Spin-selective generation of triplet nitrenes: Olefin aziridination through visible-light photosensitization of azidoformates

Article abstract of DOI:10.1002/anie.201510868

Azidoformates are interesting potential nitrene precursors, but their direct photochemical activation can result in competitive formation of aziridination and allylic amination products. Herein, we show that visible-light-activated transition-metal comple

Full text of DOI:10.1002/anie.201510868

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201510868
German Edition: DOI: 10.1002/ange.201510868
Photochemistry
Spin-Selective Generation of Triplet Nitrenes: Olefin Aziridination
through Visible-Light Photosensitization of Azidoformates
Spencer O. Scholz, Elliot P. Farney, Sangyun Kim, Desiree M. Bates, and Tehshik P. Yoon*
Abstract: Azidoformates are interesting potential nitrene
precursors, but their direct photochemical activation can
result in competitive formation of aziridination and allylic
amination products. Herein, we show that visible-light-acti-
vated transition-metal complexes can be triplet sensitizers that
selectively produce aziridines through the spin-selective photo-
generation of triplet nitrenes from azidoformates. This
approach enables the aziridination of a wide range of alkenes
and the formal oxyamination of enol ethers using the alkene as
the limiting reagent. Preparative-scale aziridinations can be
easily achieved under continuous-flow conditions.
Seminal studies by Lwowski et al., however, demonstrated
that while singlet carbethoxynitrenes competitively undergo
both amination and aziridination reactions, triplet
carbethoxynitrenes react selectively with alkenes to afford
À
aziridines with comparatively slow reaction with allylic C H
bonds.[10] Thus the fundamental challenge in photochemical
aziridination reactions appears not to be the absolute
reactivity of free nitrenes but rather the unselective produc-
tion of both singlet and triplet nitrenes from direct photolysis
of azides.
We wondered if chemoselective photochemical aziridina-
tion reactions could be achieved through triplet sensitization,
which would produce nitrenes selectively in the triplet state.
Our laboratory previously studied the use of visible-light-
absorbing transition-metal complexes to sensitize vinyl azides
towards intramolecular heterocyclic ring-closing reactions
(Scheme 1).[11] Quite recently, Kçnig and co-workers reported
A
ziridines are versatile intermediates for the synthesis of
nitrogen-containing compounds.[1] Many important natural
products also feature aziridines as their principal bioactive
functionality.[2] However, methods for aziridine synthesis are
somewhat underdeveloped,[3] particularly in comparison to
the wealth of methods available for epoxide synthesis. The
most widely utilized methods for alkene aziridination involve
the generation of metallonitrenes from iminoiodinane
reagents.[4] These methods, unfortunately, produce stoichio-
metric haloarene byproducts, and there has consequently
been significant interest in the use of alternate nitrene
precursors for aziridination reactions.[5] Organic azides
appear particularly attractive in this regard because they
generate nitrenes by expelling dinitrogen as the sole stoichio-
metric byproduct. Several laboratories, including notably the
Zhang[6] and Katsuki[7] groups, have reported pioneering
advances in catalytic aziridination with organoazides. Never-
theless, these processes often require a large excess of alkene,
and many methods are limited to styrenic olefins. Thus, there
remains a need for new approaches to aziridination that
utilize organoazides as nitrene precursors.
Scheme 1. Photocatalytic activation of azides by visible-light triplet
sensitization.
Photochemical activation offers one potential solution.
Electronically excited organic azides rapidly decompose to
form reactive free nitrenes.[8] However, attempts to perform
intermolecular aziridination reactions by direct photolysis of
azidoformates typically produce complex mixtures containing
both aziridines and allylic amination products,[9] a result that
has led to the pervasive notion that free nitrenes are too
reactive to provide synthetically useful chemoselectivities.
an intriguing method for photocatalytic amidation of elec-
tron-rich heterocycles, the key step of which was proposed to
involve triplet sensitization of a benzoyl azide.[12] To the best
of our knowledge, however, the use of triplet sensitizers to
promote chemoselective intermolecular alkene aziridination
reactions has not previously been described. Herein, we
demonstrate that visible-light triplet sensitization of azido-
formates enables the preparation of a range of structurally
diverse aziridines. Notably, this method utilizes the alkene as
the limiting reagent, provides high yields for both aliphatic
and aromatic alkenes, and exhibits excellent selectivity for
aziridination over allylic amination.
[*] S. O. Scholz, E. P. Farney, S. Kim, D. M. Bates, Prof. T. P. Yoon
Department of Chemistry
University of Wisconsin-Madison
1101 University Avenue, Madison, WI 53706 (USA)
E-mail: tyoon@chem.wisc.edu
Supporting information and ORCID(s) from the author(s) for this
article are available on the WWW under http://dx.doi.org/10.1002/
anie.201510868.
We elected to focus our investigations on azidoformates as
nitrene precursors, based on several considerations. First,
Angew. Chem. Int. Ed. 2016, 55, 2239 –2242
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2239

Angewandte
Communications
Chemie
unlike acyl azides, nitrenes derived from azidoformates have
little proclivity to undergo Curtius-type rearrangements that
might be competitive with bimolecular aziridination reac-
tions.[13] Second, azidoformates are less easily reduced than
acyl azides, which should disfavor undesired photoreductive
pathways for azide decomposition via non-nitrene-mediated
pathways.[14] Finally, we computationally estimated the
lowest-energy triplet excited state of ethyl azidoformate to
be approximately 54 kcalmolÀ1,[15] thermodynamically within
a range accessible by the Ir complexes we have previously
exploited as energy-transfer photocatalysts.[11,16]
We began by examining the reaction of ethyl azidofor-
mate (3a; 0.5 mmol) with cyclohexene (0.1 mmol) in the
presence of a range of photoactive RuII and IrIII complexes
(Table 1, entries 1–4). From this study, [Ir(ppy)2(dttbpy)]PF6
(1c) emerged as the most effective photocatalyst, providing
17% yield of aziridine after 4 h of irradiation with a 15 W blue
LED lamp (Table 1, entry 3). Notably, we detected neither
allylic amination nor the formation of other products
characteristic of uncontrolled nitrene insertion. A screen of
azidoformates revealed that introduction of electron-with-
drawing substituents increased the rate of aziridination
(entries 5 and 6); when 2,2,2-trichloroethyl azidoformate
(TrocN3) was used, the consumption of cyclohexene was
complete after 4 h, and the aziridine was formed in 77% yield
(entry 6).[17] Interestingly, an excess of the azide was required
only for the optimal rate; lowering the azide concentration
resulted in a slower reaction, but the mass balance of the azide
remained high (entries 7 and 8). Similarly, increasing the
reaction scale to 0.4 mmol resulted in partial conversion into
the aziridine at 4 h (entry 9), although full conversion could
be restored by increasing the reaction time to 20 h (entry 10).
These results indicate the reaction is likely photon-limited,
consistent with the high molar absorptivity of this class of
photocatalysts.[18] Control studies showed that both catalyst
and light (entries 11 and 12) are essential for the reaction to
proceed. Finally, in line with our original hypothesis, we
studied the direct UV-promoted photochemical reaction
(lexc = 310 nm) of 2 with 3c and detected an unselective
mixture of allylic amination and aziridination products,
highlighting the value of triplet sensitization for chemo-
selective aziridination.[19]
Table 1: Optimization studies for photocatalytic aziridination.[a]
Figure 1 summarizes the scope of the photocatalytic
aziridination. A variety of cyclic alkenes undergo aziridina-
tion smoothly under optimized conditions (5–8). The method
Entry
Catalyst
Azidoformate (R)
Equiv. azide
Yield[b]
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1c
1c
1c
1c
1c
1c
1c
none
3a (CH2CH3)
3a (CH2CH3)
3a (CH2CH3)
3a (CH2CH3)
3b (CH2CH2Cl)
3c (CH2CCl3)
3c (CH2CCl3)
3c (CH2CCl3)
3c (CH2CCl3)
3c (CH2CCl3)
3c (CH2CCl3)
3c (CH2CCl3)
5
5
5
5
5
5
3
1
5
5
5
5
0%
<2%
17%
<2%
50%
77%
65%
41%
50%
75%[e]
0%
9[c]
10[c,d]
11[f]
12
0%
Figure 1. Scope of aziridination products available by photocatalytic
activation. Reactions were conducted using alkene (0.40 mmol) in
degassed solvent and irradiated with a 15 W blue LED flood lamp for
20 h. Yields represent the averaged values from two reproduced
experiments. [a] Isolated as a 13:1 mixture of aziridine to imine (see
the Supporting Information for more information). [b] Reaction irradi-
ated for 30 h. Troc=2,2,2-trichloroethoxycarbonyl; Ts=tosylate;
TBS=tert-butyldimethylsilyl.
[a] Unless otherwise noted, all reactions were conducted using
0.10 mmol of 2 in degassed solvent and irradiated with a 15 W blue LED
flood lamp for 4 h. [b] Yields determined by using 1H NMR analysis using
1,4-bis(trimethylsilyl)benzene as an internal standard (unless otherwise
noted). [c] Reaction conducted on 0.40 mmol scale with respect to 2.
[d] Reaction irradiated for 20 h. [e] Yield of isolated product. [f] Reaction
conducted in the dark.
2240 www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 2239 –2242

Angewandte
Communications
Chemie
provides good yields of aziridines derived from both aliphatic
and styrenic alkenes (9 and 10), and internal olefins (11 and
12) undergo aziridination as successfully as terminal alkenes.
A variety of functional groups are tolerated, including
protected amines and alcohols (13–15), potentially photo-
sensitive halides (16), and Lewis basic heterocycles (17).
Electron-deficient alkenes react at a slower rate, consistent
with the electrophilic nature of the nitrene intermediate,
often requiring somewhat extended reaction times (18 and
19). Consistent with this rate difference, a substrate bearing
both an a,b-unsaturated ketone and an aliphatic trisubstituted
alkene undergoes selective aziridination at the electron-rich
alkene (20).
Reactions of alkenes bearing electron-releasing moieties
were also explored (Scheme 2). However, the resulting
aziridines were formed with varying amounts of an isomeric
Scheme 3. Studies supporting a stepwise diradical mechanism for
aziridination via a triplet nitrene.
considered the possibility that the aziridination was not
proceeding via a nitrene intermediate but rather by a [3+2]
cycloaddition of the carbamoyl azide followed by a photo-
catalytic decomposition of a 1,2,3-triazoline intermediate.[24]
To test this hypothesis, we prepared norbornene-derived
triazoline 31 and irradiated it for 4 h in the presence and
absence of 1c (Scheme 3C). Although we detected slow
conversion of 31 to aziridine 8, the rate of this spontaneous
ring contraction is not influenced by the presence of photo-
catalyst 1c and is kinetically incompetent to be a significant
contributor to the formation of aziridine under catalytic
conditions. Hence, all available evidence is consistent with the
triplet nitrene postulated in our reaction design.
One notable result of our optimization studies was the
observation that the rate of azide consumption decreased as
the scale of the reaction increased. This phenomenon is
consistent with a photon-limited process, which is common in
photoreactions that involve chromophores with high molar
absorptivity. Such reactions often benefit from flow reactors
whose reduced dimensionality can increase the photon flux
and allow for efficient scale-up.[25] We thus examined the
aziridination of cyclohexene in a flow reactor and detected
significantly increased efficiency, with a high yield and short
residence time (tR) of 2.3 h (Scheme 4). More importantly, we
also found that the concentration of azide could be lowered to
2.0 equiv in flow, albeit with a somewhat longer residence
time (tR = 8 h). Finally, to demonstrate the scalability of the
aziridination reaction in flow, we conducted an aziridination
reaction on preparative scale and produced 785 mg of pure
aziridine over a 16 h period which was isolated in 70% yield.
In conclusion, we have demonstrated that photocatalytic
aziridinations can be achieved by using a visible-light-
Scheme 2. Scope of oxazoline products available by photocatalytic
activation (R=CH2CCl3).
oxazoline. Anethole, for example, yielded a 9:1 mixture of
aziridine 22 and oxazoline 23 in a combined yield of 79%. We
wondered if more electron-rich substrates such as enol ethers
might favor the formation of these formal oxyamination
products.[20] Indeed, the reaction of benzofuran afforded
oxazoline 24 exclusively,[21] with no azirdine detectable in the
reaction mixture. Glycals similarly underwent exclusive oxy-
amination, yielding 25 and 26 with excellent d.r. values.
Our reaction design was based on the hypothesis that
triplet nitrenes would be selectively formed through Dexter
energy transfer from the triplet excited state of a transition-
metal photocatalyst.[22] Aziridination reactions of triplet
nitrenes should proceed in a stepwise fashion involving
electrophilic radical addition to an alkene followed by
relatively slow intersystem crossing and fast ring closure of
the resulting singlet 1,3-diradical. Consistent with this expect-
ation, aziridinations of stereodefined cis and trans alkenes are
not stereospecific (Scheme 3A). The radical nature of the
intermediate was further verified by subjecting cyclopropyl
alkene 29 to the aziridination conditions (Scheme 3B). This
reaction resulted in the scrambling of the cyclopropane
stereochemistry in the aziridinated products 30, a result that is
only consistent with the reversible ring fragmentation of
a cyclopropyl carbinyl radical intermediate.[23] Finally, we
Angew. Chem. Int. Ed. 2016, 55, 2239 –2242
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2241

Angewandte
Communications
Chemie
Ruppel, S. Zhu, J. A. Perman, X. P. Zhang, Chem. Commun.
2009, 4266 – 4268.
[7] a) K. Omura, M. Murakami, T. Uchida, R. Irie, T. Katsuki,
Chem. Lett. 2003, 32, 354 – 355; b) K. Omura, T. Uchida, R. Irie,
T. Katsuki, Chem. Commun. 2004, 2060 – 2061; c) H. Kawabata,
K. Omura, T. Katsuki, Tetrahedron Lett. 2006, 47, 1571 – 1574;
d) H. Kawabata, K. Omura, T. Katsuki, Chem. Asian J. 2007, 2,
248 – 256.
[8] Azides and Nitrenes, Reactivity and Utility (Ed.: E. F. V. Scriven),
Academic Press, New York, 1984, pp. 1 – 291.
Scheme 4. Aziridination of 2 in flow.
[9] W. Lwowski, T. Mattingly, J. Am. Chem. Soc. 1965, 87, 1947 –
1958.
[10] a) W. Lwowski, F. P. Woerner, J. Am. Chem. Soc. 1965, 87, 5491 –
5492; b) W. Lwowski, Angew. Chem. Int. Ed. Engl. 1967, 6, 897 –
906; Angew. Chem. 1967, 79, 922 – 931.
[11] E. P. Farney, T. P. Yoon, Angew. Chem. Int. Ed. 2014, 53, 793 –
797; Angew. Chem. 2014, 126, 812 – 816.
[12] E. Brachet, T. Ghosh, I. Ghosh, B. Kçnig, Chem. Sci. 2015, 6,
987 – 992.
[13] J. Liu, S. Mandel, C. M. Hadad, M. S. Platz, J. Org. Chem. 2004,
69, 8583 – 8593.
absorbing transition-metal photosensitizer for the spin-selec-
tive generation of triplet nitrenes from azidoformates. A wide
range of alkenes can be aziridinated or oxyaminated under
operationally facile batch conditions, and the use of a flow
reactor enables the reaction to be conducted on a preparative
scale. Further efforts in our lab will continue to develop
visible-light triplet sensitization as a general, conceptually
À
novel strategy for C N bond forming reactions using organo-
[14] Y. Chen, A. S. Kamlet, J. B. Steinman, D. R. Liu, Nat. Chem.
2011, 3, 146 – 153.
azides as the nitrene precursors.
[15] See the Supporting Information for further information regard-
ing the molecular energies and coordinates of these computa-
tional studies.
[16] a) Z. Lu, T. P. Yoon, Angew. Chem. Int. Ed. 2012, 51, 10329 –
10332; Angew. Chem. 2012, 124, 10475 – 10478; b) A. E. Hurtley,
Z. Lu, T. P. Yoon, Angew. Chem. Int. Ed. 2014, 53, 8991 – 8994;
Angew. Chem. 2014, 126, 9137 – 9140.
[17] Organoazide compounds are potentially explosive, and proper
safety procedures should always be used in their manipulation.
However, samples of TrocN3 stored at 08C in our laboratory
show no signs of decomposition after several months, and DSC
analysis indicates that TrocN3 is stable to 1578C. See: H. Lus, V.
Subbarayan, J. Tao, X. P. Zhang, Organometallics 2010, 29, 389 –
393.
Acknowledgements
We are grateful to Dale Kreitler and Prof. Sam Gellman for
the loan of the HPLC pump used in all flow experiments. We
thank Yuliya Preger for helpful discussions. We gratefully
acknowledge the NIH NIGMS for financial support
(GM095666) and Sigma–Aldrich for a generous gift of
iridium(III) chloride.
Keywords: azides · aziridination · chemoselectivity · nitrenes ·
photocatalysis
[18] L. Flamigni, A. Barbieri, C. Sabatini, B. Ventura, F. Barigelletti,
Top. Curr. Chem. 2007, 281, 143 – 203.
How to cite: Angew. Chem. Int. Ed. 2016, 55, 2239–2242
Angew. Chem. 2016, 128, 2279–2282
[19] See the Supporting Information for details of these experiments.
[20] J. Du Bois, C. S. Tomooka, J. Hong, E. M. Carreira, J. Am. Chem.
Soc. 1997, 119, 3179 – 3180.
[21] Kçnig et al. reported the formation of an analogous oxazoline
product in two examples. In the reaction of benzoyl azide with
benzofuran, the oxazoline is produced as a 2:1 mixture with the
[1] a) X. E. Hu, Tetrahedron 2004, 60, 2701 – 2743; b) C.-Y. Huang,
A. G. Doyle, Chem. Rev. 2014, 114, 8153 – 8198; c) Aziridines
and Epoxides in Organic Synthesis (Ed.: A. K. Yudin), Wiley-
VCH, Weinheim, 2006.
À
C H amination product in 50% overall yield. See Ref. [12].
[22] Addition of anthracene to an aziridination reaction completely
inhibits consumption of the azidoformate, indicating that the
excited Ir triplet complex reacts more rapidly with this triplet
quencher than with azidoformate. This experiment, unfortu-
nately, does not elucidate the mechanism by which azidoformate
and the photocatalyst interact in the productive reaction.
[23] a) P. Müller, C. Baud, Y. Jacquier, Can. J. Chem. 1998, 76, 738;
b) T. Benkovics, J. Du, I. A. Guzei, T. P. Yoon, J. Org. Chem.
2009, 74, 5545 – 5552; c) J. F. Van Humbeck, S. P. Simonovich,
R. R. Knowles, D. W. C. MacMillan, J. Am. Chem. Soc. 2010,
132, 10012 – 10014.
[24] For an example of the formation of aziridines by decomposition
of triazolines, see: P. Scheiner, J. Am. Chem. Soc. 1966, 88, 4759 –
4760.
[25] For a review, see: Z. J. Garlets, J. D. Nguyen, C. R. J. Stephen-
son, Isr. J. Chem. 2014, 54, 351 – 360.
[2] a) T. Kametani, T. Honda, Adv. Heterocycl. Chem. 1986, 39,
3373 – 3387; b) C. Botuha, F. Chemla, F. Ferreira, A. PØrez-Luna
in Heterocycles in Natural Product Synthesis (Eds.: K. C.
Majumdar, S. K. Chattopadhyay), Wiley-VCH, Weinheim,
2011, pp. 1 – 39.
[3] a) J. B. Sweeney, Chem. Soc. Rev. 2002, 31, 247 – 258; b) I. D. G.
Watson, L. Yu, A. K. Yudin, Acc. Chem. Res. 2006, 39, 194 – 206.
[4] a) P. Müller, C. Fruit, Chem. Rev. 2003, 103, 2905 – 2920; b) L.
Degennaro, P. Trinchera, R. Luisi, Chem. Rev. 2014, 114, 7881 –
7929.
[5] For reviews, see: a) T. Katsuki, Chem. Lett. 2005, 34, 1304 – 1309;
b) S. Fantauzzi, A. Caselli, E. Gallo, Dalton Trans. 2009, 5434 –
5443; c) T. G. Driver, Org. Biomol. Chem. 2010, 8, 3831 – 3846;
d) H. Lu, X. P. Zhang, Chem. Soc. Rev. 2011, 40, 1899 – 1909.
[6] a) G.-Y. Gao, J. E. Jones, R. Vyas, J. D. Harden, X. P. Zhang, J.
Org. Chem. 2006, 71, 6655 – 6658; b) J. V. Ruppel, J. E. Jones,
C. A. Huff, R. M. Kamble, Y. Chen, X. P. Zhang, Org. Lett. 2008,
10, 1995 – 1998; c) L.-M. Jin, X. Xu, H. Lu, X. Xui, L. Wojtas,
X. P. Zhang, Angew. Chem. Int. Ed. 2013, 52, 5309 – 5313;
Angew. Chem. 2013, 125, 5417 – 5421; d) V. Subbarayan, J. V.
Received: November 23, 2015
Published online: January 6, 2016
2242 www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 2239 –2242