Reaction control in synthetic organic photochemistry: Switching between [5+2] and [2+2] Modes of Cycloaddition

Article abstract of DOI:10.1002/anie.200904059

Split personality: Reaction conditions that enable powerful control over the mode of photocycloaddition in maleimides have been developed. Direct irradiation favors the [5+2] mode whereas sensitized irradiation allows a complete switch to the [2+2] mode (see scheme; TBS=tertbutyldimethylsilyl).

Full text of DOI:10.1002/anie.200904059

Communications
DOI: 10.1002/anie.200904059
Reaction Control
Reaction Control in Synthetic Organic Photochemistry: Switching
between [5+2] and [2+2] Modes of Cycloaddition**
Claudio Roscini, Kara L. Cubbage, Malcolm Berry, Andrew J. Orr-Ewing,* and
Kevin I. Booker-Milburn*
Modern synthetic chemistry has evolved to such an extent
that exquisite levels of chemo-, regio-, and stereocontrol can
be readily achieved within complex, multifunctional mole-
cules. Reaction control in synthetic photochemistry is a much
more difficult prospect where the basic bond-forming step is
generally controlled by the lifetime of a key excited state, the
formation of which is often dominated by complex structural
and photophysical issues. Different reaction pathways of a
=
single chromophore (e.g. C C bond) can be observed by
populating either singlet or triplet states.[1] Further limited
modes of selection can be achieved by irradiation at user-
selected wavelengths,[2] through higher exited states from
multiphoton absorption,[3] and by feedback-based optical
control using evolutionary algorithms and pulse shapers.[4]
More recently our research group has demonstrated that
photon flux can be a useful parameter in controlling reaction
pathways, albeit within the scale limitations imposed by
tunable dye lasers.[5]
Scheme 1. Contrasting intermolecular [2+2] and intramolecular
[5+2] photocycloaddition behavior of imide compounds.
of the C N bond).[7] This observation has since proved to be a
ꢀ
general trend with maleimide derivatives: [2+2] photocy-
cloaddition is observed with intermolecular reactions and
[5+2] photocycloaddition with intramolecular variants.
More recently we have focused on exploring the scope of
this [5+2] cycloaddition and its application in natural product
synthesis.[8] During model studies towards the synthesis of the
Stemona alkaloids we observed that an N-alkenyl maleimide
displayed atypical [5+2] behavior (Scheme 2). Irradiation of
Herein we demonstrate how sensitized and nonsensitized
reactions of N-alkenyl maleimides lead to a selective reaction
ꢀ
=
from different bonds (C N vs. C C) within the same
molecule. This selection enables either a [5+2] or [2+2] cyclo-
addition pathway to be chosen, thus providing selective
synthesis of complex 7,5-fused azepines or cyclobutanes,
respectively.
We previously described the intermolecular [2+2] photo-
cycloaddition reactions of tetrahydrophthalimide anhydride 1
(Scheme 1) with alkenol and alkynol partners.[6] We found
these reactions to be remarkably efficient, with the
[2+2] cycloadducts (e.g. 2) being formed in high yields and
with excellent stereoselectivity. In contrast, the attempted
intramolecular [2+2] photocycloaddition of the pentenyl-
substituted imide 3 yielded the tricyclic azepine 5 exclusively
in excellent yield by a [5+2] pathway (attributed to cleavage
[*] C. Roscini, K. L. Cubbage, Prof. Dr. A. J. Orr-Ewing,
Prof. Dr. K. I. Booker-Milburn
Scheme 2. Selective [2+2] vs. [5+2] cycloaddition by sensitization.
a) hn (125 W medium-pressure mercury lamp, Pyrex), MeCN, 7 h;
b) hn (125 W medium-pressure mercury lamp, Pyrex), benzophenone
(BP; 1 equiv), MeCN, 1 h.
School of Chemistry, University of Bristol
Cantock’s Close, Bristol, BS8 1TS (UK)
Fax: (+44)117-929-8611
E-mail: k.booker-milburn@bristol.ac.uk
a.orr-ewing@bristol.ac.uk
Dr. M. Berry
the maleimide system 6 (containing an exocyclic alkene unit)
led to a high yield of cycloaddition products, but for the first
time the intramolecular [2+2] adduct dominated. Although
chemically efficient, the low quantum yields (F[2+2] = 0.018
and F[5+2] = 0.009) illustrate photochemically inefficient
bond-forming processes. This observation was initially dis-
regarded as an anomaly; however, recent similar results have
Medicines Research Centre, GlaxoSmithKline
Gunnels Wood Road, Stevenage, Herts, SG1 2NY (UK)
[**] We thank the University of Bristol for a University Scholarship
(K.C.), EPSRC (GR/S25593) and G.S.K. for funding, and Dr. M.
Haddow for X-ray crystallographic analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200904059.
8716
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8716 –8720

Angewandte
Chemie
prompted us to reinvestigate this system. From a mechanistic
The study also highlighted some interesting stereoselec-
point of view we proposed that the [5+2] reaction proceeds
tivity differences between the [5+2] and [2+2] pathways. In
entries 1–4 the [5+2] azepine products were always formed as
two, often separable diastereomers (see the Supporting
Information for ratios). In contrast, their corresponding
[2+2] cyclobutanes were always isolated as a single diaste-
reomer (see the Supporting Information for the X-ray
structure of 12).[11] This result, we believe, is a clear reflection
of the relative rates of reaction between a singlet and triplet
process, where in the latter the lifetime of the excited state is
much longer, thus allowing time for the substituted alkenyl
side chain to adopt the lowest energy conformation for
[2+2] cycloaddition (see below).
From the results in Table 1 two key mechanistic pathways
can be proposed (Scheme 3). Direct irradiation leads to S1
which can undergo subsequent a cleavage to diradical 14
followed by [5+2] cycloaddition forming the azepine product
(k4). Recombination of 14 could, however, also occur to
ꢀ
from cleavage of the singlet N C bond and subsequent
addition of the resultant diradical to the pendant alkene
unit.[9] We assumed that the unusual intramolecular
[2+2] reaction was a result of intersystem crossing (ISC)
=
from the initially formed singlet (C O, n !p*) to form the
=
triplet diradical of the maleimide C C bond. We then
embarked upon a study to find a triplet quencher that
would shut down this unwanted pathway without perturbing
the key [5+2] singlet reaction.
Previously, time-dependent density functional theory
(TD-DFT) studies[9] indicated that the T1 energies for a
variety of substituted maleimides were between 266–
294 kJmolꢀ1. After screening a large number of common
triplet quenchers,[10] with T1 energies in the range of 148–
311 kJmolꢀ1, we were unable to shut down the [2+2] mode of
cyclization. Surprisingly, however, when triphenylene (ET =
280 kJmolꢀ1),
acetophenone
(ET =
310 kJmolꢀ1), and benzophenone (ET =
288 kJmolꢀ1) were employed, the cyclo-
butane 8 was formed exclusively.
Clearly these three excited “quench-
ers” were instead acting as triplet sensi-
tizers by transferring energy to populate
=
the maleimide C C triplet state and thus
facilitating [2+2] cycloaddition only.
Quantum yield measurements of the
[2+2] cycloaddition with benzophenone
(F[2+2] = 0.33) illustrated the efficiency of
this sensitized process over direct irradi-
ation (F[2+2] = 0.018). It was then demon-
strated that this selective sensitization
Scheme 3. Reactive pathways of the [5+2] and [2+2] photocycloadditions.
could be exploited in
a preparative
manner for the exclusive formation of 8 in 91% yield
(Scheme 2). Even more importantly the application of these
findings to dichloromaleimide 9 enabled, for the first time, the
ability to choose exclusively between [5+2] or [2+2] modes of
cycloaddition to give 11 or 10, respectively.
regenerate starting maleimide 13 (k5). It is postulated that for
entries 5, 9, and 10, where [2+2] products are observed under
nonsensitized conditions the rate of cycloaddition is hindered
by steric bulk in the alkenyl chain. In these examples
recombination to give 13 is either favored or competitive
(k5 vs. k4). The consequence is a continuous recycling of S1,
thus allowing ISC (k3) to T1 and resulting in the observed
[2+2] cycloaddition for these examples. If k3 > k4 then
reaction will favor the triplet pathway and a larger ratio of
cyclobutane to azepine will be observed (compare with
entries 5, 9, and 10). The reverse scenario, where k4 > k3,
results in more [5+2] product (entry 4). Where irradiation
under nonsensitized conditions only yields the [5+2] azepines
the rate of k4 is considered to be sufficiently fast that diradical
recombination, and hence the T1 pathway is not observed.
In all these examples, sensitization with benzophenone
clearly allows for very efficient population of the maleimide
T1 state (k1 @ k2). From this point on only the [2+2] pathway
is possible.
Next, we explored the application of these observations to
a variety of maleimide derivatives that had more recently
reacted exclusively by a [5+2] pathway or had given mixed
results (Table 1). Overall, sensitization afforded the
[2+2] pathway exclusively, thus providing a route to hitherto
unobtainable molecular architectures. When the maleimides
were substituted a to the nitrogen center (entries 1–4) direct
excitation gave efficient [5+2] cycloaddition only (except
entry 4). Irradiation with two equivalents of benzophenone
gave only [2+2] cycloaddition in good to excellent yields in all
these cases. Similarly, entries 6–8, and 11 displayed complete
switching of the mode of cycloaddition upon sensitized
irradiation. Entries 4, 5, 9, and 10 (and compound 6) were
maleimides that we had recently studied in an attempted
alkaloid synthesis, but had abandoned because of low yields
of the [5+2] adduct, which resulted from competing
[2+2] cycloaddition. On reinvestigation under sensitized
conditions it was fascinating to observe the exclusive for-
mation of their respective cyclobutane derivatives in good to
excellent yields.
In addition to the quantum yield studies (Scheme 2), we
compared the relative rates of the sensitized vs. nonsensitized
reactions. Maleimide 6 was chosen as a model substrate as it
displays dual cycloaddition behavior under direct irradiation.
The rates of formation of 7 and 8 were monitored by 1H NMR
spectroscopy under the two different reaction conditions
Angew. Chem. Int. Ed. 2009, 48, 8716 –8720
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8717

Communications
Table 1: Pathway switching in photocycloaddition of N-alkenyl maleimides.
Entry
Substrate
Products
[5+2] cycloaddition [2+2] cycloaddition
Direct irradiation[a]
(yield in %)[c]
Sensitized irradiation[b]
(yield in %)[c]
[5+2] only
(69)
[2+2] only
(84)
1
[5+2] only
(81)
[2+2] only
(67)
2
[5+2] only
(70)
[2+2] only
(87)
3
4
[5+2] + [2+2]
(40) (19)
[2+2] only
(100)[d]
[5+2] + [2+2]
(20) (42)
[2+2] only
(67)
5
6
[5+2] only
(79)
[2+2] only
(42)
[5+2] only
(88)
[2+2] only
(34)
7
[5+2] only
(99)
[2+2] only
(32)
8
[5+2] + [2+2]
(8) (45)
[2+2] only
(77)
9
[5+2] + [2+2]
(3) (30)
[2+2] only
(80)
10
11
[5+2] only
(41)
[2+2] only
(20)
[a] hn (125 W medium-pressure mercury lamp), 150 mL Pyrex immersion well, MeCN (1 mmol), 1–9 h. [b] hn (125 W medium-pressure mercury lamp)
150 mL Pyrex immersion well, MeCN (1 mmol), benzophenone (2 equiv), 1 h. [c] Yield of isolated product after column chromatography. [d] Reaction
performed in MeCN (3 mL) in a quartz cuvette taped to a water-cooled Pyrex jacket containing a 125 W medium-pressure mercury lamp (30 min
irradiation); yield was determined from the 1H NMR spectrum of the evaporated photolysate.
and the results are plotted in Figure 1. The chemical efficiency
of the sensitized [2+2] cycloaddition to 8 is striking; 79 times
faster than the unsensitized reaction. This result clearly
illustrates the exceptionally efficient energy transfer
(k1) from the excited benzophenone to the maleimide
chromophore.
In summary, appropriate choice of irradiation conditions
has enabled synthetically useful reaction control in maleimide
8718
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8716 –8720

Angewandte
Chemie
2.0 Hz, NCHCH2CHH), 1.61–1.50 (1H, m, NCHCH2CHH), 1.49
(3H, d, J = 7.4 Hz, CHCH3), 1.38 (1H, qd, J = 14.9 Hz, J = 2.0 Hz,
=
=
NCHCHH), 1.25 (3H, s, C CCH3), 1.13 ppm (3H, s, C CCH3);
13C NMR (100 MHz, CDCl3): dC = 189.2 (CO), 183.1 (CO), 53.5
(CH), 52.6 (C), 45.0 (C), 38.5 (CH), 32.8 (CH2), 29.3 (CH2), 23.9
~
(CH2), 17.1 (CH3), 14.9 (CH3) 11.3 ppm (CH3); IR (neat): n = 2950
(w), 1774 (w), 1700 (s), 1447 (m), 1020 cmꢀ1 (m); HRMS (CI) calcd
for C12H18NO2 ([MH+]): 208.1259; found: 208.1337.
Received: July 22, 2009
Published online: October 13, 2009
Keywords: alkenes · cycloaddition · maleimides ·
photochemistry · sensitizers
.
Figure 1. Rate of product formation of 7 and 8 for direct and
sensitized irradiations of maleimide 6.
[1] a) K. N. Houk, D. J. Northington, R. E. Duke, J. Am. Chem. Soc.
1972, 94, 6233; b) H. E. Zimmerman, A. C. Pratt, J. Am. Chem.
Soc. 1970, 92, 6267; c) H. E. Zimmerman, J. A. Pincock, J. Am.
Chem. Soc. 1973, 95, 2957; d) J. S. Swenton, J. A. Hyatt, T. J.
Walker, A. L. Crumrine, J. Am. Chem. Soc. 1971, 93, 4808;
e) N. J. Turro, Modern Molecular Photochemistry, Benjamin
Cummings, California, 1978; f) A. Gilbert, J. Baggot, Essentials
of Molecular Photochemistry, Blackwell Scientific, Oxford, 1991.
[2] a) L. D. Elliott, M. Berry, A. J. Orr-Ewing, K. I. Booker-Mil-
burn, J. Am. Chem. Soc. 2007, 129, 3078; b) K. B. Eisenthal, N. J.
Turro, C. G. Dupuy, D. A. Hrovat, J. Langan, T. A. Jenny, E. V.
Sitzmann, J. Phys. Chem. 1986, 90, 5168; c) W. G. Dauben, B.
Disanayaka, D. J. H. Funhoff, B. E. Kohler, D. E. Schilke, B.
Zhou, J. Am. Chem. Soc. 1991, 113, 8367; d) W. G. Dauben, B.
Zhou, J. Y. L. Lam, J. Org. Chem. 1997, 62, 9005; e) V. Jakfflbek,
A. J. Lees, Inorg. Chem. 2000, 39, 5779.
photochemistry to be demonstrated for the first time.
Significantly we have shown that for a molecule containing
a complex chromophore, two different modes of cycloaddi-
tion can be selected by judicious choice of reaction conditions.
The sensitized conditions have enabled routine synthesis of
cyclobutane adducts that were previously inaccessible
because of the unimodal [5+2] behavior typically observed
for N-alkenyl maleimides. These observations will greatly
extend the scope of maleimide photochemistry in the syn-
thesis of complex polycyclic molecules.
Experimental Section
Reaction conditions for [5+2] cycloaddition (Table 1, entry 1): A
solution of S11 (see the Supporting Information) (207 mg, 1.0 mmol)
in degassed MeCN (150 mL) was irradiated using a 125 W medium-
pressure lamp in a Pyrex immersion well for 1 h. Purification by
column chromatography (60% EtOAc in petroleum ether) yielded
the azepine S27 (see the Supporting Information) as two inseparable
diastereomers (1:1.27, 142 mg, 69%); Rf = 0.09 (30% EtOAc in
petroleum ether); H NMR (400 MHz, CDCl3): dH = 4.34–4.09 (4H,
m, NCHCH3 and NCHCH2, both isomers), 2.81–2.43 (4H, m,
COCH2, both isomers), 2.31–1.52 (12H, m, CH3CHCH2CH2CH2,
[3] a) R. M. Wilson, K. A. Schnapp, Chem. Rev. 1993, 93, 223;
b) E. V. Khoroshilova, N. P. Kuzmina, V. S. Letokhov, Y. A.
Matveetz, Appl. Phys. B 1983, 31, 145; c) M. Hara, S. Tojo, T.
Majima, J. Phys. Chem. A 2003, 107, 4778; d) N. J. Turro, M.
Aikawa, I. R. Gould, J. Am. Chem. Soc. 1982, 104, 856; e) N. J.
Turro, M. Aikawa, J. A. Butcher, Jr., G. W. Griffin, J. Am. Chem.
Soc. 1980, 102, 5127; f) Y. Kajii, T. Suzuki, Y. Takatori, K.
Shibuya, K. Obi, Bull. Chem. Soc. Jpn. 1992, 65, 1349; g) W. C.
McGimpsey, J. C. Scaiano, Chem. Phys. Lett. 1987, 138, 13; h) J.-
y. Kohno, F. Mafunꢀ, T. Kondow, Bull. Chem. Soc. Jpn. 2005, 78,
957; i) T. I. Sølling, E. W.-G. Diau, C. Kꢁtting, S. de Feyter, A. H.
Zewail, ChemPhysChem 2002, 3, 79; j) L. J. Johnston, J. C.
Scaiano, J. Am. Chem. Soc. 1987, 109, 5487; k) H. Baumann, C.
Merkel, H.-J. Timpe, A. Graness, J. Kleinschmidt, I. R. Gould,
N. J. Turro, Chem. Phys. Lett. 1984, 103, 497; l) J. A. Butcher, Jr.,
H. R. Hinz, N.-h. Tsou, S. Shah, Tetrahedron Lett. 1984, 25, 5483;
m) H. Murai, K. Obi, J. Phys. Chem. 1975, 79, 2446.
1
=
both isomers), 1.99 (3H, d, J = 1.1 Hz, C CCH3, minor isomer), 1.98
=
(3H, d, J = 1.1 Hz C CCH3, major isomer), 1.86 (3H, d, J = 1.1 Hz,
=
C CCH3, major isomer), 1.83 (3H, d, J = 1.1 Hz, minor isomer), 1.21
(3H, d, J = 6.4 Hz, CHCH3, major isomer), 1.15 ppm (3H, d, J =
6.4 Hz, CHCH3, minor isomer); 13C NMR (100 MHz, CDCl3): dC =
202.4 (CO), 202.3 (C), 166.4 (CO), 166.2 (CO), 140.1 (C), 138.8 (C),
138.3 (C), 136.9 (C), 54.1 (CH), 54.1 (CH), 53.4 (CH2), 53.2 (CH), 52.8
(CH), 51.5 (CH2), 31.0 (CH2), 29.4 (CH2), 29.4 (CH2), 28.3 (CH2), 20.3
(CH3), 18.8 (CH3), 17.9 (CH3), 17.5 (CH3), 16.1 (CH3), 15.5 ppm
[4] T. Brixer, G. Gerber, ChemPhysChem 2003, 4, 418.
[5] C. Roscini, D. M. E. Davis, M. Berry, A. J. Orr-Ewing, K. I.
Booker-Milburn, Angew. Chem. 2008, 120, 2315 – 2318; Angew.
Chem. Int. Ed. 2008, 47, 2283.
[6] a) K. I. Booker, K. Cowell, A. Sharpe, F. D. Jimenez, Chem.
Commun. 1996, 249; b) K. I. Booker-Milburn, S. Gluten, A.
Sharpe, Chem. Commun. 1997, 1385; c) K. I. Booker-Milburn,
J. K. Cowell, F. D. Jimenez, A. Sharpe, A. J. White, Tetrahedron
1999, 55, 5875.
[7] a) K. I. Booker-Milburn, N. J. Costin, R. F. Dainty, D. Patel, A.
Sharpe, Tetrahedron Lett. 1998, 39, 7423; b) K. I. Booker-
Milburn, C. E. Anson, C. Clissold, N. J. Costin, R. F. Dainty,
M. Murray, D. Patel, A. Sharpe, Eur. J. Org. Chem. 2001, 1473.
[8] a) M. D. Lainchbury, M. I. Medley, P. M. Taylor, P. Hirst, W.
Dohle, K. I. Booker-Milburn, J. Org. Chem. 2008, 73, 6497;
b) K. I. Booker-Milburn, L. F. Dudin, C. E. Anson, S. D. Guile,
Org. Lett. 2001, 3, 3005.
~
(CH3); IR (neat): n = 2967 (w), 1671 (m), 1630 (s), 1602 (s), 1417 (s),
1284 (m), 937 (m), 755 cmꢀ1 (s). All data is in accordance with
literature values.[7b]
Reaction conditions for [2+2] cycloaddition (Table 1, entry 1): A
solution of S11 (see the Supporting Information) (207 mg, 1.0 mmol)
and benzophenone (2 equiv, 364 mg, 2.0 mmol) dissolved in degassed
acetonitrile (150 mL) was irradiated using a 125 W medium-pressure
mercury lamp in a Pyrex immersion well for 1 h. Purification by
column chromatography (0–15% EtOAc in petroleum ether) yielded
the cyclobutane S28 (see the Supporting Information) (174 mg, 84%)
as a colorless oil; Rf = 0.36 (30% EtOAc in petroleum ether);
1H NMR (300 MHz, CDCl3): dH = 4.35 (1H, quint, J = 7.4 Hz, NCH),
2.42 (1H, dd, J = 12.3 Hz, J = 10.3 Hz, NCOCCHH), 2.29–2.20 (1H,
m, NCOCCH), 2.20–2.05 (1H, m, NCHCHH), 2.01 (1H, dd, J =
12.3 Hz, J = 2.6 Hz, NCOCCHH), 1.86 (1H, tt, J = 14.9 Hz, J =
Angew. Chem. Int. Ed. 2009, 48, 8716 –8720
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8719

Communications
[9] D. Davies, C. Murray, M. Berry, A. J. Orr-Ewing, K. I. Booker-
289 kJmolꢀ1); indazole (ET = 284 kJmolꢀ1); benzotriazole
(ET = 295 kJmolꢀ1); benzophenone (ET = 287–289 kJmolꢀ1);
acetophenone (ET = 310–311 kJmolꢀ1); from S. L. Murov, I.
Carmichael, G. L. Hug, Handbook of Photochemistry, Marcel
Dekker, New York, 1993, pp. 56 – 97.
Milburn, J. Org. Chem. 2006, 72, 1449.
[10] Sensitizers tested were: perylene (ET = 148–151 kJmolꢀ1); iso-
prene
(ET = 251 kJmolꢀ1);
naphthalene
(ET = 253–
255 kJmolꢀ1); quinoline (ET = 258–261 kJmolꢀ1); styrene (ET =
258 kJmolꢀ1); a-methyl styrene (ET = 260 kJmolꢀ1); indene
(ET = 264 kJmolꢀ1); phenaxanthin (ET = 267 kJmolꢀ1); m-ter-
phenyl (ET = 269 kJmolꢀ1); biphenyl (ET = 274 kJmolꢀ1); tri-
[11] CCDC 739183 (12) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
phenylene
(ET = 280 kJmolꢀ1);
fluorene
(ET = 282–
8720 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8716 –8720