Investigating the photochemical properties of an arenyl dienol

Article abstract of DOI:10.1021/ol302718p

Upon irradiation with UV light, an arenyl dienol was transformed into linear and angular meta photocycloadducts and ortho derived photoadducts. Extended exposure to UV radiation resulted in the formation of other degradation products, which shed light on the chemical processes taking place. One of the linear meta photocycloadducts was thermally unstable and underwent further thermal and photochemical transformation, while the ortho-derived photocycloadducts ring-opened and eliminated methanol to afford a cyclooctadienone product.

Full text of DOI:10.1021/ol302718p

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Investigating the Photochemical
Properties of an Arenyl Dienol
Clive S. Penkett*,† and Jason A. Woolford†,‡,§
clive.penkett@googlemail.com
Received October 2, 2012
ABSTRACT
Upon irradiation with UV light, an arenyl dienol was transformed into linear and angular meta photocycloadducts and ortho derived photoadducts.
Extended exposure to UV radiation resulted in the formation of other degradation products, which shed light on the chemical processes taking
place. One of the linear meta photocycloadducts was thermally unstable and underwent further thermal and photochemical transformation, while
the ortho-derived photocycloadducts ring-opened and eliminated methanol to afford a cyclooctadienone product.
Organic photochemistry has a remarkable propensity
for creating dramatic increases in molecular complexity
startingfromsimplecompounds.1,2 A good example of this
is the double [3 þ 2] photocycloaddition reaction3 which
can assemble the fenestrane 1 in a one-pot procedure
simply by irradiating the arenyl diene 2. Further investi-
gation of this reaction revealed that the efficiency of
fenestrane formation could be improved by isolating the
initially formed meta photocycloadduct2 3 and then irra-
diating 3 in the presence of a triplet sensitizer during a
second step (Scheme 1). To our knowledge there has been
very limited research into the photochemical properties
of arenyl dienes, and hence we report some interesting
observations involving the photochemistry of the arenyl
dienol substrate 5.
† Research undertaken at Department of Chemistry, University of Sussex,
Brighton BN1 9QJ, U.K.
Scheme 1. Double [3 þ 2] Photocycloaddition Reaction
Involving Acetal 23
‡ The chemistry herein has previously been described in Woolford, J. A.
The Double [3 þ 2] Photocycloaddition Reaction; Springer: 2011. ISBN
978-3-642-22859-9, a part of the Springer Theses Series, recognizing out-
standing Ph.D. research and is presented with the kind permission of Springer
ScienceþBusiness Media.
§ Present address: Department of Chemistry, Imperial College London
SW7 2AZ, U.K.
ꢀ
(1) (a) Klan, P.; Wirz, J. Photochemistry of Organic Compounds:
From Concepts to Practice; Wiley-Blackwell: Chichester, U.K., 2009. (b)
Coyle, J. D. Introduction to Organic Photochemistry; John Wiley & Sons:
Chichester, U.K.: 1991. (c) Horspool, W.; Armesto, D. Organic Photo-
chemistry: A Comprehensive Treatment; Ellis Horwood, Prentice Hall:
New York, 1992.
The photosubstrate 5 was prepared in 75% yield by the
addition of 2 equiv of butenylmagnesium bromide to
2-methoxy methylbenzoate 4 (Scheme 2).
(2) (a) Wilzbach, K. E.; Kaplan, L. J. Am. Chem. Soc. 1966, 88, 2066.
(b) Bryce-Smith, D.; Gilbert, A.; Orger, B. H. J. Chem. Soc., Chem.
Commun. 1966, 512. (c) Wender, P. A.; Howbert, J. J. J. Am. Chem. Soc.
1981, 103, 688. (d) Wender, P. A.; Siggel, L.; Nuss, J. M. In Organic
Photochemistry, Vol. 10; Padwa, A., Ed.; Marcel-Dekker: New York,
1989; Chapter 4, p 357. (e) Wender, P. A.; Ternansky, R.; deLong, M.;
Singh, S.; Olivero, A.; Rice, K. Pure Appl. Chem. 1990, 62, 1597. (f)
Wender, P. A.; Siggel, L.; Nuss, J. M. In Comprehensive Organic
Synthesis, Vol. 5; Trost, B. M., Fleming, I., Eds.; Pergamon Press:
Oxford, U.K., 1991; p 645. (g) Cornelisse, J. Chem. Rev. 1993, 93, 615.
(h) Hoffmann, N. Synthesis 2004, 481. (i) Chappell, D.; Russell, A. T.
Org. Biomol. Chem. 2006, 4, 4409. (j) Mattay, J. Angew. Chem., Int. Ed.
2007, 46, 663. (k) Streit, U.; Bochet, C. G. Chimia 2008, 62, 962. (l) Streit,
U.; Bochet, C. G. Beilstein J. Org. Chem. 2011, 7, 525.
Scheme 2. Formation of Photosubstrate 5
A solution of arenyl dienol 5 in cyclohexane was irra-
diated with 254 nm UV light using a 16 W low-pressure
Hg vapor lamp until all the starting material had been
(3) (a) Penkett, C. S.; Woolford, J. A.; Day, I. J.; Coles, M. P. J. Am.
Chem. Soc. 2010, 132, 4. (b) Penkett, C. S.; Woolford, J. A.; Read, T. W.;
Kahan, R. J. J. Org. Chem. 2011, 76, 1295.
r
10.1021/ol302718p
XXXX American Chemical Society

consumed. TLC analysis of the crude photolysis residue
identified the formation of five major products, which
were separated by flash column chromatography. Initially
yields were difficult to establish due to a high number of
impurities, some of which arose from the instability of the
linear meta adducts toward acid induced fragmentation.
This complicated flash column chromatography on a
larger scale, but the acid induced degradation could be
combated by adding 0.1% triethylamine during initial
silica slurry formation and yields were eventually deter-
mined to be 19%, 10%, <10%, 12%, and 19% for 6r, 6β,
7β, 8a, and 8b respectively with an irradiation time of 10 h
for 1.84 g of 5 (Scheme 3).
ring-opening, and finally a photochemically induced four-
electron electrocyclic disrotatory ring-closure (Scheme 4).
The origin of the high diastereoselectivity during intra-
molecular meta photocyclization is a well-established
process,2c when there is a methyne CꢀH group on the
tether adjacent to the aromatic ring (see 3 being prepared
from 2). However the diastereoselectivity is less clearly
defined when there are two groups of similar bulkiness
situated on the tethering chain adjacent to the aromatic
ring. A hydroxyl group created considerably more steric
bulk than a hydrogen atom, which meant that both
diastereoisomers of the linear meta adduct (6r and 6β)
were formed to a significant degree (Scheme 5). Aside from
the steric bulk argument a case could be made that a
hydrogen-bonding interaction during the photocycloaddi-
tion of alcohol 5 would favor the formation of 6β; however
Cornelisse5 has shown this not to be the case with his study
on 5-(20-methoxyphenyl)-pent-1-ene derivatives. It was
interesting that only one angular meta adduct (7β) was
formed during the primary irradiation step and we only
managed to isolate it in a pure form by separately irradiat-
ing purified 6β to give 7β in 10% yield. This favorable
photolytic interchange between 6β and 7β may partly
explain the greater prevalence of 6r over 6β, although 6β
was also found to be thermally unstable.
Scheme 3. Principal Photoadducts Generated from 5
Only three meta photocycloadducts were formed upon
initial irradiation; two linear forms 6r and 6β (with R and
βoriented hydroxyl groups respectively) and one angular form
6β (with a β hydroxyl group). Two other tricyclic compounds
8a and 8b were isolated, which were formed by intramolecular
ortho cyclization4 of one alkene across the arene ring, then
a thermally induced six-electron electrocyclic disrotatory
Scheme 5. Different Conformations of 5 Leading to 6r and 6β
Scheme 4. Formation of Ortho-Adduct 8 from 5
The diastereomeric assignment of the linear meta photo-
cycloadducts 6r and 6β was highly significant to any
subsequent chemical transformations that were to be
undertaken. Unfortunately unambiguous confirmation
could not be provided using ROESY correlation spectra,
and hence the stereochemical assignment was based on a
stereoelectronic analysis of the chemical reactivities of the
two diastereoisomers (Scheme 6). It was noted that the
β-hydroxy substituted analogue 6β underwent thermal
degradation at an accelerated rate compared to the
R-hydroxy variant 6r. This greater instability was attrib-
uted to stereoelectronic factors6 relating to subtle structur-
al differences between 6r and 6β. Acetal photoadduct
3 and alcohol photoadduct 6r both showed little sign of
decomposition after being dissolved for several days in
(mildly acidic) deuterated chloroform that had not been
neutralized by passing through neutral alumina; however a
€
(4) (a) Ayer, D. E.; Buchi, G. U.S. Patent 2,805,252, 1957. (b)
Atkinson, J. E.; Ayer, D. E.; B€uchi, G.; Robb, E. W. J. Am. Chem. Soc.
1963, 85, 2257. (c) Al-Qaradawi, S. Y.; Cosstick, K. B.; Gilbert, A.
J. Chem. Soc., Perkin Trans. 1 1992, 1145. (d) Van Der Eycken, E.; De
Keukeleire, D.; De Bruyn, A. Tetrahedron Lett. 1995, 36, 3573. (e)
Hoffmann, N.; Pete, J.-P. J. Org. Chem. 1997, 62, 6952. (f) Avent, A. G.;
(5) Timmermans, J. L.; Wamelink, M. P.; Lodder, G.; Cornelisse, J.
Eur. J. Org. Chem. 1999, 463.
(6) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry;
Baldwin, J. E., Ed.; Organic Chemistry Series; Pergamon Press: Oxford,
U.K., 1983.
ꢀ
Byrne, P. W.; Penkett, C. S. Org. Lett. 1999, 1, 2073. (g) Vızvardi, K.;
´
ꢀ
Toppet, S.; Hoornaert, G. J.; De Keukeleire, D.; Bako, P.; Van der
Eycken, E. J. Photochem. Photobiol. A: Chem. 2000, 133, 135. (h)
Hoffmann, N.; Pete, J.-P.; Inoue, Y.; Mori, T. J. Org. Chem. 2002, 67,
2315. (i) Wagner, P. J.; Lee, J.-I. Tetrahedron Lett. 2002, 43, 3569.
B
Org. Lett., Vol. XX, No. XX, XXXX

considerable proportion of 6β decomposed after sitting for
a few hours in the same solvent. The key difference was
related to the orientation of the CꢀO bond of the hydroxyl
moiety with respect to the CꢀC bond of the cyclopropane
ring in the linear meta photoadduct framework of 6r and
6β. If the CꢀO bond of the hydroxyl moiety and the CꢀC
bond of the cyclopropane ring were aligned in a syn-
periplanar fashion, the methoxy lone pair would assist in
the fragmentation of the cyclopropane ring and cause the
elimination of hydroxide. The insipient oxonium ion 11
would then be hydrolyzed by the hydroxide ion to give
ketone 12 and a molecule of methanol. Compound 6r was
relatively stable due to there being near orthogonal orbital
alignment with the pseudoequatorial CꢀO bond of its
hydroxyl group and the adjacent cyclopropane bond,
which would hinder any fragmentation/elimination pro-
cess. By contrast compound 6β was unstable due to there
being almost syn-periplanar orbital alignment between the
pseudoaxial CꢀO bond of the hydroxyl group and the
CꢀC bond of the cyclopropane. A similar degradation
process would not be possible in the angular case for 7β.
Scheme 7. Thermally Induced Transformation of 3 into 143b
additional minor products, although many of these re-
mained uncharacterized due to an inability to separate
themfrom polymericdegradation material. Itwas however
possible to isolate three compounds, 15, 16, and 17, albeit
in very low yield, which provided an insight into the
chemical processes that were taking place in the reaction
vessel (Figure 1).
Figure 1. Degradation products 15, 16, and 17.
Scheme 6. Comparison of Thermal Stability of 6r and 6β
The bicyclic trienone compound 15 was probably the
product of a decomposition pathway involving the ortho
addition compounds 8a and 8b. We ascertained that this
decomposition pathway was not thermal, because adducts
8a and 8b showed limited thermal decomposition even
when heated above 200 °C for several hours. This sug-
gested that the process was occurring through a photolytic
reopening of the cyclobutene ring to give the methyl trienol
ether 10, which then eliminated the tertiary alcohol moiety
and was hydrolyzed to 15 (Scheme 8). The existence of
eight-membered ring intermediates has long been postu-
lated to occur during the ortho mode of cycloaddition;4
however isolation of such structures has proved very diffi-
cult. Although not generated directly from the primary
photolysis stage, compound 15 revealed that isolating such
intermediates may be possible by careful choice of substrate.
Further evidencetosubstantiate the theory thatcleavage
of similar pseudoequatorial CꢀO bonds in related com-
pounds was a disfavored process (see compound 6r) was
provided by the thermolysis reaction of acetal 3.3b Like
compound 6r, acetal 3 had a pseudoequatorial substituent
that was attached via a CꢀO bond. Indeed there were two
CꢀO bonds attached at the cyclopropylcarbinyl position
of 3 in the form of an acetal and hence there was a choice of
two bonds that in principle could break when acetal 3 was
heated to 200 °C. Neither ofthe acetal CꢀO bonds of 3 had
syn-periplanar orbital alignment withthe CꢀC bond of the
cyclopropane; hence it was comparatively stable at room
temperature. When bond fragmentation did finally occur
after heating to 200 °C, it wasthe CꢀO bond situatedinthe
ring rather than the CꢀO bond residing in a pseudoequa-
torial position that broke preferentially and gave rise to the
tricyclic acetal 14 (Scheme 7).
Scheme 8. Proposed Mechanism for Formation of 15
One of the key challenges often facing synthetic organic
photochemistry is that photochemical processes often lead
to multiple products that are themselves photochemically
or thermally reactive. Irradiation of 5 beyond its total
consumption (up to 36 h) resulted in the creation of
It is likely that compounds 16 and 17 were both obtained
from the common intermediate 11, which could be
Org. Lett., Vol. XX, No. XX, XXXX
C

Scheme 9. Formation of Ketone 12 and Acetal 16
Scheme 11. De Keukeleire’s Norrish I Observation8
De Keukeleire8 reported a similar example of this
Norrish I chemistry in the case of cyclopentenone deriva-
tive 21 (Scheme 11). Interestingly two diastereomers (22r
and 22β) were formed, although we only managed to
isolate one (17) and its stereochemistry could not be
unambiguously defined using ROESY correlation spectra.
In summary, we have shown how the arenyl dienol 5
generated two linear meta photoadducts (6r and 6β),
one angular meta photoadduct (7β), and two ortho
derived photoadducts (8a and 8b). The two linear meta
photoadducts demonstrated markedly different thermal
stability, with one (6β) readily undergoing fragmentive
elimination of methanol to give ketone 12. Although it
could not be isolated, ketone 12 revealed its existence by
the presence of acetal 16 and the Norrish I elimination
product 17. One further cyclooctadienone compound
(15) was identified from the reaction mixture as being a
ring-opened product of the ortho derived photoadducts
8 that had eliminated a molecule of methanol.
Scheme 10. Norrish Type I Process Involving Ketone 12
derived from the decomposition of compounds 6r and
6β (Scheme 6). Oxonium ion 11 would either undergo
hydrolysis to give ketone 12 and eliminate a molecule
of methanol or combine with a molecule of methanol
to give acetal 16 and eliminate a molecule of water
(Scheme 9). For the formation of acetal 16 to be a viable
process, a source of methanol was required, and this
would have been supplied during the generation of
trienone 15 and ketone 12.
Although there was much evidence for the existence of
ketone 12 within the reaction mixture, it was not possible
to isolate it as a pure sample. This was probably due to it
being rapidly converted to other compounds, before a
significant concentration had accumulated. The existence
of tricyclic [4.5.5] ring system 17 provided a clue to the
most likely decomposition pathway for ketone 12, which
involved a Norrish Type I photochemical reaction.7
Absorption of a photon of UV light would cause the
excitation of the carbonyl moiety (nfπ*) of ketone 12,
which would be followed by R-bond cleavage to give
diradical 19, extrusion of carbon monoxide, and finally
radicalꢀradical recombination to afford 17 (Scheme 10).
Acknowledgment. The authors would like to dedicate
this paper to Professor Philip Parsons of Imperial College
London in celebration of his 60th birthday. This work was
supported by the EPSRC, and we thank Dr. Iain Day of
Sussex University for NMR support.
Supporting Information Available. Complete experimen-
tal procedures, full characterization including 1H, 13C NMR
spectra and where relevant COSY or DQF-COSY, multi-
plicity-edited HSQC, standard HMBC, and ROESY corre-
lation spectra for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
(7) (a) Norrish, R. G. W.; Bamford, C. H. Nature 1936, 138, 1016. (b)
Norrish, R. G. W.; Bamford, C. H. Nature 1937, 140, 195.
(8) De Keukeleire, D. Aldrichimica Acta 1994, 27, 59.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX