A chiral thioxanthone as an organocatalyst for enantioselective [2+2] photocycloaddition reactions induced by visible light

Article abstract of DOI:10.1002/anie.201310997

Thioxanthone 1, which was synthesized in a concise fashion from methyl thiosalicylate, exhibits a significant absorption in the visible light region. It allows for an efficient enantioselective catalysis of intramolecular [2+2] photocycloaddition reactions presumably by triplet energy transfer. Lighting up S: The thioxanthone 1, which was synthesized in a concise fashion from methyl thiosalicylate, exhibits a significant absorption in the visible-light region. It allows for an efficient enantioselective catalysis of intramolecular [2+2] photocycloaddition reactions, presumably by triplet-energy transfer. Ts=4-toluenesulfonyl.

Full text of DOI:10.1002/anie.201310997

Angewandte
Chemie
DOI: 10.1002/anie.201310997
Photochemistry
A Chiral Thioxanthone as an Organocatalyst for Enantioselective
[2+2] Photocycloaddition Reactions Induced by Visible Light**
Rafael Alonso and Thorsten Bach*
Abstract: Thioxanthone 1, which was synthesized in a concise
fashion from methyl thiosalicylate, exhibits a significant
absorption in the visible light region. It allows for an efficient
enantioselective catalysis of intramolecular [2+2] photocy-
cloaddition reactions presumably by triplet energy transfer.
lenge[6] given that these reactions proceed at an energy
hypersurface which is high above the ground state and which
does not involve significant activation barriers. We now report
on the synthesis of the chiral thioxanthone 1 (for structure see
Scheme 1), which absorbs visible light and which has the
potential to act as an effective catalyst for triplet-sensitized
reactions. It was shown that [2+2] photocycloaddition reac-
tions of 4-(pent-4-enyl)quinolones and their heteroatom
analogues proceed with high enantioselectivity[7] in the
presence of this catalyst.
Compared to xanthones, thioxanthones show a bathochro-
mic shift in their UV/Vis spectra. Typically, their absorption
maximum is centered at lmax ꢁ 390 nm with a significant
absorption in the visible region (see below). The triplet
energy of parent thioxanthone has been determined as
264 kJmolÀ1 versus 310 kJmolÀ1 for xanthone.[8] Based on
the above-mentioned considerations and based on our
successful work with a chiral xanthone as sensitizer,[9] it
seemed promising to prepare a chiral thioxanthone which
would incorporate a related hydrogen-bonding motif. Starting
from the thioether 2, which was accessible by nucleophilic
substitution of readily available 3-isopropoxy-4-nitrofluoro-
benzene[10] with methyl thiosalicylate,[11] the thioxanthone
ring was closed upon saponification of the ester group by an
intramolecular Friedel–Crafts acylation (Scheme 1).[9a]
Removal of the isopropyl group converted the arylether 3
into alcohol 4, which was linked to the activated acid rac-5 via
its mixed anhydride. Cyclization of the aromatic ortho-
nitroester rac-6 was achieved with thionyl chloride after
reduction to the respective aniline with tin(II) chloride.
Eventually, separation of the two enantiomers, 1 and ent-1,
was performed by semipreparative HPLC using a chiral
stationary phase.
One of the most fascinating aspects of photochemistry is the
fact that visible light can potentially be used to drive chemical
transformations. Many photochemical textbook reactions, for
example, the [2+2] photocycloaddition[1] or the Paternꢀ–
Bꢁchi reaction,[2] originate from irradiation experiments
performed with sunlight. With the advent of artificial light
sources, however, UV irradiation became common practice in
photochemistry and replaced visible light irradiation sour-
ces.[3] Indeed, for many important chromophores (carbonyl
compounds, enones, arenes, etc.) direct excitation requires
a wavelength region of l = 250–350 nm, which is nicely
covered by medium-pressure mercury lamps. Despite the
high efficiency of photochemical reactions occurring in the
UV region, it is important to realize that the energy of a l =
400 nm photon corresponds to about 300 kJmolÀ1, which is
significantly higher than the triplet energy of many photo-
chemical substrates. If suitable sensitizers are found, and they
are competent of transferring light energy to these substrates,
visible-light irradiation can be similarly or even more
effective than direct UV excitation. Along these lines, Yoon
et al. have recently reported that an iridium complex with
a triplet energy (ET) of 255 kJmolÀ1 is capable of sensitizing
the intramolecular [2+2] photocycloaddition reaction of
certain styrenes (ET ꢀ 250 kJmolÀ1).[4] The catalyst loading
was low (1 mol%) and reaction yields were high if the
reaction was performed in a polar solvent (DMSO). A
drawback of this specific transformation—as of many other
[2+2] photocycloaddition reactions[5]—is the fact that it
produces exclusively racemic products from prochiral sub-
strates. Indeed, the control of absolute product configuration
in photochemical reactions remains a considerable chal-
The thioxanthone 1 is a yellow-colored compound, the
UV/Vis spectrum of which exhibits a maximum at lmax
=
387 nm with a molar absorption coefficient e = 4540mÀ1 cmÀ1
(Figure 1). Although its triplet energy has not yet been
determined, it is evident from photostability studies that the
triplet state of 1 is less aggressive towards hydrogen abstrac-
tion[12] than the triplet state of its xanthone analogue. While
the latter compound decomposed instantaneously (< 10 min)
upon irradiation (l = 366 nm) in toluene,[13] 1 showed a detect-
able lifetime at l = 400–700 nm. After 60 minutes approxi-
mately 35% of the compound was degraded. In trifluoroto-
luene (PhCF3) the stability was even higher. After 60 minutes,
less than 10% of the material was decomposed based on UV/
Vis measurements (see the Supporting Information). The
thioxanthone 1 therefore seemed well suited to catalyze
photochemical reactions[14] in trifluorotoluene while employ-
ing a visible-light source.
[*] Dr. R. Alonso, Prof. Dr. T. Bach
Lehrstuhl fꢀr Organische Chemie I and Catalysis Research Center
(CRC), Technische Universitꢁt Mꢀnchen
Lichtenbergstr. 4, 85747 Garching (Germany)
E-mail: thorsten.bach@ch.tum.de
Homepage: http://www.oc1.ch.tum.de/home_en/
[**] This project was supported by the Deutsche Forschungsgemein-
schaft (DFG) and by the Fonds der Chemischen Industrie. R.A. is
the recipient of an Alexander von Humboldt research fellowship. We
thank Olaf Ackermann and Marcus Wegmann for help with the
HPLC analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201310997.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 1: Variation of reaction conditions for the enantioselective intra-
molecular [2+2] photocycloaddition of 7 catalyzed by 1.
Entry[a] c [mm] equiv of 1 T [8C] t [min][b] Yield [%][c] ee [%][d]
1
2
3
4
5
6
7
2.5
2.5
2.5
2.5
2.5
1.0
5.0
–
À25
À25
RT
À25
À25
À25
À25
60
60
60
30
75
30
75
–
88
83
86[e]
90
99
86
–
0.1
0.1
0.1
0.05
0.1
0.1
90
84
92
87
89
91
Scheme 1. Synthesis of 1. Reaction conditions: a) KOH (10 equiv),
MeOH, reflux, 5 h, 62%; b) TFA (24.3 equiv), TFAA (26.7 equiv),
CH2Cl2, RT, 16 h, 60%; c) TFA, reflux, 3 d, 87%; d) rac-5 (1.1 equiv),
tBuOCOCl (1.1 equiv), NEt3 (1.2 equiv), CH2Cl2, 08C, 5 h; 4
(1.0 equiv), DMAP (0.1 equiv), THF, RT, 16 h, 93%; e) SnCl2
(5.0 equiv), THF, reflux, 16 h; f) SOCl2 (3.0 equiv), py (6.5 equiv), PhH,
reflux, 6 h, 52% over two steps; g) ChiralPak AD-H (250ꢂ4.6 mm) n-
hexane/iPrOH 50:50, 1 mLminÀ1, tR(1)=5.7 min, tR(ent-1)=46.9 min.
DMAP=4-(N,N-dimethylamino)pyridine, TFAA=trifluoroacetic anhy-
dride, TFA=trifluoroacetic acid, THF=tetrahydrofuran.
[a] All reactions were conducted using a Rayonet RPR-100 reactor with 16
Osram L 8W/640 cool white lamps (Duran apparatus) as the irradiation
source in deaerated trifluorotoluene. [b] Elapsed reaction time to achieve
100% conversion of the starting material. [c] Yield of isolated product
after chromatographic purification. [d] The enantiomeric excess (ee) was
calculated from the enantiomeric ratio, which was determined by HPLC
analysis using a chiral stationary phase (see the Supporting Informa-
tion). [e] 90% conversion.
enantioselective reactions under ambient conditions by solar
irradiation.[16] Since the catalyst is not completely stable in
trifluorotoluene it was probed whether shorter reaction times
lead to higher enantioselectivities, and it was indeed found to
be the case (entry 4). Lowering the amount of catalyst led to
a slightly increased reaction time and to a minimal decrease in
enantioselectivity (entry 5). The substrate concentration
influenced the time required for complete conversion but
not the enantioselectivity (entries 6 and 7).
Additional experiments showed that—as indicated by the
UV/Vis studies—catalyst decomposition is not severe. After
an irradiation time of 60 minutes (Table 1, entry 2), catalyst
1
recovery was (85 Æ 10)% and a clean H NMR spectrum of
the catalyst was obtained (see the Supporting Information).
Moreover, since the employed visible-light source showed
minimal emission in the UV range, we also used another set of
lamps (RPR-4190 ꢂ), which do not emit any UV light. Under
conditions otherwise identical to those of entry 4 in Table 1,
conversion was complete and 8 was obtained in 84% yield
and with 96% ee. The catalyst recovery was also high in this
experiment [(85 Æ 10)%].
Figure 1. UV/Vis spectra of 1 (c=0.5 mm) and 7 (c=0.5 mm) in
trifluorotoluene (PhCF3).
The quinolone 7 was selected as a test substrate because it
shows—compared to other quinolones—a relative long wave-
length absorption, thus indicating that the energy of its triplet
excited state might be lower than the triplet state energy of
quinolone (ET = 276 kJmolÀ1).[15] As expected from its spec-
tral properties (Figure 1), the compound is readily converted
into 8 upon irradiation at l = 300 nm.[9b] With a visible-light
source, however, there is no reaction to be observed after
60 minutes in trifluorotoluene at À258C (Table 1, entry 1). To
our delight, addition of 10 mol% of 1 to the reaction mixture
led, with all other parameters unchanged, to a clean intra-
molecular [2+2] photocycloaddition which generated 8 in
88% yield with 90% ee (entry 2). The reaction remained
highly enantioselective even if performed at room temper-
ature (entry 3), a feature which may eventually allow highly
Given the success encountered in the enantioselective
reaction of the test substrate 7, related quinolones were
prepared and subjected to the optimal reaction conditions
(Table 1, entry 2: c = 2.5 mm, 10 mol% catalyst, 60 min irra-
diation time at À258C). The synthesis of the starting materials
was straightforward (see the Supporting Information). Com-
pounds with an all-carbon tether were prepared from
commercially available 4-methylquinolone by deprotonation
and subsequent alkylation.[17] Oxygen and nitrogen analogues
were synthesized by nucleophilic substitution of 4-bromome-
thylquinolone[18] with the respective allylic alcohol or amine.
Irradiation experiments proceeded smoothly and deliv-
ered the expected tetracyclic products 9–14 in excellent yields
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
(79–97%) and with high enantioselectivities (87–94% ee;
Figure 2). No other regio- or diastereoisomers were detect-
able. The products 11 and 13, having a newly generated all-
carbon bicyclo[3.2.0]heptane skeleton, showed the highest
Figure 3. Mechanistic model for sensitization and enantioface differ-
entiation in the complex 1·7. The unshaded arrow indicates the
approach of the tethered alkene onto the excited quinolone double
bond.
antenna, which transmits the energy of the absorbed photon
to the quinolone, presumably by triplet-energy transfer.[22]
Enantioface differentiation is provided by the planar thio-
xanthone so that attack at the quinolone double bond occurs
with high selectivity. The deviation in enantioselectivity when
comparing products 9–14 (Figure 2) does not necessarily
reflect the binding properties of the individual substrates but
may also be related to the rate of intramolecular addition (to
a 1,4-diradical) versus dissociation of the complex 1·7.[9c]
In summary, the chiral thioxanthone 1 was synthesized,
and was shown to act as an efficient sensititzer for enantio-
selective intramolecular [2+2] photocycloaddition reactions.
To the best of our knowledge, 1 is the first chiral photocatalyst
capable of processing visible light for a highly enantioselec-
tive photochemical reaction. Its relatively high triplet energy
should enable other triplet-sensitized reactions to occur in an
enantioselective fashion.
Figure 2. The products 9–14 of the enantioselective intramolecular
[2+2] photocycloaddition of various quinolones catalyzed by 1.
ee values, which even exceeded the ee value of the unsub-
stituted product 8 (Table 1). The enantioselectivity recorded
for the heteroatom analogues 9, 10, 12, and 14 were slightly
lower. The time required for the complete conversion into 10
was four hours and it was 90 minutes for conversion into 12.
All other reactions went to completion after 60 minutes.
Regarding the olefin configuration at the reacting alkenyl
chain, it was recognized in earlier studies that a non-
stereospecific reaction course is typical for reactions proceed-
ing via triplet intermediates.[19] Intermediate 1,4-diradicals
exhibit a significantly long lifetime to allow free rotation
before final ring closure.[20] In agreement with the postulated
triplet-energy transfer by 1, it was found that indeed (E)-4-
(hept-4-enyl)quinolone (15) gave a mixture of two products
(16), in which the respective cis-isomer 16a prevailed
(Scheme 2). The trans-isomer 16b, which would be the
Received: December 19, 2013
Published online: && &&, &&&&
Keywords: cycloaddition · enantioselectivity · hydrogen bonds ·
.
organocatalysis · photochemistry
[1] a) G. Ciamician, P. Silber, Ber. Dtsch. Chem. Ges. 1908, 41, 1928 –
1935; b) G. Bꢁchi, I. M. Goldman, J. Am. Chem. Soc. 1957, 79,
4741 – 4748.
[2] a) E. Paternꢀ, G. Chieffi, Gazz. Chim. Ital. 1909, 39 – 1, 341 –
361; b) G. Bꢁchi, C. G. Inman, E. S. Lipinsky, J. Am. Chem. Soc.
1954, 76, 4327 – 4331.
[3] H.-D. Scharf, J. Fleischhauer, J. Aretz in Methoden der
Organischen Chemie (Houben-Weyl), 4th ed., Vol. IV/5a (Ed.:
E. Mꢁller), Thieme, Stuttgart, 1975, pp. 41 – 89.
[4] Z. Lu, T. P. Yoon, Angew. Chem. 2012, 124, 10475 – 10478;
Angew. Chem. Int. Ed. 2012, 51, 10329 – 10332.
Scheme 2. Intramolecular enantioselective [2+2] photocycloaddition of
(E)-4-(hept-4-enyl)quinolone (15).
[5] General reviews on [2+2] photocycloaddition reactions: a) J. P.
Hehn, C. Mꢁller, T. Bach in Handbook of Synthetic Photo-
chemistry (Eds.: A. Albini, M. Fagnoni), Wiley-VCH, Weinheim,
2009, pp. 171 – 215; b) S. A. Fleming in Molecular and Supra-
molecular Photochemistry, Vol. 12 (Eds.: A. G. Griesbeck, J.
Mattay), Marcel Dekker, New York, 2005, pp. 141 – 160; c) P.
Margaretha in Molecular and Supramolecular Photochemistry,
Vol. 12 (Eds.: A. G. Griesbeck, J. Mattay), Marcel Dekker, New
York, 2005, pp. 211 – 237; d) J.-P. Pete in CRC Handbook of
Organic Photochemistry and Photobiology, 2nd ed. (Eds.: W.
Horspool, F. Lenci), CRC, Boca Raton, 2004, pp. 71/1 – 71/14;
e) T. Bach, Synthesis 1998, 683 – 703; f) J.-P. Pete, Adv. Photo-
chem. 1996, 21, 135 – 216; g) J. Mattay, R. Conrads, R. Hoffmann
in Methoden der Organischen Chemie (Houben-Weyl) 4th ed.,
expected product of a stereospecific reaction, was found to
be the minor diastereoisomer. The enantioselectivity of the
intramolecular [2+2] photocycloaddition was again excellent
(94% and 89% ee).
The hydrogen-bonding properties of the 1,5,7-trimethyl-3-
azabicyclo[3.3.1]nonan-2-one skeleton, to which the thioxan-
thone part of 1 is connected, have been established[21] and it
appears safe to assume that a precoordination of quinolone
substrates occurs, as depicted in Figure 3, to give the putative
complex 1·7. The thioxanthone acts as light-harvesting
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Vol. E 21c (Eds.: G. Helmchen, R. W. Hoffmann, J. Mulzer, E.
Schaumann), Thieme, Stuttgart, 1995, pp. 3085 – 3132; h) M. T.
Crimmins, T. L. Reinhold, Org. React. 1993, 44, 297 – 588.
[6] Reviews: a) C. Yang, Y. Inoue, Chem. Soc. Rev. 2014, DOI:
10.1039/c3cs60339c; b) C. Mꢁller, T. Bach, Aust. J. Chem. 2008,
61, 557 – 564.
2345 – 2347; b) P. C. Wong, D. R. Arnold, Can. J. Chem. 1979, 57,
1037 – 1049; c) A. Padwa, T. J. Blacklock, R. Loza, R. Polnias-
zek, J. Org. Chem. 1980, 45, 2181 – 2189; d) A. Padwa, M. W.
Wannmaker, Tetrahedron Lett. 1986, 27, 5817 – 5820.
[15] S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photo-
chemistry, 2nd ed., Dekker, New York, 1993, p. 82.
[7] For a different approach to enantioselective [2+2] photocy-
cloaddition reactions, see: R. Brimioulle, T. Bach, Science 2013,
342, 840 – 843.
[16] Reviews: a) P. Esser, B. Pohlmann, H.-D. Scharf, Angew. Chem.
1994, 106, 2093 – 2108; Angew. Chem. Int. Ed. Engl. 1994, 33,
2009 – 2023; b) M. Oelgemçller, C. Jung, J. Mattay, Pure Appl.
Chem. 2007, 79, 1939 – 1947; c) S. Protti, M. Fagnoni, Photochem.
Photobiol. Sci. 2009, 8, 1499 – 1516; d) T. P. Yoon, M. A. Ischay,
J. Du, Nat. Chem. 2010, 2, 527 – 532.
[17] a) O. Martin, E. de La Cuesta, C. AvendaÇo, Tetrahedron 1995,
51, 7547 – 7554; b) M. Dressel, T. Bach, Org. Lett. 2006, 8, 3145 –
3147; c) P. Selig, T. Bach, J. Org. Chem. 2006, 71, 5662 – 5673.
[18] R. G. Kalkhambkar, G. M. Kulkarni, C. M. Kamanavalli, N.
Premkumar, S. M. B. Asdaq, C. M. Sun, Eur. J. Med. Chem. 2008,
43, 2178 – 2188.
[19] a) D. Becker, M. Nagler, S. Hirsh, J. Ramun, J. Chem. Soc. Chem.
Commun. 1983, 371 – 373; b) D. Becker, M. Nagler, Y. Sahali, N.
Haddad, J. Org. Chem. 1991, 56, 4537 – 4543; c) D. Becker, N.
Klimovich, Tetrahedron Lett. 1994, 35, 261 – 264.
[20] D. I. Schuster, G. Lem, N. A. Kaprinidis, Chem. Rev. 1993, 93, 3 –
22.
[8] S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photo-
chemistry, 2nd ed., Dekker, New York, 1993, p. 80, 91.
[9] a) C. Mꢁller, A. Bauer, T. Bach, Angew. Chem. 2009, 121, 6767 –
6769; Angew. Chem. Int. Ed. 2009, 48, 6640 – 6642; b) C. Mꢁller,
M. M. Maturi, A. Bauer, M. C. Cuquerella, M. A. Miranda, T.
Bach, J. Am. Chem. Soc. 2011, 133, 16689 – 16697; c) M. M.
Maturi, M. Wenninger, R. Alonso, A. Bauer, A. Pçthig, E.
Riedle, T. Bach, Chem. Eur. J. 2013, 19, 7461 – 7472.
[10] A. B. Reitz, M. E. McDonnell, E. W. Baxter, E. E. Codd, W.-N.
Wu, Bioorg. Med. Chem. Lett. 1997, 7, 1927 – 1930.
[11] X. Du, H. Liu, D.-M. Du, Tetrahedron: Asymmetry 2010, 21,
241 – 246.
[12] a) P. J. Wagner, B.-S. Park, Org. Photochem. 1991, 11, 227 – 3667;
b) E. C. Lathioor, W. J. Leigh, Photochem. Photobiol. 2006, 82,
291 – 300.
[13] A. Bauer, Ph.D. thesis, TU Mꢁnchen, 2006.
[14] For thioxanthone-sensitized [2+2] photocycloaddition reactions,
see: a) A. Padwa, T. J. Blacklock, J. Am. Chem. Soc. 1977, 99,
[21] A. Bakowski, M. Dressel, A. Bauer, T. Bach, Org. Biomol.
Chem. 2011, 9, 3516 – 3529.
[22] D. L. Dexter, J. Chem. Phys. 1953, 21, 836 – 850.
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Photochemistry
R. Alonso, T. Bach*
&&&& — &&&&
A Chiral Thioxanthone as an
Organocatalyst for Enantioselective [2+2]
Photocycloaddition Reactions Induced by
Visible Light
Lighting up S: The thioxanthone 1, which
was synthesized in a concise fashion
from methyl thiosalicylate, exhibits a sig-
nificant absorption in the visible-light
region. It allows for an efficient enantio-
selective catalysis of intramolecular [2+2]
photocycloaddition reactions, presuma-
bly by triplet-energy transfer. Ts=4-tol-
uenesulfonyl.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!