Lewis acid-promoted ketene-alkene [2 + 2] cycloadditions

Article abstract of DOI:10.1021/ja3103007

Described are the first examples of ketene-alkene [2 + 2] cycloadditions promoted by Lewis acids. Notable features of this method include (1) substantial rate acceleration relative to traditional thermal reactions, (2) good diastereoselectivities and yields for the formation of the cyclobutanone products, and (3) inverse diastereoselectivity compared with related thermal cycloadditions for many examples. These studies not only provide access to synthetically versatile cyclobutanones that cannot be prepared by traditional thermal cycloadditions but also address important mechanistic questions regarding ketene-alkene [2 + 2] cycloaddition reactions.

Full text of DOI:10.1021/ja3103007

Communication
pubs.acs.org/JACS
Lewis Acid-Promoted Ketene−Alkene [2 + 2] Cycloadditions
Christopher M. Rasik and M. Kevin Brown*
Department of Chemistry, Indiana University, 800 East Kirkwood Avenue, Bloomington, Indiana 47401, United States
S
* Supporting Information
a
Table 1. Initial Evaluation of Reaction Conditions
ABSTRACT: Described are the first examples of ketene−
alkene [2 + 2] cycloadditions promoted by Lewis acids.
Notable features of this method include (1) substantial
rate acceleration relative to traditional thermal reactions,
(2) good diastereoselectivities and yields for the formation
of the cyclobutanone products, and (3) inverse diaster-
eoselectivity compared with related thermal cycloadditions
for many examples. These studies not only provide access
to synthetically versatile cyclobutanones that cannot be
prepared by traditional thermal cycloadditions but also
address important mechanistic questions regarding
ketene−alkene [2 + 2] cycloaddition reactions.
temp
time
(h)
equiv
yield
b
c
entry
(°C)
of 3
Lewis acid
(%)
dr
1
2
3
120
180
48
48
1
10
10
2
−
−
<2
∼10
70
−
d
∼1:1
13:1
−78
EtAlCl2
(2.5 equiv)
a
b
c
See the Supporting Information (SI) for experimental details.
Determined by 1H NMR analysis with an internal standard.
d
Determined by 1H NMR analysis. 1H NMR analysis of the
etene−alkene [2 + 2] cycloadditions are an important
class of transformations for chemical synthesis.1,2 This is
largely due to the synthetic utility of the resultant cyclo-
butanone products and is evidenced by the numerous reports
regarding their use in total synthesis.3,4 Many modifications and
improvements introduced over the last several decades have
extended the versatility of these cycloadditions.1 Despite such
advances, significant limitations still remain in regard to the
range of ketenes and alkenes capable of undergoing [2 + 2]
cycloadditions. For example, current methods for thermal
ketene−alkene [2 + 2] cycloadditions between relatively
unreactive reaction partners typically require forcing conditions
(>120 °C, >24 h, large excess of alkene or ketene) that at best
result in low isolated yields of the desired products (eq 1 and
unpurified reaction mixture showed a complex mixture of products.
K
Nearly all ketene−alkene [2 + 2] cycloadditions rely solely
on the inherent reactivity of the reacting partners. Modulation
of this reactivity can be achieved by variation of the reaction
solvent7 and/or temperature8 with moderate degrees of success.
In contrast, reagent- or catalyst-controlled variants of these
reactions have largely been neglected in the chemical literature.
The introduction of such an advance holds the potential to
transform the way ketene−alkene [2 + 2] cycloadditions are
carried out and could significantly expand the scope of the
reaction. To the best of our knowledge, only one example
regarding the use of reagent or catalyst control in a ketene−
alkene [2 + 2] cycloaddition has been reported. Dickinson and
co-workers demonstrated that Pd(II) salts [although Pd(0) was
implicated as the active catalyst] can slightly influence the
diastereoselectivity and improve the yield of ketene−alkene [2
+ 2] cycloadditions relative to thermal reactions.9 However, the
method appears to be limited to reactions between very specific
n-alkyl/bromo-substituted ketenes and cyclopentadiene and
provides access to cyclobutanones that can be prepared readily
under traditional thermal conditions.
Herein we introduce a new paradigm in ketene−alkene [2 +
2] cycloadditions that relies on the use of Lewis acid-based
reagent control (eq 2). This method is notable for not only the
high yields and diastereoselectivities obtained for the cyclo-
butanone product but also the inverse selectivity relative to
traditional thermal cycloadditions observed for many ketene−
alkene cycloadditions. Furthermore, this method provides
access to synthetically versatile, highly substituted cyclo-
butanones that are difficult to access through current methods.4
Table 1, entries 1 and 2).5 Such reaction conditions are likely
detrimental in cases where diastereomers are produced,
ultimately leading to poor control of the selectivity.5b As part
of a program aimed at developing diastereo- and enantiose-
lective reactions involving allenes and heteroallenes,6 we
became interested in extending the utility of ketene−alkene
[2 + 2] cycloadditions by specifically addressing some of the
deficiencies outlined above.
Received: October 18, 2012
Published: January 28, 2013
© 2013 American Chemical Society
1673
dx.doi.org/10.1021/ja3103007 | J. Am. Chem. Soc. 2013, 135, 1673−1676

Journal of the American Chemical Society
Communication
a
Contemporary mechanistic data regarding thermal ketene−
alkene [2 + 2] cycloadditions point toward a reaction that is
likely either stepwise, with involvement of a dipolar
intermediate, or concerted yet highly asynchronous.1,2,10,11
On the basis of these potential reaction pathways, it seemed
plausible that Lewis acids could accelerate the cyclo-
addition.12−14 We envisioned that interaction of a Lewis acid
with the electrophilic ketene would effectively lower the energy
of the lowest unoccupied molecular orbital and render these
adducts more reactive.15,16 To the best of our knowledge,
definitive Lewis acid acceleration of ketene−alkene [2 + 2]
cycloadditions has not previously been established.17
Chart 1. Substrate Scope
To initiate our studies, the cycloaddition of ketene 2
(generated in situ by treatment of acid chloride 1 with Et3N)
and cyclohexene (3) was examined. In the traditional thermal
approach, forcing conditions (180 °C, 48 h, 10 equiv of alkene)
were needed to generate cyclobutanone 4, albeit in low yield
(∼10%) and with poor diastereoselectivity (∼1:1) (Table 1,
entry 2). The reaction in the presence of the Lewis acid
ethylaluminum dichloride (EtAlCl2, 5) gave rise to a substantial
rate enhancement, resulting in the formation of 4 in <1 h at
−78 °C in 70% yield with 13:1 dr using only 2.0 equiv of alkene
(Table 1, entry 3). Several points regarding this result warrant
further mention: (1) Al-based Lewis acids are uniquely effective
at promoting this reaction.18 (2) At present, excess Lewis acid
(2.5 equiv) is required because of product inhibition18 and
unavoidable reaction with Et3N·HCl (a byproduct of ketene
generation). Fortunately, 5 is inexpensive, readily available, and
easy to use. (3) Ketene 2 can be prepared and used in situ
(without removal of Et3N·HCl). This procedure avoids the
cumbersome isolation of the ketene.19
The scope of this process with respect to both the ketene and
the alkene was examined (Chart 1). Several points regarding
the ketene substrate scope are noteworthy: (1) Diaryl, dialkyl,
and alkyl/aryl ketenes undergo the reaction to provide the
desired products in good to moderate yields and diaster-
eoselectivities. (2) Reactions with monosubstituted ketenes do
not provide any desired product, likely because of the instability
associated with these highly reactive ketenes. (3) The reaction
of 3 with Me/Ph ketene is less selective (2:1 dr, 6) than the
reactions with Et/Ph (13:1 dr, 4), Bn/Ph (8:1 dr, 7), and i-Pr/
Ph (9:1 dr, 8) ketenes. (4) Electron-rich (e.g., 10), electron-
deficient (e.g., 11), and sterically hindered aryl ketenes (e.g., 9)
undergo the reaction in good yield and selectivity. (5) Because
of difficulties associated with the in situ preparation of i-Pr/Ph
ketene, 2-ClC6H4/Et ketene, and all of the dialkyl ketenes,
these were isolated and purified prior to use.18 (6) The
synthesis of 4 can be carried out on a gram scale with only 1.1
equiv of nondistilled 3 in 85% yield with 13:1 dr.
A wide range of cyclic disubstituted alkenes and dienes
underwent the reaction to provide the cyclobutanone products
(13−21 and 24−27) with uniformly good diastereoselectivities
(>7:1) and yields (>60%) (Chart 1). A limitation of this
method is that reactions with trisubstituted alkenes gave
product yields of <5% under a variety of conditions. This is
likely due to the severe steric interaction that results when
vicinal quaternary carbons are generated. Reactions with
terminal alkenes led to the formation of 22 and 23 in good
to moderate yields, but the diastereoselectivities of the reactions
suffered (2:1−3:1). Acyclic disubstituted alkenes are also
suitable substrates for this process (eqs 3 and 4). For example,
the reaction of diphenyl ketene (28) with cis-4-octene (29) led
to the formation of both 30 and 31 (7:1) (eq 3), while the
a
See the SI for experimental details. Isolated yields are shown, unless
otherwise indicated. Diastereomeric ratios (dr) were determined by 1H
b
NMR analysis of the unpurified reaction mixtures. Determined by 1H
NMR analysis of the unpurified reaction mixture with an internal
c
standard. Regioisomeric ratio (rr) determined by 1H NMR analysis of
d
the unpurified reaction mixtures. Formation of ∼15% of the
regioisomeric product was observed (see the SI for details). Reaction
was kept at −78 °C for 3 h. The corresponding ketene was isolated
and purified prior to use.
e
f
reaction with trans-4-octene (32) generated 31 exclusively
(>20:1) (eq 4). The formation of trans-cyclobutanone 31 by
the reaction with cis-alkene 29 (eq 3) may be the result of a
Lewis acid-promoted E/Z alkene isomerization or product
equilibration rather than a stepwise cycloaddition reaction.18
Finally, the regioselectivity of cycloadditions with unsym-
metrical alkenes can be controlled not only through electronic
activation (e.g., cyclopentadiene) but also with sterics. For
example, the synthesis of 15 under Lewis acid-promoted
conditions resulted in the formation of a single regioisomer
1674
dx.doi.org/10.1021/ja3103007 | J. Am. Chem. Soc. 2013, 135, 1673−1676

Journal of the American Chemical Society
Communication
Scheme 1. Model for Reversal of Diastereoselectivity in the
Lewis Acid-Promoted Reaction
(>20:1). Thermal cycloaddition to access 15 requires heating of
the corresponding alkene and 28 for 10 days at 100 °C and
delivers the product in 90% yield as a 6:1 mixture of
regioisomers.20
A notable aspect of this study is the inversion of
diastereoselectivity for the Lewis acid-promoted reactions of
alkyl/aryl ketenes in comparison with the thermal reactions (eqs
5 and 6).21 When cyclopentadiene (34) was allowed to react
ethyl substituent (relative to phenyl) generates dipolar
intermediate 38. Direct ring closure of this intermediate gives
rise to the observed endo-Ph diastereomer 37.1c,d,f
The Lewis acid-mediated ketene−alkene [2 + 2] cyclo-
addition likely is mechanistically similar to the thermal process,
with the primary difference being that the alkene now adds to a
Lewis acid-activated ketene (Scheme 1b). The observed
inversion in diastereoselectivity relative to the thermal reaction
may be the result of a “syn-to-Ph” approach of the alkene
(Scheme 1b). A necessary requirement of this model is that the
Ph group must rotate out of conjugation to allow for approach
of the alkene. To test this hypothesis, we carried out the Lewis
acid-promoted reaction with an aryl ketene incapable of out-of-
plane rotation (i.e., 40), which would block the approach of the
alkene syn to the aryl group and favor the formation of the
endo-aryl diastereomer (Scheme 1c). Indeed, the Lewis acid-
promoted [2 + 2] cycloaddition of 40 and 3 generated endo-aryl
isomer 41 with good diastereoselectivity (10:1), thus
supporting the model proposed in Scheme 1b. While this
model is certainly consistent with the data, the mechanistic
subtleties of this process are likely complex because of the
possibility of bifurcated transition states11e,f and C2-bound
Lewis acid−ketene complexes.16 Further studies will be
necessary to delineate fully these interesting possibilities.
In summary, the first definitive examples of a Lewis acid-
promoted ketene−alkene [2 + 2] cycloaddition reaction have
been demonstrated. We have discovered that this reaction not
only results in improved yields and diastereoselectivities but
also furnishes inverse selectivity relative to thermal reactions for
many of the substrates investigated. Furthermore, these studies
have established that rate acceleration of a ketene−alkene [2 +
2] cycloaddition is possible and sets the stage for the
development of related catalytic as well as catalytic
enantioselective variants. Studies along these lines are currently
in progress.
with ketene 33 at 22 °C, a 1:6 mixture of diastereomers 20 and
35 was generated (eq 5). With most ketene−alkene [2 + 2]
cycloaddition reactions, the diastereomer in which the larger
group is in the endo position is preferentially generated.1
Remarkably, under Lewis acid-promoted conditions, the opposite
diastereomer 20 was generated with good selectivity (7:1 dr) (eq
5).22 This behavior was not limited to reactions of dienes but
was observed for all of the “activated” alkenes23 investigated.18
For example, the Lewis acid-promoted cycloaddition of ketene
2 and indene (36) resulted in the formation of 21 in reasonable
yield (65%) with excellent diastereoselectivity (13:1 dr) (eq 6).
The corresponding thermal reaction provides the opposite
diastereomer 37 as the major product, albeit with low
diastereoselectivity (1:2 dr) (eq 6).7
A rationale that accounts for the reversal of diastereose-
lectivity is presented in Scheme 1. Thermal ketene−alkene [2 +
2] cycloadditions with “activated” alkenes23 likely proceed by a
stepwise mechanism, and the diastereoselectivity of such
processes has been explained according to the model illustrated
in Scheme 1a.1c,d,f Addition of the alkene orthogonal to the
C1−C2 π system of ketene 2 and syn to the sterically smaller
1675
dx.doi.org/10.1021/ja3103007 | J. Am. Chem. Soc. 2013, 135, 1673−1676

Journal of the American Chemical Society
Communication
from those presented here. For a review, see: (a) Orr, R. K.; Calter, M.
A. Tetrahedron 2003, 59, 3545. For a recent report regarding the
mechanism of ketene−aldehyde cycloaddition reactions, see: (b) Sin-
gleton, D. A.; Yang, H. W.; Romo, D. R. Angew. Chem., Int. Ed. 2002,
45, 1572.
(13) For Lewis acid-catalyzed [2 + 2] cycloadditions of allenic esters
(or ketones) and alkenes, see: (a) Snider, B. B.; Spindell, D. K. J. Org.
Chem. 1980, 45, 5017. (b) Snider, B. B.; Ron, E. J. Org. Chem. 1986,
51, 3643. (c) Zhao, J.-F.; Loh, T.-P. Angew. Chem., Int. Ed. 2009, 48,
7232.
(14) Lewis acid-promoted formal [2 + 2] cycloadditions between
electron-deficient and electron-rich alkenes are known. For example,
see: (a) Boxer, M. B.; Yamamoto, H. Org. Lett. 2005, 7, 3127.
(b) Canales, E.; Corey, E. J. J. Am. Chem. Soc. 2007, 129, 12686.
(15) Yates, P.; Eaton, P. J. Am. Chem. Soc. 1960, 82, 4436.
(16) (a) Prakash, G. K. S.; Bae, C.; Rasul, G.; Olah, G. A. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 6251. (b) Gronert, S.; Keefe, J. R. J. Org.
Chem. 2007, 72, 6343.
(17) Brady and Lloyd suggested that a Lewis acid might accelerate
one particular ketene−alkene [2 + 2] cycloaddition but did not report
the confirmation of this observation with control experiments. See:
Brady, W. T.; Lloyd, R. M. J. Org. Chem. 1980, 45, 2025.
(18) See the Supporting Information (SI) for details.
(19) For example, see: Hodous, B. L.; Fu, G. C. J. Am. Chem. Soc.
2002, 124, 10006.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures, analytical data for all compounds, and
crystallographic data (CIF). This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
brownmkb@indiana.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Indiana University for financial support. Dr. Maren
Pink (X-ray), Dr. Jonathan A. Karty and Angela M. Hanson
(mass spectrometry), and Dr. Frank Gao (NMR) are
acknowledged for their support.
REFERENCES
■
(1) For excellent reviews, see: (a) Snider, B. B. Chem. Rev. 1988, 88,
793. (b) Hyatt, J. A.; Raynolds, P. W. In Organic Reactions, Vol. 45;
Wiley: New York, 1994. (c) Tidwell, T. T. Ketenes; Wiley: New York,
1995. (d) Tidwell, T. T. Ketenes II; Wiley-Interscience: Hoboken, NJ,
2006. (e) Tidwell, T. T. Eur. J. Org. Chem. 2006, 563. (f) Three
Carbon−Heteroatom Bonds: Ketenes and Derivatives; Danheiser, R. L.,
Ed.; Science of Synthesis, Vol. 23; Thieme: Stuttgart, Germany, 2006,
(2) (a) Staudinger, H. Chem. Ber. 1905, 38, 1735. (b) Staudinger, H.
Chem. Ber. 1907, 40, 1145.
(3) For selected recent examples of ketene−alkene [2 + 2]
cycloadditions in total synthesis, see: (a) Boulton, L. T.; Brick, D.;
Fox, M. E.; Jackson, M.; Lennon, I. C.; McCague, R.; Parkin, R.;
Rhodes, D.; Ruecroft, G. Org. Process Res. Dev. 2002, 6, 138.
(b) Reisman, S. E.; Ready, J. M.; Weiss, M. M.; Hasuoka, A.; Hirata,
M.; Tamaki, K.; Ovaska, T. V.; Smith, C. J.; Wood, J. L. J. Am. Chem.
Soc. 2008, 130, 2087. (c) Wang, Q.; Chen, C. Org. Lett. 2008, 10,
1223. (d) Fracet, J.-B.; Himmelbauer, M.; Mulzer, J. Org. Lett. 2012,
14, 2195.
(20) Hassner, A.; Cory, R.; Sartoris, N. J. J. Am. Chem. Soc. 1976, 98,
7698.
(21) Lewis acid-promoted reactions with dialkyl ketenes provided the
same major diastereomers as the analogous thermal processes. The
diastereoselectivities and yields, however, were significantly higher
under the Lewis acid-promoted conditions. For example, the synthesis
of 25 under thermal conditions required heating for 48 h at 60 °C to
give a yield of 42% with 2:1 dr. See the SI for details.
(22) Control experiments confirmed the observed selectivity to be
the result of a Lewis acid-promoted cycloaddition reaction rather than
a Lewis acid-promoted isomerization between diastereomers. See the
SI for details.
(23) Here “activated” alkenes are dienes, styrenes, and allylsilanes.
(4) (a) Lee-Ruff, E.; Mladenova, G. Chem. Rev. 2003, 103, 1449.
(b) Namyslo, J. C.; Kaufmann, D. E. Chem. Rev. 2003, 103, 1485.
(5) For examples, see: (a) Martin, J. C.; Gott, P. G.; Goodlett, V. W.;
Hasek, R. H. J. Org. Chem. 1965, 30, 4175. (b) Brady, W. T.; Parry, F.
H., III; Stockton, J. D. J. Org. Chem. 1971, 36, 1486. (c) Hassner, A.;
Cory, R. M.; Sartoris, N. J. Am. Chem. Soc. 1976, 98, 7698.
(6) For enantioselective ketene−alkene [2 + 2] cycloadditions
utilizing chiral auxiliaries, see: (a) Houge, C.; Frisque-Hesbain, A. M.;
Mockel, A.; Ghosez, L. J. Am. Chem. Soc. 1982, 104, 2920. (b) Darses,
B.; Greene, A. E.; Poisson, J.-F. J. Org. Chem. 2012, 77, 1710.
(7) Brady, W. T.; Roe, R., Jr.; Hoff, E. F.; Parry, F. H., III .J. Am.
Chem. Soc. 1970, 92, 146.
(8) Al-Husaini, A. H.; Moore, H. W. J. Org. Chem. 1985, 50, 2597.
(9) Fairlamb, I. J. S.; Dickinson, J. M.; Cristea, L. M. Tetrahedron
2001, 57, 2237.
(10) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl.
1969, 8, 781.
(11) For selected recent studies, see: (a) Yamabe, S.; Dai, T.; Minato,
T.; Machiguchi, T.; Hasegawa, T. J. Am. Chem. Soc. 1996, 118, 6518.
(b) Machiguchi, T.; Hasegawa, T.; Ishiwata, A.; Terashima, S.;
Yamabe, S.; Minato, T. J. Am. Chem. Soc. 1999, 121, 4771.
(c) Machiguchi, T.; Okamoto, J.; Takachi, J.; Hasegawa, T.; Yamabe,
S.; Minato, T. J. Am. Chem. Soc. 2003, 125, 14446. (d) Belanger, G.;
Levesque, F.; Paquet, J.; Barbe, G. J. Org. Chem. 2005, 70, 291.
(e) Ussing, B. R.; Hang, C.; Singleton, D. A. J. Am. Chem. Soc. 2006,
128, 7594. (f) Gonzalez-James, O. M.; Kwan, E. E.; Singleton, D. A. J.
Am. Chem. Soc. 2012, 134, 1914.
(12) Lewis acid-catalyzed ketene−aldehyde cycloadditions are well-
established. These reactions are undoubtedly mechanistically distinct
1676
dx.doi.org/10.1021/ja3103007 | J. Am. Chem. Soc. 2013, 135, 1673−1676