Functional Groups; Wiley: Chichester, U.K.,
1980; Parts 1 and 2.
(19) Gosez, L.; de Perez, C. Angew. Chem., Int. Ed. Engl. 1971, 10,
184-185.
(20) Marchand-Brynaert, J.; Ghosez, L. Tetrahedron Lett. 1974, 377-
(27) Cycloadditions of ketenes with a large and a small substituents (L
and S) to 2 give B regio- and stereoselectively: The predominance of the
adduct B with the larger substituent in the endo position was regarded as
evidence for the [π2s + π2a] pathway in Scheme 2.
380.
(21) (a) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl.
1969, 8, 781-853. (b) Woodward, R. B.; Hoffmann, R. The ConserVation
of Orbital Symmetry; Verlag Chemie: Weinheim, Germany, 1970.
(28) For an autobiographic account, see ref 1.

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4773
Scheme 3. The Exceptional Examples, 1,4-Cycloadditions, in
Ketene-Diene Reactions
Scheme 4. Four Cycloadditions Dealt with in This Worka-c
a Equations 1-3 are the known reactions, whereas eq 4 is an
b
unknown reaction treated here for the first time. For reaction 1, see
refs 7, 9, 10, 11e, and 15. For reaction 2, see refs 9 and 39d. For reaction
3, see refs 9a, 15, and 41. cReaction 4 is predicted to give cyclobutanone
9 judging from major examples of the conventional results from the
reactions of a ketene with butadiene (see ref 42).
We have become aware that the above exceptions in Scheme
3 may reflect the general feature of the ketene reactions with
1,3-dienes. As a naive idea, the carbonyl group in the ketene
should be a good dienophile (with the low-lying π*CdO energy
level) for the normal-electron demand in Diels-Alder reactions.
Thus, the conVentional and traditional results (two-step mech-
anism Via zwitterionic intermediate in i of Scheme 2) of the
reaction between diphenylketene and cyclopentadiene has been
reVised to a two-step mixed reaction mechanism of a combina-
tion of a [4 + 2] addition and the subsequent [3,3] sigmatropic
rearrangement in our preliminary report.40 In the first step, the
frontier molecular orbital (FMO) theory has supported the
Diels-Alder cycloaddition over the [2 + 2] addition. Second,
ab initio calculations have confirmed the FMO prediction and
have revealed that the [4 + 2] cycloadduct should be converted
to the [2 + 2] adduct. The latter adduct is thermodynamically
more stable than the former. Third, through low-temperature
NMR monitoring, the R-methylenepyran-type [4 + 2] adduct
has been detected successfully. Thus, a new stepwise mechanism
([4 + 2] cycloaddition f Claisen shift) has been proposed in
the reaction between diphenylketene (1) and cyclopentadiene
(2).40
But, there remain several problems in the new mechanism.
First, cyclopentadiene has a small distance (2.363 Å) between
two terminal (1 and 4) carbon atoms which is conveniently fit
for the Diels-Alder reaction. Open-chain dienes have larger
distances [e.g., 3.042 Å in 2,3-dimethyl-1,3-butadiene (4)] which
are less fit for the reaction. In these acyclic dienes, [4 + 2] and
[2 + 2] cycloadditions might be competitive. Second, one might
criticize that our new stepwise route is minor and that the direct
cyclobutanone formation is still major. Third, a parent ketene
has been adopted in ab initio calculations, while low-temperature
experiments have been carried out by the use of diphenylketene.
To adjust the theoretical model to experimental reactants,
diphenylketene must be employed in ab initio calculations.
Phenyl groups might impose steric hindrance on reacting
systems. To solve those problems and to get a decisive and
general picture of the reaction mechanism, we would like to
describe full accounts of ketene-diene reactions here. Four
reactions (eq 1, 2, 3, and 4) in Scheme 4 will be examined both
experimentally and theoretically. Historically ambiguous selec-
tivities of ketene-diene reactions will be elucidated.
a The first sign of a 1,4-cycloaddition of a ketene (refs 29 and 30).
c
bThe first detection of a 1,4-cycloadduct (ref 32b). Results of 1,4-
addition of diphenylketene (1) to butadienes. For the formation of a
mixture of [4 + 2] and [2 + 2] products, see ref 34 (trans-1-ethoxy-
or trans-1-thioethylbutadiene). Exclusive formations of [4 + 2] adducts,
see ref 34 (2-ethoxybutadiene, trans-1-ethoxy-2,3-dimethylbutadiene,
etc.) and ref 33 (trans-1-tert-butoxybutadiene).
It is a primitive question why ketenes do not undergo [4 +
2] cycloadditions in most cases. There are various examples of
those reactions across the CdO bond of ketene.13,14 However,
those Diels-Alder-type reactions ([4 + 2] in Scheme 1) are
rare and have been regarded as exceptions. Scheme 3 illustrates
the exceptions, i.e., the reactions of ketenes with conjugated
dienes giving 1,4-cycloadducts or via [4 + 2] cycloadditions.
The first symptom of the 1,4-addition of a ketene toward a diene
was reported in 1965 (Scheme 3a).29,30 An electron-deficient
bis(trifluoromethyl)ketene31 reacts with butadiene to form a 1,4-
cycloadduct across the CdO bond of ketene (Scheme 3b).32
When an electron-donating (alkoxy or trimethylsiloxy) group
is introduced to 1,3-butadiene, 1,4-cycloadducts33 are obtained
exceptionally (Scheme 3c) in the mixture of [4 + 2] and [2 +
2] cycloadducts.34 There also have been other examples of
exceptional Diels-Alder reactions.35-39
(29) (a) Martin, J. C.; Gott, P. G.; Goodlett, V. W.; Hasek, R. H. J. Org.
Chem. 1965, 30, 4175-4180. (b) Hasek, R. H.; Gott, P. G.; Martin, J. C.
J. Org. Chem. 1964, 29, 2510-2513.
(30) In Scheme 3a, the reaction of diphenylketene (1) with 2-methox-
ybutadiene leads to the isolation of an isomer of the real intermediate. The
product is thought to be formed via a [4 + 2] cycloadduct and its
isomerization, hydrogen 1,3-shift.
(31) Similarly, a 1,4-cycloadduct between bis(trifluoromethyl)ketene and
1,3-butadiene isomerizes easily to give a pyran (Scheme 3b). England, D.
C.; Krespan, C. G. J. Am. Chem. Soc. 1965, 87, 4019-4020. England, D.
C.; Krespan, C. G. J. Am. Chem. Soc. 1966, 88, 5582-5587.
(32) (a) Cheburkov, Y. A.; Mukhamadaliev, N.; Knunyants, I. L.
Tetrahedron 1968, 24, 1341-1356. (b) England, D. C.; Krespan, C. G. J.
Org. Chem. 1970, 35, 3300-3307.
(33) Clennan, E. L.; Lewis, K. K. J. Am. Chem. Soc. 1987, 109, 2475-
2478.
(34) (a) Gouesnard, J. P. C. R. Acad. Sci., Ser. C 1973, 277, 883-886.
(b) Gouesnard, J. P. Tetrahedron 1974, 30, 3113-3117.
(35) (a) Brady, W. T.; Agho, M. O. Synthesis 1982, 500-502. (b) Brady,
W. T.; Agho, M. O. J. Heterocycl. Chem. 1983, 20, 501-506.
(36) (a) Ito, T.; Aoyama, T.; Shioiri, T. Tetrahedron Lett. 1993, 34,
6583-6586. (b) Takaoka, K.; Aoyama, T.; Shioiri, T. Tetrahedron Lett.
1996, 37, 4973-4976. (c) Takaoka, K.; Aoyama, T.; Shioiri, T. Tetrahedron
Lett. 1996, 37, 4977-4980.
Reactions 1 and 2 have cyclic 1,3-dienes, and reactions 3
and 4 do open-chain 1,3-dienes. Is our new mechanism40 specific
only for reaction 1?
(37) Merino, I.; Hegedus, L. S. Organometallics 1995, 14, 2522-2531.
(38) Mayr, H.; Heigl, W. J. Chem. Soc., Chem. Commun. 1987, 1804-
1805.
(39) Roberts, S. M.; Sutton, P. W.; Wright, L. J. Chem. Soc., Perkin
Trans. 1 1996, 1157-1165.
(40) Yamabe, S.; Dai, T.; Minato, T.; Machiguchi, T.; Hasegawa, T. J.
Am. Chem. Soc. 1996, 118, 6518-6519.

4774 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
Scheme 5. Product Analyses for Diphenylketene-1,3-Diene
Reactions (Reactions 1-4 in Scheme 4): Cyclic 1,3-Dienes
Give Normal Staudinger-Type [2 + 2] Products, whereas
Open-Chain 1,3-Dienes Give Diels-Alder-Type [4 + 2]
Productsa,b
cycloadducts were not obtained as the final products. Reaction
3 proceeds very slowly and takes a very long reaction time for
complete consumption of diphenylketene 1. Reaction 3 gave
the [4 + 2] product 12 eventually (iii in Scheme 5). Similarly,
we have found that reaction 4 affords the [4 + 2] product 13
exclusively (iv in Scheme 5). The IR spectra of the products
(12 and 13) have revealed the strong C-O stretching vibrations
for 12 (νmax 1242 and 1028 cm-1) and 13 (νmax 1236 and 1012
cm-1) and no CdO stretching vibrations due to cyclobutanone
1
structures. H and 13C NMR spectroscopies have clarified that
the new products, 12 and 13, are not the traditional Staudinger-
type [2 + 2] cycloadducts but the Diels-Alder-type [4 + 2]
ones. For spectral detail, see the Supporting Information.
2. Reaction-Intermediate Search. To clarify the reaction
mechanism, we have searched all the reactions (eqs 1-4 in
Scheme 5) for the transient reaction intermediates by NMR
spectroscopies. Choosing experimental conditions (Scheme 6)
precisely, we have succeeded in detecting and isolating all the
intermediates concerned with the elementary processes. All three
experimental parametersstemperature, time, and concentrationss
were strictly tested. (The NMR monitoring results in all the
reactions are shown in the Supporting Information.)
3. Analyses of Elementary Processes and Intermediates
of Cyclic 1,3-Dienes (A). For reactions 1 and 2, we have found
that the reactions proceed in a unique reaction pathway
(reactants f [4 + 2] intermediate f [2 + 2] product). Scheme
6A summarizes experimental results for cyclic 1,3-dienes. Two
reactions 1 and 2 proceed initially by [4 + 2] (Diels-Alder)
cycloadditions to form intermediates 10 and 11. They are
followed by [3,3] sigmatropic (Claisen) rearrangements to give
[2 + 2]-type (Staudinger) products, cyclobutanones (6 and 7),
eventually.
a Conditions for isolation experiments of the final products, Staudinger-
type [2 + 2] cycloadducts (6 and 7). (i) 1 + 2 f 6: CH2Cl2, room
temperature, 3 h, 98%; toluene, room temperature, 3 h, 98%. (ii) 1 +
3 f 7: benzene, 80 °C, 2 days, 96%. bConditions for isolation
experiments of the final products, Diels-Alder-type [4 + 2] cycload-
ducts (12 and 13). (iii) 1 + 4 f 12: CHCl3, 40 °C, 32 days, 97%; no
solvent, 40 °C, 40 days, 94%. (iv) 1 + 5 f 13: CHCl3, room
temperature, 20 days, 98%; benzene, room temperature, 30 days, 96%.
1
Figure 1 illustrates a low-temperature H NMR and Fourier
transform (FT) IR spectroscopic monitoring result of reaction
1. The figure shows that the reaction proceeds very readily even
at low temperatures. Figure 2 shows that reaction 2 requires
higher temperatures and longer reaction times than those in
reaction 1. Both reactions give the [2 + 2] products (6 and 7)
exclusively as the final form.
3.1. Diels-Alder Reaction, the First Step, and Isolation
of [4 + 2] Cycloadducts. Figure 1a illustrates the low-
temperature 1H NMR monitoring result of reaction 1. This figure
reveals the presence of a rapidly formed intermediate (10) of
which the concentration exceeds surprisingly more than 40%
of the reaction mixture at a stage (-20 °C, 3 h) of the reaction.
Figure 3 illustrates the 1H NMR for the reaction mixture at that
stage. The spectra show that the transient intermediate 10 is a
[4 + 2] cycloadduct between 1 and 2. The reaction proceeds
very smoothly at low temperatures (v in Scheme 6) and is
completed within 3 h at 25 °C. FT-IR spectroscopic monitoring
(Figure 1a, bottom) of reaction 1 has shown drastic changes of
CdO stretching vibrations [ketene 1 f (a mixture of 1 + 10
+ 6) f cyclobutanone 6].43,44
II. Experimental Results
We have examined the four reactions (eqs 1-4) in Scheme
4. First in the experiments, we have analyzed four products
precisely. Second, we have searched for reaction intermediates
thoroughly to clarify the reaction mechanisms by H NMR
monitorings.
1
1. Analyses of Final Products. For cyclic 1,3-dienes [cy-
clopentadiene (2) and cyclohexa-1,3-diene (3)], we have traced
reactions 1 and 2 in Scheme 4 at the traditional conditions. We
have observed that reaction 1 proceeds in dichloromethane or
benzene very readily even at room temperature and is completed
within 3 h. Reactions 1 and 2 led to smooth and exclusive
formation of the [2 + 2] products 6 and 7 in quantitative yields
(i and ii in Scheme 5). We have confirmed that both reactions
1 and 2 give the Staudinger-type [2 + 2] cycloadducts, 6 and
7, respectively, as Schemes 4A and 5A show. Their IR spectra
of the products indicate the very strong CdO stretching
vibrations for 6 (νmax 1772 cm-1) and 7 (νmax 1770 cm-1) of
cyclobutanones, respectively.
For open-chain 1,3-dienes (4 and 5), we have carried out
reactions 3 and 4 at 40 °C and room temperature, respectively.
In contrast to the above conventional results of cyclic 1,3-dienes,
we have newly found that both reactions 3 and 4 give
EXCLUSIVE formation of the Diels-Alder-type [4 + 2]
cycloadducts, 12 and 13, respectively. Staudinger-type [2 + 2]
In reaction 1, we have made careful experiments with low-
temperature column chromatography followed by recrystalli-
zation at -78 °C. The intermediate 10 could be isolated as a
labile crystalline material [t1/2(0 °C, solid) ) ca. 6 h] with the
melting point of 38 °C (v in Scheme 6). We have applied various
(41) Farooq, M. O.; Vahidy, T. H.; Husain, S. M. Bull. Soc. Chim. Fr.
1958, 830-832.
(42) For example, the reaction of dimethylketene with trans-1-methoxy-
butadiene (5) gives the corresponding cyclobutanone, 2,2-dimethyl-3-(trans-
2′-methoxyvinyl)cyclobutan-1-one.29a Also, reactions between diphenyl-
ketene (1) with butadienes (butadiene,11d,41 trans-1-acetoxy-butadiene,34 cis-
1-cyanobutadiene,15,34
trans-1-methylbutadiene,9a,15,34
cis-1-methyl-
butadiene,15,34b and 2-methylbutadiene,11d,15,34b etc.) afford cyclobutanones.

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4775
Scheme 6. Low-Temperature Experimental Results for 1,3-Dienes (Reactions 1-4 in Scheme 4): (A) Cyclic 1,3-Dienes and
(B) Open-Chain 1,3-Dienes
a Conditions for isolation experiments of [4 + 2] (Diels-Alder) intermediates (10 and 11). (v) 1 + 2 f 10: CH2Cl2, -20 °C, 3 h, low temperature
b
(-50 °C), column chromatography, 28%. (vi) 1 + 3 f 11: CHCl3, room temperature, 8 days, column chromatography, 31%. Conditions for the
Claisen rearrangements of the [4 + 2] cycloadducts, 10 and 11: (vii) 10 f 6 (solution state): CHCl3, 0 °C, 5 h, 98%; CHCl3, -10 °C, 12 h, 96%.
10 f 6 (solid state): 0 °C, 2.5 days, 100%. Half-life times: t1/2(-10 °C, CD2Cl2) ) ca. 2 h; t1/2 (0 °C, solid) ) ca. 6 h. (viii) 11 f 7: benzene,
c
1
80 °C, 2 days, 98%. Half-life time: t1/2(80 °C, benzene-d6) ) 4.1 h. Conditions for H NMR detection of the retro-Diels-Alder reactions of the
[4 + 2] adducts, 10 and 11. (ix) 10 f 1 + 2. Maximum dissociation: CD2Cl2, -10 °C, 4 h; 1 (8%) + 2 (8%) + 6 (53%) + 10 (14%). (x) 11 f
1 + 3. Maximum dissociation: benzene-d6, 80 °C, 10 h; 1 (2%) + 3 (2%) + 7 (78%) + 11 (18%). The adduct 11 remained unchanged in CDCl3
(25 °C, 24 h; 40 °C, 24 h) or in benzene-d6 at 60 °C for 24 h (no reaction occurred). dConditions for nonoccurrence of retro-Claisen rearrangement.
(Cyclobutanones 6 and 7 remained unchanged against the rearrangement.) (xi) 6 f 10: No reaction occurred, CD2Cl2, CDCl3, or benzene-d6, 25
°C, 15 days. (xii) 7 f 11: No reaction occurred, CDCl3, 40 °C, 15 days; benzene-d6, 80 °C, 2 days. eConditions for isolation of [2 + 2] cycloadducts
(8 and 9): (xiii) 1 + 4 f 8: CHCl3, 25 °C, 15 days, 28%; no solvent, 25 °C, 25 days, 25%. (xiv) 1 + 5 f 9: CHCl3, 0 °C, 5 days, 62%.
fConditions for retro-Claisen rearrangements of cyclobutanones (8 and 9): (xv) 8 f 12: benzene, 40 °C, 17 days, 96%; CHCl3, 40 °C, 12 days,
97%. Half-life times: t1/2(40 °C, benzene-d6) ) ca. 4 days; t1/2(25 °C, benzene-d6) ) ca. 20 days; t1/2(40 °C, CDCl3) ) 1.7 days. (xvi) 9 f 13:
benzene, 25 °C, 24 days, 94%; CHCl3, 25 °C, 16 days, 98%. Half-life times: t1/2(25 °C, benzene-d6) ) 3.4 days; t1/2(25 °C, CDCl3) ) 2.0 days.
gConditions for nonoccurrence of Claisen rearrangement as well as of retro-Diels-Alder dissociation in the [4 + 2] cycloadducts, 12 and 13. (xvii)
12 f 8 and 12 f 1 + 4: No reaction occurred and the cycloadduct 12 remained unchanged; benzene-d6, 60 °C, 2 days; CDCl3, 40 °C, 2 days.
(xviii) 13 f 9 and 13 f 1 + 5: No reaction occurred and the cycloadduct 13 remained unchanged; benzene-d6, 60 °C, 2 days; CDCl3, 40 °C, 2
days.
modern NMR techniques45,46 to the intermediate 10.47 Those
spectral data have given us satisfactory information to solve
the spin networks of hydrogens of the structures as a [4 +
2]-type cycloadduct48 (R-methylenedihydropyran). We have
used the one-dimensional (1D) 1H NMR enhanced with the sine-
bell wind function (Figure 3, lower) and two-dimensional (2D)
NMR (1H and 13C) spectra. By 13C NMR measurements, the
13C-1H long-range spin network through JC-H proved the
(43) The IR spectra of a reactant, diphenylketene (1), and the cyclobu-
tanones (6-9) display very strong stretching vibrations of the carbonyl group
at 2096 and around 1770 cm-1, respectively. On the other hand, the [4 +
2]-type cycloadducts (10-13) indicate strong vibrations of C-O stretching
at around 1240-1140 [νas (C-O)] and 1030-1010 [νs (C-O)] cm-1. The
FT-IR spectroscopy shows that the fragile intermediate 10 is an ether (νC-O
1138 and 1033 cm-1). The 13C NMR chemical shifts of the products (6-
9) appear at around δC ) 210 ppm, indicating that they are cyclobutanones.
Those of the [4 + 2] adducts 10-13 appear downfield around δC ) 150
and 70-90 ppm, indicating an enol carbon and methine (or methylene)
ones, respectively.
(44) At the initial stage of the reaction, i.e., immediately after mixing
the reactants (1 and 2) at -80 °C, a strong CdO stretching vibration of the
ketene appeared at 2096 cm-1 (Figure 1a, bottom left). At an intermediate
stage (-20 °C, 3 h) of the reaction, the FT IR showed that the carbonyl
absorption of ketene 1 was weakened and that new absorptions of the strong
C-O stretching vibrations of 10 appeared at 1138 and 1033 cm-1 (Figure
1a, bottom center). At the final stage (0 °C, 5 h) of the reaction, the FT IR
indicated a new absorption due to the strong CdO stretching vibration of
cyclobutanone 6 at 1772 cm-1 (Figure 1a, bottom right). At the intermediate
stage, the concentration of the Diels-Alder-type [4 + 2] cycloadduct 10
reached the maximum molar concentration of the total of the mixture by
1H NMR monitoring.
(46) (a) Derome, A. E. In Modern NMR Techniques for Chemistry
Research; Baldwin, J. E., Ed.; Organic Chemistry Series 6; Pergamon:
Oxford, 1987. (b) Friebolin, H. Basic One- and Two-Dimensional NMR
Spectroscopy (English Translation), 2nd enlarged ed.; VCH: New York,
1993. (c) van de Ven, F. J. M. Multidimensional NMR in Liquids: Basic
Principles and Experimental Methods; VCH: New York, 1995.
(47) The NMR (13C and 1H) spectral data show that the [4 + 2]
cycloadduct 10 is formed by attack of diphenylketene to the terminal 1 and
4 positions of the diene 2.
1
(48) In the H NMR spectrum of 10, two upfield resonances at δ 1.64
(dd) and 1.96 (dt) are assigned to aliphatic methylene protons (H-7exo and
H-7endo, respectively) which couple each other with J7,7 ) 8.5 Hz and
further with two methine signals at δ 3.62 (H-4) and δ 5.28 (H-1) with J
) 1.5 Hz. The latter methine (H-1) signal shifts to downfield by influence
of a linkage with the oxygen atom (O-2) directly attached to the methine
carbon, C-1. The two methine signals of H-1 and H-4 are disposed in a
W-type arrangement with a small coupling of J1,4 ) 1.5 Hz. They link
further with two olefinic protons at δ 6.37 (H-6) and δ 6.58 (H-5) with
vicinal couplings (J1,6 ) 2.5, J4,5 ) 3.0 Hz). These results indicate that the
cycloadduct 10 is formed by attack at the C-1 position by the oxygen of
the ketene 1 as well as at the C-4 position of 10 by the carbonyl carbon of
1. 2D NMR 1H-1H COSY also demonstrates this proton spin network.
Long-range interactions with small magnitudes between H-1 and H-4, H-1
and H-5, and H-4 and H-6 are observed in both the 2D NMR 1H-1H
COLOC and homonuclear decoupling experiments. The 1H-1H relationship
demonstrates that the structure of 10 has a norbornene moiety.
(45) To accomplish a full assignment and accurate analysis we have used
various modern techniques of the following NMR (1H and 13C) spec-
troscopies.46 1D NMR: Resolution-enhanced spectra with sine-bell wind
1
function. 2D NMR: 1H-1H shift-correlation spectroscopy (COSY), H-
1H shift-correlation spectroscopy by long-range couplings (COLOC), 13C-
1H COSY, 13C-1H COLOC, with an aid of composite proton decoupling
(CPD), gated decoupling, 13C-1H selective decoupling (SEL), and 13C-
1H long-range selective decoupling (LSPD).

4776 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
Figure 2. 1H NMR spectroscopic monitoring of reaction 2. (a) Plots
of molar ratio of the reaction components versus reaction times of the
reaction between diphenylketene (1) and cyclohexa-1,3-diene (3) in
toluene-d8. The reaction was carried out in an NMR tube in the
temperature range. (b) Monitoring result of the Claisen rearrangement
(11 f 7) together with slight retro-Diels-Alder dissociation (11 f 1
+ 3). A solution of pure isolated 11 in benzene-d6 was heated in an
NMR tube.
complete consumption of the reactants (1 + 3) and gives a
mixture of [4 + 2] cycloadduct 11 (38%) and [2 + 2] product
1
7 (62%) in vi of Scheme 6. Figure 2a illustrates the H NMR
monitoring result and shows that reaction 2 is not so ready for
completion as reaction 1. The low reactivity of the subsequent
Claisen shift leads to the easy isolation of intermediate 11.50
We have isolated the [4 + 2] intermediate, 11, as a stable
crystalline compound. The initial [4 + 2] cycloadduct 11 is
stable enough to be isolated and can be handled at room
temperature.
3.2. Claisen Rearrangement of [4 + 2]-Type Intermedi-
ates, the Second Step. Figure 1b shows that isolated [4 + 2]
intermediate 10 is unstable and that the Claisen rearrangement
10 f 6 takes place easily even at 0 °C to give [2 + 2] product
Figure 1. Spectroscopic monitoring of reaction 1. (a) Upper: 1H NMR
monitoring result of the reaction between diphenylketene (1) and
cyclopentadiene (2) in CD2Cl2 by plots of molar ratio of the reaction
components versus reaction times. The reaction was carried out in a
sealed NMR tube. The reaction temperatures were regulated in an NMR
probe. Lower: Selected FT-IR charts in the monitoring of the reaction
[left, the initial stage (-80 °C) of the reaction immediately after mixing
both reactants (1 and 2); center, a medium stage of the optimum
1
concentration of 10 (-20 °C); right, the final stage (0 °C)]. (b) H
(49) In the 13C NMR of 10, an upfield signal at δ 53.80 (t) is assigned
to the methylene carbon C-7 due to off-resonance experiment. Of the two
methine-carbon signals at δ 49.78 (d) and 85.62 (d), the latter downfield-
shifted signal is assigned to C-1 which directly links with the oxygen atom
(O-1). The 13C NMR confirms the existence of one exocyclic C(3)dC(8)
bond, which linked with C-4, O-1, and two phenyl groups, whose chemical
shifts appear at δ 150.85 (s, C-3) and 110.30 (s, C-8) (the atom numbering
is shown in 10 of Scheme 6A). The C-3 and C-8 are connected with H-4
and o-phenyl protons, respectively, with long-range JCH couplings [3JC(3)H(1)
) 3JC(3)H(7) ) 6.6 Hz, 3JC(8)H(4) < 1.0 Hz, 4JC(8)H(o-Ph) ) 3.5 Hz, 4JC(8)H(o′-Ph)
) 3.5 Hz]. The 13C-1H relationship of homonuclear long-range couplings
builds up the carbon framework of compound 10 as an R-diphenylmeth-
ylene-dihydropyran structure. Both C-1 and H-1 are key atoms to decipher
the 1H and 13C spectra, since the 13C and 1H signals shift to downfield
prominently.
NMR monitoring result of the solution-state Claisen rearrangement (10
f 6) in CDCl3 together with retro-Diels-Alder dissociation (10 f 1
+ 2). The dissociation occurs above -15 °C. (c) 1H NMR monitoring
result of the solid-state Claisen rearrangement at 0 °C in the dark. The
monitoring was performed by intermittent dissolution of the sample
into CDCl3. The measurements were made at the NMR probe
temperature of -40 °C.
linkages between the ketene and diene frames in 10.49 (All the
data as well as the selected NMR charts for all the products
and intermediates are included in the Supporting Information.)
Reaction 2 gives exclusively [2 + 2] product 7 at conventional
high-temperature conditions (ii in Scheme 5). Under mild
conditions, it takes a long reaction time (8 days at 25 °C) for
(50) Roberts and co-workers recently reported that the reaction gave an
equilibrium mixture of 7 and 11. See ref 39.

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4777
Figure 3. 400-MHz 1H NMR (-20 °C, CD2Cl2) spectrum of the [4 + 2]-type transient intermediate 10 in the reaction between diphenylketene (1)
and cyclopentadiene (2). Upper: Ordinary (Lorentz-transformed) spectrum. Lower: The spectrum resolution is enhanced with the sine-bell wind
function. A molar concentration of the [4 + 2]-type intermediate 10 reaches 43% of the total reaction mixture (v in Scheme 6). For the selected
1
1D and 2D H and 13C NMR charts of pure isolated 10, see the Supporting Information.
6 (vii in Scheme 6). Also, the dissociation, 10 f 1 + 2, is
induced by a retro-Diels-Alder reaction (ix in Scheme 6).
Figure 1c shows that the Claisen rearrangement 10 f 6 occurs
eVen in the solid state at 0 °C (below the melting temperature,
38 °C) without any sign of melting or retrogression51 to the
reactants (1 + 2) (vii in Scheme 6). The solid-state reaction of
10 gives the product 6 finally. Thus, the Claisen shift 10 f 6
in reaction 1 has been proved experimentally. In reaction 2,
the Claisen shift 11 f 7 is very slow and requires higher
temperatures (viii in Scheme 6 and Figure 2b) than those in
reaction 1.
3.3. Retro-Diels-Alder Dissociation of [4 + 2]-Type
Intermediates. Dissolution of the pure isolated 10 in toluene-
d6 or CD2Cl2 at the temperature above 0 °C causes a retrogres-
sion (retro-Diels-Alder reaction) to the reactants (1 + 2) (ix
in Scheme 6 and Figure 1b). However, the smooth Claisen
conversion 10 f 6 gives final [2 + 2] product 6 (i in Scheme
5). Figure 2b shows that the intermediate 11 in reaction 2
retrogresses slightly at temperatures higher than those in reaction
1 (x in Scheme 6).
were subject to standing for a long time, the products were
recovered unchanged (xi and xii in Scheme 6).
4. Analysis of Elementary Processes and Intermediates
of Open-Chain 1,3-Dienes (B). In contrast to cyclic 1,3-dienes
proceeding in a single reaction pathway, we have found that
the open-chain 1,3-dienes react with diphenylketene in a dual
pathway through both [2 + 2] (Staudinger) and [4 + 2] (Diels-
Alder) reactions. Reactions 3 and 4 give finally and exclusively
[4 + 2] cycloadducts (R-methylenedihydropyrans) according to
ready retro-Claisen rearrangements ([2 + 2] intermediates f
[4 + 2] products). Scheme 6B summarizes those experimental
results.
4.1. Competition of [4 + 2] (Diels-Alder) and [2 + 2]
1
(Staudinger) Reactions. Figure 4a shows the H NMR moni-
toring result of reaction 3 at 40 °C in CDCl3. The figure
demonstrates that the reaction affords simultaneously both [2
+ 2] 8 and [4 + 2] 12 cycloadducts. It also shows preferential
[2 + 2] cycloaddition to [4 + 2] cycloaddition initially. It is
shown explicitly that the formation of 8 is superior to that of
12 when the reaction was carried out at 25 °C (for details, see
plotting of the reaction monitoring in the Supporting Informa-
tion). At an initial stage (7 days) of the reaction at 25 °C, we
have detected predominant formation of [2 + 2] cycloadduct 8
(33%) along with the minor formation of 12 (18%), while 49%
of the reactants (1 + 4) remain unreacted. At 40 °C after 5
weeks, the reaction is completed by the consumption of the
reactants and gives [4 + 2] cycloadduct 12 as the sole product.
When the reactants (1 + 4) are mixed with no solvent at room
3.4. Nonoccurrence of Retro-Claisen Rearrangement in
[2 + 2]-Type Products. Two [2 + 2] products (6 and 7) of
cyclic 1,3-dienes do not reveal any retro-Claisen rearrangement
at room temperature. Although the [2 + 2] products, 6 and 7,
(51) The 1H NMR monitoring has been carried out by intermittent
dissolution of the crystals in CDCl3. The measurements were performed at
-60 °C. This conversion proceeds without any sign of melting as a crystal-
to-crystal phase reaction. Cf.: Machiguchi, T.; Hasegawa, T.; Ito, S.;
Mizuno, H. J. Am. Chem. Soc. 1989, 111, 1920-1921.

4778 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
Figure 4. 1H NMR spectroscopic monitoring of reaction 3. (a) Plots
of molar ratio of the reaction components versus reaction times. The
reaction was carried out in an NMR tube using diphenylketene (1) and
2,3-dimethylbutadiene (4) in CDCl3 at 40 °C. (b) Monitoring result of
the retro-Claisen rearrangement (8 f 12). A solution of pure isolated
cyclobutanone 8 in CDCl3 was heated at 40 °C in an NMR tube.
Figure 5. 1H NMR spectroscopic monitoring of reaction 4. (a) Plots
of molar ratio of the reaction components versus reaction times by 1H
NMR spectroscopic monitoring result of the reaction between diphenyl-
ketene (1) and trans-1-methoxybutadiene (5) in CDCl3. The reaction
was carried out in an NMR tube. (b) Monitoring result of the retro-
Claisen rearrangement (9 f 13) of the [2 + 2] intermediate 9 to [4 +
2] product 13. A solution of pure isolated cyclobutanone 9 in CDCl3
was heated in an NMR tube.
temperature, the reaction proceeds very slowly (14 weeks15b
without solvent effect (vide infra).
)
Figure 5a shows that reaction 4 proceeds smoothly at low
temperatures for complete consumption of the reactants (0 °C,
5 days, CDCl3; 25 °C, 2 days, benzene-d6). The reaction yields
simultaneously [2 + 2] cycloadduct 9 and [4 + 2] product 13,
which is similar to reaction 3. Reaction 4 eventually affords [4
+ 2] product 13 exclusively, when the reaction is carried out
for long times (iv in Scheme 5).
4.2. Isolation of Reaction Intermediates. Reaction 3 gives
a mixture of the [4 + 2] and [2 + 2] cycloadducts (8 + 12) in
the approximate molar ratio of 38:19% (xiii in Scheme 6).
Reaction 4 is terminated with the complete consumption of the
reactants at intermediate conditions (0 °C, 5 days, CHCl3). The
reaction gives [2 + 2] cycloadduct 9 (70%) predominantly with
a minor [4 + 2] product 13 (30%) (xiv in Scheme 6). We have
isolated both the [2 + 2]-type intermediates, 8 and 9, from the
reaction mixtures (xiii and xiv in Scheme 6) in reactions 3 and
4, respectively, as stable compounds which can be handled at
room temperature.
4.3. Resistance to Claisen Rearrangement and to Retro-
gression to the Reactants of [4 + 2]-Type Products. The pure
isolated [4 + 2] products, 12 and 13, were subject to heating.
In contrast to the easy Claisen rearrangement in cyclic 1,3-dienes
(reactions 1 and 2), we have found that [4 + 2] cycloadducts
12 and 13 resist the Claisen reactions (12 f 8 and 13 f 9)
even at high-temperature conditions at 60 °C (xvii and xviii in
Scheme 6). At this temperature, the [4 + 2] products do not
retrogress at all to the reactants, and the products (12 and 13)
were recovered unchanged. In reaction 4, neither the Claisen
shift of the [4 + 2] product (13 f 9) nor the retro-Diels-Alder
dissociation (13 f 1 + 5) takes place even at 60 °C (xviii in
Scheme 6). In open-chain diene cases, thus, the [4 + 2]
cycloadducts (12, 13) do not show any sign of retrogression to
the reactants (1 and 4-5) at room temperature. [4 + 2] products
12 and 13 are more stable than [2 + 2] cycloadducts 8 and 9,
respectively.
4.4. Ready Retro-Claisen Rearrangement of [2 + 2]-Type
Intermediates. The pure isolated intermediates, 8 and 9, were
subject to warming. In contrast to the nonoccurrence of Claisen
rearrangement in 12, we have newly found that the retro-Claisen
reaction of [2 + 2] cycloadduct 8 occurs gradually and gives a
mixture of 8 + 12 after 15 days (25 °C, benzene-d6). The Claisen
step 12 f 8 does not occur at 25 °C and the reaction reaches
an equilibrium mixture of 12 with 8 in the ratio of 71:29 at this
temperature. At higher temperature (40 °C), the retro-Claisen
rearrangement of 8 gives smoothly (Figure 4b) [4 + 2] product
12, when the reaction of 8 was maintained at 40 °C for an
additional 12 days (xv in Scheme 6). We have observed the
similar retro-Claisen shift of [2 + 2] cycloadduct 9 to 13 in
reaction 4 (xvi in Scheme 6). The 9 f 13 shift occurs smoothly
at 25 °C to afford [4 + 2] product 13.
5. Summary of the Experiments. The diphenylketene reacts
with cyclic (i.e., s-cis) dienes to give [2 + 2] (Staudinger)
products, cyclobutanones, exclusively under the traditional
conditions at high temperatures and long reaction times. In
contrast, we have found in this work that ketene reacts with
open-chain (i.e., s-trans) dienes to give [4 + 2] (Diels-Alder)
products, R-methylenedihydropyrans (not Staudinger products),
under these conditions. 1H NMR monitorings have clarified that
ketene-cyclic diene reactions proceed in a unique pathway
(initial Diels-Alder reaction followed by Claisen rearrangement)

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4779
Scheme 7. Orientations of Two Cycloadditionsa
a Upward arrows indicate dominant charge-transfer (CT) interactions
and broken downward arrows do minor (back CT) interactions.
one phenyl group and the cyclopentadiene ring is avoided by
the parallel orientation. The other (remote) phenyl group is
conjugated to the ketene fragment. The distance between the
carbonyl oxygen and one diene carbon is large, 2.406 Å.
After the Diels-Alder TS, the transient intermediate, 10, is
attained. There is a ring strain in the bicyclo species (for
structural details, see the Supporting Information). At two
bridgehead carbon atoms, sp3 bond angles are ca. 100°, which
is smaller than 109.5°. The long C(1)-O(2) bond of 10 (1.489
Å) indicates their ready cleavage for [3,3] sigmatropic rear-
rangements. To escape from the ring strain of 10, a [3,3]
sigmatropic rearrangement takes place. In Figure 6, the TS
geometry is shown. One may suspect that this TS is for [2 +
2] cycloaddition. However, the reaction-coordinate vector
corresponding to νq ) 156.9i cm-1 indicates clearly the motion
of the Claisen shift.
The distance of the uncleaved C-C bond, 1.598 Å, in the
Claisen TS (10 f 6) is slightly larger than the standard covalent-
bond one, 1.52 Å. Phenyl groups do not give steric crowding
during the rearrangement. After the Claisen TS, the cyclobu-
tanone product, 6, is obtained. There is a concerted [2 + 2]
addition path shown in Figure 6. The path is also free from
steric crowding due to phenyl groups. The crowd can be avoided
because of the [2 + 2] TS one-center-like geometry.52,53 The
selectivity of [4 + 2] and [2 + 2] cycloadditions comes from
orientations of ketene and diene planes (Scheme 7).
Figure 6. Geometries optimized with RHF/3-21G in the reaction
between diphenylketene (1) and cyclopentadiene (2) (reaction 1).
Distances are in Å. For transition states, sole imaginary frequencies in
cm-1 are shown. Reaction-coordinate vectors of the Claisen (or retro-
Claisen) transition state are also sketched.
as Scheme 6A shows. In contrast, we have newly found that
ketene-acyclic diene reactions proceed in a dual pathway. It
is a parallel reaction of Diels-Alder and Staudinger reactions.
The latter is followed by retro-Claisen rearrangement of the [2
+ 2] intermediate as Scheme 6B shows. In short, ketene
recognizes s-cis and s-trans 1,3-dienes distinctly. Under mild
conditions (0-30 °C) for short reaction times, all four reactions
form mixtures of both [2 + 2] and [4 + 2] cycloadducts. Both
the intermediates and products can often be isolated as stable
compounds and can be handled at room temperature (except
the labile intermediate 10 in reaction 1). At a glimpse, such
formation of a mixture seems to be the common type of reaction.
But, long reaction times transform such a mixture into the final
product. They allow the exclusive formation of [2 + 2] products
for cyclic 1,3-dienes by the Claisen rearrangement and the
formation of [4 + 2] products for open-chain dienes by retro-
Claisen rearrangement. The present experimental results may
rationalize the conventional results (Scheme 2) and exceptions
(Scheme 3) of ketene-diene reactions. Mixtures of [4 + 2] and
[2 + 2] products reported so far most likely come from
insufficient reaction times.
Figure 7 shows the process of reaction 2. The geometry of
the [4 + 2] addition TS is similar to that in reaction 1. The
C‚‚‚C distance, 1.801 Å, of reaction 2 is slightly smaller than
that, 1.871 Å, of reaction 1. On the contrary, the C‚‚‚O distance
of reaction 2 is larger. These distance differences demonstrate
that the [4 + 2] TS of reaction 2 is more asynchronous due to
the more distant terminal 1,4 carbon atoms in cyclohexadiene
III. Computational Results
1. Structures. Four reactions are examined theoretically.
Figure 6 exhibits geometries of five stationary points of reaction
1. The [4 + 2] transition state (TS) demonstrates that the
carbonyl group of 1 works effectively as a dienophile. The
bending and orientation of the ketene cumulene bond enhances
the charge transfer (CT) interaction. The steric crowd between
(52) Wang, X.; Houk, K. N. J. Am. Chem. Soc. 1990, 112, 1754-1756.
(53) Salzner, U.; Bachrach, S. M. J. Org. Chem. 1996, 61, 237-242.

4780 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
Figure 8. TS structures and changes of Gibbs free energies (T ) 298.15
K) of [2 + 2] cycloadditions of diphenylketene (1) to s-cis- and s-trans
open-chain dienes (4 and 5) (reactions 3 and 4). Atomic bond
populations were calculated with RHF/STO-3G.
Figure 7. Geometries in the reaction between diphenylketene (1) and
cyclohexadiene (3) (reaction 2).
Scheme 8. Distances (R’s) between 1- and 4-Carbon Atoms
and Dihedral Angles (θ’s) as Indices for [4 + 2]
Cycloadditions
are not fit to Diels-Alder reactions. The cis-trans energy
differences in 4 and 5 are examined. Trans isomers of 4 and 5
are only more stable by 1.5 and 2.2 kcal/mol (MP2/6-31G*//
RHF/3-21G) than cis ones, respectively. Rotational energy
barrier heights are also small, ∆Gq(trans f TS) ) 2.9 (4) and
6.0 kcal/mol (5). Thus, dienes 4 and 5 undergo almost free
rotations, and their cis and trans isomers are practically
indistinguishable. Figure 8 exhibits TS structures of [2 + 2]
cycloadditions. Surprisingly, those of cis dienes are more stable
(with larger total energies) than those of trans dienes. A
secondary interaction enhances the cis [2 + 2] addition (Scheme
9). Diphenylketene prefers cis dienes over trans dienes. Here-
after, only cis dienes are considered.
Figure 9 shows the process of reaction 3. In the [4 + 2] TS,
clearly, the size of the ketene carbonyl group does not match
that between two terminal methylene groups of 4 (Scheme 8).
That is, the line of the incipient C‚‚‚O bond deviates significantly
from the direction of lumo (π*CdO). A large activation energy
of the [4 + 2] TS of reaction 3 is predicted. But, the [4 + 2]
adduct, 12, has little ring strain and steric crowding (the ether
bond C-O-C ) 115.2°). Figure 9 shows the other process
of reaction 3, (1 + 4) f 8 f 12. A steric repulsion between a
phenyl group and a methyl group is noticed in 8. Owing to this
steric crowding, one C-C distance is 1.613 Å in the cyclobu-
tanone ring. [2 + 2] adduct 8 is thought to be less stable than
[4 + 2] adduct 12. A retro-Claisen TS (8 f 12) has been found.
The reaction-coordinate vector demonstrates the isomerization
path. The steric congestion is somewhat relaxed by the shift, 8
f 12.
(3) (Scheme 8). In contrast to the difficulty in [4 + 2]
cycloaddition, its resultant cycloadduct, 11, does not suffer the
ring strain as much as 10. The bicyclo compound 11 suffers
only eclipsed conformation in the ethylene group. In Figure 7,
intermediate 11 is isomerized to the [2 + 2] cycloadduct, 7,
via the Claisen TS. This TS is somewhat earlier than that in
reaction 1 due to the flexibility of the hexadiene ring.
Cyclic dienes (A) are restricted to cis geometrical isomers.
But, open-chain dienes (B) prefer naturally trans forms, which
Figure 10 exhibits the process of reaction 4. Diene 5 is
nucleophilic at the C-4 methylene group. In accordance with

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4781
Scheme 9. An “Ortho Secondary Orbital Interaction” To
Make the cis-Diene More Reactive than the trans-Diene for
[2 + 2] Cycloadditionsa
a The attractive interaction along the broken line is verified by the
RHF/3-21G bonding population, 0.046.
Figure 10. Geometries in the reaction between diphenylketene (1) and
trans-1-methoxybutadiene (5) (reaction 4).
stability similar to that of reactants (1 + 2). If it were not for
the Claisen shift, an equilibrium of similar molar concentrations
between reactants and 10 would be attained. The second
activation barrier (∆Gq ) 19.1 kcal/mol) is also small and
similar to the first one. The Claisen shift should take place
readily. The [2 + 2] product, 6, is the most stable. It is clear
that [4 + 2] intermediate 10 is eventually converted to 6. The
low-temperature experiment has been needed for us to trap the
intermediate, 10 (see ∆G° ) -0.6 kcal/mol at T ) 300 K and
∆G° ) -5.4 kcal/mol at T ) 200 K). Owing to those small
activation energies, it is understandable that reaction 1 has been
thought to be of the concerted [2 + 2] process historically.
Second, energetics of reaction 2 are examined in Figure 11.
The value of ∆Gq of the [4 + 2] addition is larger than that of
reaction 1. The reaction-center two-carbon atoms of cyclohexa-
diene (3) are more distant with each other and are less suitable
for the Diels-Alder reaction than those of cyclopentadiene (2)
(see Scheme 8). On the other hand, the [4 + 2] adduct (11) is
more stable than the reactants (1 + 3). Adduct 11 has the small
ring strain. The Claisen TS has a lower energy level than the
Diels-Alder TS, and the [2 + 2] product (7) is most stable.
Viewing the energy diagram of reaction 2, one may confirm
that reaction 2 is slower than reaction 1 and that the [4 + 2]
adduct intermediate 11 is trapped more easily than that (10) of
reaction 1. Of course, the concerted [2 + 2] path (1 + 3 f 7,
the broken curve) is ruled out energetically.
Figure 9. Geometries in the reaction between diphenylketene (1) and
2,3-dimethylbutadiene (4) (reaction 3).
this high reactivity of 5, the [4 + 2] addition TS (1 + 5 f 13)
is very early in the figure. A small activation energy of the
Diels-Alder reaction is predicted. The [4 + 2] adduct, 13, is
of the small ring strain. Figure 10 shows the other process of
reaction 4, (1 + 5) f 9 f 13. The retro-Claisen TS is verified
by reaction-coordinate vectors.
2. Energy Diagrams. Figure 11 shows Gibbs free-free energy
diagrams including the self-consistent-reaction-field (SCRF)
solvent effect of reactions 1 and 2. First, energetics of reaction
1 are examined. The Diels-Alder activation free energy, ∆Gq
) 19.7 kcal/mol (T ) 300 K), is found to be small. The ketene
carbonyl group is an excellent dienophile. For instance, the
ethylene-cyclopentadiene Diels-Alder reaction has ∆Gq )
24.7 kcal/mol. The [4 + 2]-cycloadduct (10) intermediate has
Figure 12 exhibits energy diagrams of reactions 3 and 4. The
[4 + 2] TS of reaction 3 has the largest activation free energy,
∆Gq ) 30.3 kcal/mol (T ) 300 K). A very slow reaction is
expected. Noteworthy is that [4 + 2] cycloadduct 12 is more
stable than [2 + 2] adduct, 8. The exceptional examples
displayed in Scheme 3 and the present results in Scheme 6 are
rationalized.
The mixture of 12 and 8 is the temporary product in reaction
3. Kinetically 8 is generated, and thermodynamically 12 is

4782 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
Figure 11. Gibbs free-energy diagrams of reactions 1 (upper) and 2
(lower) in kcal/mol. Electronic energies are obtained by MP2/6-31G*
(frozen core) single-point calculations on the RHF/3-21G geometries
in Figures 6 and 7 (MP2/6-31G*//RHF/3-21G). Thermal correction and
entropies are evaluated via RHF/3-21G vibrational analyses (T ) 200
and 300 K). The solvent effect is included as a RHF/3-21G SCRF
correction to the MP2/6-31G* electronic energy. The dielectric constant
ꢀ is taken to be 6.0 for chloroform (CHCl3).
Figure 12. Gibbs free-energy diagrams of reactions 3 (upper) and 4
(lower).
mental and theoretical verification of the new mechanism gives
us a unified point of view as to the ketene-diene reaction. While
the ketene carbonyl group (>CdCdO) has been regarded as
an inViolable substituent, this work has demonstrated that the
group is the reaction center and an excellent dienophile.
Therefore, the primary cycloadditions between ketenes and
dienes should be Diels-Alder reactions. However, when diene
terminal carbon atoms are distant and two vinyl bonds are
nonplaner, [2 + 2] cycloadditions are favored specifically.
Cyclic dienes (A) have unique reaction channels, [4 + 2]
cycloaddition f [3,3] Claisen rearrangement to [2 + 2]
cycloadducts. Open-chain dienes (B) are reactive in their s-cis
forms for [2 + 2] cycloadditions as well as for [4 + 2]
cycloadditions. They have dual channels, direct [4 + 2] reaction,
and [2 + 2] cycloaddition and subsequent retro-Claisen shift.
Scheme 10 summarizes the present result, which seems to be
an innovation in ketene history.
produced. Reaction 4 will occur readily because of the small
values of ∆Gq ) 19.4 kcal/mol (T ) 300 K) for [4 + 2] and
18.3 kcal/mol for [2 + 2] in Figure 12. The [4 + 2] adduct
(13) is much more stable than the [2 + 2] adduct (9), and 13
should not be exceptional but be dominant. Primarily, two
adducts 9 and 13 are afforded. Eventually, 13 is obtained where
9 is isomerized to 13 by the retro-Claisen rearrangement.
Energetic results in Figures 11 and 12 are summarized. [4 +
2] cycloaddition and the subsequent Claisen shift are likely
processes for cyclic dienes (A). The [2 + 2] products (6 and 7)
are not generated via the concerted [2 + 2] cycloaddition.
Contrary to the superiority or inferiority of [4 + 2] and [2 + 2]
cycloadditions of cyclic dienes (A), the latter reaction is favored
slightly for open-chain dienes (B). The [4 + 2] adducts 12 and
13 are final products. [2 + 2] adducts 8 and 9 are intermediates
which are subject to facile retro-Claisen shifts.
V. Experimental Section
1. General. The starting materials, diphenylketene54 (1), cyclopen-
tadiene55 (2), and cyclohexa-1,3-diene (3),56 were prepared according
to methods reported in the literature. Freshly distilled 1-3, 2,3-
IV. Concluding Remarks
In this work, four ketene-diene reactions have been inves-
tigated experimentally and theoretically. Precise and low-
temperature experiments have been made to compare their
results with computational data. Consistency between two
independent approaches has been proved. The present experi-
(54) Smith, L. I.; Hoehn, H. H. Organic Synthesis; Wiley: New York,
1955; Collect. Vol. 3, pp 356-358.
(55) Moffett, R. B. Organic Synthesis; Wiley: New York, 1963; Collect.
Vol. 4, pp 238-241.
(56) Schaefer, J. P.; Endres, L. Organic Synthesis; Wiley: New York,
1973; Collect. Vol. 5, pp 285-288.

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4783
Scheme 10. Summary of Ketene-Diene Reactions by Pictorial Energy Levelsa
a Larger balls denote more stable species.
dimethylbutadiene (4) (Aldrich), and trans-1-methoxybutadiene (5)
(Aldrich) were used for the reactions. The solvents used for the reactions
were freshly distilled under nitrogen from appropriate drying agents,
and all acids were removed carefully.57 Methylene-d2 chloride and
chloroform-d3 (Aldrich) were filtered through activity grade 1 basic
alumina prior to use. For isolation of the intermediates and products,
6-13, silica gel (0.063-0.200 mm) (Merck Kieselgel 60) was packed
and eluted with pentane-ether (9:1). Preparative layer chromatography
(PLC) was employed with Merck Kieselgel 60 (70-230 mesh). Thin
layer chromatographic (TLC) analyses were performed on silica gel
(Merck Kieselgel 60 GF254) with a 0.2-mm-layer thickness. Rf values
were measured on silica gel in pentane-ether (9:1) unless otherwise
noted.
Melting points were determined on a Bu¨chi 511 apparatus in open
capillary tubes and are uncorrected. Ordinary IR spectra were recorded
on a Hitachi 260-50 spectrometer using CH2Cl2 solution unless
otherwise noted. Low-temperature FT-IR spectral monitoring was
performed on a JEOL JIR-100 FT-IR instrument. Electron impact (EI)
mass spectra (MS) were obtained with a JEOL DM-303 high-resolution
(HR) spectrometer at 70 eV using direct inlet. The values of m/z of
significant ions are reported with relative intensities in parentheses
(percent for the base peak) for low-resolution analyses. NMR (13C and
1H) spectra were recorded on a Bruker AM-400 instrument (100.6 and
400 MHz for 13C and 1H nuclei, respectively) in CDCl3, unless otherwise
specified, with Me4Si as the internal standard. The assignments of NMR
spectra are based on 1H{1H} homonuclear decoupling and 13C{1H}
heteronuclear 1D and 2D NMR (1H-1H COSY, 1H-1H COLOC, 13C-
1H COSY, and 13C-1H COLOC) experiments. (The experimental details
for each type of 2D NMR spectroscopy are described on page 5 of the
Supporting Information.)
check. Solvent removal gave a residue which was purified (for the
products, 6-7 and 12-13) or separated into each component (for
isolation of the intermediates, 10-11 and 8-9) by chromatography.
Analytical samples were obtained by recrystallization from cold
1
pentane-ether (9:1) unless otherwise noted. The IR and H NMR of
the obtained samples of the known compounds 6, 7, 8, and 11 were
identical with those of authentic samples.58,59
1
(ii) H NMR Monitoring of the Reactions. 1 (0.60 mmol), diene
(2-5) (0.60 mmol), and 0.40 mL of a solvent (benzene-d6, CD2Cl2,
CDCl3 or toluene-d8) were vacuum transferred into an NMR tube. The
temperatures of the mixture were regulated in an NMR probe within
the range of (1 °C. The monitoring of the reactions was carried out
periodically at various temperatures. For long-time reactions, the
samples stood at a desired temperature in a thermostated bath in the
dark during intervals of the measurements. These analytical results are
displayed in the Supporting Information.
3.1. Reaction 1. (i) Under Room-Temperature Conditions To
Give [2 + 2] Product 6. The general procedure was followed, using
1 (1.16 g) and 2 (436 mg) in CH2Cl2 or toluene at room temperature
for 3 h to give 7,7-diphenylbicyclo[3.2.0]hept-2-en-6-one (6) (1.53 g,
98%).60
(ii) Under Low-Temperature Conditions To Isolate [4 + 2]
Intermediate 10. The general procedure was followed, using 1 (2.33
g) and 2 (0.88 g) in CH2Cl2 at -20 °C for 3 h. Volatile material was
removed under reduced pressure (oil-pump vacuum) at -40 °C to leave
colorless glassy solid including a mixture of 10 (ca. 43%) and 6 (ca.
11%). Low-temperature medium-pressure column chromatography
(-60 °C) followed by PLC (-50 °C) separation gave each component
which was collected to give 2-oxa-3,3-diphenylmethylidenebicyclo-
[2.2.1]hept-5-ene (10)61 (880 mg, 28%) and 6 (252 mg, 8%).
2. Manipulations. All reactions and their NMR monitorings were
performed using either an inert atmosphere or a high-vacuum line
(<10-4 Torr) technique and were degassed by three successive freeze-
pump-thaw cycles to 10-4 mmHg. NMR tubes for sealed tube
experiments were flame-dried under vacuum immediately prior to the
experiments. Low-temperature chromatographies were carried out in a
cooling glovebox (Suns Engineering L-131) under a cold nitrogen
stream. All the reactions with solvents CHCl3, CDCl3, CH2Cl2, and
CD2Cl2 were carried out in the dark to prevent generation of
hydrochloric acid under exposure to light.
3. General Procedure for the Reactions of Diphenylketene (1)
with 1,3-Dienes (2-5). (i) Under Preparative Conditions: In a typical
case, into a solution of 1 (1.16 g, 6.0 mmol) in 5 mL of an organic
solvent (benzene, CHCl3, CH2Cl2, or toluene) was added dropwise at
0 °C a solution of a diene (6.6 mmol) in the same solvent (5 mL). The
reaction mixture was stirred until the reaction was completed by TLC
(58) The authentic samples were obtained by the literature methods for
6,11d 7, and 11.39
(59) During the reexamination to obtain an authentic sample of the
cyclobutanone 8 in Scheme 4, 8 was not obtained according to the Farooq’s
experimental method9b,41 (heating conditions at 100 °C, overnight, no
solvent). Instead of the isolation of 8, the reaction (100 °C, 4 h) brought us
the predominant formation of an unknown R-methylenedihydropyran 12
(82%), which has been reported in the present paper. The reaction was
completed to consume the ketene within 3 h under the conditions. In a
short reaction time (10 min, 100 °C), the reaction led to the isolation of the
cyclobutanone 8 [mp 79-80 °C (not 118-119 °C9b,15b,41)] in low yield
(21%) accompanied by the dihydropyran 12 (12%) and most of the unreacted
ketene 1. When reaction 3 was carried out under mild conditions [room
temperature, 3 months, no solvent (see the Supporting Information); 40
°C, 40 days, no solvent (iii in Scheme 5)], the product obtained exclusively
has been identified to be R-methylenedihydropyran 12. Cyclobutanone 8
was isolated as the intermediate of the reaction (xiii in Scheme 6).
(60) Cyclobutanone 6: colorless prisms (MeOH), mp 89-90 °C; Rf )
0.34; IR νmax 1772 (vs) cm-1; detailed NMR (1H and 13C) and MS data are
in the Supporting Information.
(57) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Pergamon: Oxford, U.K., 1988.

4784 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
10: colorless needles, mp 37-38 °C; Rf ) 0.43 (-50 °C); FT-IR
(dd, 1 H, J ) 2.3, 1.0 Hz, H-2a); 13C NMR δ 14.14 (q, 4-CH3), 18.08
(q, 3-CH3), 32.62 (t, C-5), 70.86 (t, C-2), 119.86 (s, C-7), 122.91 (s,
C-4), 123.27 (s, C-3), 147.52 (s, C-6); EI-MS m/z 276 (M+, 66), 274
(M+ - H2, 37), 194 (M+ - C6H10, 81), 165 (100), 81 (18). HR-MS
calcd mass for C20H20O 276.1514, found 276.1506. Anal. Calcd for
C20H20O: C, 86.92; H, 7.29. Found: C, 86.73; H, 7.27.
ν
max (-60 °C) 1138 (s), 1033 (s) cm-1; 1H NMR (CD2Cl2, -60 °C) δ
1.64 (dd, J ) 8.5, 1.5 Hz, H-7exo), 1.96 (dt, J ) 8.5, 1.5 Hz, H-7endo),
3.62 (dq, J ) 3.0, 1.5 Hz, H-4), 5.28 (ddt, J ) 2.5, 1.7, 1.5 Hz, H-1),
6.37 (dd, J ) 5.3, 2.5 Hz, H-6), 6.58 (ddd, J ) 5.3, 3.0, 1.7 Hz, H-5);
13C NMR (CD2Cl2, -60 °C) δ 49.78 (d, C-4), 53.80 (t, C-7), 85.62 (d,
C-1), 111.30 (s, C-8), 150.85 (s, C-3); EI-MS m/z 260 (M+, 22), 194
(M+ - C5H6, 100), 165 (92), 66 (26). HR-MS calcd mass for C19H16O
260.1202, found 260.1198. Anal. Calcd for C19H16O: C, 87.66; H, 6.19.
Found: C, 87.58; H, 6.34.
(iii) Low-Temperature Monitorings of Reaction 1. (a) NMR (1H
and 13C) Spectroscopy. The general procedure was followed, using 1
and 2 in CD2Cl2 and toluene-d8, separately. After 3 h at -20 °C, the
reaction mixture was confirmed to consist of the highest concentration
of [4 + 2] intermediate 10. The 13C NMR measurements were
performed at a probe temperature of -60 °C.62 The monitoring results
exhibited similar changes in the following molar-% ratios (10:6). Run
1 (CD2Cl2): See Figure 1a. Run 2 (toluene-d8): 4:1 (-20 °C, 0 h),
30:9 (-20 °C, 4 h), 42:18 (-20 °C, 8 h), and 90:3 (0 °C, 2 h). (b)
FT-IR Spectroscopy. The above procedure was followed, using 163a
(58 mg) and 263b (20 mg) in CH2Cl2 (0.4 mL). The measurements were
performed at -60 °C by intermittent dissolution of a small amount of
the reaction mixture in cold CH2Cl2 (selected charts are displayed in
Figure 1a).
(ii) Under Room-Temperature Conditions To Isolate [2 + 2]
Intermediate 8. The general procedure was followed, using 1 (2.50 g)
and 4 (2.21 g) in CHCl3 at 25 °C for 15 days to give a mixture of 12
(0.677 g, 19%) and 2,2-diphenyl-3-isopropenyl-3-methylcyclobutan-
1-one (8) (1.35 g, 38%), with recovery of 4. After removing volatile
material under vacuum, column chromatography [benzene-hexane (9:
1)] followed by PLC separation gave both component 12 (0.357 g,
10%) and 8 (0.998 g, 28%).68
1
(iii) H NMR Monitoring of Reaction 3. The general procedure
was followed, using 1 and 4 in CDCl3 at 40 (run 1) and 25 °C (run 2).
Monitorings of the reaction without solvent were also carried out at
40 (run 3) and 25 °C (run 4), separately. All the results demonstrated
an initial formation of [2 + 2] intermediate 8 preferentially to that of
[4 + 2] cycloadduct 12. The results exhibited similar changes in the
following molar-% ratios (8:12). Run 1 (40 °C, CDCl3): see Figure
4a. Run 2 (25 °C, CDCl3): 22:5 (5 days), 38:19 (15), 18:78 (40), and
0:100 (84). Run 3 (40 °C, no solvent): 22:12 (2 days), 18:38 (5), 2:81
(15), and 0:100 (40). Run 4 (25 °C, no solvent): 17:3 (5 days), 22:10
(10), 38:28 (25), 35:38 (30), 19:71 (50), and 4:96 (105).
3.2. Reaction 2. (i) Under Heating Conditions To Give [2 + 2]
Product 7. The general procedure was followed, using 1 (2.19 g) and
3 (1.00 g) in refluxing benzene (10 mL) at 80 °C for 2 days to give
8,8-diphenylbicyclo[4.2.0]oct-2-en-7-one (7) (2.98 g, 96%).64
(ii) Under Room-Temperature Conditions To Isolate [4 + 2]
Intermediate 11. The general procedure was followed, using 1 (2.43
g) and 3 (1.10 g) in CH2Cl2 at room temperature for 8 days to give a
mixture of 2-oxa-3,3-diphenylmethylidenebicyclo[2.2.2]oct-5-ene (11)
and 7 in the ratio of 38:62. Column chromatography [benzene-hexane
(9:1)] followed by PLC separation gave each component, 11(1.06 g,
31%) and 7 (1.99 g, 58%).65
3.4. Reaction 4. (i) Under Conditions for Long Reaction Time
To Give [4 + 2] Product 13. The general procedure was followed,
using 1 (1.18 g) and 5 (0.563 g) in CH2Cl2 or benzene at room
(66) We have carried out reaction 3 under acidic conditions in CDCl3
containing a small amount of hydrochloric acid or trifluoroacetic acid under
room light (not in the dark). After a long reaction time (1 month, room
temperature), we observed the predominant formation of the [4 + 2]
cycloadduct 12 accompanied by a small amount of its isomeric product
14.67 A similar conversion of 8 to 14 has also been confirmed under the
acidic conditions (isolated yield 92-96%). A thermal conversion of 12 f
14 has also been monitored under heating conditions (100 °C, 12 h) in a
sealed NMR tube. However, the thermal conversion 12 f 14 did not occur
1
(iii) H NMR Monitoring of Reaction 2. The general procedure
was followed, using 1 (116 mg) and 3 (48 mg) in benzene-d6 and
CDCl3, separately. The results exhibited changes in the following
molar-% ratios (11:7). Run 1 (toluene-d8, 80 °C, 52 h): see Figure 2a.
Run 2 (benzene-d6, 25 °C, 8 days): 14:32 (1 h), 23:48 (2 h), and 38:
62 (8 h). Further heating of the mixture at 80 °C for an additional day
exhibited a conversion to 7. Run 3 (CDCl3, 25 °C, 8 days): 14:32 (1
day), 23:48 (2 days), and 38:62 (8 days).
3.3. Reaction 3. (i) Under Conditions at 40 °C To Give [4 + 2]
Product 12. The general procedure was followed, using 1 (2.07 g)
and 4 (1.83 g) in CHCl3 at 40 °C for 32 days in the dark.66 Removal
of volatile material under vacuum left viscous liquid. Chromatographic
purification gave 3,4-dimethyl-6-diphenylmethylidene-5,6-dihydropyran
(12) (2.87 g, 97%) as pale yellowish oil. A similar result was obtained
using 1 and 4 without solvent at 40 °C for 40 days to give 12 (2.78 g,
94%).
in a short reaction time (100 °C, 3 h) (see also ref 59). 14: colorless needles,
mp 78.5-79.5 °C [pentane-ether (9:1)]; Rf ) 0.31 (CCl4); IR νmax 1635
(s), 1584 (vs), 1227 (vs), 760 (vs), 698 (vs) cm-1; 1H NMR δ 1.12 (d, 3 H,
J ) 7.1 Hz, 5-CH3), 1.76 (t, 3 H, J ) 1.6 Hz, 4-CH3), 2.26 (qddqd, 1 H,
J ) 7.1, 4.7, 3.9, 1.2, 1.0 Hz, H-5), 3.84 (dd, 1 H, J ) 10.4, 4.7 Hz, H-6),
4.07 (dd, 1 H, J ) 10.4, 3.9 Hz, H-6), 5.89 (qd, 1 H, J ) 1.2, 1.0 Hz, H-3);
13C NMR δ 15.17 (q, 5-CH3), 21.31 (q, 4-CH3), 33.43 (d, C-5), 70.76 (t,
C-6), 117.72 (d, C-3), 118.35 (s, C-7), 140.26 (s, C-4), 147.25 (s, C-2);
EI-MS m/z 276 (M+, 100%), 261 (17), 194 (13), 166 (23), 165 (30), 109
(14), 81 (12), 77 (10). Anal. Calcd for C20H20O: C, 86.92; H, 7.29. Found:
C, 86.75; H, 7.29.
12: colorless liquid [colorless cubes, mp 10-12 °C (cold pentane-
ether)]; Rf ) 0.44; IR νmax (neat) 1242 (s), 1028 (s) cm-1; 1H NMR δ
1.60 (dt, 3 H, J ) 1.6, 0.9 Hz, 4-CH3), 1.63 (dt, 3 H, J ) 1.3, 1.0 Hz,
3-CH3), 2.87 (ddd, 1 H, J ) 1.9, 1.3, 0.9 Hz, H-5b), 2.88 (dd, 1 H, J
) 2.3, 0.9 Hz, H-5a), 4.35 (ddd, 1 H, J ) 1.9, 1.6, 1.0 Hz, H-2b), 4.36
(67) Compound 14 is also obtained from the reaction of 1 with 4 under
photoirradiation (Otto, P. Ph.D. Dissertation, University of Munich, 1970).
We thank Prof. Rolf Huisgen for his kind information.
(68) Cyclobutanone 8: colorless prisms (cold MeOH or ether-pentane),
mp 81-82 °C (lit.9b,41 mp 118-119 °C, lit.15b mp 117-118 °C); Rf ) 0.32;
IR νmax (KBr) 1763 (vs) cm-1; 1H NMR δ 1.23 (d, 3 H, J ) 1.0 Hz, 3-CH3),
1.71 (d, 3 H, J ) 1.6 Hz, dC-CH3), 2.85 (d, J ) 17.2 Hz, H-4b), 3.64
(dd, J ) 17.2, 1.0 Hz, H-4a), 5.05 (br s, 1 H, H-6d), 5.06 (q, J ) 1.6 Hz,
H-6c); 13C NMR δ 21.64 (q, dC-CH3), 28.31 (q, 3-CH3), 45.65 (s, C-3),
55.27 (t, C-4), 79.64 (s, C-2), 113.45 (t, C-6), 148.78 (s, C-5), 208.22 (s,
CdO). The IR (KBr; see Table 5 in Supporting Information) and 1H NMR
data of the [2 + 2] cycloadduct 8 in the literature15b are identical with those
obtained by the present work. However, the melting point data (82 °C, from
cold MeOH or ether-pentane) from the present work are different from
those (118 °C, from ether-petroleum ether) in the literature.9b,15b,41 The
reason is not clear. Regarding this discrepancy, we have repeatedly examined
reaction 3 with special care in two research groups (Saitama University
and Sagami Chemical Research Center) independently. The IR spectrum
and the melting point data of cyclobutanone 8 (mp 82-83 °C, measured
on a Yanako MP-3 micro-melting point apparatus) obtained at the latter
group were identical with those at the former group. By our hands, crystals
of cyclobutanone 8 were found to show neither polymorphism nor formation
of any hydrates.
(61) Recrystallization from cold pentane-ether gave pure crystals with
the melting point. The crystals thus obtained remained unchanged at least
over half a year at -80 °C. On standing at 0 °C below the melting
temperature, however, the crystals of 10 were converted spontaneously to
a solid yielding exclusively a [2 + 2]-type cycloadduct 6 (vii in Scheme
6).
(62) No change of the reaction was confirmed by 1H NMR spectroscopy
before and after the 13C NMR measurement.
(63) (a) 1: IR (CH2Cl2) νmax (CdO) 2096 (vs) cm-1. (b) 2: IR (CH2-
Cl2) νmax (CdC) 1632 (vs), 1597 (vs) cm-1
.
(64) Cyclobutanone 7: colorless prisms (MeOH), mp 132-133 °C; Rf
) 0.38; IR νmax 1770 (vs) cm-1; detailed NMR (1H and 13C) and MS data
are in the Supporting Information.
(65) Dihydropyran 11: colorless prisms [MeOH-ether (5:1)], mp 115-
116 °C; Rf ) 0.42; IR νmax (KBr) 1208 (vs), 1018 (vs) cm-1; detailed NMR
(1H and 13C) and MS data are in the Supporting Information. Anal. Calcd
for C20H18O: C, 87.56; H, 6.61. Found: C, 87.47; H, 6.62.

Detection of Diels-Alder Cycloadducts by NMR
J. Am. Chem. Soc., Vol. 121, No. 20, 1999 4785
temperature for 20 or 30 days, respectively, to give 2-methoxy-6-
diphenylmethylidene-2,5-dihydropyran (13) (1.66 g, 98%) as pale
yellowish oil.
Retrogression. A solution of pure isolated sample of 10 (ca. 4 mg) in
0.05 mL of CD2Cl2 was dissolved.69 The reaction monitoring was
performed at the temperature between -20 and 0 °C. The Claisen
rearrangement was observed predominantly by a formation of [2 + 2]
product 6 with a slight dissociation to the reactants, 1 and 2, in the
maximum molar concentration of 16% (0 °C, 2 h). The result is
displayed in Figure 1b [t1/2(0 °C, CD2Cl2) ) 40 min].
4.2. In Reaction 2. (i) Under Preparative Conditions. The general
procedure was followed, using 11 (274 mg) in 20 mL of benzene. The
mixture was refluxed at 80 °C for 2 days until the TLC spot of 11
indicated complete disappearance. Solvent removal followed by short
column chromatography gave [2 + 2] product 7 (263 mg, 96%). (ii)
1H NMR Monitoring with a Slight Retrogression. The general
procedure was followed, using 11 in CDCl3 at 25 and 40 °C, separately,
for 24 h or in benzene-d6 (60 °C, 24 h). Any conversion was not
observed by 1H NMR monitoring. However, a solution of 11 in
benzene-d6 (80 °C, 10 h) in a sealed NMR tube indicated a slight
retrogression (11 f 1 + 3) in the maximum concentration (4%) of the
reactants, 1 + 3, with a predominant formation of the Claisen-shift
product 7 (78%) and 11 (18%) remained unchanged. Entire conversion
of the mixture to the final product 7 was achieved at 80 °C for 2 days.
The result is exhibited in Figure 2b [t1/2(80 °C) ) 4 h].
5. General Procedure for Nonoccurrence of Claisen Rearrange-
ment as Well as of Retro-Diels-Alder Dissociation in the [4 + 2]
Cycloadducts (12, 13) Originating from Open-Chain 1,3-Dienes (4,
5). (i) Under Preparative Conditions. A solution of a [4 + 2]
cycloadduct (1.0 mmol) in 10 mL of a solvent was allowed to stand in
the dark or was heated for 48 h. Solvent removal gave the cycloadducts
recovered unchanged. (ii) 1H NMR Monitoring. The monitoring was
performed using a sample of [4 + 2] cycloadduct (4 mg) at desired
temperatures in 0.4 mL of an organic solvent (benzene-d6, CDCl3 or
toluene-d8). These analytical results are displayed in the Supporting
Information.
13: colorless liquid [colorless prisms, mp 2-4 °C (cold pentane-
ether)]; Rf ) 0.40; IR νmax (neat) 1236 (s), 1012 (s) cm-1; 1H NMR δ
2.89 (dddd, 1 H, J ) 21.0, 4.4, 2.4, 1.0 Hz, H-5b), 2.96 (dddd, 1 H, J
) 21.0, 3.4, 2.4, 1.0 Hz, H-5a), 3.23 (s, 3 H, OCH3), 5.13 (dq, J )
3.1, 1.0 Hz, H-2), 5.76 (ddt, J ) 10.3, 3.1, 2.4 Hz, H-3), 5.89 (dddd,
J ) 10.3, 4.4, 3.4, 1.0 Hz, H-4); 13C NMR δ 26.95 (t, C-5), 56.14 (q,
OCH3), 98.03 (d, C-2), 123.29 (s, C-7), 124.12 (d, C-3), 127.54 (d,
C-4), 143.46 (s, C-6); EI-MS m/z 278 (M+, 18), 246 (M+ - MeOH,
81), 194 (M+ - C4H5OMe, 100), 165 (79), 84 (37); HR-MS calcd mass
for C19H18O2 278.1307, found 278.1309. Anal. Calcd for C19H18O2:
C, 81.99; H, 6.52. Found: C, 81.87; H, 6.68.
(ii) Under Conditions for Short Reaction Time To Isolate [2 +
2] Intermediate 9. The general procedure was followed, using 1 (1.22
g) and 5 (0.581 g) in CHCl3 at 0 °C for 5 days to give a mixture of 13
(30%) and 2,2-diphenyl-3-(trans-2′-methoxyvinyl)cyclobutan-1-one (9)
(70%). Column chromatography followed by PLC separation gave each
component which was collected to give 13 (454 mg, 26%) and 9 (1.08
g, 62%) as colorless crystals.
9: colorless prisms, mp 75-76 °C (cold MeOH); Rf ) 0.34; IR
1
νmax 1776 (vs) cm-1; H NMR δ 3.03 (dd, J ) 17.6, 9.0 Hz, H-4a),
3.28 (dd, J ) 17.6, 9.0 Hz, H-4b), 3.38 (s, 3 H, OCH3), 3.78 (td, J )
9.0, 8.4 Hz, H-3), 4.36 (dd, J ) 11.3, 8.4 Hz, H-5), 6.58 (d, J ) 11.3
Hz, H-6); 13C NMR δ 35.20 (d, C-3), 50.73 (t, C-4), 55.75 (q, OMe),
78.11 (s, C-2), 103.40 (d, C-5), 147.77 (d, C-6), 209.13 (s, CdO);
EI-MS m/z 278 (M+, 8), 246 (25), 194 (100), 165 (66), 84 (8), 77 (4),
69 (4). Quantitative difference NOE experiment: irr δ 3.79 (H-3),
enhanced δ 3.28 (19%, H-4b); irr δ 6.68 (H-6), enhanced δ 3.08 (16%,
H-4a); irr δ 7.45 (o-Ph), enhanced δ 4.30 (13%, H-5). Anal. Calcd for
C19H18O2: C, 81.99; H, 6.52. Found: C, 81.81; H, 6.64.
1
(iii) H NMR Monitoring of Reaction 4. The general procedure
was followed, using 1 (102 mg) and 5 (44.2 mg) at 25 °C in benzene-
d6 (run 1) and at 0-25 °C in CDCl3 (run 2), separately. The latter
sample at 0 °C was then warmed to 25 °C for an additional 15 days
until the completion of the reaction. The results exhibited the changes
in the following molar-% ratios (9:13). Run 1 (25 °C, benzene-d6):
see Figure 5a. Run 2 (0-25 °C, CDCl3): 22:3 (0 °C, 2 h), 52:21 (0
°C, 1 day), 30:70 (0 °C, 5 days, entire consumption of the reactants),
32:68 (0 °C, 7 days), 62:38 (25 °C, 1 day), 8:92 (25 °C, 4 days), 0:100
(25 °C, 8 days).
5.1. In Reaction 3. (i) Under Preparative Conditions. The general
procedure was followed, using 12 at 40 °C for 2 days in CHCl3 and at
60 °C for 2 days in benzene, separately. Each sample was recovered
unchanged. (ii) 1H NMR Monitoring. The general procedure was
followed, using 12 under two conditions (40 °C, CDCl3, 2 days; 60
°C, benzene-d6, 2 days). Any conversions of 12 were not observed
throughout the monitorings.
5.2. In Reaction 4. (i) Under Preparative Conditions. The general
procedure was followed, using 13 in CHCl3 (40 °C) and benzene (60
°C), separately, for 2 days. Each sample was recovered unchanged.
(ii) 1H NMR Monitoring. The general procedure was followed, using
13 under the above conditions (40 °C, CDCl3, 2 days; 60 °C, benzene-
d6, 2 days). No conversions were observed throughout the monitorings.
4. General Procedure for Claisen Rearrangement of the [4 + 2]
Cycloadducts (10, 11) Originating from Cyclic 1,3-Dienes (2, 3) to
Their [2 + 2] Products (6, 7). (i) Under Preparative Conditions.
Crystalline sample or a solution of a [4 + 2] cycloadduct (1.0 mmol)
in 10 mL of a solvent was allowed to stand in the dark or was heated
6. General Procedure for Retro-Claisen Rearrangement of the
[2 + 2] Intermediates (8, 9) Originating from Open-Chain 1,3-
Dienes. (i) Under Preparative Conditions. A solution of the [2 + 2]
cycloadduct (100 mg) in 5 mL of a solvent was left to stand at room
temperature or was heated at 40 °C until the TLC indicated complete
disappearance of the starting material. Solvent removal gave the [4 +
2] cycloadducts, 12 and 13, quantitatively. The samples obtained were
identified by comparison of their 1H NMR data with those of authentic
samples. (ii) 1H NMR Monitoring. A solution of 4 mg of the [2 + 2]
intermediate (8, 9) in 0.40 mL of a solvent (CD2Cl2 or benzene-d6) in
an NMR tube was allowed to stand at 25 ( 1 °C in 0.4 mL of CDCl3
and benzene-d6, separately, in the dark. The monitorings of the reactions
1
until the H NMR indicated disappearance of the signals of the [4 +
2] cycloadduct. Solvent removal gave a [2 + 2]-type cycloadduct. The
1
samples obtained were identified by comparison of IR and H NMR
data with those of authentic samples.
1
(ii) H NMR Monitoring of the Claisen Reaction with a Slight
Retro-Diels-Alder Dissociation. A solution of 10 mg of the [4 + 2]
intermediate (10, 11) in 0.40 mL of solvent (benzene-d6 or CD2Cl2) in
an NMR tube was allowed to stand at various temperatures in the dark.
The monitoring of the reactions was performed periodically at the
desired temperature by 1H NMR. These analytical results are exhibited
in the Supporting Information.
4.1. In Reaction 1. (i) Under Preparative Conditions. (a) Solid-
State Conversion. The general procedure was followed, using pure
isolated crystals of 10 (65 mg) at 0 °C for 2.5 days to give
cyclobutanone 6 quantitatively. The resulting solid samples were
observed to be transformed completely to 6 by 1H NMR. (b) Solution-
State Conversion. The general procedure was followed, using 10 (130
1
were performed periodically at this temperature by H NMR. These
analytical results are exhibited in the Supporting Information.
6.1. In Reaction 3. (i) Under Preparative Conditions. The general
procedure was followed, using 8 in benzene at 40 °C for 17 days to
give 12 in 96% yield. A similar experiment in CHCl3 at 40 °C for 12
days gave 12 quantitatively. (ii) 1H NMR Monitoring. The NMR
monitoring confirmed that cyclobutanone 8 was converted to an
equilibrium mixture of 8 and 12 after 15 days at 25 °C in the ratio of
1
mg) in CHCl3 at 0 °C for 5 h to give 6 (127 mg, 98%). (ii) H NMR
Monitorings. (a) Solid-State Conversion. Pure isolated crystals of
10 were left at 0 °C in the dark. Periodical monitoring of the reaction
with 1H NMR (measured at -40 °C) by intermittent dissolution of the
solid sample exhibited a complete conversion to cyclobutanone 6
without any sign of the retrogression to the reactants, 1 and 2 (Figure
1c) [t1/2(0 °C) ) ca. 6 h]. (b) Solution-State Conversion with a Slight
(69) The measurements were performed using a 2.0-mm (i.d.) capillary
(Japan Precision Instrument, N-502A), which was loaded inside of a 5φ-
NMR tube (N-5P), to keep a concentration similar to that under preparative
conditions.

4786 J. Am. Chem. Soc., Vol. 121, No. 20, 1999
Machiguchi et al.
71:29. Heating of the sample at 40 °C for an additional 12 days
indicated the complete conversion to 12.
ꢀ of CHCl3 was taken to be 6.0. Thermal corrections and entropies
were evaluated at two temperatures, T ) 200 and 300 K. These
temperatures follow nearly the experimental conditions (the monitoring
of reactions between -80 and 20 °C). Gibbs free-energy (G) diagrams
were drawn at those two temperature. Thus, the energy G is composed
of the MP2/6-31G* electronic energy, RHF/3-21G SCRF and thermal
corrections, and RHF/3-21G entropies. All the calculations were carried
out, using the GAUSSIAN94 (G94)71 program. G94 was installed both
at CONVEX spp1200/XA (the Information Processing Center, Nara
University of Education) and at CONVEX spp1600/XA (Computer
Center, Nara University).
6.2. In Reaction 4. (i) Under Preparative Conditions. The general
procedure was followed, using 9 (60 mg) in CHCl3 for 16 days or in
benzene at room temperature for 24 days to give 13 in 98 and 94%
yield, respectively. (ii) 1H NMR Monitoring. The 1H NMR monitoring
result (25 °C, CDCl3 or benzene-d6) indicated the retro-Claisen
rearrangement to the [4 + 2] adduct 13 in the ratios of 81:19 (17 h),
74:26 (24 h), 46:54 (2 days), 14:86 (6 days), and 0:100 (16 days).
Likewise, a monitoring in benzene-d6 demonstrated the conversion in
the ratio of 82:18 (1 day), 23:77 (6 days), 6:94 (14 days), and 0:100
(24 days).
Acknowledgment. Dedicated to Professor Emeritus Rolf
Huisgen, Universita¨t Mu¨nchen, for his life-long career in and
brilliant contributions to organic chemistry. For his biography,
see ref 1. The authors gratefully thank Professor Rolf Huisgen,
who kindly disclosed unpublished experimental data related to
the present work, for his continuous encouragement, helpful
suggestions, and stimulating discussions. This work was sup-
ported by a Grant-in-Aid for Scientific Research from the
Ministry of Education, Science and Culture, Japan.
7. General Procedure for Nonoccurrence of Retro-Claisen
Rearrangement of [2 + 2] Cycloadducts (6, 7) Originating from
Cyclic 1,3-Dienes (2, 3). (i) Under Preparative Conditions. A solution
of [2 + 2] cycloadduct (100 mg) in 5 mL of a solvent was heated at
the desired temperature. Solvent removal left solid which was identified
to remain unchanged by 1H NMR with those of authentic samples. (ii)
1H NMR Monitoring. A solution of 4 mg of [2 + 2] intermediate (6,
7) in 0.40 mL of a solvent (CD2Cl2 or benzene-d6) in an NMR tube
was heated in 0.4 mL of CDCl3 and benzene-d6, separately, in the dark.
The monitorings of the reactions were performed periodically. These
analytical results and the representative charts are exhibited in the
Supporting Information.
7.1. In Reaction 1. (i) Under Preparative Conditions. The general
procedure was followed, using 6 (100 mg) in CHCl3 and benzene,
separately, at 25 °C for 15 days. Each sample used was recovered
unchanged quantitatively. (ii) 1H NMR Monitoring. The general
procedure was followed, using 6 (100 mg) in CDCl3 or benzene-d6.
No conversion was observed throughout the monitorings.
Supporting Information Available: List of experiments,
1
tables of full spectral data [IR, MS, and NMR (13C and H)],
selected NMR spectral charts [400-MHz 1D 1H NMR (ordinary
Lorentz-transformed spectra and those resolution-enhanced with
1
1
sine-bell wind function), 2D NMR H-1H COSY and H-1H
COLOC, 100.6-MHz 1D 13C NMR with CPD, gated decoupling,
and 2D NMR 13C-1H COSY and 13C-1H COLOC spectra]
for the reported compounds 6-13, plots of reaction monitoring
experiments, Cartesian coordinates (Chem3D drawing input
data) of optimized TS geometries shown in Figures 6-10, and
tables of energetics (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
7.2. In Reaction 2. (i) Under Preparative Conditions. The general
procedure was followed, using 7 (100 mg) in CHCl3 (40 °C, 15 days)
and benzene (80 °C, 2 days), separately. Each sample used was
recovered unchanged. (ii) 1H NMR Monitoring. The general procedure
was followed, using 7 (100 mg) under the above conditions (40 °C,
CDCl3, 15 days, 80 °C, benzene-d6, 2 days). No conversion was
observed throughout the monitorings.
JA990072U
8. Computational Methods. Geometries of reactants (1-5), inter-
mediates and products (6-13), and transition states were optimized
with the RHF/3-21G method. RHF/3-21G vibrational analyses were
made to judge whether the optimized geometries were at stable points
or at saddle points. Electronic energies were obtained with single-point
calculations of the MP2/6-31G* (frozen core) method on the RHF/3-
21G geometries, MP2/6-31G*//RHF/3-21G. The solvent (chloroform,
CHCl3) effect is included by the self-consistent reaction field (SCRF)
method of the Onsager reaction field model.70 The SCRF correction
was added to the MP2/6-31G* electronic energy. The dielectric constant
(70) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486-1493.
(71) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, 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 94, ReVision C.4;
Gaussian, Inc.: Pittsburgh, PA, 1995.