Anionic halocuprate(II) complexes as catalysts for the oxaziridine-mediated aminohydroxylation of olefins

Article abstract of DOI:10.1021/jo900902k

(Chemical Equation Presented) We have discovered that the oxaziridine-mediated copper-catalyzed aminohydroxylation reaction recently discovered in our laboratories is dramatically accelerated in the presence of halide additives. The use of this more activ

Full text of DOI:10.1021/jo900902k

pubs.acs.org/joc
Anionic Halocuprate(II) Complexes as Catalysts for the
Oxaziridine-Mediated Aminohydroxylation of Olefins
Tamas Benkovics, Juana Du, Ilia A. Guzei, and Tehshik P. Yoon*
Department of Chemistry, University of Wisconsin–Madison, 1101 University Avenue, Madison,
Wisconsin 53706
tyoon@chem.wisc.edu
Received April 30, 2009
We have discovered that the oxaziridine-mediated copper-catalyzed aminohydroxylation reaction
recently discovered in our laboratories is dramatically accelerated in the presence of halide additives.
The use of this more active catalyst system enables the efficient aminohydroxylation of electronically
and sterically deactivated styrenes and also enables the use of nonstereogenic 3,3-dialkyl oxaziridines
as terminal oxidants in the aminohydroxylation reaction. We present evidence that anionic
halocuprate(II) complexes are the catalytically active species responsible for the increased reactivity
under these conditions. This unexpected observation has led us to re-evaluate our mechanistic
understanding of this reaction. On the basis of the results of a variety of radical trapping experiments,
we propose a modified mechanism that involves a homolytic reaction of the olefin with a copper(II)-
activated oxaziridine. Together, the observation that anionic additives significantly increase the
oxidizing ability of oxaziridines and the recognition of the radical nature of reactions of oxaziridines
under these conditions suggest that a variety of new oxidative transformations catalyzed by
halocuprate(II) complexes should be possible.
Introduction
halocuprate(II) complexes have been published.4,5 Some of
the most intriguing of these papers, beginning with a report
in the Japanese patent literature, describe the beneficial effect
of lithium and ammonium chloride salts in CuCl2-catalyzed
Halocuprates(II) are anionic cupric halide complexes that
typically possess the empirical formula [ACuX3] or
[A2CuX4], where X is a halide anion and A is a cationic
metal or ammonium counterion. Due to the constitutional
and stereochemical diversity of their solid-state structures1
and their unusual magnetic properties,2 the physical and
electronic structure properties of these compounds have
been investigated in great depth.3 In contrast, only a hand-
ful of reports describing organic reactions catalyzed by
(4) (a) Tamura, M.; Kochi, J. K. Bull. Chem. Soc. Jpn. 1971, 44, 3063–
3073.bAsakawa, Y.; Matsuda, R.; Hashimoto, T. Jpn. Patent 39847 (1984);
Chem. Abstr. 1984, 101, 130395x. (c) Takehira, K.; Shimizu, M.; Watanabe,
Y.; Orita, H.; Hayakawa, T. Tetrahedron Lett. 1989, 30, 6691–6692.
(d) Shimizu, M.; Watanabe, Y.; Orita, H.; Hyakawa, T.; Takehira, K. Bull.
Chem. Soc. Jpn. 1992, 65, 1522–1526. (e) Sasaki, T.; Zhong, C.; Tada, M.;
Iwasawa, Y. Chem. Commun. 2005, 2506–2508.
(5) Halocuprate(I) compounds, on the other hand, are used synthetically
and have been shown to be convenient sources of nucleophilic chloride. See:
(a) Muscio, O. J. Jr.; Jun, Y. M.; Philip, J. B. Jr. Tetrahedron Lett. 1978, 19,
2379–2382. (b) Elsevier, C. J.; Meijer, J.; Tadema, G.; Stehouwer, P. M.; Bos,
H. J. T.; Vermeer, P.; Runge, W. J. Org. Chem. 1982, 47, 2194–2196.
(c) Elsevier, C. J.; Bos, H. J. T.; Vermeer, P.; Spek, A. L.; Kroon, J. J. Org.
Chem. 1984, 49, 379–381. (d) Elsevier, C. J.; Vermeer, P. J. Org. Chem. 1984,
49, 1649–1650. (e) Liedholm, B.; Nilsson, M. Acta Chem. Scand. A. 1984, 38,
555–562.
(1) Arnby, C. H.; Hagner, S.; Dance, I. CrystEngComm 2004, 6, 257–275.
(2) (a) O’Brien, S.; Gaura, R. M.; Landee, C. P.; Ramakhrishna, B. L.;
Willett, R. D. Inorg. Chim. Acta 1988, 141, 83–89. (b) Willett, R. D.;
Twamley, B.; Montfrooij, W.; Granroth, G. E.; Nagler, S. E.; Hall, D. W.;
Park, J.-H.; Watson, B. C.; Meisel, M. W.; Talham, D. R. Inorg. Chem. 2006,
45, 7689–7697.
(3) (a) Smith, D. W. Coord. Chem. Rev. 1976, 21, 93–158. (b) Willett,
R. D. Coord. Chem. Rev. 1991, 109, 181–205.
DOI: 10.1021/jo900902k
r
Published on Web 06/09/2009
J. Org. Chem. 2009, 74, 5545–5552 5545
2009 American Chemical Society

Benkovics et al.
JOCArticle
aerobic oxidations of phenols and speculate that the forma-
tion of halocuprate(II) complexes is responsible for the
observed rate accelerations.4b-4d To the best of our knowl-
edge, however, the effect of halide additives in other copper-
catalyzed oxidation reactions has not been extensively in-
vestigated, despite the community’s increasing interest in
the development of new copper-catalyzed oxidative trans-
formations.6 In this paper, we report that the oxaziridine-
mediated aminohydroxylation reaction recently developed
in our laboratory7 is also significantly accelerated in the
presence of anionic halide additives, and we present evidence
that a halocuprate(II) complex formed in situ is responsible
for this effect.
The need to develop a more reactive catalyst system
became apparent to us in the course of our investigations
toward a highly enantioselective version of our amino-
hydroxylation protocol. We recently reported our initial
efforts toward an asymmetric aminohydroxylation.8 Synthe-
tically useful levels of enantioselectivity (up to 86% ee)
can be achieved in the aminohydroxylation of styrenes with
N-sulfonyl oxaziridine 1 (“Davis’ oxaziridine”)9 when a
copper(II) bis(oxazoline) complex is used as the catalyst
(eq 1), and the enantiopurity of the amino alcohol products
can be increased to very high levels (>99% ee) upon
recrystallization. Thus, this method represents a useful
osmium-free complement to the highly enantioselective
Sharpless aminohydroxylation reaction10 that remains the
state-of-the-art method for olefin oxyamination.
benzylidene aminal moiety makes purification difficult and
complicates spectroscopic analysis of the reaction mixture.
More critically, the chirality of oxaziridine 1 is a significant
impediment to further improving the enantioselectivity of
the asymmetric process. Coordination of racemic 1 to a
chiral copper catalyst results in the formation of a matched
and mismatched set of diastereomeric oxaziridine-metal
complexes, each of which would be expected to react at
different rates and with different stereoselectivities in the
aminohydroxylation reaction. Thus, the stereoinduction in
this transformation consists of two separate stereodifferen-
tiating steps: (1) initial coordination of the racemic oxazir-
idine to the enantiopure chiral copper catalyst, followed by
(2) the aminohydroxylation process itself (Scheme 1). These
two steps are not stereochemically well-coupled, and during
the course of our optimization studies, we found that varia-
tions of the reaction conditions that improved the selectivity
of one step often decreased the selectivity of the other.
SCHEME 1
A possible approach to address the complexity of this
system would be to utilize oxaziridines that are not
stereogenic at carbon (i.e., symmetrically 3,3-disubstituted
oxaziridines). This approach would simplify the system by
avoiding the problematic formation of diastereomeric
copper-oxaziridine complexes in the enantioselective
reaction altogether. In addition, the aminohydroxylation
reaction would not produce mixtures of diastereomeric
products, and thus the nonstereogenic aminal moiety could
be used as a convenient protecting group during a multistep
synthetic sequence. Unfortunately, 3,3-dialkyloxaziridines
are much less reactive oxidants than monosubstituted 1-aryl
oxaziridines such as 1 and fail to produce useful yields of
aminohydroxylation products with any combination of
bis(oxazoline) ligand and copper precursor we have investi-
gated to date. Thus, the development of a highly enantiose-
lective aminohydroxylation protocol is predicated on the
discovery of a more reactive catalyst system capable of
engaging these sterically and electronically deactivated
oxaziridines in the aminohydroxylation reaction.
We quickly recognized, however, that the use of a chiral,
racemic oxaziridine as the terminal oxidant for the copper-
(II)-catalyzed asymmetric aminohydroxylation was non-
ideal, for several reasons. First, we have been unable to
control the orientation of the benzylidene aminal stereocen-
ter in any oxaziridine-mediated aminohydroxylations we
have studied to date. Fortunately, both diastereomers of
the product produced in the enantioselective reaction de-
picted in eq 1 possess the same orientation at the benzylic
amine stereocenter, and the same major enantiomer of the
N-sulfonyl aminoalcohol is revealed upon acid-catalyzed
hydrolysis of either diastereomer. Nevertheless, the product
mixture arising from the lack of control over the stereogenic
(6) For recent reviews, see: (a) Punniyamurthy, T.; Rout, L. Coord.
Chem. Rev. 2008, 252, 134–154. (b) van der Vlught, J. I.; Meyer, F. Top.
Organomet. Chem. 2007, 22, 191–240.
(7) (a) Michaelis, D. J.; Shaffer, C. J.; Yoon, T. P. J. Am. Chem. Soc.
2007, 129, 1866–1867. (b) Michaelis, D. J.; Ischay, M. A.; Yoon, T. P. J. Am.
Chem. Soc. 2008, 130, 6610–6615.
(8) Michaelis, D. J.; Williamson, K. S.; Yoon, T. P. Tetrahedron 2009,
65, 5118-5124.
(9) (a) Davis, F. A.; Nadir, U. K.; Kluger, E. W. J. Chem. Soc., Chem.
Commun. 1977, 25–26. (b) Davis, F. A.; Jenkins, R.; Yocklovich, S. G.
Tetrahedron Lett. 1978, 52, 5171–5174.
In this paper, we report that an anionic halocuprate(II)
complex is a significantly more active catalyst for the oxazir-
idine-mediated aminohydroxylation of olefins. While the initial
goal for this research was the development of conditions that
(10) Sharpless, K. B.; Chong, A. O.; Oshima, J. J. Org. Chem. 1976, 41,
177–179.
5546 J. Org. Chem. Vol. 74, No. 15, 2009

Benkovics et al.
JOCArticle
enable the use of 3,3-dialkyl oxaziridines as the terminal
oxidant for this reaction, we show that these improved
reaction conditions also dramatically accelerate the reac-
tions of a variety of sterically and electronically deactivated
styrene substrates, as well. Moreover, the surprising obser-
vation that anionic copper complexes provide dramatically
superior rates of reaction led us to probe the mechanism
of this transformation in greater depth, and we also describe
the results of mechanistic studies that suggest that the
oxaziridine-mediated aminohydroxylation involves a radical
intermediate, rather than the cationic intermediate we had
originally proposed. Together, the discovery of these more
reactive reaction conditions coupled with our understanding
of the unique ability of halocuprate(II) complexes to initiate
one-electron oxidation reactions of oxaziridines should en-
able us to design new approaches to other copper-catalyzed
oxidative functionalization reactions.
TABLE 1. Effect of Additives on Copper-Catalyzed Aminohydroxyla-
tions Using Oxaziridine 2a
entry
copper source
additive
yield
1
2
Cu(TFA)2 (2 mol %)
CuCl2 (5 mol %)
CuCl2 (5 mol %)
CuCl2 (5 mol %)
CuCl2 (5 mol %)
none
CuCl2 (5 mol %)
CuCl2 (5 mol %)
CuCl2 (5 mol %)
HMPA (10 mol %)
HMPA (10 mol %)
IPr Cl (3a) (6 mol %)
<5%
19%
82%
<5%
80%
<5%
<5%
59%
3
3
4b
5
IPr BF4 (3b) (6 mol %)
3
Bu4N+Cl- (6 mol %)
Bu4N+Cl- (6 mol %)
Bu4N+ClO4- (6 mol %)
Bu4N+Br- (6 mol %)
Bu4N+OAc- (6 mol %)
Results and Discussion
6
7b
8
A. Anionic Halocuprate(II) Complexes as Catalysts for
Aminohydroxylation. Figure 1 outlines the features that we
considered to be desirable in the design of an ideal terminal
oxidant for the copper-catalyzed aminohydroxylation reac-
tion (e.g., 2). As outlined above, the most critical element of
our design plan is the absence of a stereogenic carbon in the
oxaziridine ring. We considered the use of 3,3-unsubstituted
and 3,3-diaryl oxaziridines, but the former could not be
synthesized in useful yields, and the latter decompose readily
and are not suitable for long-term storage. Thus, we focused
our attention upon aminohydroxylations using 3,3-dialkyl
oxaziridines, which can be easily synthesized using the
method of Jennings11 and appear to be indefinitely stable
when stored in the refrigerator. Second, we were interested in
the reactivity of oxaziridines bearing N-substituents that
can be removed from the 1,3-oxazolidine products under
mild conditions. Oxaziridines bearing N-carbamate12 and
N-phosphinoyl13 groups failed to afford aminohydroxyla-
tion products. On the other hand, N-nosyl oxaziridines
possess reactivity similar to that of the analogous N-benze-
nesulfonyl oxaziridines and afford products that can be
deprotected under conditions much milder than the dissol-
ving metal reduction conditions required for removal of an
N-benzenesulfonyl group.
9
70%
aReactions were conducted using 1.5 equiv of oxaziridine in CH2Cl2
(2 M) under argon. bCuCl2 was insoluble under these conditions.
the Cu(TFA)2/HMPA catalyst system that we previously
reported (entry 1). Extensive optimization of copper source
and reaction conditions resulted in only modest improve-
ment in the reactivity when CuCl2 was used as the copper
source instead of Cu(TFA)2 (entry 2). Thus, we performed a
survey of ligands that might modify the reactivity of the
copper catalyst, which revealed a significant increase in
the efficiency of aminohydroxylation upon the addition of
a variety of imidazolium chloride salts (e.g., IPr Cl 3a,
3
entry 3), precursors to the corresponding N-heterocyclic
carbene (NHC) ligands.14 We were able to obtain diffract-
able crystals of a putative catalyst complex upon combining
dihydroimidazolium chloride 4 (SIPr Cl) with CuCl2 in
3
deuterochloroform. Surprisingly, the X-ray structure of
these crystals revealed that a chloride-bridged copper dimer
had been formed rather than the expected NHC-copper
complex (see Supporting Information). We wondered, there-
fore, if the observed increase in catalytic efficiency was
attributable to the presence of an anionic chloride additive,
rather than a carbene ligand. Indeed, the addition of imida-
zolium salts bearing noncoordinating counterions (e.g., IPr
3
BF4 3b, entry 4) failed to accelerate the reaction. On the other
hand, tetrabutylammonium chloride had the same dramatic
effect on the rate of the aminohydroxylation as imidazolium
chloride 3a (entry 5).
FIGURE 1. Design considerations for oxaziridines used in amino-
hydroxylation. Bs = benzenesulfonyl; Ns = 4-nitrobenzenesulfonyl.
Further control experiments showed that no reaction occurs
in the absence of copper(II) additives (entry 6), which rules
out the possibility that the aminohydroxylation could be
catalyzed by the halide alone.15 We also found that reactions
using tetrabutylammonium salts of noncoordinating anions
Table 1 summarizes our experiments probing the utility of
oxaziridine 2 in the copper-catalyzed aminohydroxylation
reaction. As we stated previously, 3,3-dialkyloxaziridines are
significantly poorer oxidants than 3-phenyloxaziridine 1,
and no aminohydroxylation products are observed using
(14) The direct formation of metal-NHC complexes from the reaction
of imidazolium salts with metal complexes bearing basic ligands is well-
(11) Jennings, W. B.; Watson, S. P.; Boyd, D. R. J. Chem. Soc., Chem.
Commun. 1988, 931–932.
€
precedented: Herrmann, W. A.; Elison, M.; Fischer, J.; Kocher, C.; Artus,
(12) Vidal, J.; Damestoy, S.; Guy, L.; Hannachi, J.-C.; Aubry, A.; Collet,
A. Chem.;Eur. J. 1997, 3, 1691–1709.
(13) Boyd, D. R.; Jennings, W. B.; McGuckin, R. M.; Rutherford, M.;
Saket, B. M. J. Chem. Soc., Chem. Commun. 1985, 582–583.
G. R. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 2371–2374.
(15) For a rare example of halogen-catalyzed aziridination, see: Jeong,
J. U.; Tao, B.; Sagasser, I.; Henniges, H.; Sharpless, K. B. J. Am. Chem. Soc.
1998, 120, 6844–6845.
J. Org. Chem. Vol. 74, No. 15, 2009 5547

Benkovics et al.
JOCArticle
(e.g., Bu4N+ClO4-) were inefficient (entry 7). On the other
hand, a variety of additives possessing coordinating anions,
including Bu4N+Br- and Bu4N+OAc-, were also effective
at increasing the rate of aminohydroxylation (entries 8
and 9). Together, these results suggest that an anionic
halocuprate(II) complex formed upon coordination of the
halide or pseudohalide additive to CuCl2 constitutes a more
reactive catalyst for the aminohydroxylation than the neu-
tral CuCl2 species itself.
An important initial goal for this study was to demonstrate
that dimethyloxaziridine 2 is indeed a competent terminal
oxidant for the aminohydroxylation using these halocuprate
catalysts for a range of olefinic substrates, which is demon-
strated by the experiments summarized in Table 2. As was the
case with our previous aminohydroxylation protocol that
utilized the more reactive oxaziridine 1, aminohydroxylations
with 2 require some electron-stabilizing group in order to
provide reasonable product yields. In addition to styrene itself
(entry 1), styrenes that are substituted at various positions on
the ring (entries 2-4) and styrenic olefin (entry 5) undergo
efficient aminohydroxylation. Similarly, styrenes bearing
electron-donating and -withdrawing substituents are also
well-tolerated (entries 6-8). Finally, other stabilized olefinic
substrates such as dienes (entry 9) and vinyl ethers (entry 10)
are also competent substrates for aminohydroxylation with
oxaziridine 2 using the more reactive CuCl2/Bu4N+Cl-
catalyst system.
Unlike the products of aminohydroxylations using 3-
phenyloxaziridine 1, the aminohydroxylation products
shown in Table 2 are formed as single diastereomers, which
simplifies the purification and analysis of these reactions
considerably. Moreover, the isopropylidene aminal moiety
can be quantitatively removed under standard acidic hydro-
lysis conditions, and the N-nosyl group is also efficiently
removed using a protocol (PhSH, K2CO3)16 that is much
milder than the strongly reducing conditions required for
cleavage of other benzenesulfonyl groups. Both deprotec-
tions are clean and high yielding, and they can be conducted
on gram scale without the need for chromatographic purifi-
cation after either step (Scheme 2, 91% yield).
TABLE 2. Aminohydroxylations Using Oxaziridine 2a
SCHEME 2
The effect of the halocuprate(II) catalyst on the efficiency
of the reaction is also dramatically evident in aminohydrox-
ylations that utilize the more reactive 3-phenyloxaziridine
1 as the terminal oxidant (Table 3). We previously reported
that styrene is efficiently aminohydroxylated by 1 in the
presence of Cu(TFA)2/HMPA system, and that the reaction
requires 11 h to proceed to completion (entry 1). Using the
CuCl2/Bu4N+Cl- system, however, the reaction is complete
in just 2 h, and the yield of the reaction is not negatively
affected (entry 2). Similarly, p-(trifluoromethyl)styrene,
which previously required 36 h to proceed to completion,
affords good yields of the aminohydroxylation product
under these improved conditions after 3 h (entries 3 and 4).
Given the dramatic rate increase we observed in the reaction
of this electron-deficient styrene, we wondered if the halo-
cuprate(II) catalyst would enable successful aminohydroxy-
lation of styrenes bearing strongly electron-withdrawing
groups, which were not useful substrates using our original
system. Indeed, the reaction of p-nitrostyrene, which did not
proceed to completion even after extended reaction times
using the Cu(TFA)2/HMPA system (entry 5), requires
only 4.5 h to generate the aminohydroxylated product in
76% yield (entry 6), and the rate of reaction with ethyl
p-methoxycinnamate is also increased to synthetically useful
levels (entries 7 and 8). Sterically encumbered styrenes also
react efficiently under these improved conditions; in re-
examining the aminohydroxylation of cis-stilbene, we found
that the yield was improved and the reaction time was
aUnlessotherwisenoted, all reactionswere performedusing5 mol% of
CuCl2, 5 mol % of Bu4N+Cl-, and 1.5 equiv of oxaziridine in CH2Cl2
(2 M). bYields are the averaged results of two reproducible experiments.
cReaction performed using 10 mol % of CuCl2 and 10 mol % of
Bu4N+Cl- and at 1 M. d>20:1 trans/cis. eReaction performed using
2 mol % CuCl2 and 2 mol % Bu4N+Cl-. f>20:1 olefin selectivity.
(16) Maligres, P. E.; See, M. M.; Askin, D.; Reider, P. J. Tetrahedron
Lett. 1997, 38, 5253–5256.
5548 J. Org. Chem. Vol. 74, No. 15, 2009

Benkovics et al.
JOCArticle
TABLE 3. Comparison of Catalyst Systems with Less Reactive Sub-
stratesa
the efficiency of aminohydroxylations using sterically and
electronically deactivated styrenes and also enable the use of
nonstereogenic 3,3-dimethyl oxaziridine 2 as the terminal
oxidant for the aminohydroxylation, which is an important
step toward the development of a highly enantioselective
aminohydroxylation protocol. Finally, examples of synthe-
tically useful reactions catalyzed by halocuprate(II) com-
plexes are relatively rare, and while the beneficial effects of
halide additives in palladium-catalyzed reactions17 are well-
appreciated, their effect in copper-catalyzed reactions has
not been systematically explored. The possibility of devel-
oping a general understanding of halide effects in copper-
catalyzed oxidations prompted us to examine the mechanism
of the aminohydroxylation in greater detail.
B. Revision of the Mechanism of Aminohydroxylation. Our
previous investigations of the mechanism of the amino-
hydroxylation under our original Cu(TFA)2/HMPA condi-
tions had led us to propose a two-step, cationic mechanism
(Scheme 3, Mechanism A) in which the copper catalyst serves
as a Lewis acid that activates the oxaziridine toward nucleo-
philic attack by the styrenic olefin. This hypothesis was based
on several observations, including the following: (1) styrenes
bearing electron-donating groups react more rapidly than
those bearing electron-withdrawing groups; (2) amino-
hydroxylations of cis- and trans-olefins are stereoconver-
gent, indicating an intermediate that enables free rotation
about the bond arising from the alkene; and (3) vinyl ethers
and dienes that would be expected to give rise to stabilized
radical or cationic intermediates are excellent substrates for
aminohydroxylation. However, it seems unlikely that the
aReactions conducted using catalyst system A were performed using 2
mol % of Cu(TFA)2, 10 mol % of HMPA, and 1.5 equiv of 1 in CH2Cl2
(0.5 M). Reactions conducted using catalyst system B were performed
using 2 mol % of CuCl2, 2 mol % of Bu4N+Cl-, and 1.5 equiv of 1 in
CH2Cl2 (0.5 M). bData reported are from ref 7a.
-
anionic CuCl3 fragment would be a more powerful Lewis
acid than a neutral CuCl2•HMPA complex. Thus, the results
of these studies seem to be at odds with our original
mechanistic hypothesis. Instead, we wondered if our obser-
vation of the anion acceleration effect suggested a non-
redox-innocent role for the copper catalyst.
reduced even when the loading of the copper catalyst was
reduced from 10 to 2 mol % (entries 9 and 10). As was the case
with the original Cu(TFA)2/HMPA catalyst system, this reac-
tion is not stereospecific, and the cis-olefin exclusively generates
products in which the two vicinal phenyl substituents are trans
on the oxazolidine ring. Finally, we also observed a rate
acceleration using non-aromatic substrates, but the effect was
less pronounced. The aminohydroxylations of strained aliphatic
olefins such as norbornene are faster using the halocuprate(II)
catalyst (entries 11 and 12), but reactions involving less reactive
primary aliphatic olefins are not high-yielding, and further
explorations will be required to increase the reactivity of these
substrates to synthetically useful levels. These studies will benefit
from a deeper understanding of the origins of the rate accelera-
tion in the presence of halide additives. Our recent investigations
have therefore been focused on a detailed examination of the
mechanism of this reaction, and these studies are described in
the next section.
To summarize this section, halocuprate(II) complexes
generated in situ by coordination of exogenous anionic
halide additives to copper salts are superior catalysts for
the oxaziridine-mediated aminohydroxylation developed in
our laboratories. This discovery is significant for a number of
reasons. First, the halocuprate(II) complex is freely soluble
in many common organic solvents, and thus reactions
conducted using the CuCl2/Bu4N+Cl- catalyst system do
not require the use of HMPA as a solubilizing additive.
Second, these more powerfully oxidizing conditions increase
SCHEME 3
It seems unlikely, given the strongly oxidizing conditions of the
aminohydroxylation, that a copper(I) salt would be the catalyti-
cally relevant species in this process.18,19 Evidence against this
possibility is provided by the results of stoichiometric experiments.
(17) (a) Jeffery, T. Tetrahedron 1996, 52, 10113–10130. (b) Amatore, C.;
Jutand, A. Acc. Chem. Res. 2000, 33, 314–321. (c) Fairlamb, I. J. S.; Taylor,
R. J. K.; Serrano, J. L.; Sanchez, G. New J. Chem. 2006, 30, 1695–1704.
(18) Radical rearrangements of N-alkyloxaziridines initiated upon one-
ꢀ
electron reduction by copper(I) catalysts have been studied in detail by Aube.
See: (a) Aube, J.; Ghosh, S.; Tanol, M. J. Am. Chem. Soc. 1994, 116, 9009–
ꢀ
9018. (b) Aube, J. Chem. Soc. Rev. 1997, 26, 269–277.
(19) An intramolecular aminohydroxylation of olefins using N-benzoy-
loxyamines catalyzed by copper(I) salts has been reported: Noack, M.;
€
Gottlich, R. Chem. Commun. 2002, 536–537.
J. Org. Chem. Vol. 74, No. 15, 2009 5549

Benkovics et al.
JOCArticle
First, we showed that styrene reacts with oxaziridine 1 in the
presence of 1 equiv of Cu(OTf)2 in acetonitrile to produce the
expected aminohydroxylation product, albeit in modest
yield (eq 4, 41% yield). In contrast, no aminohydroxylation
is observed when 1 equiv of CuOTf is used as the stoichio-
metric promoter (eq 5).20 Instead, the oxaziridine is rapidly
consumed upon addition, and the colorless copper(I) solu-
tion immediately turns blue-green, suggesting that copper(I)
is oxidized to copper(II) by the oxaziridine under the reac-
tion conditions. Thus, it seems very unlikely that copper(I)
generated upon disproportionation of copper(II) is the
mechanistically relevant resting state of the catalyst.
intermediate. The strongly oxidizing conditions of the amino-
hydroxylation limited the choice of radical trap; most standard
trapping reagents such as TEMPO and trialkyltin hydrides
are rapidly oxidized by the oxaziridine under the reaction
conditions. Ultimately, we discovered that carbon tetrabro-
mide, although stable to the oxaziridine and copper(II)
catalyst alone, could inhibit formation of the aminohydroxy-
lation product. In the presence of excess CBr4, we observed
exclusive formation of Kharasch addition product 5 (eq 6).26
No reaction was observed in the absence of the oxaziridine.
These data are consistent with interception of a short-lived
benzylic radical intermediate by CBr4, which would generate
tribromomethyl radical and initiate the radical chain addi-
tion of CBr4 across the styrenic bond. Thus, this experiment
provides strong evidence that a radical species is indeed
generated under the reaction conditions, although we could
not demonstrate that this radical was a productive inter-
mediate en route to the aminohydroxylation product.
Thus, we propose instead that this reaction proceeds via a
copper(II/III) catalytic cycle. Our modified mechanistic
hypothesis is depicted in Scheme 3, Mechanism B, in which
the copper-activated oxaziridine undergoes an initial homo-
lytic reaction with the alkene substrate. The intermediate of
this reaction would involve a transient copper(III) sulfon-
amide that reacts with a benzylic radical to produce the cyclic
aminal product and regenerate the copper(II) catalyst.21
Previous computational studies of oxaziridine reactivity
provide support for a radical mechanism. Houk reported
that the transition state for oxaziridine-mediated epoxida-
tions is concerted but features a significant buildup of
unpaired spin density on one of the carbons of the alkene and
on the oxaziridine nitrogen.22 Thus, Mechanism B is consistent
with the propensity of oxaziridines to react with signi-
ficant homolytic character. In addition, a key feature of this
mechanism is the generation of a highly oxidized copper(III)
intermediate. We would expect that an anionic halocuprate(II)
would be more easily oxidized than a neutral copper-HMPA
complex and could thus be a more active catalyst.23-25
We also investigated the aminohydroxylation of cyclopropane-
substituted alkenes in an attempt to observe products result-
ing from ring opening of a cyclopropylcarbinyl radical
(Scheme 4). cis-1,2-Diphenyl-3-vinylcyclopropane 6 proved
to be unreactive under our optimized reaction conditions.
On the other hand, the analogous 1-(cis-2,3-diphenylcyclopro-
pyl)butadiene 7 reacted smoothly under the same conditions
and produced a 43% isolated yield of aminohydroxylation
SCHEME 4
In order to obtain evidence to support Mechanism B,
we first investigated the possibility that an exogenous trap-
ping reagent might intercept the putative benzylic radical
(20) These stoichiometric studies were conducted using copper(I) and
copper(II) triflate because of the limited solubility of the chloride and TFA
salts in methylene chloride.
(21) This mechanism would be similar to the “molecule-induced homo-
lysis” pathway proposed by Minisci to explain dioxirane-mediated epoxida-
tions: Bravo, A.; Fontana, F.; Fronza, G.; Minisci, F.; Zhao, L. J. Org.
Chem. 1998, 63, 254–263.
(22) Houk, K. N.; Liu, J.; DeMello, N. C.; Condroski, K. R. J. Am.
Chem. Soc. 1997, 119, 10147–10152.
(23) Margerum’s studies on copper(II)-peptide complexes indicate that
multiple anionic ligands facilitate electrochemical oxidation of the metal by
stabilizing the copper(III) oxidation state: Bossu, F. P.; Chellappa, K. L.;
Margerum, D. W. J. Am. Chem. Soc. 1977, 99, 2195–2203.
(24) A copper(III) intermediate has been implicated in the decomposition
of hypochlorite and hypobromite salts by anionic hydroxycuprate(II) cata-
lysts: Gray, E. T. Jr.; Taylor, R. W.; Margerum, D. W. Inorg. Chem. 1977, 16,
3047–3055.
(25) We have also ruled out a mechanism involving initial homolytic ring
opening of the oxaziridine. One-electron reduction of the oxaziridine would
be expected to produce a copper alkoxide and a nitrogen-centered radical, in
18
ꢀ
analogy to Aube’s studies with copper(I), which would give rise to the
opposite regiochemical outcome. Moreover, as shown in eq 5, copper(I)
complexes reduce the oxaziridine but fail to produce any aminohydroxyla-
tion products, which suggests that reductive cleavage of the N-O bond is not
the initial step of the mechanism.
(26) Kharasch, M. S.; Jensen, E. V.; Urry, W. H. J. Am. Chem. Soc. 1946,
68, 154–155.
5550 J. Org. Chem. Vol. 74, No. 15, 2009

Benkovics et al.
JOCArticle
products. Surprisingly, we obtained a 2:3:3 mixture of three
diastereomeric products (8a-c) possessing varying stereo-
chemistry about the cyclopropane. Control reactions indi-
cated that the stereochemical fidelity of the cis-diphenyl-
cyclopropane unit was preserved when diene substrate 7 was
subjected to the copper catalyst and when the isomerically pure
product 8a was isolated and resubjected to the reaction con-
ditions. Thus, both the substrate and the product are stereo-
chemically stable to the reaction conditions; stereomutation of
the cyclopropane must therefore occur at an intermediate en
route to the aminohydroxylation products.
required for this reaction to occur. Similarly, 1,4-hexadiene
is oxidized to benzene in 20 min under the same conditions
(eq 8), presumably by oxidation of the doubly allylic
C-H bond. This substrate is also much less reactive using
the Cu(TFA)2/HMPA catalyst system (8% yield). These
intriguing results suggested that we should be able to design
synthetically useful oxidative functionalization reactions
initiated by oxaziridine-mediated C-H bond abstraction.
We have made significant progress along this line of investi-
gation, and these results will be reported shortly.
These observations provide strong evidence for a benzylic
radical intermediate, as proposed in Mechanism B. The dia-
stereomeric mixture of products obtained in this experiment
is consistent with ring-opening of a cyclopropylcarbinyl
radical to afford an acyclic intermediate where the stereo-
chemical fidelity of the cyclopropane ring is lost. On the
other hand, cyclopropylcarbinyl cations are nonclassical,27
and their rearrangements are stereospecific in nature.28
Therefore, the lack of stereochemical fidelity in this experi-
ment seems inconsistent with cationic Mechanism A and is
better explained by the radical Mechanism B.29
Thus, our investigations into the oxaziridine-mediated
aminohydroxylation under the influence of halocuprate(II)
catalysts have two principal conclusions. First, the reactivity
of oxaziridines is dramatically increased in the presence of
halocuprate(II) catalysts, enabling both less reactive oxazir-
idines and less electron-rich olefins to participate readily in
the aminohydroxylation reaction. Second, our mechanistic
investigations suggest that the aminohydroxylation involves
a radical mechanism, rather than the cationic mechanism we
originally proposed. Together, these observations prompted
us to explore whether we might be able to design new
oxaziridine-mediated oxidation reactions of less reactive,
non-olefinic functional groups using halocuprate(II) cata-
lysts. In particular, we wondered if activated C-H bonds
might be prone to oxidation under these conditions.
Conclusion
In the context of our research program on the activation of
oxaziridines, the discovery that anionic halocuprate(II) com-
plexes serve as significantly more active catalysts for amino-
hydroxylation than neutral copper(II) salts is significant for
several reasons. First, these more reactive conditions engage
less reactive oxaziridines, 3,3-dimethyl oxaziridines that are
completely unreactive under our original conditions are
suitable terminal oxidants when the CuCl2/Bu4N+Cl- cata-
lyst system is used. The ability to utilize these nonstereogenic
oxaziridines is an important prerequisite for our plans to
develop a highly asymmetric aminohydroxylation reaction,
and the greater ease of deconvoluting the product mixtures
enabled us to conduct a more thorough investigation of the
reaction mechanism. Second, the efficiency of aminohydrox-
ylation reactions using a variety of sterically and electro-
nically deactivated styrenes is dramatically increased.
Substrates bearing very electron-withdrawing groups that
failed to proceed to completion under our original condi-
tions even with extended reaction times can now be amino-
hydroxylated in a matter of hours. Finally, our efforts to
understand the origins of this effect led us to significantly
revise our understanding of the mechanism of this transfor-
mation. The surprising discovery that an anionic halocup-
rate(II) complex is a superior catalyst for activation of
oxaziridines toward homolytic reactions suggests that other
copper-catalyzed oxidative functionalization processes
might also be designed that take these observations into
account.
In addition to these considerations, the rate acceleration we
observe for the aminohydroxylation in the presence of halide
additives seems similar to the acceleration previously observed
in copper-catalyzed aerobic oxidations of phenols.4b-4d These
observations raise the intriguing possibility that halocuprate(II)
complexes, whose reactivity as catalysts has not been exten-
sively studied to date, may be suitable catalysts for a broader
range of other copper-catalyzed oxidations that employ a
variety of terminal oxidants. Thus, the studies reported in this
paper should have broad significance in light of the chemistry
community’s increased interest in the development of new
oxidation reactions mediated by inexpensive and relatively
nontoxic copper catalysts.
In our preliminary studies toward this goal, we found that
sec-phenethylalcohol is rapidly and quantitatively oxidized
to acetophenone within 4 h in the presence of oxaziridine 2
and CuCl2/Bu4N+Cl- under an atmosphere of argon (eq 7).
Only a trace of acetophenone (2% yield) is observed using
the Cu(TFA)2/HMPA system, and control experiments in-
dicate that the oxaziridine and the copper catalyst are both
(27) For reviews, see: (a) Richey, H. G. In Carbonium Ions; Olah, G. A.,
Schleyer, P. v. R., Eds.; John Wiley: New York, 1972; Vol. III, pp 1201-
1294. (b) Wiberg, K. B.; Hess, B. A., Jr.; Ashe, A. J. In Carbonium Ions;
Olah, G. A., Schleyer, P. v. R., Eds.; John Wiley: New York, 1972; Vol. III, pp
1295-1346. (c) Olah, G. A.; Reddy, V. P.; Prakash, G. K. S. Chem. Rev. 1992, 92,
69–95.
(28) (a) Wiberg, K. B.; Szeimies, G. J. Am. Chem. Soc. 1968, 90, 4195–
4196. (b) Wiberg, K. B.; Szeimies, G. J. Am. Chem. Soc. 1970, 92, 571–579.
(c) Majerski, Z.; Schleyer, P. v. R. J. Am. Chem. Soc. 1971, 93, 665–671.
(29) The radical mechanism may also help to explain the observed
reactivity of allyl silanes. Previously, we found that allyltriisopropylsilane
reacts with oxaziridine 1 in the presence of Cu(TFA)2/HMPA to afford the
expected aminohydroxylation product in 66% yield. However, this reaction
required extended reaction times (36 h) to proceed to completion. The ability
of β-silicon groups to stabilize both cations and radicals is well-known, but
the former stabilization energy has been estimated to be 29-30 kcal/mol,
while the latter is closer to 3-5 kcal/mol. Thus the observation that allyl
silanes react more readily than primary aliphatic olefins but still require
longer reaction times than styrenes is more consistent with a radical
mechanism than with our originally proposed cationic mechanism. For a
recent review of R- and β-silicon effects in organic synthesis, see: Chabaud,
L.; James, P.; Landais, Y. Eur. J. Org. Chem. 2004, 3173–3199.
J. Org. Chem. Vol. 74, No. 15, 2009 5551

Benkovics et al.
JOCArticle
Experimental Section
of oxaziridine 2, and 0.25 mL of CH2Cl2. Isolated 132 mg
(0.365 mmol, 74% yield): IR (neat) 3005, 1535, 1346, 1155;
1H NMR (500 MHz, CDCl3) δ 7.94 (dt, J = 9.5, 2.4 Hz, 2H),
7.48 (dt, J = 9.5, 2.4 Hz, 2H), 7.16 (m,1H), 7.10 (dt, J = 4.3 Hz,
4H), 4.94 (dd, J = 7.0, 2.7 Hz, 1H), 4.44 (dd, J = 9.1, 7.0 Hz, 1H),
3.94 (dd, J = 9.2, 2.7 Hz, 1H), 1.84 (s, 3H), 1.82 (s, 3H); 13C NMR
(125 MHz, CDCl3) δ 149.4, 146.9, 139.5, 128.8, 128.7, 128.4, 127.7,
123.5, 99.4, 71.7, 62.9, 27.2, 26.6; HRMS (EI+) calcd for
[C16H15N2O5S]+ (M - CH3)+ requires m/z 347.0697, found m/z
347.0700 (mp = 103-105 °C).
Full experimental details for this work are available in the
Supporting Information.
General Producedure for Aminohydroxylations Using Dimethyl-
oxaziridine 2. In a nitrogen atmosphere glovebox, copper(II)
chloride and tetrabutylammonium chloride were placed in a
2 dram vial with a stirbar. The vial was capped with a septum
and transferred out of the glovebox. The vessel was charged with
CH2Cl2, and the mixture was stirred under argon for 30 min.
Styrene was then added by syringe. The septum was removed,
oxaziridine 2 was quickly added, and the vessel was sealed
and flushed with argon. The reaction progress was monitored
by 1H NMR or TLC. Upon completion of the reaction, the
remaining oxaziridine was quenched with dimethylsulfide, the
solvent was removed by rotary evaporation, and the remaining
residue was loaded directly onto silica for purification by flash
column chromatography.
N-(4-Nitrobenzenesulfonyl)-2,2-dimethyl-4-phenyl-1,3-oxazo-
lane (Table 2, entry 1). Prepared according to the general
procedure using 52.3 mg (0.502 mmol) of styrene, 3.6 mg
(0.025 mmol) of CuCl2, 7.0 mg (0.025 mmol) of Bu4N+Cl-,
191 mg (0.75 mmol) of oxaziridine 2, and 0.25 mL of CH2Cl2.
Reaction time was 2.5 h. The silica gel was loaded using 6:1 hexane/
acetone, and the product was eluted using 5:1 hexane/
acetone. Isolated 134 mg (0.371 mmol, 74% yield) white solid.
Yield 2: 51.2 mg (0.492 mmol) of styrene, 3.5 mg (0.025 mmol)
of CuCl2, 7.1 mg (0.025 mmol) of Bu4N+Cl-, 191 mg (0.75 mmol)
Acknowledgment. We thank Prof. Clark Landis and Chris
Shaffer for their insights on the mechanistic studies described
in this paper, and David Michaelis and Justin Woods for
optimizing the large-scale preparation of oxaziridine 2.
Financial support from the NSF (CHE-0645447), the NIH
(R01-GM084022), and the ACS Petroleum Research Fund
(44783-G1) is gratefully acknowledged. The NMR spectro-
scopy facility at UW-Madison is funded by the NIH (S10
RR04981-01) and NSF (CHE-9629688).
Supporting Information Available: Experimental proce-
dures and spectral data for all new compounds are provided.
This material is available free of charge via the Internet at http://
pubs.acs.org.
5552 J. Org. Chem. Vol. 74, No. 15, 2009