Copper-Mediated [3+2] Annulation of 3-N-Hydroxyallylamines with Nitrosoarenes

Article abstract of DOI:10.1002/chem.201504784

Cu-mediated annulations of N-hydroxyallylamines with nitrosoarenes proceed through unprecedented formal [3+2] cycloadditions of N-hydroxyaminoallyl radicals with nitrosoarenes. Our mechanistic analysis opposes a 5-endo-trig cyclization involved in the final ring-closure step. To manifest the reaction utility, chemical elaborations of resulting isoxazolidinyl products into 2- or 3-substituted quinoline N-oxides and acyclic 1,3-diamino-2-ols are also described. Radical initiative! Cu-mediated annulations of N-hydroxyallylamines with nitrosoarenes proceed through unprecedented formal [3+2] cycloaddition reactions (see scheme). To manifest the reaction utility, chemical elaborations of resulting isoxazolidinyl products into 2- or 3-substituted quinoline N-oxides and acyclic 1,3-diamino-2-ols have been described.

Full text of DOI:10.1002/chem.201504784

DOI: 10.1002/chem.201504784
Communication
&
Synthetic Methods
Copper-Mediated [3+2] Annulation of 3-N-Hydroxyallylamines
with Nitrosoarenes
Satish Ghorpade, Prakash D. Jadhav, and Rai-Shung Liu*[a]
3 into highly functionalized N- and O-containing compounds
Abstract: Cu-mediated annulations of N-hydroxyallyl-
will be described.
amines with nitrosoarenes proceed through unprecedent-
ed formal [3+2] cycloadditions of N-hydroxyaminoallyl
radicals with nitrosoarenes. Our mechanistic analysis op-
poses a 5-endo-trig cyclization involved in the final ring-
closure step. To manifest the reaction utility, chemical
elaborations of resulting isoxazolidinyl products into 2- or
3-substituted quinoline N-oxides and acyclic 1,3-diamino-
2-ols are also described.
[3+2] Cycloaddition reactions are powerful tools to access five-
membered carbo- or heterocyclic compounds.[1] Allyl radicals
are versatile to react with various p-bond motifs to implement
formation of a CÀX (X=C, N) bond.[2–4] In the gaseous phase,
allylic radicals can undergo [3+2] cycloadditions with butadi-
enes,[3a] acetylenes,[3b] alkenes,[3c] allenes,[3d] and indenes,[3e]
Table 1 shows the optimizations of [3+2] annulation reac-
albeit with poor chemoselectivity. Such radical [3+2] cycloaddi-
tions are unlikely to occur in solution, because a final 5-endo-
trig cyclization is a difficult process according to Baldwin’s
rule.[5] We are aware of no example of [3+2] cycloadditions of
allyl radicals with any kind of 2p-bond motif, although 5-endo-
trig cyclizations have been reported for alkenes of special
types.[6] The development of such radical [3+2] cycloadditions
is highly desirable in allyl radical chemistry.
tions of 3-N-hydroxyallylamine 1a (1.0 equiv) with nitrosoben-
zene (2a; 2 equiv) using various Cu salts, oxidants, and sol-
vents. We initially tested the reaction with CuBr2, CuCl2, CuBr,
and CuCl (5 mol%) in toluene/O2 at 258C, affording isoxazo-
lidinyl product 3a in 24–58% yields (entries 1–4) that indicated
the CuI is more efficient than the CuII. The diazene oxide 2a’
was isolated in small proportions (18–20%). Interestingly, the
We reported Cu-catalyzed aerobic oxidations of 3-N-hydroxy-
allylamines to form initially nitroxy radicals I,[7,8] which were
convertible to 3-N-hydroxyaminoallylic radicals I’, possibly by
an intermolecular hydrogen transfer. The dimerization of I’
yielded no common 1,5-hexadiene derivatives, but afforded
1,4-dihydroxy-2,3-diaminocyclohexanes I’2 efficiently [Eq. (1)].[9]
Herein, we report Cu-mediated [3+2] annulation reactions of
3-N-hydroxyallylamines with nitrosoarenes to form isoxazo-
lidin-5-yl products 3 [Eq. (2)] with excellent diastereoselectivi-
ties (d.r. >25:1). Notably, the overall transformation represents
a 3-amino-1,2-dioxygenation of an allylamine skeleton. Our
mechanistic analysis opposes a 5-endo-trig radical cyclization;
instead, the success of these radical [3+2] annulations is attri-
buted to the relay action of allyl radical I’ and its N-hydroxy-
amino functionality. Chemical elaborations of resulting product
Table 1. Optimization of reaction conditions.
Entry Initiator [mol%] Oxidant
Solvent[a]
Products Yield [%][b]
1a
3a
2a’
1
2
3
4
5
6
7
8
CuBr2 (5)
CuCl2 (5)
CuBr (5)
CuCl (5)
[(IPr)CuCl] (5)
[(IPr)CuCl] (5)
[(IPr)CuCl] (5)
[(IPr)CuCl] (5)
[(IPr)CuCl] (5)
[(IPr)CuCl] (5)
–
O2
O2
O2
O2
O2
H2O2
toluene
toluene
toluene
toluene
toluene
toluene
–
–
–
–
–
–
–
–
–
74
–
–
24
27
58
48
73
54
56
62
20
26
18
22
traces
10
4
6
3
25
15
5
[c]
tBuOOH[c] toluene
O2
O2
N2
O2
N2
DCE
THF
toluene
toluene
toluene
9
67
10
11
12
traces
traces
60
[a] Dr. S. Ghorpade, P. D. Jadhav, Prof. Dr. R.-S. Liu
Department of Chemistry, National Tsing-Hua University
101, Sec. 2, Kuang-Fu Rd., Hsinchu, 30043, Taiwan (P.R. China)
E-mail: rsliu@mx.nthu.edu.tw
TEMPO (10)
[a] 1a (0.15m, 1 equiv) and 2a (0.30m, 2 equiv). [b] Product yields are re-
ported after purification through a neutral alumina column. [c] Molecular
sieve 4Š was added.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201504784.
Chem. Eur. J. 2016, 22, 2915 – 2919
2915
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
presence of the nitrosobenzene completely suppressed the
formation of oxidative dimerization product I’2. We envisage
that the acidic CuII complex likely coordinates with N-hydroxyl-
amine 1a to inhibit the formation of nitroxy radicals.[7] Accord-
ingly, we tested electron-rich [(IPr)CuCl] (IPr=1,3-bis(diisopro-
pylphenyl)imidazol-2-ylidene), which greatly increased the
yield of desired 3a (up to 73%) with diazene oxide 2a’ ob-
tained in a negligible amount. With [(IPr)CuCl] (5 mol%), other
oxidants, such as H2O2 and tert-butyl hydroperoxide (TBHP), af-
forded desired 3a in 54% and 56% yields, respectively (en-
tries 6–7). Other solvents, like 1,2-dichloroethane and THF, af-
forded compound 3a in 62% and 67% yields, respectively (en-
tries 8 and 9). In the absence of O2, no reaction occurred, and
initial 1a was recovered in 74% yield (entry 10). Likewise, the
reaction led to a mixture of complicated products if no Cu cat-
alysts were employed (entry 11). The use of 2,2,6,6-tetramethyl-
piperidine N-oxide (TEMPO) under N2 also implemented this
annulation, yielding desired 3a in 60% yield (entry 12). These
data suggest that both Cu/O2 and TEMPO serve as radical ini-
tiators. The isoxazolidinyl framework of compound 3a was in-
ferred from an X-ray diffraction of its relative 4 f (Scheme 1).[10]
Table 2. Tests on 3-N-hydroxyallylamines.
[a] 1b (0.14m, 1 equiv) and 2a (0.30m, 2 equiv). [b] Product yields are re-
ported after purification through a neutral alumina column.
The Cu-mediated reactions of C2-unsubstituted N-hydroxyal-
lylamines 1l and 1m with nitrosobenzene (2a) proceeded
through a distinct aerobic oxidation, such that O2 was an oxi-
dant rather than a radical activator; the optimized yields of de-
sired 3l and 3m were 74% and 67%, respectively, with
a molar ratio of amine/2a=2.0:1.1 [Eq. (3)]. Herein, initial N-hy-
droxyallylamine 1l acted as a nucleophile, which was replace-
able by other N-hydroxyallylamines. We successfully employed
N-hydroxyaminopropane (4 equiv) as a nucleophile, affording
compound 3l’ in 78% yield [Eq. (4)].
Scheme 1. Reaction scope of nitrosoarenes.
Table 2 shows the generalization of this [3+2] annulation
using various 3-N-hydroxyallylamines 1b–1k (1.0 equiv) and ni-
trosobenzene (2a; 2.0 equiv). The reactions were mediated
with [(IPr)CuCl] (5 mol%) under O2 (1 atm.) in toluene (258C,
0.5–1.0 h), yielding 3b–3k as single diastereomeric products
(d.r. >25:1). We tested the reactions on N-hydroxyallylamines
1b–1e bearing electron-rich and -deficient aniline substituents
(X=CH3, tBu, Cl, and F; entries 1–4), the resulting products
3b–3e were obtained in satisfactory yields (68–76%). The reac-
tion is extensible to additional substrates 1 f–1h bearing vari-
ous aryl groups at the alkenyl C2-carbons (Ar=4-Me-Ph, 4-Cl-
Ph, 2-thienyl), yielding desired compounds 3 f–3h in 71–74%
yields (entries 5–7). To our delight, these annulations were
compatible with substrates 1i–1k bearing various alkyl groups
at the alkenyl C2-carbons (R=Me, Et, isopentyl); their desired
products 3i–3k were obtained in reasonable yields (54–67%;
entries 8–10). A complicated mixture of products was obtained
when we attempted the reaction on tert-butyl-substituted N-
hydroxyallylamine.
The reaction scope was expanded with various nitrosoarenes
(Scheme 1). The reaction of 4-methyl-substituted nitrosoben-
zene (R=Me) gave isoxazolidin-5-yl species 4a in 61% yield.
The reaction became more efficient with electron-deficient ni-
trosoarenes (R=Cl, Br, NO2, and CO2Et), affording products
4b–4e in 74–80% yields. Electron-deficient nitrosoarenes are
more favorable for this reaction presumably because they are
efficient electron acceptors.[11] We finally synthesized com-
pound 4 f, which had good crystallinity for X-ray diffraction to
confirm the isoxazolidinyl framework.[10]
Scheme 2 shows the elaboration of resulting isoxazolidin-5-
yl species 3a into useful N- and O-containing compounds.
Treatment of species 3a with p-TSA (10 mol%) efficiently yield-
ed substituted quinoline N-oxide product 5a and diazene
Chem. Eur. J. 2016, 22, 2915 – 2919
www.chemeurj.org
2916
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
tries 4–7). For unsubstituted and C3-substituted acetates 1l’
and 1m’–1o’, resulting products 5i and 5j–5k were obtained
in 65–70% yields (entries 8–10).
Table 4 shows the applicability of our new method to the ef-
ficient synthesis of 1,3-diamino-2-ols through reductive NÀO
cleavages of compounds 3; herein, these reactions represents
triple N- and O-functionalizations of readily available allylic ace-
tates. The reactions were performed with Pd/C and H2 (1 atm)
Scheme 2. Chemical elaboration.
Table 4. Synthesis of 1,3-diamino-2-ols by NÀO cleavage.
oxide in 69% and 6% yields, respectively. Importantly, product
5a was directly accessible by heating a mixture of nitrosoben-
zene (2a; 2 equiv) and 3-N-hydroxyallylamines 1a (1 equiv)
with [(IPr)CuCl]/O2 (5 mol%) and p-TSA (10 mol%) in toluene;
the yield of resulting 5a was 72% along with diazine oxide 2a’
at 10% yield. To highlight the value of intermediate 3a, we de-
veloped a two-step one-pot cascade reaction involving an ini-
tial Pd-catalyzed amination of readily available allylic acetate
1a’ in toluene for 1 h, followed by heating the same solution
with nitrosobenzene (2a; 1.2 equiv), [(IPr)CuCl], and oxygen to
yield 3-phenyl quinoline N-oxide 5a efficiently.
Entry
R
R’
Compounds (t [h], yield [%])[b]
1
2
3
4
5
Ph
4-Me-Ph
CH3
isopentyl
H
H (3a)[a]
H (3 f)
H (3i)
H (3j)
allyl (3l)
6a (8, 71)
6b (5, 76)
6c (6, 74)
6d (5, 68)
6e (6, 69)
[a] 3a (0.045m). [b] Product yields are reported after purification through
a silica gel column.
Unlike 2-substituted analogues,[12a] the synthesis of 3-substi-
tuted quinoline N-oxides requires prior synthesis of their quin-
oline derivatives, which are usefully prepared from the conden-
sation of expensive 2-aminobenzaldehydes with ketones or al-
dehydes.[12b–d] In this regard, the synthesis of quinoline N-oxide
5a in a Pd/Cu/H+ relay catalysis (Scheme 2) is appealing, be-
cause cheap nitrosoarenes and allylic acetates are employed in
this one-pot synthesis. The reaction scope was generalized
with various N-hydroxyaniline 2a“, allylic acetates 1’, and nitro-
soarenes 2 in a 1:1:1.2 molar proportions (Table 3).[13] With elec-
tron-deficient nitrosobenzenes (X=Cl, Br, and CO2Et), Cu-medi-
ated annulations with species 1a’ provided 3-substituted quin-
oline N-oxides 5b–5d in satisfactory yields (>71%, entries 1–
3). This catalytic synthesis was extendible to additional allylic
acetates bearing various aryl and alkyl groups at their C2-car-
bons (R=4-tolyl, 4-chlorophenyl, methyl, and isopentyl), yield-
ing 3-substituted quinoline oxides 5e–5h in 61-70% yield (en-
in MeOH, yielding desired products 6a–6e exceeding 68%
yields (entries 1–5).
Control experiments [Eqs. (5)–(7)] were carried out to eluci-
date the mechanism for our major reactions depicted in
Tables 1–2 and Scheme 1. We tested a reaction on 2-cycloprop-
yl-substituted substrate 1n, giving compound 3n of the same
type [Eq. (5)]; the radical II was thus excluded, because the cy-
clopropane ring of 3n was not cleaved.[6] To identify the
source of the C2-amino group of compound 3a, compound
1a, N-hydroxyaniline (2a“; 2 equiv), and C6D5NO (d5-2a) were
tested for a brief period (10 min) to attain a 32% conversion of
the reaction [Eq. (6)]. The resulting product d10-3a contained
mainly C6D5NHÀOÀ at its C2-carbon (>90%), indicating that
the C2-amino group appeared primarily from C6D5NO. The for-
mation of carbocation II’ was unlikely to occur here because
potent nucleophile 2a” did not participate in the reaction. In
contrast, the reactions depicted in Equations (3) and (4), likely
involve nitrone 7 according to the results in Equation (7),
which gave the same product 3l’ as in Equation (4). Such ni-
trone species are unlikely to form for 2-substituted N-hydroxy-
allylamines according to a separate experiment.[14]
Table 3. Catalytic synthesis of quinoline oxides.
[a] 1’ (0.3m) and 2a’’ (0.3m). [b] Product yields are reported after purifica-
tion through a silica gel column.
Chem. Eur. J. 2016, 22, 2915 – 2919
www.chemeurj.org
2917
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
cation II, and its attack on this cation will form single diaste-
reomeric products.
Prior to this work, intermolecular [3+2] cycloadditions of
allyl radicals with p-bond motifs remained a formidable task in
solution chemistry. In summary, we reported the success of
such formal cycloadditions in Cu-mediated [3+2] annulations
of 3-N-hydroxyallylamines with nitrosoarenes. These transfor-
mations represent remarkable 3-amino-1,2-dioxygenations of
allylamines. Notably, the resulting isoxazolidin-5-yl products
can be transformed into 2- or 3-substituted quinoline oxides in
the presence of a Brønsted acid. We subsequently employed
the Pd/Cu/H+ relay catalysis to access these quinoline oxides
directly from cheap allylic acetates. Our mechanistic analysis in-
dicated that the addition of allyl radicals to nitrosoarenes is
feasible only with 2-substituted allyl radicals; nitrone species
formed in the case of unsubstituted and 3-substituted allyl-
amines.
Scheme 3 shows a plausible mechanism to rationalize the
major annulations, in which [(IPr)CuCl]/O2 acts as a radical ini-
tiator The key step of this mechanism involves an addition of
the nitrosobenzene to N-hydroxyaminoallyl radical I’ to form
new nitroxy radical III.[11] Notably, such a radical/nitroso addi-
Acknowledgements
The authors thank the National Science Council and the Minis-
try of Education, Taiwan, for supporting this work.
Keywords: annulation · copper · nitrosoarenes · oxidants ·
radicals
Scheme 3. A postulated mechanism for [3+2] annulations.
[1] a) F. Cardona, A. Goti, Angew. Chem. Int. Ed. 2005, 44, 7832–7835;
Angew. Chem. 2005, 117, 8042–8045; b) L. M. Stanley, M. P. Sibi, Chem.
Rev. 2008, 108, 2887–2902; c) T. Hashimoto, K. Maruoka, Chem. Rev.
2015, 115, 5366–5412.
tion is operable only for 2-substituted allyl radicals (R=alkyl,
aryl), otherwise the nitrone species will form in the case of un-
substituted allylamine 1l [Eq. (7)]. A 5-endo-trig cyclization of
radical III is excluded because the experiment in Equation (5)
opposes radical intermediate II. Herein, we postulate that spe-
cies III abstracts a hydrogen from initial 1 to form species IV
and initial nitroxy radical I, thus completing a radical chain re-
action. Accordingly, the generation of nitroxy radical 1a in cat-
alytic amounts is sufficient for the reaction. In species IV, we
envisage that the N-hydroxy group forms a hydrogen bond
with the nitroso to increase its electrophilicity, leading to the
formation of a CÀO bond as depicted by intermediate V. This
acid-mediated O-nitroso alkylation by enamine nucleophiles
has a literature example.[15] The 5-exo-trig cyclization of result-
ing species VI is known to be feasible to yield observed prod-
uct 3a. Equation (6) excludes the intermediacy of tertiary car-
bocation II’. To examine the reactivity of this carbocation, we
treated compound 3 with p-TSA to form 2,5-dihydroisoxazole
intermediate VII, which subsequently underwent an acid-cata-
lyzed ring opening to induce a skeletal rearrangement, yielding
observed product 4 through an intramolecular cyclization of
resulting intermediates VIII or VIII’. In the preceding VI!3
transformation, resulting 2,5-dihydroisoxazole species 3 likely
exists in two diastereomeric forms; we postulate that the
Brønsted acid is likely to exist in this Cu/O2 system to promote
an interconversion between species 3 and carbocation II’. The
released N-hydroxyaniline will be closely associated with carbo-
[2] a) D. P. Curran, N. A. Porter, B. Giese, Stereochemistry of Radical Reactions,
Wiley-VCH, Weinheim, 1996; b) Radicals in Organic Synthesis (Eds.: P.
Renaud, M. P. Sibi), Wiley-VCH, Weinheim, 2001; c) A. J. McCarroll, J. C.
Walton, Angew. Chem. Int. Ed. 2001, 40, 2224–2248; Angew. Chem.
2001, 113, 2282–2307; d) G. Stork, M. E. Reynolds, J. Am. Chem. Soc.
1988, 110, 6911.
[3] a) D. Nohara, T. Sakai, Ind. Eng. Chem. Res. 1988, 27, 1925–1929; b) D.
Nohara, T. Sakai, Ind. Eng. Chem. Fundam. 1980, 19, 340–344; c) W. A.
Bryce, D. J. Ruzicka, Can. J. Chem. 1960, 38, 835–844; d) D. Nohara, T.
Sakai, Sekiyu Gakkaishi 1992, 35, 163–168; e) D. Nohara, T. Sakai, Sekiyu
Gakkaishi 1988, 31, 403–409.
[4] Additions to imines, see selected examples: a) M. Aldemir, F. W. Heine-
mann, H. Kisch, Photochem. Photobiol. Sci. 2012, 11, 908–913; b) H.
Keck, W. Schindler, F. Knoch, H. Kisch, Chem. Eur. J. 1997, 3, 1638–1645.
[5] J. E. Baldwin, J. Cutting, W. Dupont, L. Kruse, L. Silberman, R. C. Thomas,
J. Chem. Soc. Chem. Commun. 1976, 736–738.
[6] a) C. Chatgilialoglu, C. Ferreri, M. Guerra, V. Timokhin, G. Froudakis, T. Gi-
misis, J. Am. Chem. Soc. 2002, 124, 10765–10772; b) S. Bogen, M. Malac-
ria, J. Am. Chem. Soc. 1996, 118, 3992–3993; c) Y. Yamamoto, M. Ohno,
S. Eguchi, J. Org. Chem. 1996, 61, 9264–9271; d) K. Kobayashi, M.
Suzuki, H. Suginome, J. Org. Chem. 1992, 57, 599–606.
[7] a) V. A. Schmidt, E. J. Alexanian, J. Am. Chem. Soc. 2011, 133, 11402–
11405; b) B. C. Giglio, V. A. Schmidt, E. J. Alexanian, J. Am. Chem. Soc.
2011, 133, 13320–13322; c) V. A. Schmidt, E. J. Alexanian, Angew. Chem.
Int. Ed. 2010, 49, 4491–4494; Angew. Chem. 2010, 122, 4593–4596;
d) V. A. Schmidt, E. J. Alexanian, Chem. Sci. 2012, 3, 1672–1674; e) R. K.
Quinn, V. A. Schmidt, E. J. Alexanian, Chem. Sci. 2013, 4, 4030–4034.
[8] a) Q. Lu, Z. Liu, Y. Luo, G. Zhang, Z. Huang, H. Wang, C. Liu, J. T. Miller, A.
Lei, Org. Lett. 2015, 17, 3402–3405; b) A. Studer, Angew. Chem. Int. Ed.
2000, 39, 1108–1111; Angew. Chem. 2000, 112, 1157–1160; c) P. Sharma,
R.-S. Liu, Chem. Eur. J. 2015, 21, 4590–4594; d) X.-Y. Jiang, F.-L. Qing,
Angew. Chem. Int. Ed. 2013, 52, 14177–14180; Angew. Chem. 2013, 125,
Chem. Eur. J. 2016, 22, 2915 – 2919
www.chemeurj.org
2918
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
14427–14430; e) A. A. Andia, M. R. Miner, K. A. Woerpel, Org. Lett. 2015,
17, 2704–2707.
[9] S. Ghorpade, R.-S. Liu, Angew. Chem. Int. Ed. 2014, 53, 12885–12888;
Angew. Chem. 2014, 126, 13099–13102.
[14] Treatment of N-hydoxyaniline, 2-phenyl-2-buten-1-al, and nitrosoben-
zene produced 3-phenylquinoline in 48% yield; in this case nitrosoben-
zene did not participate in the reaction. .
[10] CCDC 1429299 (4 f) contain the supplementary crystallographic data for
this paper. These data are provided free of charge by The Cambridge
Crystallographic Data Centre.
[11] For the reactions of carbon-based radicals with nitrosoarenes, see se-
lected examples: a) D. J. Fisher, L. G. Burnett, R. Velasco, J. R. d. Alaniz, J.
Am. Chem. Soc. 2015, 137, 11614–11617; b) Y. Miura, T. Tomimura, N.
Matsuba, R. Tanaka, M. Nakatsuji, J. Org. Chem. 2001, 66, 7456–7463;
c) W. Adam, S. E. Bottle, K. Paters, Tetrahedron Lett. 1991, 32, 4283–
4286.
[15] Y. Yamamoto, N. Nomiyama, H. Yamamoto, J. Am. Chem. Soc. 2004, 126,
5962–5963.
[12] a) L.-C. Campeau, D. R. Stuart, J.-P. Leclerc, M. Bertrand-Laperle, E. Ville-
mure, H.-Y. Sun, S. Lasserre, N. Guimond, M. Lecavallier, K. Fagnou, J.
Am. Chem. Soc. 2009, 131, 3291–3306; b) R. Martínez, D. Ramon, M.
Yus, J. Org. Chem. 2008, 73, 9778–9780; c) W. Luo, Q, Mu, W. Qiu, T.
Liu, F. Yang, X. Liu, J. Tang, Tetrahedron 2011, 67, 7090–7095; d) M. P.
Maguire, K. R. Sheets, K. Mcvety, A. P. Spada, A. Zilberstein, J. Med.
Chem. 1994, 37, 2129–2137.
[16] Generation of oxygen-centered radicals with Cu/O2, see ref. [8] and a) R.
Bag, D. Sar, T. Punniyamurthy, Org. Lett. 2015, 17, 2010–2013; b) Q. Lu,
P. Peng, Y. Luo, Y. Zhao, M. Zhou, A. Lei, Chem. Eur. J. 2015, 21, 18580–
18583.
[13] In the mechanism depicted in Scheme 3, N-hydroxyaniline will be re-
leased and reoxidized by O2 in the transformation of 3!4, and this
equimolar ratio is sufficient for the reaction.
Received: November 27, 2015
Published online on January 25, 2016
Chem. Eur. J. 2016, 22, 2915 – 2919
www.chemeurj.org
2919
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim