Ring strain release as a strategy to enable the singlet state photodecarbonylation of crystalline 1,4-cyclobutanediones

Article abstract of DOI:10.1002/poc.1902

Challenging most chemists' intuition, highly reactive dialkyl biradicals can be reliably generated in the solid state by taking advantage of the photodecarbonylation of cyclic ketones. However, it has been shown that radical stabilizing groups with resonance-delocalizing abilities at the α-carbons of the precursor are required to facilitate the α-cleavage reaction, and that triplet state reactivity is essential to slow down the combination of the intermediate acyl-alkyl biradical back to the starting ketone. Relatively long triplet acyl-alkyl biradical lifetimes give a chance for the loss of CO to occur. Looking for additional strategies to generate transient biradicals in solids, we studied the solid state photochemistry of four aliphatic, dispiro-substituted 1,4-cyclobutandiones (1a-d) that were expected to react from the singlet state. We hypothesized that the release of ring strain from the small ring carbonyl would make the reverse acyl-alkyl combination disfavored, allowing for the loss of CO to occur efficiently and irreversibly. We report here the results of studies carried out in solution, bulk (powder) crystals, and nanocrystalline photochemistry. We have recently shown that excitation of dispirocyclohexyl-1,3-cyclobutanedione 1c led to the trapping of the intermediate oxyallyl with a half life of about 42 min. Our studies with the three other crystalline derivatives revealed that, while all react efficiently, the remarkably long lifetime of oxyallyl is unique to crystals of 1c.

Full text of DOI:10.1002/poc.1902

Short Communication
Received: 8 March 2011,
Revised: 27 May 2011,
Accepted: 31 May 2011,
Published online in Wiley Online Library: 21 July 2011
(wileyonlinelibrary.com) DOI 10.1002/poc.1902
Ring strain release as a strategy to enable the
singlet state photodecarbonylation of
crystalline 1,4-cyclobutanediones
Gregory Kuzmanicha and Miguel A. Garcia-Garibaya*
Challenging most chemists’ intuition, highly reactive dialkyl biradicals can be reliably generated in the solid state by
taking advantage of the photodecarbonylation of cyclic ketones. However, it has been shown that radical stabilizing
groups with resonance-delocalizing abilities at the a-carbons of the precursor are required to facilitate the a-cleavage
reaction, and that triplet state reactivity is essential to slow down the combination of the intermediate acyl–alkyl
biradical back to the starting ketone. Relatively long triplet acyl–alkyl biradical lifetimes give a chance for the loss of
CO to occur. Looking for additional strategies to generate transient biradicals in solids, we studied the solid state
photochemistry of four aliphatic, dispiro-substituted 1,4-cyclobutandiones (1a–d) that were expected to react from
the singlet state. We hypothesized that the release of ring strain from the small ring carbonyl would make the
reverse acyl–alkyl combination disfavored, allowing for the loss of CO to occur efficiently and irreversibly. We report
here the results of studies carried out in solution, bulk (powder) crystals, and nanocrystalline photochemistry. We
have recently shown that excitation of dispirocyclohexyl-1,3-cyclobutanedione 1c led to the trapping of the interme-
diate oxyallyl with a half life of about 42 min. Our studies with the three other crystalline derivatives revealed that,
while all react efficiently, the remarkably long lifetime of oxyallyl is unique to crystals of 1c. Copyright © 2011
John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this paper.
Keywords: crystalline photochemistry; nanocrystals; oxyallyl; photochemistry; photodecarbonylation
despite lacking the radical-stabilizing substituents that would be
needed for noncyclic aliphatic ketones. In order to confirm the
INTRODUCTION
generality of the cyclobutanedione reaction and to explore the
effects of spiro ring substitution on the singlet state photodecar-
bonylation, we synthesized four 1,3-spirocyclobutanediones
(1a–d, Scheme 1) .
The emergence of robust models of chemical reactivity in rigid
media[1–9] have led to new developments in solid state photo-
chemistry.[10–13] Although the topochemical postulate[14–16] sug-
gests that preorganization in the crystal lattice controls the
outcome of photochemical reactions in crystals, it is the intrinsic
properties of the chromophore that ultimately determine the
feasibility of a given photoreaction.[4,6,17] We have previously
suggested that Norrish Type I decarbonylation of crystalline
ketones to give radical intermediates is a stepwise process con-
tingent on: (1) a substituents that lower the bond dissociation
energy of the two a-bonds to be cleaved and (2) excited state re-
activity occurring in the triplet manifold to slow down reversible
bond formation long enough for the loss of CO to occur before
the first acyl–alkyl radical pair returns back to the starting mate-
rial.[18] Recently, we extended the scope of reactivity with exam-
ples that involve singlet state photodecarbonylations in the
crystalline solid state.[19,20] As opposed to slowing down revers-
ible bond formation by taking advantage of a spin-forbidden re-
combination, a promising route is to take advantage of ring
strain. If the first a-cleavage reaction occurs with commitment
release of ring strain, bond formation back to the starting ketone
is unlikely to occur, even in the singlet manifold. Furthermore, it
is well known that strained cyclic ketones generally undergo
decarbonylation from the singlet excited state[21,22] and we have
recently reported the photochemistry of crystalline diphenylcyclo-
propenone[19,20] and dispiro[5.1.5.1]tetradecane-8,16-dione,[23]
both of which undergo an efficient singlet state decarbonylation
During our investigation of the singlet state photodecarbony-
lation reaction of dispiro[5.1.5.1]tetradecane-8,16-dione 1c, we
discovered that bulk crystals exposed to UV light acquired a
bright blue color with a half-life of about 42min at 298K (Fig. 1). Us-
ing a combination of chemical trapping, spectroscopic tech-
niques, and theoretical calculations, we were able to assign the
species responsible for the blue color to the elusive reactive in-
termediate oxyallyl (Scheme 1). Although ubiquitous throughout
the chemical literature,[24–39] oxyallyl had eluded spectroscopic
characterization and isolation,[23] except for recent photoelec-
tron spectroscopic studies in the gas phase.[40] The spectroscopic
and computational evidence obtained on the blue transient in-
cluded matching femtosecond transient spectra measured in so-
lution that matched calculations that helped assigned it as an
open shell singlet.[23,41,42] It is worth noting that crystals of 1c
* Correspondence to: M. A. Garcia-Garibay, Department of Chemistry and
Biochemistry, University of California, Los Angeles, CA 90024-1569, U.S.A.
E-mail: mgg@chem.ucla.edu
a G. Kuzmanich, M. A. Garcia-Garibay
Department of Chemistry and Biochemistry, University of California, Los
Angeles, CA, 90024-1569, U.S.A.
J. Phys. Org. Chem. 2011, 24 883–888
Copyright © 2011 John Wiley & Sons, Ltd.

G. KUZMANICH AND M. A. GARCIA-GARIBAY
Scheme 1. Photochemistry of 1,3-spirocyclobutanediones.
trapped the tetrasubstituted oxyallyl by extending its lifetime by
about 1014 orders of magnitude relative to the solution. During
the course of examining the singlet state photodecarbonylation
of other 1,3-cyclobutanediones it was of interest to see whether
the trapping and observation of oxyallyl would be possible, even
though lifetimes greater than a few seconds would be needed.
In this paper, we report the solution and solid state photochem-
istry of four spirocyclobutanediones. Experiments carried out in
bulk crystals and in nanocrystalline suspensions showed that
the solid state photoreaction is general, but only crystals of 1c
are able to extend the lifetime of oxyallyl to the extent that
can be visible to the eye.
mixture was then refluxed for 25 h to give crude 1a–d. The latter
was purified by column chromotagraphy, (1:5 Et2O : Hex) and
recrystallized from ethanol, cyclohexane, and then a second re-
crystallization from ethanol immediately prior to photochemical
investigations, to give the final product in 45–71% yield after
recrystallization. Compound 4,11-di-t-butyl-dispiro[5.1.5.1]
tetradecane-8,16-dione 1d formed two polymorphs, a and b,
both from ethanol. A single crystal X-ray structure determination
of the a-form was possible and the crystallographic information
files were deposited in the Cambridge crystallographic data
base. In contrast, after the initial characterization and photo-
chemical experiments, the b polymorph could not be isolated
again, despite numerous crystallizations under a wide variety of
experimental conditions.
EXPERIMENTAL
Synthesis of dispirocyclobutanediones 1a-d
Quantum yield determination in the solid state
On the basis of a known procedure,[43] the appropriate cycloal-
kylcarboxylchloride (about 7mmol) was dissolved in 10mL of dry
benzene and cooled to 0ꢀC in an oven-dried 50-mL three-neck
round bottom flask fitted with a reflux condenser under Ar. After
adding 15mL of dry triethylamine over 10 min, the reaction
Relative quantum yield determinations were preformed with
dicumyl ketone (ΦÀCO=0.18) as an internal standard in a Rayonet
photochemical reactor using 312-nm lamps. Quantum yields
were determined with equimolar, optically dense suspensions.
Optically dense suspension of 1a–d and dicumyl ketone (DCK)
were prepared by the reprecipitation method.[44] Typically, 16.7
mL of a saturated acetone solution of 1a–d was injected into 3
mL of vortexing millipore purified water to give an optically
dense, but nonscattering, suspension. Independently, 21.0 mL
of a saturated acetone solution of DCK was injected into 3mL
of vortexing millipore purified water. The suspensions were then
combined immediately prior to irradiation in a 50mL quartz
Erlenmeyer flask and irradiated. Every 150 s an aliquot was re-
moved (approximately 1/20th of the suspension). The suspen-
sions were extracted with ethyl ether (1mL), washed with brine
(2 Â 0.5mL) and dried over magnesium sulfate. Samples were
then subjected to gas chromatography–mass spectrometry to
determine the extent of photoreaction and confirm product
assignments. Quantum yields were determined in a minimum
of three independent runs.
Quantum yield determination for solution
Using both optically matched (3 experiments), and equimolar (3
experiments) solutions of the cyclobutanediones 1a–d, relative
quantum yield determinations were performed with valerophe-
none (Φ=0.33) as an internal standard in a Rayonet photochem-
ical reactor using 312nm lamps (BLE-8T312).[45] Low conversions
of less than 10% were used to minimize the possible formation
of secondary photoproducts.
Figure 1. Photolysis of white 1c at 312 nm resulted in a persistent blue
color assigned to oxyallyl with a half-life of 42 min as determined by dif-
fuse reflectance ultraviolet spectroscopy (DRUV).
wileyonlinelibrary.com/journal/poc
Copyright © 2011 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 883–888

PHOTODECARBONYLATION OF CRYSTALLINE 1,3-CYCLOBUTANEDIONES
cyclohexyl ketene 2d and cyclohexylidene alkenes 4d are
formed in a ratio of approximately 1:19, by gas chromatography.
The 4-t-butyl-cyclohexenyl ketene 2d is hydrolyzed to the car-
boxylic acid during analysis and its gas chromatographic reten-
tion time matches that of the commercially available carboxylic
acid. Identification of bis-(4-t-butylcyclohexenyl) alkene 4d was
accomplished by analysis of the gas chromatography–mass
spectrometry fragmentation pattern.[50] Although the photo-
chemistry for spirodiones 1b–1d is similar, the photochemistry
of cyclobutyl 1a is unique as only the starting material and car-
boxylic acid 7 are detected during analysis of the reaction mixture.
Krapcho had previously reported that spirocyclobutanedione 1a
reacts photochemically to form lactone 6, which is assumed to
arise from a fragmentation of 1a followed by [2+2] heterocy-
cloaddition of ketene 2a (Scheme 2). Lactone 6 was previously
analyzed by strong IR stretches at 1870 and 1740cm–1. However,
attempts to detect lactone 6 by IR measurements during photol-
ysis in solution were negative. The only product detected was
carboxylic acid 7, which can be formed by ring opening hydroly-
sis of 6 followed by ketonization. Quantum yields of reaction
were also determined for 1a–d using standard valerophenone
actinometry (SI).[52] Compounds 1a and 1b have relatively low
quantum yields between 0.03–0.04. In agreement with previous
reports, the cyclohexyl diones 1c,[46,47] and 1d, undergo efficient
photoreaction, with quantum yield of decarbonylations of 0.34
and 0.33, respectively (Table 1).
Similar to the results observed in solution, photolysis of bulk
crystals at 254 or 312nm led to the formation of the
corresponding ketene and alkene, as detected by gas chroma-
tography. The use of nanocrystalline suspensions of organic crys-
tals in water provides a simple method to measure steady state
quantum yields in the solid state. Nanocrystalline suspensions
of 2,4-spiro-1,3-cyclobutanediones 1a–d were prepared by the
reprecipitation method (SI).[44] Nanocrystalline suspensions syn-
thesized in this manner can reproducible form crystals with an
average size of about 200nm. Powder X-ray diffraction patterns
confirmed that suspensions prepared in this manner are crystal-
line, and are the same polymorph as the bulk material (SI).
Dicumylketone (ΦRxn=0.18)[53] was used to determine the quan-
tum yield of reactions as both optically matched and optically dense
suspensions (SI). As shown in Table 1, the quantum yield of reac-
tion for spirocyclobutyl 1a and spirocyclopentyl 1b are similar.
The cyclobutyl 1a has a quantum yield of decarbonylation (Φ4)
of 0.08, and a quantum yield for fragmentation (Φ2) of 0.18.
The cyclopentyl-1,3-dione derivative 1b has a similar quantum
yield of decarbonylation, Φ4=0.09, but undergoes reverse cyclo-
addition less efficiently, Φ2=0.05. In contrast to 1a and 1b,
RESULTS AND DISCUSSION
Spirocyclobutanediones 1a–d were chosen as photochemical
precursors for singlet state Norrish Type I decarbonylation in
crystals because it has been reported that most undergo decarbony-
lation upon irradiation at either 254 or 312nm in solution[46]
(Scheme 1), and all have relatively high melting points (> 75ꢀC).[47]
Spirocyclobutanediones 1a–d differ by the size of the side
rings (1a–1d), as well as by the substitution patterns (1c and
1d). For the remainder of the manuscript, compounds 1a–d
will be referred by the size of the side ring (i.e., cyclobutyl 1a,
cyclopentyl 1b, cyclohexyl 1c, and t-butylcyclohexyl 1d). At the
outset, it was assumed that these structural differences should
not affect the photochemical properties of these molecules dras-
tically, but that they may affect their packing in the crystal lattice.
Spirocyclobutanediones 1a–d were synthesized in one step from
commercially available cycloalkylcarboxylic acid chlorides via
ketene dimerization in good to moderate yields (71–45%).[48]
1
13
The diones 1a–d all displayed highly symmetric H-NMR and
C-NMR spectra, as well as the characteristic cyclobutanedione
13C-NMR carbonyl peak between 210–216ppm and infrared
stretching frequencies from 1734–1736cm–1 (SI). Compounds
1a–d formed robust crystals from hexane or ethanol, which were
characterized by powder X-ray diffraction, and for 1a, 1c, and
a-1d, X-ray crystal structures were determined from suitable
crystals grown from hexane. Two polymorphic phases of t-butyl
1d, a and b, were obtained from ethanol and had different melt-
ing points, IR spectra, and X-ray powder diffraction patterns. A
structure for the a form of spiro-(t-butylcyclohexyl)-dione 1d
was obtained by single crystal X-ray diffraction analysis (SI). Ini-
tially, the polymorphs crystallized into large single crystals, which
could be manually separated. Individual crystals were then char-
acterized by IR, powder X-ray diffraction patterns, and melting
point to determine the polymorphic phase. The second poly-
morph, b, was metastable and converted entirely into the a
form. Subsequent attempts to obtain t-butyl 1d in the b form
from a variety of solvents failed, rendering it one more case of
a disappearing polymorph.[49]
Previous work by Krapcho reported that photolysis of cyclobu-
tanediones 1a–1c in dichloromethane undergo either a formal
reverse [2+2] cycloaddition to generate the ketenes 2a–c, or
two sequential decarbonylations to form alkenes 4a–4c via the
corresponding cyclopropanones 3a–3c.[47] In agreement with
Krapcho’s previous studies, diones 1a–c undergo efficient photo-
reaction upon irradiation at 312nm in dichloromethane (DCM).
The photochemical reactions were monitored by gas chroma-
tography. The products were identified by comparison to the
chromatographic properties of commercially available standards
and analyzed by mass spectrometry.[50] As shown in Scheme 1,
all compounds underwent either reverse cycloaddition to give
the corresponding ketenes 2a–c, which readily undergo hydroly-
sis to give the corresponding carboxylic acids, or sequential de-
carbonylations to give the corresponding alkenes 4a–4c. As in
previous studies, the accumulation of cyclopropanones was not
oberved, suggesting that they react faster than the cyclobutane-
diones under the reaction conditions. Although reports of oxa-
carbene formation and ring expansion during irradiation in
methanol have been reported, no other primary photoproducts
were detected during irradiation in DCM.[51] The photochemistry
of spiro-(4-t-butylcyclohexyl)-dione 1d in solution has not been
described previously, but was similar to the unsubstituted ana-
log 1c. Upon photolysis in DCM at 312nm, a mixture of
Scheme 2. Photochemistry of 1,3-spirocyclobutanedione 1a
J. Phys. Org. Chem. 2011, 24 883–888
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

G. KUZMANICH AND M. A. GARCIA-GARIBAY
not able to extend the lifetime of oxyallyl as much as crystals
of the cyclohexyl derivative 1c. To determine if oxyallyl was
formed, but was rapidly undergoing ring closure at room tem-
perature, irradiation of crystals of all spirocyclobutanediones
1a–d was performed at 77K. After confirming that crystals of
cyclohexyl 1c acquired a pale blue coloration at 77K, we found
that crystals of 1a, 1b, and a-1d remained unchanged.
Table 1. Quantum yields of retro [2+2] cyclization (Φ2) and
decarbonylation (Φ4)
Compound
DCM
Φ4
Nanocrystals
Φ2
2:4
Φ2
0.18
Φ4
2:4
1a
1b
1c
a-1d
b-1d
0.03 0.00
0.00 0.04
0.01 0.34
0.01 0.33
100:0
0.08 69:31
0.09 36:64
Given the structural and photochemical similarities of the
spirocyclohexyl derivatives 1c and 1d, it was of interest to explore
whether substantial differences in the crystalline environment
may be responsible for the different kinetic behavior of oxyallyl.
The single crystal structure of 1c, grown from hexane and solved
in a Cmca orthorhombic space group,[56] has been previously
reported.[23] A single crystal of the a polymorph of spiro-(4-t-
butylcylcohexyl) 1d was grown from hexane and solved in the
orthorhombic space group Pmna.[57] A single crystal structure
of the disappearing b polymorph could not be acquired. As
shown in Figures 2(A) and (B), the packing arrangements of 1c
and a-1d are similar with both compounds packing in alternat-
ing layers with different lateral displacements. Each dione has
a different symmetry operation relating the two cyclohexyl
rings. Whereas cyclohexyl 1c has a C2 axis perpendicular to the
cyclobutane ring, t-butylcyclohexyl 1d has a mirror plane
through the 1,3-butadione moiety. As a result of the C2 axis of
symmetry relating the spirocyclohexyl rings in 1c, to form the
cyclopropanone 3c, one of the cyclohexyl rings seems to un-
dergo approximately a 60o angular displacement, so that the
corresponding 4-carbon atom must be displaced by about
0.137nm (Fig. 3). To accomplish the ring closure, the cyclohexyl
moiety would either have to move through space occupied
by an adjacent carbonyl moiety, or wait for the neighboring
molecules to unergo larger local displacements. In contrast,
both cyclohexyl rings are oriented in the same direction in
crystals of a-(4-t-butylcyclohexyl) 1d as a result of the orienta-
tion of the two cyclohexyl moieties. We speculate that two rings
may partake in less motion to undergo ring closure so that the
resulting cyclopropane may fit better within the original cavity
of the reactant. This suggestion is supported in part by compar-
ing the packing coefficient of spirocyclohexyl 1c and a-(t-butylcy-
clohexyl) 1d, which are 0.71 and 0.67, respectively.[58] The lower
packing density of a-(t-butylcyclohexyl) 1d indicates that there is
more void space present in the crystal lattice that might accom-
modate more local displacements. We tentatively conclude that
there is enough void space and flexibility in the lattice to allow
for ring closure. A single crystal of cyclobutyl 1a was grown from
hexane and solved in the monoclinic space group P21/c.[59] One
can see that the crystalline environment of cyclobutyl 1a is
highly congested. The crystal lattice forms alternating rows of
0:100 0.05
3:97 <0.001 <0.001
3:97 ꢁ0.002 0.01 17:83
0.03
0.10 23:77
spirocyclohexyldione 1c has a very low quantum yield of reac-
tion in the solid state (Φ2+4<0.001). The a form of spiro-(4-t-
butylcyclohexyl) derivative 1d also had very low quantum yields
of reaction as a nanocrystalline suspension, with Φ4=0.01, and
Φ2ꢁ0.002. However, the b form (4-t-butylcyclohexyl) derivative
1d had a moderate quantum yield of Φ4=0.10, and Φ2=0.03.
We previously reported that irradiation of white crystals of
cyclohexyl dione 1c take on a persistent deep blue color. This
blue color is immediately quenched by dissolving the crystals
but it decays slowly under argon or oxygen atmospheres with
a half-life of about 42min in the crystal lattice.[23] The intermedi-
ate responsible for this blue color was trapped with furan in an
interfacial liquid–solid reaction to quantitatively form the tetra-
cyclic ketoether 5 (at about 6% conv.). The trapped product
was presumed to arise via a [3+4] cycloaddition between furan
and oxyallyl.[54,55] Extended irradiations (>24h) resulted in a
photobleaching of the intermediate and the only product
detected is cyclohexylidene alkene 4c. In combination with com-
putational results, the experimental results led us to conclude
that oxyallyl was efficiently trapped in the crystal lattice.[23]
The following experiments were preformed to determine
whether oxyallyl could be trapped on timescales of a few sec-
onds in the crystal lattice of the photochemical precursors.
Attempts to detect oxyallyl upon irradiation of bulk crystals of
1a, 1b, or a-1d met with no success. In all cases, irradiation of
the bulk crystals resulted in efficient singlet state decarbonyla-
tion but there was no detectable blue coloration, either visible
to the eye or by diffuse reflectance UV–Vis. To test the possibility
that colorless oxyallyl could be forming and accumulating in
crystals of 1a, 1b, and a-1d with an absorption lmax<400
nm,[23] crystals were irradiated at 298K and immediately dis-
solved in neat furan. Noting that no trapped products analogous
to the tetracyclic ketoether 5 were detected, one may conclude
that crystals of spirocyclobutanediones 1a, 1b, and a-1d are
Figure 2. Space-filling views from the crystal structures of 1c (A) and a-1d (B) showing similar packing and packing coefficients, 0.71 and 0.67, respec-
tively. The similarity of their packing motifs is highlighted by the molecules within the blue boxes (which do not correspond to the unit cells). (C) The
highly congested crystal lattice of 1a with alternating rows of 2,4-spirocyclobutyl-1,3 cyclobutadiones
wileyonlinelibrary.com/journal/poc
Copyright © 2011 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 883–888

PHOTODECARBONYLATION OF CRYSTALLINE 1,3-CYCLOBUTANEDIONES
observed by diffuse reflectance UV–Vis upon irradiation of the
mixed crystals.[62]
CONCLUSIONS
We have gathered strong photochemical evidence suggesting
that the release of ring strain in cyclobutanediones is a viable
strategy to facilitate the formation of biradicals from crystalline
ketones that are known to react primarily from the singlet ex-
cited state. Using bulk dry powders and nanocrystalline suspen-
sions, we found that all five aliphatic diones studied in this
work undergo decarbonylation (Φ-COꢁ0.01–0.2) despite lack-
ing radical stabilizing groups. We had recently reported that oxy-
allyl is efficiently trapped in the crystal lattice of dispiro[5.1.5.1]
tetradecane-8,16-dione 1c, but all attempts to generate and trap
oxyallyl in the crystal lattices of the other precursors or in other
rigid matrixes failed. We conclude that the crystal lattice of
1c is unique in its ability to stabilize and trap oxyallyl for a time
scale of several minutes and we plan to analyze the formation
of oxyallyl as a short-lived intermediate in nanocrystals of the
other precursors.
Figure 3. Hypothetical structures of oxyally derived from the X-ray crys-
tal structures of diones 1c (A) and a-1d (B) obtained by removing one of
the carbonyl groups. The energy minimized structures of the
corresponding cyclopropanone products are shown on the right. While
it appears that one of the two rings in 1c must undergo a much greater
displacement than the other, as shown by the arrows, smaller displace-
ments in a lower density crystals may be shared by the two cyclohexyl
groups in the case of a-1d
ACKNOWLEDGEMENTS
two crystallographically independent molecules (Fig. 2(C)). Cyclo-
butyl 1a is unique among the cyclobutadiones in that it has a
strong preference for retro [2+2] cycloaddition over decarbony-
lation in both solution and the crystalline environment. Compar-
ison of the ratio of retro [2+2] cycloaddition (2a) and
decarbonylation (4a) in DCM and the crystalline environment
indicates that the decarbonylation reaction is favored in the solid
state (Table 1). Inspection of the crystal structure indicates that
retro [2+2] cycloaddition may be difficult. Numerous close con-
tacts are within the crystal lattice, which may impede the
motions needed for the ring opening process in the crystal lat-
tice. This may be the reason why spirocyclobutyl 1a has a much
greater quantum yield of decarbonylation as a nanocrystalline
suspension than in solution. A single crystal for cyclopentyl 1b
could be grown from hexane, but the crystals were of poor
quality.
This work was supported by United States National Science Foun-
dation (NSF) grants DMR0605688 and CHE0844455. G.K. acknowl-
edges the integrated graduate education and research training,
material creation and training program (IGERT MCTP) grant
DGE0114443.
REFERENCES
[1] J. R. Scheffer, Acc. Chem. Res. 1980, 13, 283.
[2] M. D. Hollingsworth, J. M. McBride, Adv. Photochem. 1990, 15, 279.
[3] H. Ihmels, J. R. Scheffer, Tetrahedron 1999, 55, 885.
[4] H. E. Zimmerman, S. Zhu, J. Am. Chem. Soc. 1995, 117, 5245.
[5] M. A. Garcia-Garibay, Acc. Chem. Res. 2003, 36, 491.
[6] M. A. Garcia-Garibay, S. Shin, C. Sanrame, Tetrahedron 2000, 56,
6729.
[7] J. Hallmann, W. Morgenroth, C. Paulmann, J. Davaasambuu, Q. Kong,
M. Wulff, S. Techert, J. Am. Chem. Soc. 2009, 131, 15018.
[8] M. A. Garcia-Garibay, Photochem. Photobiol. Sci. 2010, 9, 1574.
[9] M. Bertmer, R. C. Nieuwendaal, A. B. Barnes, S. E. Hayes, J. Phys.
Chem. B 2006, 110, 6270.
It is notable that only the crystal lattice of cyclohexyl 1c stabi-
lized oxyallyl for the accumulation of a steady state concentra-
tion that is visible to the eye. Hoping to confirm that the
crystal lattice of 1c was the stabilizing force behind the trapping
of oxyallyl, we searched for polymorphs of cyclohexyl 1c to con-
firm that the long lifetime is due to crystal packing effects. Unfor-
tunately, no new polymorphs could be grown from a variety of
organic solvents or from the melt. In addition, the photolysis of
cyclohexyl 1c was attempted in other rigid media. Samples of
1c were dissolved independently in methylcyclohexane, methyl-
tetrahydrofurn, and toluene. The solutions of about 0.01M were
cooled down to 77K to form transparent glasses and irradiated
at l=312nm. No color was present in any of the glassy matrixes
during irradiation times as long as 4 h.[52] Subsequent gas chro-
matographic analysis confirmed that the reaction in the matrixes
had taken place (about 40% conversion). Dione 1c was also em-
bedded in a rigid sucrose matrix[60,61] and irradiated both at 77
and 298K (SI). There was no detectable discoloration of the su-
crose glass during irradiation at either temperature. Finally,
mixed crystals of about 1% of cyclohexyl 1c embedded within
the crystal lattice of cyclopentyl 1b and t-butylcyclohexyl 1d
were irradiated at 312nm. No purple coloration could be
[10] V. Ramamurthy, K. Venkatesan, Chem. Rev. 1987, 87, 433.
[11] J. Sankaranarayanan, L. Bort, J. A. Krause, S. M. Mandel, P. Chen,
E. E. Brooks, P. Tsang, A. D. Gudmundsdottir, Org. Lett. 2008, 10, 937.
[12] J. Harada, K. Ogawa, Chem. Soc. Rev. 2009, 28, 2244.
[13] S. Shiraki, M. A. Garcia-Garibay, in: Handbook of Synthetic Photo-
chemistry (Eds: A. Albini, M. Fagnoni), John Wiley & Sons, Hoboken,
2010.
[14] M. D. Cohen, G. M. J. Schmidt, J. Chem. Soc. 1964, 53, 1996.
[15] M. D. Cohen, G. M. J. Schmidt, F. I. Sonntag, J. Chem. Soc. 1964, 53,
2000.
[16] G. M. J. Schmidt, J. Chem. Soc. 1964, 53, 2014.
[17] L. M. Campos, M. A. Garcia-Garibay, in: Reviews of Reactive Intermedi-
ate Chemistry (Eds: M. S. Platz, M. Jones, R. Moss), John Wiley & Sons,
Hoboken, 2007.
[18] L. M. Campos, D. Ng, Z. Yang, H. Dang, H. L. Martinez, M. A. Garcia-
Garibay, J. Org. Chem. 2002, 67, 3749.
[19] G. Kuzmanich, A. Natarajan, K. K. Chin, M. Veerman, C. J. Mortko,
M. A. Garcia-Garibay, J. Am. Chem. Soc. 2008, 130, 1140.
[20] G. Kuzmanich, M. N. Gard, M. A. Garcia-Garibay, J. Am. Chem. Soc.
2009, 131, 11606.
[21] D. Weiss, in: Organic Photochemistry, Vol. 5, (Ed.: A. Padwa), Marcel
Deker: New York, 1981, 347.
[22] E. W.-G. Diau, C. Kötting, T. I. Sølling, A. H. Zewail, Chem. Phys. Chem.
2002, 3, 57.
J. Phys. Org. Chem. 2011, 24 883–888
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

G. KUZMANICH AND M. A. GARCIA-GARIBAY
[23] G. Kuzmanich, F. Spänig, C.-K. Tsai, J. M. Um, R. M. Hoekstra, K. N. Houk,
D. M. Guldi, M. A. Garcia-Garibay, J. Am. Chem. Soc. 2001, ASAP.
[24] B. A. Hess, Jr, L. Smentek, A. R. Brash, J. K. Cha, J. Am. Chem. Soc.
1999, 121, 5603.
[25] A. R. Brash, S. W. Baertschi, C. D. Ingram, T. M. Harris, Proc. Natl. Acad.
Sci. U.S.A. 1988, 85, 3382.
[46] N. J. Turro, P. A. leermakers, H. R. Wilson , D. C. Neckers, G. W. Byers,
G. F. Vesley, J. Am. Chem. Soc. 1965, 87.
[47] A. P. Krapcho, B. Abegaz, J. Org. Chem. 1974, 39, 2251.
[48] J. L. E. Erickson, F. E. Collins, Jr., B. L. Owen, J. Org. Chem. 1966, 31,
480.
[49] J. D. Dunitz, J. Bernstein Accounts of Chemical Research 1995 28(4),
[26] R. Kolijak, O. Boutaud, B.-H. Shieh, N. Samel, A. R. Brash, Science
1997, 277, 1994.
[27] R. Lee, C. H. Yoon, J. Chem. Soc. Chem. Commun. 1994, 479.
[28] T. Noriko, S. Yamazaki, Y. Shinichi, Org. Biomol. Chem. 2008, 6, 3109.
[29] G. D. Hamblin, R. P. Jimenez, T. S. Sorensen, J. Org. Chem. 2007, 72,
9439.
[30] A. R. Matlin, P. M. Lahti, D. Appella, A. Straumanic, S. Lin, H. Patel,
K. P. Schrieber, J. Pauls, P. Paulerson, J. Am. Chem. Soc. 1999, 121,
2164.
193–200
[50] F. W. McLaffery, Wiley Registry of Mass Spectral Data, Vol. 9, John
Wiley & Sons Inc., Hoboken, 2009.
[51] K. Kumura, M. Takamura, S. Koshibe, M. Juro, Y. Fukuda, Y. Odaira,
Bull. Chem. Soc. Jpn. 1976, 49, 741.
[52] S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photochemistry,
2nd edn, Marcel Dekker, Inc., New York, 1993.
[53] M. Veerman, M. J. E. Resendiz, M. A. Garcia-Garibay, Org. Lett. 2006,
8, 2615.
[31] G. Liang, D. Trauner, J. Am. Chem. Soc. 2004, 126, 9544.
[32] J. Hauang, A. J. Frontier, J. Am. Chem. Soc. 2007, 129, 8060.
[33] W. He, I. R. Herrick, T. A. Atesin, P. A. Caruana, C. A. Kellenberger,
A. J. Frontier, J. Am. Chem. Soc. 2008, 130, 1003.
[34] B. Foehlisch, I. Herrscher, Chem. Ber. 1986, 119, 524.
[35] S. Bhargava, J. Hou, M. Parvez, T. S. Sorensen, J. Am. Chem. Soc.
2005, 127, 3704.
[36] A. G. Moiseev, M. Abe, E. O. Danilov, D. C. Neckers, J. Org. Chem.
2007, 72, 2777.
[37] H. E. Zimmerman, V. Suryanarayan, Eur. J. Org. Chem. 2007, 24, 4091.
[38] I. Gomez, S. Olivella, M. Reguero, A. Riera, A. Sol, J. Am. Chem. Soc.
2002, 124, 15375.
[39] W. K. Chung, S. K. Lam, B. Lo, L. L. Lui, W.-T. Wong, P. Chiu, J. Am.
Chem. Soc. 2009, 131, 4556.
[40] T. Ichino, S. M. Villano, A. J. Gianola, D. J. Goebbert, L. Velarde,
A. Sanov, S. J. Blanksby, X. Zhou, D. A. Hrovat, W. T. Borden,
W. C. Lineberger, Angew. Chem. Int. Ed. 2009, 48, 8509.
[41] D. Lim, D. A. Hrovat, W. T. Borden, W. L. Jorgensen, J. Am. Chem. Soc.
1994, 116, 3494.
[42] V. Mozhayskiy, D. J. Boebbert, L. Velarde, A. Sanov, A. I. Krylov,
J. Phys. Chem. A 2010, 114, 6935.
[43] J. A. Hyatt, Organic Reactions, John Wiley & Sons, Hoboken, 1994.
[44] H. Kasai, H. S. Nalwa, H. Oikawa, S. Okada, H. Matsuda, N. Minami,
A. Kakuta, K. Ono, A. Mukoh, H. Nakanishi, Jpn. J. Appl. Phys. 1992,
31, 1132.
[54] Generation of 2 by irradiation of dione 1 in neat furan gave both al-
kene 4 (95%) and 5 (5%). Examples involving sterically hindered
oxyally have shown that it is generally not isolatable: F. D. Greene,
D. B. Sclove, J. F. Pazos, R. L. Camp, J. Am. Chem. Soc. 1970, 92, 7488.
[55] Generation of 2 by irradiation of dione 1 in neat furan gave both al-
kene 4 (95%) and 5 (5%). Examples involving sterically hindered
oxyally have shown that it is generally not isolatable: M. H. J. Cordes,
J. A. Berson, J. Am. Chem. Soc. 1996, 118, 6241.
[56] Diffraction data from crystal of diketone 1c were solved in the
orthorhombic space group Cmca (C14H20O2, a=9.922, b=6.689,
c=18.285, a=b=g=90o, Volume (Å3)=1213.6, Z=4).
[57] Diffraction data from a crystal of diketone a-1d were solved in the
orthorhobric space group Pmna (C22H36O2, a=12.175, b=10.422,
c=16.366Å, a=b=g=90o, Volume (Å3)=2076.7, Z=4).
[58] Packing coefficients were defined by Kitaigorodskii as the volume of
molecules in the unit cell relative to the total volume of the unit cell.
A. I. Kitaigorodski, Molecular Crystals and Molecules, Academic Press,
New York, 1973.
[59] Diffraction data from crystal of diketone 1a were solved in the
monoclinic space group P21/C (C10H12O2, a=8.276, b=5.608,
c=9.710, a=g=90o, b=111.928o, Volume (Å3)=418.04, Z=2).
[60] T. Choi, D. Cizmeciyan, S. I. Kahn, M. A. Garcia-Garibay, J. Am. Chem.
Soc. 1995, 117, 12893.
[61] W. W. Wright, J. C. Baez, J. M. Vanderkooi, Anal. Biochem. 2002, 307,
167.
[45] J. J. Kuhn, S. E. Braslavsky, R. Schmidt, Pure Appl. Chem. 2004, 76,
[62] The possibility that the concentration of trapped oxyallyl was too
low to detect in the mixed crystals cannot be eliminated.
2105.
wileyonlinelibrary.com/journal/poc
Copyright © 2011 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2011, 24 883–888