Hydrazines and azides via the metal-catalyzed hydrohydrazination and hydroazidation of olefins

Article abstract of DOI:10.1021/ja062355+

The discovery, study, and implementation of the Co- and Mn-catalyzed hydrohydrazination and hydroazidation reactions of olefins are reported. These reactions are equivalent to direct hydroaminations of C-C double bonds with protected hydrazines or hydrazoic acid but are based on a different concept in which the H and the N atoms come from two different reagents, a silane and an oxidizing nitrogen source (azodicarboxylate or sulfonyl azide). The hydrohydrazination reaction using di-tert-butyl azodicarboxylate is characterized by its ease of use, large functional group tolerance, and broad scope, including mono-, di-, tri-, and tetrasubstituted olefins. Key to the development of the hydroazidation reaction was the use of sulfonyl azides as nitrogen sources and the activating effect of tert-butyl hydroperoxide. The reaction was found to be efficient for the functionalization of mono-, di-, and trisubstituted olefins, and only a few functional groups are not tolerated. The alkyl azides obtained are versatile intermediates and can be transformed to the free amines or triazoles without isolation of the azides. Preliminary mechanistic investigations suggest a rate-limiting hydrocobaltation of the alkene, followed by an amination reaction. Radical intermediates cannot be ruled out and may be involved.

Full text of DOI:10.1021/ja062355+

Published on Web 08/10/2006
Hydrazines and Azides via the Metal-Catalyzed
Hydrohydrazination and Hydroazidation of Olefins
Je´roˆme Waser, Boris Gaspar, Hisanori Nambu, and Erick M. Carreira*
Contribution from the Laboratorium fu¨r Organische Chemie, ETH Zu¨rich, Ho¨nggerberg
HCI-H335, 8093 Zu¨rich, Switzerland
Received April 6, 2006; E-mail: carreira@org.chem.ethz.ch
Abstract: The discovery, study, and implementation of the Co- and Mn-catalyzed hydrohydrazination and
hydroazidation reactions of olefins are reported. These reactions are equivalent to direct hydroaminations
of C-C double bonds with protected hydrazines or hydrazoic acid but are based on a different concept in
which the H and the N atoms come from two different reagents, a silane and an oxidizing nitrogen source
(azodicarboxylate or sulfonyl azide). The hydrohydrazination reaction using di-tert-butyl azodicarboxylate
is characterized by its ease of use, large functional group tolerance, and broad scope, including mono-,
di-, tri-, and tetrasubstituted olefins. Key to the development of the hydroazidation reaction was the use of
sulfonyl azides as nitrogen sources and the activating effect of tert-butyl hydroperoxide. The reaction was
found to be efficient for the functionalization of mono-, di-, and trisubstituted olefins, and only a few functional
groups are not tolerated. The alkyl azides obtained are versatile intermediates and can be transformed to
the free amines or triazoles without isolation of the azides. Preliminary mechanistic investigations suggest
a rate-limiting hydrocobaltation of the alkene, followed by an amination reaction. Radical intermediates
cannot be ruled out and may be involved.
barrier,6e and only the reaction with certain electron deficient
olefins occurs spontaneously.7 Whereas heterogeneous catalysts
Introduction
Olefins are inexpensive and readily available starting materials
for organic synthesis. For this reason, the direct heterofunc-
tionalization of the C-C double bond has been the focus of
research for many years.1 Most of the early methods were based
on the use of stoichiometric reagents. The advances of catalysis
in the past decades have allowed the development of economi-
cally and environmentally improved processes,2 with particular
success in the introduction of oxygen functionalities on alkenes,
as examplified by Sharpless’ epoxidation3 and dihydroxylation4
reactions. Progress in the amination of olefins has been much
slower, and, thus, this transformation is much less selective and
efficient than the corresponding oxyfunctionalization reactions.1b
Because of the ubiquity of amine-derived functional groups in
natural products and pharmaceuticals,5 a fast and general access
to amines would be highly desirable.
are used in industrial processes,6b the extreme conditions needed
(>100 bar ammonia, >200 °C) demand special equipment.
Furthermore, any effective hydroamination method must deal
with the problem of regio- and stereoselectivity.8 Given these
limitations, the majority of academic research has focused on
well-defined homogeneous systems. The first report on a
hydroamination reaction of olefins was published in 1932 by
Hickinbottom, who described the direct reaction of arylamines
with cyclohexene, styrene, and certain dienes at high temperature
(250-300 °C) in very low yields (<20%).9 Later, strong bases
were found to catalyze the hydroamination reaction.10 The first
transition-metal-catalyzed olefin hydroamination was reported
in 1971 by Coulson using Rh salts.11 Since this initial report,
(6) For a review on early work, see: (a) Hegedus, L. In ComprehensiVe Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Elmsford, NY,
1991; Vol. 4, p 551. For reviews on catalytic hydroamination, see: (b)
Brunet, J. J.; Neibecker, D. In Catalytic Heterofunctionalization: From
Hydroamination to Hydrozirconation; Togni, A., Gru¨tzmacher, H., Eds.;
Wiley-VCH: New York, 2001; p 91. (c) Nobis, M.; Driessen-Holscher,
B. Angew. Chem., Int. Ed. 2001, 40, 398. For recent reviews on asymmetric
hydroamination, see: (d) Roesky, P. W.; Muller, T. E. Angew. Chem., Int.
Ed. 2003, 42, 2708. (e) Hultzsch, K. C. AdV. Synth. Catal. 2005, 347, 367.
(7) Jung, M. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I.,
Eds.; Pergamon Press: Elmsford, NY, 1991; Vol. 4, p 1.
(8) For a general discussion on the regioselectivity for the functionalization of
olefins, see: Beller, M.; Seayad, J.; Tillack, A.; Jiao, H. Angew. Chem.,
Int. Ed. 2004, 43, 3368.
(9) (a) Hickinbottom, W. J. J. Chem. Soc. 1932, 2644. (b) Hickinbottom, W.
J. J. Chem. Soc. 1934, 319. (c) Hickinbottom, W. J. J. Chem. Soc. 1934,
1981.
The olefin-hydroamination reaction, the direct addition of
amines across C-C double bonds, represents a seemingly
straightforward process to access amines.6 Although the addition
of amines to alkenes is thermodynamically neutral or exother-
mic, this reaction typically displays a relatively high activation
(1) (a) The Chemistry of Double-Bonded Functional Groups; Patai, S., Ed.;
Wiley & Sons, Ltd.: New York, 1997; Supplement A3. (b) Catalytic
Heterofunctionalization: From Hydroamination to Hydrozirconation;
Togni, A., Gru¨tzmacher, H., Eds.; Wiley-VCH: New York, 2001.
(2) (a) Trost, B. M. Science 1991, 254, 1471. (b) Trost, B. M. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 259.
(3) Johnson, R. A.; Sharpless, K. B. In Catalytic Asymmetric Synthesis; Ojima,
I., Ed.; VCH: New York, 1993; p 103.
(10) For a review, see: (a) Seayad, J.; Tillack, A.; Hartung, C. G.; Beller, M.
AdV. Synth. Catal. 2002, 344, 795. For a recent example of an application
of base-promoted hydroamination in total synthesis, see: (b) Trost, B. M.;
Tang, W. P. J. Am. Chem. Soc. 2002, 124, 14542.
(4) Kolb, H. C.; Vannieuwenhze, M. S.; Sharpless, K. B. Chem. ReV. 1994,
94, 2483.
(5) Brown, B. R. The Organic Chemistry of Aliphatic Nitrogen Compounds;
Oxford University: New York, 1994.
(11) Coulson, D. R. Tetrahedron Lett. 1971, 12, 429.
9
10.1021/ja062355+ CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 11693-11712
11693

A R T I C L E S
Waser et al.
Scheme 1. Hydration, Hydrohydrazination, and Hydroazidation of
Olefins
two catalyst classes have proven efficient in the hydroamination
reaction: lanthanide and early metal-based catalysts are typically
used for intramolecular hydroaminations,12 whereas for late
transition metals, the preferred substrates are Michael accep-
tors,13 olefins activated via conjugation (vinylarenes and dienes)
or ring-strain (norbornene) for intermolecular hydroaminations.14
Despite encouraging recent progress,15 the search for a truly
general method for the hydroamination of olefins is still ongoing.
To avoid the difficulties inherent to the hydroamination
approach, more active electrophilic amination reagents had been
devised, which allow the functionalization of simple carbon
nucleophiles under mild conditions.16 Among these reagents,
nitroso compounds,17 azides,18 and especially azodicarboxy-
lates16,19 have emerged as useful nitrogen sources. However,
with the exception of Diels-Alder20a-d and ene reactions,20e-g
such reagents cannot be used for the direct functionalization of
nonnucleophilic C-C double bonds.
However, there are no really general methods for the direct
addition of an O-H bond onto a C-C double bond.21 In 1989,
Mukaiyama and Isayama developed a new concept for the
hydration of olefins: instead of activating a single O-H bond
for addition, they decided to use simultaneously a hydride source
and an oxygen source.22 This stepwise approach allows the use
of much more reactive reagents, with the hope of overriding
the kinetic barrier to the functionalization of nonactivated
alkenes. However, this gain goes together with formidable
challenges for the development of an efficient process. Three
undesired pathways must be taken into account: overoxidation
to ketone, over-reduction to the alkane, and direct reaction of
the oxidant with the reductant. With the choice of silanes or
alcohols as reductants, molecular oxygen as oxidant, and Co or
Mn as catalyst, Mukaiyama and Isayama were able to manage
these difficulties (Scheme 1A).
The hydration of olefins21 is the oxygen analogue of the
hydroamination reaction and could as such serve as a source of
inspiration for the development of new hydroamination methods.
(12) (a) Gagne´, M. R.; Marks, T. J. J. Am. Chem. Soc. 1989, 111, 4108. For a
review, see: (b) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673.
(13) (a) Kawatsura, M.; Hartwig, J. F. Organometallics 2001, 20, 1960. (b)
Zhuang, W.; Hazell, R. G.; Jorgensen, K. A. Chem. Commun. 2001, 1240.
(c) Fadini, L.; Togni, A. Chem. Commun. 2003, 30. (d) Fadini, L.; Togni,
A. Chimia 2004, 58, 208.
(14) For a review, see: (a) Hartwig, J. F. Pure Appl. Chem. 2004, 76, 507. For
styrene hydroamination with late metal, see: (b) Kawatsura, M.; Hartwig,
J. F. J. Am. Chem. Soc. 2000, 122, 9546. (c) Nettekoven, U.; Hartwig, J.
F. J. Am. Chem. Soc. 2002, 124, 1166. (d) Utsunomiya, M.; Hartwig, J. F.
J. Am. Chem. Soc. 2003, 125, 14286. (e) Utsunomiya, M.; Kuwano, R.;
Kawatsura, M.; Hartwig, J. F. J. Am. Chem. Soc. 2003, 125, 5608. (f)
Utsunomiya, M.; Hartwig, J. F. J. Am. Chem. Soc. 2004, 126, 2702. (g)
Takaya, J.; Hartwig, J. F. J. Am. Chem. Soc. 2005, 127, 5756. (h) Johns,
A. M.; Utsunomiya, M.; Incarvito, C. D.; Hartwig, J. F. J. Am. Chem. Soc.
2006, 128, 1828. (i) Beller, M.; Breindl, C.; Eichberger, M.; Hartung, C.
G.; Seayad, J.; Thiel, O. R.; Tillack, A.; Trauthwein, H. Synlett 2002, 1579.
For diene hydroamination, see: (j) Lober, O.; Kawatsura, M.; Hartwig, J.
F. J. Am. Chem. Soc. 2001, 123, 4366. (k) Pawlas, J.; Nakao, Y.; Kawatsura,
M.; Hartwig, J. F. J. Am. Chem. Soc. 2002, 124, 3669. (l) Qin, H.;
Yamagiwa, N.; Matsunaga, S.; Shibasaki, M. J. Am. Chem. Soc. 2006, 128,
1611. For nobornene hydroamination, see: (m) Casalnuovo, A. L.;
Calabrese, J. C.; Milstein, D. J. Am. Chem. Soc. 1988, 110, 6738. (n) Dorta,
R.; Egli, P.; Zurcher, F.; Togni, A. J. Am. Chem. Soc. 1997, 119, 10857.
(15) (a) Crimmin, M. R.; Casely, I. J.; Hill, M. S. J. Am. Chem. Soc. 2005, 127,
2042. (b) Qian, H.; Widenhoefer, R. A. Org. Lett. 2005, 7, 2635. (c) Bender,
C. F.; Widenhoefer, R. A. J. Am. Chem. Soc. 2005, 127, 1070. (d) Karshtedt,
D.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2005, 127, 12640. (e)
Zulys, A.; Dochnahl, M.; Hollmann, D.; Lo¨hnwitz, K.; Herrmann, J. S.;
Roesky, P. W.; Blechert, S. Angew. Chem., Int. Ed. 2005, 44, 7794. (f)
Timokhin, V. I.; Stahl, S. S. J. Am. Chem. Soc. 2005, 127, 17888. (g) Zhang,
J. L.; Yang, C. G.; He, C. J. Am. Chem. Soc. 2006, 128, 1798. (h) Michael,
F. E.; Cochran, B. M. J. Am. Chem. Soc. 2006, 128, 4246.
Upon examination of Mukaiyama’s work, a question arises:
would it be possible to combine the metal-mediated activation
of otherwise nonactivated olefins with an electrophilic/oxidizing
nitrogen source, while avoiding direct reduction of the nitrogen-
transfer reagent?23 Indeed, we have developed such an approach
both successful with azodicarboxylates and sulfonyl azides as
nitrogen sources, allowing the development of the hydrohy-
drazination24 (Scheme 1B) and hydroazidation25 (Scheme 1C)
of olefins. Herein, we provide a full account of our work,
including effect of catalyst structure, reaction conditions, process
optimization, and expanded scope. Furthermore, investigations
on the reaction mechanism are described.
(16) For a general review on amination, see: (a) Ricci, A. Modern Amination
Methods; Wiley-VCH: Weinheim, Germany, 2000. For reviews on
electrophilic amination, see: (b) Erdik, E.; Ay, M. Chem. ReV. 1989, 89,
1947. (c) Greck, C.; Genet, J. P. Synlett 1997, 741. (d) Dembech, P.; Seconi,
G.; Ricci, A. Chem.sEur. J. 2000, 6, 1281. (e) Erdik, E. Tetrahedron 2004,
60, 8747. (f) Greck, C.; Drouillat, B.; Thomassigny, C. Eur. J. Org. Chem.
2004, 1377. (e) Janey, J. M. Angew. Chem., Int. Ed. 2005, 44, 4292.
(17) (a) Oppolzer, W.; Tamura, O.; Sundarababu, G.; Signer, M. J. Am. Chem.
Soc. 1992, 114, 5900. (b) Yamamoto, H.; Momiyama, N. Chem. Commun.
2005, 3514.
Results and Discussion
Co-Catalyzed Hydrohydrazination Reaction. Alkyl hydra-
zines are useful precursors to amines in the assembly of synthetic
building blocks and have been broadly used for the elaboration
of heterocycles omnipresent in pharmaceuticals.26 The substitu-
(18) (a) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am. Chem.
Soc. 1990, 112, 4011. (b) Ollivier, C.; Renaud, P. J. Am. Chem. Soc. 2000,
122, 6496. (c) Panchaud, P.; Chabaud, L.; Landais, Y.; Ollivier, C.; Renaud,
P.; Zigmantas, S. Chem.sEur. J. 2004, 10, 3606.
(19) (a) Duthaler, R. O. Angew. Chem., Int. Ed. 2003, 42, 975. For recent
applications, see: (b) Saaby, S.; Bella, M.; Jørgensen, K. A. J. Am. Chem.
Soc. 2004, 126, 8120. (c) Liu, X. F.; Li, H. M.; Deng, L. Org. Lett. 2005,
7, 167. (d) Poulsen, T. B.; Alemparte, C.; Jørgensen, K. A. J. Am. Chem.
Soc. 2005, 127, 11614.
(22) (a) Mukaiyama, T.; Isayama, S.; Inoki, S.; Kato, K.; Yamada, T.; Takai,
T. Chem. Lett. 1989, 449. (b) Inoki, S.; Kato, K.; Takai, T.; Isayama, S.;
Yamada, T.; Mukaiyama, T. Chem. Lett. 1989, 515. (c) Kato, K.; Yamada,
T.; Takai, T.; Inoki, S.; Isayama, S. Bull. Chem. Soc. Jpn. 1990, 63, 179.
(d) Isayama, S.; Mukaiyama, T. Chem. Lett. 1989, 1071. (e) Isayama, S.;
Mukaiyama, T. Chem. Lett. 1989, 569. (f) Isayama, S.; Mukaiyama, T.
Chem. Lett. 1989, 573. (g) Isayama, S. Bull. Chem. Soc. Jpn. 1990, 63,
1305.
(20) (a) Yamamoto, Y.; Yamamoto, H. J. Am. Chem. Soc. 2004, 126, 4128. (b)
Yamamoto, Y.; Momiyama, N.; Yamamoto, H. J. Am. Chem. Soc. 2004,
126, 5962. (c) Yamamoto, Y.; Yamamoto, H. Angew. Chem., Int. Ed. 2005,
44, 7082. (d) Chow, C. P.; Shea, K. J. J. Am. Chem. Soc. 2005, 127, 3678.
(e) Brimble, M. A.; Heathcock, C. H. J. Org. Chem. 1993, 58, 5261. (f)
Baran, P. S.; Guerrero, C. A.; Corey, E. J. Org. Lett. 2003, 5, 1999. (g)
Srivastava, R. S.; Nicholas, K. M. J. Am. Chem. Soc. 1997, 119, 3302.
(21) Tani, K.; Kataoka, Y. In Catalytic Heterofunctionalization; Togni, A.,
Gru¨tzmacher, H., Eds.; Wiley-VCH: New York, 2001; p 171.
(23) The oxidative amination of R,â-unsaturated carbonyl compounds with
nitrites has been reported by Mukaiyama: (a) Kato, K.; Mukaiyama, T.
Chem. Lett. 1990, 1395. (b) Kato, K.; Mukaiyama, T. Chem. Lett. 1990,
1917. (c) Kato, K.; Mukaiyama, T. Bull. Chem. Soc. Jpn. 1991, 64, 2948.
(24) (a) Waser, J.; Carreira, E. M. J. Am. Chem. Soc. 2004, 126, 5676. (b) Waser,
J.; Carreira, E. M. Angew. Chem., Int. Ed. 2004, 43, 4099. (c) Waser, J.;
Gonza´lez-Go´mez, J. C.; Nambu, H.; Huber, P.; Carreira, E. M. Org. Lett.
2005, 7, 4249.
(25) Waser, J.; Nambu, H.; Carreira, E. M. J. Am. Chem. Soc. 2005, 127, 8294.
9
11694 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 2. Influence of Ligand Structure in the Hydration of 531
Table 1. Simple Co Complexes Tested in the Hydrohydrazination
of 4-Phenylbutene 1
entry
complex
timea
yield of 3b
1
2
3
4
5
6
7
none
CoCl2‚6H2O
Co(OAc)2
Co(NO3)2‚6H2O
Co(acac)2
Co(modp)2
Co(dpm)2
48 h
4 h
2 h
8 h
8 h
0%
3%
<5%
<5%
5%
<5%
<5%
48 h
4 h
a The reaction was stopped when all di-tert-butylazodicarboxylate (2)
was consumed. b Standard conditions: 0.50 mmol 4-phenylbutene 1, 0.50
mmol PhSiH3, 0.75 mmol di-tert-butylazodicarboxylate (2), 5 mol %
catalyst, and 2.5 mL of EtOH at 23 °C under argon.
tion of alkyl halides with hydrazine has been used for a long
time for the synthesis of alkylated hydrazines.26 More recently,
other approaches, including asymmetric variations such as
nucleophilic additions to hydrazones27 or azodicarboxylates16
and conjugate additions of hydrazines to Michael acceptors have
been developed. However, the formation of hydrazines from
nonactivated olefins have been limited to cycloaddition or ene
reactions with azodicarboxylates,20e-f and the synthesis of
N-alkyl hydrazides by direct functionalization of the C-C
double bond of unactivated olefins was unprecedented prior to
this work.
At the beginning of our studies, we decided to examine the
functionalization of 4-phenylbutene (1) as a prototype for
monosubstituted unactivated olefins, as the presence of the
phenyl group in a remote position was expected to show no
significant effect on the reaction but confers UV activity and a
lower volatility to this substrate. Moreover, 1 was known to be
a good substrate in Mukaiyama’s hydration reaction.22 One of
the major hurdles in the development of an efficient hydrohy-
drazination reaction becomes apparent when considering the
effect of simple Co salts in the hydrohydrazination of 4-phe-
nylbutene 1 with di-tert-butyl azodicarboxylate (2; Table 1). In
the absence of catalyst, no reaction occurs (entry 1), but simple
Co salts all promote the direct reduction of azodicarboxylate 2
to hydrazine 4 instead of the desired hydrohydrazination reaction
(entries 2-4). Moreover, Co(acac)2, Co(modp)2, and Co-
(dpm)2,28 which are efficient catalysts for the related hydration
reaction,22 also favor the formation of 4 (entries 5-7). Traces
(5%) of the desired hydrohydrazination product 3 could be
isolated only in the case of Co(acac)2 (entry 5). Thus, despite
the seeming similarity between the hydrohydrazination and
Mukaiyama hydration, the parallels are limited.
a Two equiv of ligands 7-11 and 16 were mixed with 1 equiv of CoCl2
in acetonitrile for 20 h, followed by solvent removal. Ligands 12-16 were
formed in situ from the amino acid, salicylaldehyde, and KOH in EtOH;
addition of 0.50 equiv of Co(NO3)2‚6H2O, oxidation with H2O2, and solvent
removal gave the Co catalyst. b R ) 2-hydroxyphenyl. c Standard condi-
tions: 0.04 mmol enyne 5, 0.12 mmol PhSiH3, 12 mol % catalyst, 1 atm
O2, and 0.7 mL of EtOH at 23 °C.
promote the epoxidation and allylic oxidation of olefins with
oxygen29 and is easily obtained by simply mixing ligand 7 and
CoCl2 in CH3CN. This complex had proven efficient in the
hydration reaction but was unstable under the reactions requiring
high catalyst loadings (>10%) to achieve full conversion.
Subsequent systematic variation of the amine moiety of the
ligand showed a strong effect on the outcome of the reaction
(Table 2). Modification of the side chain of the amino acid
(entries 2 and 3) or removal of the ester group (entries 4 and 5)
did not lead to more active or stable catalysts, but ligands 12-
16, bearing a free carboxylate group, were stable under the
reaction conditions (entries 6-10). Among those, the use of
Schiff-base ligand 16 led to unique reactivity, allowing full
conversion of enyne 5 to alcohol 6 in 30 min.
The behavior of the Co system derived from 16 was
peculiar: all the Co(II) complexes synthesized from ligands
7-11 were green. In contrast, the solution of 16 and Co(NO3)2‚
6H2O turned partially dark-red during its synthesis. We hy-
pothesized that this change of color was an indication of a
change in the oxidation state of Co from +II to +III, thus, the
synthetic protocol was altered to obtain the Co(III) mixture
From these results, it became clear that the acetylacetonate
ligand scaffold was not well-suited for the hydrohydrazination
reaction. In ongoing investigations of the Mukaiyama hydration
of enynes, we had found that Schiff-base ligands were also
efficient catalysts (Table 2). The Co(II) complex derived from
ligand 7 had been introduced by Iqbal and co-workers to
(29) Punniyamurthy, T.; Bhatia, B.; Reddy, M. M.; Maikap, G. C.; Iqbal, J.
Tetrahedron 1997, 53, 7649.
(30) Burrows, R. C.; Bailar, J. C. J. Am. Chem. Soc. 1966, 88, 4150.
(31) The choice of a para-methoxybenzoyl protecting group was motivated by
its favorable properties for monitoring the reaction (high UV activity, low
volatility).
(26) Ragnarsson, U. Chem. Soc. ReV. 2001, 30, 205.
(27) (a) Kobayashi, S.; Ishitani, H. Chem. ReV. 1999, 99, 1069. (b) Sugiura,
M.; Kobayashi, S. Angew. Chem., Int. Ed. 2005, 44, 5176.
(28) acac ) acetylacetonato, dpm ) dipivaloylmethanato, modp ) 1-morpholi-
nocarbamoyl-4,4-dimethyl-1,3-pentanedionato.
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11695

A R T I C L E S
Waser et al.
directly. The first protocol we examined to effect the synthesis
of the Co(III) complexes was based on a reported procedure of
Bailar and Burrows for the synthesis of Co(III) coordination
complexes with ligands derived from salicylaldehyde (17) and
amino acids.30 In the Bailar one-pot procedure, the template
condensation of salicylaldehyde 17 and an R-monosubstituted
amino acid (in eq 1, R1 ) H and R2 ) alkyl) in the presence of
Co(II) salts, base, and hydrogen peroxide led to the formation
of the coordinatively saturated cobaltate complexes (eq 1). It
was not, however, clear if this procedure could be extended to
the system of interest in our study involving the more sterically
hindered R,R-dimethyl-substituted amino acid 18 in order to
synthesize the targeted complex 19.
In the experiment, we indeed isolated a dark red-brown solid,
which when employed in the hydration reaction afforded
complete conversion. However, the isolated solid did not consist
exclusively of 19, although its presence in the mixture was
confirmed by mass spectrometry. The NMR data were inde-
terminate due to the presence of remaining Co(II) impurities
but suggested the presence of another unidentified complex
(referred to as 20). Importantly, the red-brown solid proved
uniquely able to promote the hydrohydrazination of 4-phenyl-
butene 1, and hydrohydrazination product 3 was obtained in
85% yield in 4 h on a 0.50 mmol scale (eq 2).
Figure 1. Proposed structure and 1H NMR spectrum of complex 20 in
comparison with 19 (the assignment of A and B is arbitrary).
this mixture could be purified by simple column chromatography
on silica gel using MeOH/CH2Cl2 solvent mixtures without
decomposition. This new complex accounted in fact for about
60-80% of the crude mass and was pure enough to allow us
to provide spectra displaying well-defined sharp signals. Careful
comparison of the 1H NMR spectra of 19 and the new complex
20 allowed us to propose a possible structure for 20 (Figure 1).
The highly symmetrical complex 19 displays a single peak for
the imino H at 8.65 ppm and only two signals for the methyl
groups; the ratio of aromatic/aliphatic H is 2/3, as expected. In
the spectrum of 20, a ratio of aromatic/aliphatic H of 1/2.8 is
found (expected: 1/3). The appearance of a doubling of all peaks
suggests the presence of two structural isomers A and B in a
ratio of 2 to 1. Clearly, two sets of four nonequivalent methyl
groups are present in the molecule (A and B) between 1 and 2
ppm (Figure 1). This is consistent with structure 20, if the fact
that two different structural isomers A and B are possible is
taken into account. For the sake of simplicity, only one of the
isomers will be portrayed in this work, as no attempt was made
to determine which structure was the major and which was the
minor isomer. Apart from the tridentate Schiff-base ligand 16
and 2-amino-isobutyric acid 18, the ligand sphere of the complex
is further completed by a solvent molecule. The presence of
this potentially free-binding site on complex 20 could be the
reason of its enhanced reactivity.
At this stage of development, the enhanced reactivity
displayed by the red-brown mixture of Co complexes 19 and
20 was puzzling for us, as we had assumed that complex 19
was the active catalyst and there were no reasons to expect
higher reactivity for 19, especially when considering that 19 is
coordinatively saturated. As the results obtained using the
mixture of 19 and 20 were consistent from batch to batch, this
had no practical consequences of our studies on the hydrohy-
drazination of alkenes.24a However, a deeper knowledge of the
catalyst structure was deemed essential for further improve-
ments. Subsequently, after examination of the scope of the
hydrohydrazination reaction, we made use of another procedure
reported by Bailar and Burrows and were finally able to obtain
clean samples of 19 starting directly with Co(OH)3.30 The
structure of 19 could be then further confirmed via 1H and 13
C
NMR, IR, and high-resolution mass spectrometry. To our
dismay, however, complex 19 was not an effective catalyst for
hydration and hydrohydrazination, showing only a few conver-
sions (<20%) after 2 days for the hydrohydrazination of
4-phenylbutene 1. At this point, it was clear the active catalyst
was not 19, but most probably the nonidentified complex 20.
Consequently, we intensified our efforts toward the purifica-
tion of the mixture obtained using our previous procedure (eq
1). After several thwarted attempts at recrystallization, we were
pleasantly surprised to find out that the major component of
A confirmation of the structure of 20 could be obtained via
high-resolution mass spectrometry. Initially, as we were target-
9
11696 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 3. Co-Catalyzed Hydrohydrazination of Monosubstituted
Olefins
ing the anionic complex 19, we had only detected negative mass
in our measurements. However, under these conditions, the
neutral complex 20 could not be detected. With our new
structure proposal, we extended our measurements to include
positive mass signals. Indeed, high-resolution mass peaks were
obtained for both [M + H]+ and [M + Na]+ in the case of
neutral complex 20.
With a new structure proposal for the active catalyst, we
designed a two-step synthesis for complex 20 that gives higher
yield and purity (eq 3). A mixture of salicylaldehyde 17, amino
acid 18, and Co(OAc)2‚H2O provides Co(II) complex as a
yellow-orange precipitate. After filtration, suspension of this
precipitate in ethanol and stirring in air in the presence of amino
acid 18 provides the dark red complex 20. The fact that complex
20 is obtained in higher yield and purity when introducing the
tridentate and the bidentate ligands in a two-steps procedure
suggests that the proposed structure is indeed correct.
Without doubt, the discovery of catalyst 20 was the key for
the development of an efficient hydrohydrazination reaction.
Nevertheless, other reaction parameters also play an important
role in the hydrohydrazination of 4-phenylbutene 1 as test
substrate. For example, the use of diethyl azodicarboxylate
(DEAD) afforded the corresponding hydrazine in only 34%
yield, probably because DEAD lacks the steric hindrance of 2,
allowing a faster competitive reduction of the N-N double
bond. On the other hand, hydrohydrazination product 3 could
be obtained in 72% yield using only 1 equiv of azodicarboxylate
2. In contrast to Mukaiyama hydration reaction, the hydrohy-
drazination reaction displayed a strong solvent effect: alcoholic
solvents (methanol, ethanol, 2-propanol) proceeded equally well,
whereas acetonitrile gave low yields of hydrohydrazination
product and less-polar solvents (CH2Cl2, THF) lead to no
conversion at all.
Next, several other hydride sources were tested in the reaction.
Trialkylsilanes, sodium borohydride, or hydrogen gas could not
be used, as they led either to no conversion (hydrogen gas,
trialkylsilanes) or favored the direct reduction of azodicarboxy-
late 2 (sodium borohydride), but poly(methylhydrosiloxane)
(PMHS) and tetramethylhydrosiloxane (TMDSO) led to the
formation of hydrazine 3. In the case of PMHS, conversion and
reaction rate were too low to be useful (20% over 24 h), but
TMDSO was as efficient as phenylsilane, affording 3 in 86%
in 4 h. As this silane is inexpensive and widely available, this
constitutes an improvement when compared to phenysilane.32
Finally, it was established that the reaction can be run on a
larger scale by scaling it up 10-fold (5 mmol). Using only 2.5
mol % catalyst 20, the hydrohydrazination product 3 was
obtained from 4-phenylbutene 1 in 94% yield (1.75 g, starting
from 0.67 g of 1).
a Standard conditions: 0.50 mmol alkene, 0.50 mmol PhSiH3, 0.75 mmol
2, 5 mol % catalyst 20, and 2.5 mL of ethanol at 23 °C under argon. b Alkene
(0.50 mmol) was added as a solution in 1 mL of CH2Cl2, using 1.5 mol %
catalyst 20.
necessary: although the reaction time was indeed shortened
(1.5×), no difference in yield was observed compared to that
of the mixture obtained using our previously published
procedure.24a Key features of the Co-catalyzed hydrohydra-
zination of monosubstituted olefins are the following: (1) high
Markovnikov selectivity with nearly exclusive formation of the
secondary hydrazine product; (2) a strong activating and
R-directing effect of a phenyl group (entry 2), in this case, slow
substrate addition and low catalyst loadings were needed to
prevent competitive polymerization (for acrylate derivatives, no
product could be isolated due to extensive polymerization); and
(3) broad functional group tolerance, including free alcohols
(entries 3 and 4), an ether (entry 5), an acetal (entry 6), a ketone
(entry 7) and a bromide (entry 8).
The large functional group tolerance observed is one of the
major advantages of the hydrohydrazination reaction. To further
extend the scope of the reaction, we decided to examine the
functionalization of vinyl-substituted heterocycles. Heterocycles
are ubiquitous in organic chemistry: in 2003, from more than
20 million compounds registered, about one-half featured
heterocycles.33 They have found applications in chemistry,
biology, agriculture, and material science and are the constituents
of many natural products, pharmaceuticals, herbicides, dyes, and
other products of technical importance.33,34 For these reasons,
the synthesis, elaboration, and functionalization of heterocycles
The scope of the Co-catalyzed hydrohydrazination reaction
with monosubstituted olefins was then examined (Table 3).
Importantly, the use of thoroughly purified complex 20 was not
(33) Eicher, T.; Siegfried, H. The Chemistry of Heterocycles: Structure,
Reactions, Syntheses, and Applications; Wiley-VCH: Weinheim, Ger-
many, 2003.
(34) Pozharskii, A. F.; Soldatenkov, A. T.; Katritzky, A. R. Heterocycles in
Life and Society: An Introduction to Heterocyclic Chemistry and Bio-
chemistry and the Role of Heterocycles in Science, Technology, Medicine
and Agriculture; John Wiley & Sons, Ltd.: Chichester, 1997.
(32) Typical prices for laboratory-scale quantities: phenylsilane, 7.50 CHF/g,
271 CHF/mol(H); PMHS, 0.20 CHF/g, 12 CHF/mol(H); TMDSO, 1.3 CHF/
g, 87 CHF/mol(H).
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11697

A R T I C L E S
Waser et al.
Table 4. Hydrohydrazination of Vinyl-Substituted Heterocycles
is an intense field of research in organic chemistry. However,
when dealing with heterocycles, any catalytic process faces a
great challenge. Many heterocycles are also good ligands for
transition metals and can poison the catalyst. Some basic
heterocycles preclude the use of Brønsted or Lewis acids as
catalysts. Consequently, high catalytic loadings and harsh
reaction conditions are often needed for catalytic methods if
heterocycles are present.
In this context, the hydroamination of vinyl-substituted
heterocycles would be an interesting method to access densely
functionalized building blocks. To the best of our knowledge,
the full potential of this class of compounds has not been
examined for the hydroamination reaction. Marks reported that
pyridines were not compatible with lanthanide catalysts.12 Beller
introduced Rh-based catalysts for the anti-Markovnikov hy-
droamination of vinylpyridines,35 but no other heterocyclic
substrates were reported. Hartwig documented the use of amino
pyridine for the hydroamination of cyclohexadienes at elevated
temperature.14j
We began our studies with the examination of simple
unprotected vinyl- substituted heterocycles (Table 4, entries
1-8). Both 2- and 3-vinylfurans could be aminated in useful
yields (entries 1 and 2). This result is noteworthy, as vinylfurans
are otherwise prone to polymerization. The higher yields
observed for 2-vinylthiophene (entry 3) reflect the higher
stability of this compound. Among nitrogen heterocycles,
2-vinylpyrrole (entry 4) did not give the desired product, as we
were not able to suppress polymerization. However, N-methyl-
2-vinylimidazole could be functionalized in moderate yields
(entry 5). The fact that our reaction is compatible with the
relatively high basicity of N-methyl-2-vinylimidazole (pKa )
7) is interesting. Six-membered ring nitrogen heterocycles could
also be functionalized in useful yields, for example, 2-vinylpy-
ridine (entry 6) or the electron-poor 2-vinylpyrazine (entry 7).
Finally, in contrast to 2-vinyl-pyrrole, 2-vinylindole (entry 8)
gives the desired amination product in good yield. However,
the corresponding 3-vinylindole (not shown) polymerized before
a reaction could take place. The successful functionalization of
2-vinylindole showed the tolerance of the reaction toward a free
indole N-H bond. In this context, 4-vinyl aniline (entry 9) was
examined. The yields were moderate in this case. However,
simple protection of the aniline nitrogen was sufficient to afford
quantitative yield of the hydrohydrazination product (entry 10).
The FMOC protecting group appeared particularly well-suited,
as protection and subsequent deprotection were easy and high
yielding, finally furnishing the hydrohydrazination product with
a free amino group in 83% yield over four steps.
The heterocycles that we were unable to hydroaminate,
2-vinylpyrrole and 3-vinylindole, are among the more frequently
used heterocycles. The instability of these compounds is due
mostly to the high donor ability of the nitrogen atom. For this
reason, we subsequently protected the nitrogen atom to diminish
the electron-density of the heterocycles. We were pleased to
find that both Boc- and tosyl-protected 2-vinylpyrroles (entries
11 and 12) and 3-vinylindoles (entries 13 and 14) react with
useful yields in the hydrohydrazination reaction.
a General procedure: 0.50 mmol alkene was added as a solution in 1.0
mL of CH2Cl2 and 1.0 mL of ethanol to a solution of 0.50 mmol PhSiH3,
0.75 mmol di-tert-butyl azodicarboxylate (2), 2.5 mol % catalyst 20, and
2.5 mL of ethanol at 23 °C under argon. b The desired product could not
be separated from nonidentified impurities.
and directing effect of a phenyl group was observed, allowing
the functionalization of di- and trisubstituted styrene derivatives
in high yield, with exclusive formation of the hydrazine product
at the benzylic position (entries 1-5). In the case of heterocyles,
both R- and â-methylvinylpyridines react in the hydrohydra-
zination reaction with moderate yields (entries 6 and 7).
R-Methyl vinylthiophene (entry 8) could also be functionalized,
Next, we examined the hydrohydrazination reaction of di-
and trisubstituted olefins (Table 5). Again, a strong activating
(35) Beller, M.; Trauthwein, H.; Eichberger, M.; Breindl, C.; Mu¨ller, T. E. Eur.
J. Inorg. Chem. 1999, 1121.
9
11698 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 5. Hydrohydrazination of Di- and Trisubstituted Olefins
but R-methyl 2-vinylbenzothiazole (entry 9) gives only very
low yield.
For the R-substituted heterocycles, the lower yields observed
in comparison with unsubstituted substrates are partially due
to dimerization of the starting material. In the case of R-methyl
2-vinylbenzothiazole (entry 9), this becomes the main pathway.
To override the tendency of R-methyl 2-vinylbenzothiazole to
dimerize, we ran the reaction with a large excess of the more
reactive DEAD and at a lower catalyst loading (eq 4). Under
these conditions, it was possible to obtain the desired amination
product in moderate yields, but the dimerization of the starting
material could not be suppressed completely.
When no aromatic activating group was present, an important
difference in reactivity became apparent: R,R-disubstituted
olefins were good substrates, giving the tertiary hydrazine
products exclusively in good yields (Table 5, entries 10-13),
but R,â-disubstituted olefins react only in low yields (<50%)
in the hydrohydrazination reaction. Useful yields could be
obtained only in the case of an R,â-unsaturated ester (entry 14)
and some endocyclic alkenes (entries 16-18), using a larger
excess of azodicarboxylate 2 (2 equiv). For this class of
substrates, the hydrohydrazination reaction is much slower,
and hydrazine 3 resulting from the direct reduction of 2
becomes the major product. The examination of a trisubstituted
alkene revealed an intermediate reactivity between R,â-disub-
stituted and R,R-disusbstituted olefins, allowing useful yields
(entry 15).
Mn-Catalyzed Hydrohydrazination Reaction. Though highly
active for most substrates, Co catalyst 20 was not able to make
hydrohydrazination competitive with azodicarboxylate reduction
in the case of sterically hindered substrates. No Co complex
tested proved more active than 20. Therefore, we decided to
expand our study to other metals. Mukaiyama has reported the
hydration of olefins using Mn(dpm)2.36 Indeed, a first attempt
for the hydrohydrazination of 4-phenylbutene 1 with Mn(dpm)2
was promising: the amination of 1 was quantitative in 5 min
instead of the 4 h required by catalyst 20. However, this
increase in reactivity comes at the cost of the selectivity: the
ratio of Markovnikov/anti-Markovnikov was only 3:1. Another
convenient feature of the Mn catalyst is that the progress of
the reaction can be monitored by the color of the reaction
mixture: as long as azodicarboxylate 2 is present, the reaction
mixture is brown-green. When 2 is consumed, the color changes
to light yellow. To improve upon this initial result, we first
changed the catalyst to the Mn(III) complex Mn(dpm)3 (23).
The activity of the Mn(II) complex Mn(dpm)2 and 23 is
identical, but Mn(dpm)2 is unstable in air and slowly oxidizes
to give a mixture of Mn(II) and Mn(III) compounds.37 On the
other hand, Mn(III) complex 23 is bench-stable and requires
no special care in storage. Diminishing the catalyst loading to
2 mol % and the temperature to 0 °C finally allowed us to
increase the Markovnikov/anti-Markovnikov selectivity to 5.5:1
(eq 5).
a Standard conditions: 0.50 mmol alkene, 0.50 mmol PhSiH3, 0.75 mmol
2, 5 mol % catalyst 20, and 2.5 mL of ethanol at 23 °C under argon. b Alkene
(0.50 mmol) was added as a solution in 1 mL of CH2Cl2 using 2.5 mol %
catalyst 20. c Alkene (0.50 mmol), 0.75 mmol PhSiH3, and 1.0 mmol 2
were used. d Dimerization at the tertiary center was observed.
(36) Inoki, S.; Kato, K.; Isayama, S.; Mukaiyama, T. Chem. Lett. 1990, 1869.
(37) Cotton, F. A.; Soderberg, R. H. Inorg. Chem. 1964, 3, 1.
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11699

A R T I C L E S
Waser et al.
Table 6. Mn-Catalyzed Hydrohydrazination of Olefins
Next, substrates which gave low yields with Co catalyst 20
were examined (Table 6). Homoallyllic alcohol could be
aminated in much better yield (72%, 22% with Co catalyst 20;
entry 1). R,â-Disubstituted acyclic olefins (entries 2-5) give
good yields with catalyst 20, except for crotonyl nitrile (entry
4). However, the regioselectivity was low (entry 2). Cyclic
substrates are particularly well-suited for the new Mn-catalyzed
protocol (entries 6-10), showing increased yields for all
substrates compared with 20. A cyclic ether was also well
tolerated (entry 10).
Mn catalyst 23 was also effective for the functionalization
of vinyl-substituted heterocycles, giving yields comparable to
those obtained with Co catalyst 20. However, in the case of
N-methyl-2-vinylimidazole (entry 11) and 4-vinylaniline (entry
12), a significant increase in yield was observed (83% vs 60%
for N-methyl-2-vinylimidazole and 60-80% vs 20-40% for
4-vinylaniline). These results together with the improvement
observed for homoallylic alcohol (entry 1) suggest that Mn
catalyst 23 is more tolerant toward coordinating groups, which
slow the reaction for Co catalyst 20. Finally, we turned to the
most challenging substrates for Co catalyst 20: tetrasubstituted
olefins (entries 13-16). Although the yields were very low with
20 (<20%), useful yields were obtained with Mn catalyst 23
(51-79%).
To get more detailed information concerning the relative
activity of the two catalytic systems, reactions with low catalytic
loadings (0.1 mol %) were run with 4-phenylbutene 1 and
tetramethylethylene. In the case of olefin 1, a 10-fold increase
of turnover number38 was observed with 23 in comparison with
20 (430 vs 44). For tetramethylethylene, this difference even
rises to 80 (240 vs 3).
The increased reactivity of Mn catalyst 23 also makes the
use of other silanes as a reductant viable. Diphenylsilane was
found to be an efficient hydride source. More interestingly,
PMHS could be used in the hydrohydrazination of 4-phenyl-
butene 1. Hydrazine 3 was obtained in 88% yield, although the
reaction must be run at room temperature and was slower (15
h). The result with TMDSO (full conversion in 24 h at 23 °C)
is intriguing: in the case of Mn catalyst 23, the reaction with
this silane was even slower than with PMHS, in contrast with
what was observed with Co catalyst 23, where it displays
reaction rates comparable to phenylsilane. Other silanes or
sodium borohydride could not be used as hydride sources, as
they led either to no conversion (trialkylsilanes) or favored the
direct reduction of azodicarboxylate 2 (sodium borohydride).
PMHS could also be used in the case of more challenging
substrates such as 4-butenol or tetramethylethylene, and the
desired product was obtained in useful yields (60% and 60%,
respectively), although the yields were consistently lower than
those obtained with the use of phenylsilane (entries 1 and 11,
Table 6).
a General procedure: 0.50 mmol alkene, 0.50 mmol PhSiH3, 0.75 mmol
di-tert-butyl azodicarboxylate (2), 2 mol % catalyst 23, and 2.5 mL of
i-PrOH at 0 °C under argon. b Major product drawn, the minor compound
results from the hydrohydrazination next to the hydroxy group. c The desired
product could not be separated from nonidentified impurities.
Chiral esters and imide substrates were examined to find out
if diastereoselective reactions were possible with Co catalyst
20 or Mn catalyst 23. Whereas menthyl and oxazolidinone
auxiliaries did not show any significant diastereoselectivity,
crotyl ester 25 derived from pantolactone showed asymmetric
(38) The “turnover numbers" obtained this way are not really accurate because
they reflect the activity of the catalysts for two reactions at least
(hydrohydrazination reaction and reduction of azodicarboxylate). In the case
of 1, moreover, some side products resulting from a direct ene reaction
were also isolated. Nevertheless, these numbers allow a good comparison
for practical purposes.
9
11700 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Figure 2. Complexes and ligands tested for the hydroazidation of 4-phenylbutene 1.
induction (eq 6). In 2005, Yamada reported high diastereose-
lectivity in the hydrohydrazination reaction using Oppolzer
sultam auxiliary.39
or zeolite-supported NaN3.45c Simple hydroazidation of unac-
tivated, monosubstituted olefins have not been reported so far,
although several multifunctionalization methods have been
developed.45 Herein, a full report on the discovery and develop-
ment of the Co-catalyzed hydroazidation of olefins with sulfonyl
azide and silanes is presented.46
After the successful development of the hydrohydrazination
reaction, we first turned to the examination of other acceptors
for olefin heterofunctionalization. We were interested in new
nitrogen sources, as the transformation of the hydrazine deriva-
tives obtained in the hydrohydrazination reaction can be
troublesome in some cases. We were pleased to see that alkyl-
and arylsulfonyl azides were uniquely able to act as azide
sources in the amination of 4-phenylbutene with Co catalyst
20, giving the product 27 derived from the formal Markovnikov
addition of hydrazoic acid onto the C-C double bond exclu-
sively, albeit in moderate yields (40-60%; eq 7). Surprisingly,
no reaction was observed with Mn catalyst 23. This underscores
the differences between the hydrohydrazination and the hy-
droazidation reactions. With this lead result in hand, we
proceeded to the optimization of the hydroazidation reaction.
At this stage of development, the properties of the Co and
the Mn catalysts are complementary, 20 being more selective
and 23 more reactive. For example, the lower reactivity and
higher selectivity of 20 has allowed the development of an
efficient method to convert conjugated dienes and enynes to
allylic and propargylic hydrazines.24c
Hydroazidation Reaction. Azides occupy a privileged role
in organic chemistry as precursor of amines.40 More recently,
their use in cycloaddition reactions has been thoroughly
investigated for the synthesis of pharmaceutically relevant
heterocycles.41 Azides usually do not occur in natural systems,
yet they can display stability under physiological conditions:
these properties have led to increased interest for their use in
biochemistry, especially for bioconjugation via Staudinger
ligation42 or click chemistry.43 Azides are typically prepared
from the substitution of alkyl halides with sodium azide, but
this approach requires the prior synthesis of the alkyl halides
and is less successful for the synthesis of tertiary azides.44 The
hydroazidation of olefins from alkenes represents a more direct
access to these interesting compounds. However, such an
approach has been limited to alkenes, which give rise to
stabilized carbocations and require excess HN3,45a TMSN3,45b
We first examined simple variations of the reaction conditions
(stoichiometry of reagents, concentration, order of addition), but
no effects on yield or rate were observed. Increasing the
(45) (a) Hassner, A.; Fibiger, R.; Andisik, D. J. Org. Chem. 1984, 49, 4237.
(b) Breton, G. W.; Daus, K. A.; Kropp, P. J. J. Org. Chem. 1992, 57, 6646.
(c) Sreekumar, R.; Padmakumar, R.; Rugmini, P. Chem. Commun. 1997,
1133. For silyl enolate azidation, see: (d) Magnus, P.; Lacour, J.; Evans,
P. A.; Roe, M. B.; Hulme, C. J. Am. Chem. Soc. 1996, 118, 3406. For
olefin iodoazidation, see: (e) Hassner, A.; Levy, L. A. J. Am. Chem. Soc.
1965, 87, 4203. For diazidation, see: (f) Fristad, W. E.; Brandvold, T. A.;
Peterson, J. R.; Thompson, S. R. J. Org. Chem. 1985, 50, 3647. (g) Snider,
B. B.; Lin, H. Synth. Commun. 1998, 28, 1913. For olefin selenoazidation,
see (h) Tiecco, M.; Testaferri, L.; Sand, C.; Tomassini, C.; Marini, F.;
Bagnoli, L.; Temperini, A. Angew. Chem., Int. Ed. 2003, 42, 3131. For
conjugate addition to unsaturated carbonyl compounds, see: (i) Myers, J.
K.; Jacobsen, E. N. J. Am. Chem. Soc. 1999, 121, 38, 8959. (j) Horstmann,
T. E.; Guerin, D. J.; Miller, S. J. Angew. Chem., Int. Ed. 2000, 39, 3635.
For olefin aziridination using azides, see: (k) Mahoney, J. M.; Smith, C.
R.; Johnston, J. N. J. Am. Chem. Soc. 2005, 127, 1354.
(39) Sato, M.; Gunji, Y.; Ikeno, T.; Yamada, T. Chem. Lett. 2005, 34, 316.
(40) (a) Rao, H. S. P.; Siva, P. Synth. Commun. 1994, 24, 549. (b) Corey, E. J.;
Nicolaou, K. C.; Balanson, R. D.; Machida, Y. Synthesis 1975, 590.
(41) (a) Huisgen, R.; Knorr, R.; Mobius, L.; Szeimies, G. Chem. Ber. 1965, 98,
4014. (b) Demko, Z. P.; Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41,
2110. (c) Demko, Z. P.; Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41,
2113. (d) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596.
(42) Kohn, M.; Breinbauer, R. Angew. Chem., Int. Ed. 2004, 43, 3106.
(43) Kolb, H. C.; Sharpless, K. B. Drug DiscoVery Today 2003, 8, 1128.
(44) For general reviews on azides, see: (a) Scriven, E. F. V.; Turnbull, K.
Chem. ReV. 1988, 88, 297. (b) Bra¨se, S.; Gil, C.; Knepper, K.; Zimmermann,
V. Angew. Chem., Int. Ed. 2005, 44, 5188.
(46) Part of this work has been published in a previous communication.25
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11701

A R T I C L E S
Waser et al.
Table 7. Influence of Catalytic System on the Hydroazidation of
4-Phenylbutene 1 (Eq 8)
temperature led to a higher reaction rate, but the yield or
conversion were not improved. At this point, we re-examined
several Co-derived catalysts (Figure 2) for the hydroazidation
of 4-phenylbutene 1. Interestingly, simple Co catalysts 28-31
were also able to catalyze the hydroazidation reaction, although
the yields were lower (typically <40%). For these complexes,
the presence of a tert-butyl group on the aromatic ring was
necessary to have active catalysts. The best salen catalyst is
complex 28b, which afforded 45% yield of the desired azide.47
However, the reaction was not as clean compared to 20, and
the formation of several side products could not be suppressed.
The synthesis of catalyst 20 is well-suited for modifications,
and the influence of substituents on the aromatic ring were first
studied. However, none of the complexes 32a-g displayed
enhanced rate or conversion. Interestingly, the second ligand
on Co (X in 32, Figure 2) did not have any influence on the
reaction, and 2-amino-isobutyric acid (18) could be replaced
by other amino acids, acid 34 or pyridine. Furthermore, even
the Co(II) precursor obtained from the reaction of Co(II) salts
with amino acids and salicylaldehyde was an equally active
catalyst, although the catalyst was not soluble under the reaction
conditions. However, these catalysts did not offer the advantage
of purification and stability on the storage characteristic of 20.
We then turned to complexes 35 and 36 derived from R,R-
diphenylglycine.48 Complex 36 was the first to display increased
reactivity when compared to standard catalyst 20, and, in a first
run, azide 27 was obtained in 8 h in 58% yield with ethane-
sulfonyl azide from 4-phenylbutene 1. The structural features
of 36 are noteworthy; even in the presence of an excess of amino
acid, only the 1:1 complex from 37 and Co was detected by
NMR and mass spectroscopy in sharp contrast to complex 20.
This is probably due to the steric bulk of the corresponding
Schiff-base ligand 37 and could explain its increased reactivity
and lower stability. Unfortunately, we were unable to obtain
reproducible results using complex 36, as yields (40-70%) and
reaction times (8-48 h) showed a considerable batch depen-
dence. In many cases, a long initiation time was observed before
the reaction started. The reasons for the particular behavior of
36 are not clear, but the fact that we were able to obtain clean
NMR spectra of 36 only in DMSO could suggest that this
complex forms aggregates in other solvents, and the difference
in reactivity observed could be the result of different aggregation
states. The problem of slow initiation when using Co catalysts
had already been encountered by Mukaiyama and co-workers
for the hydration of certain olefins. In this case, the addition of
tert-butyl hydroperoxide as cocatalyst accelerated the reaction.
We were pleased to see a similar effect in the hydroazidation
reaction: when using catalyst 36 with ethanesulfonyl azide for
the hydoazidation of 4-phenylbutene 1, complete conversion was
observed after 2-8 h. Finally, we found that in situ formation
of complex 36 in the reaction mixture leads to reproducible
reaction times (2 h) and yield (70%).
conversion;a
time
run
catalyst
conditions
27:38b
1
2
3
4
36
36
36
t-BuOOH; 30%
t-BuOOH; 60%
t-BuOOH; 15%
>98%; 2 h
>98%; 1.5 h
>98%; 6 h
74%; 18 h
67:33
65:35
71:29
75:25
Co(NO3)2‚6H2O/
37/H2O2
5
Co(NO3)2‚6H2O/
37/H2O2
t-BuOOH; 30%
>98%; 2 h
74:26
6
7
8
9
Co(OAc)2/37/H2O2
CoCl2/37/H2O2
Co(NO3)2‚6H2O/37
Co(NO3)2‚6H2O/37
(1:2)
t-BuOOH; 30%
t-BuOOH; 30%
t-BuOOH; 30%
t-BuOOH; 30%
>98%; 2 h
50%; 24 h
>98%; 2 h
>98%; 2 h
71:29
ndc
72:28
68:32
10
Co(NO3)2‚6H2O/37
(2:1)
t-BuOOH; 30%
>98%; 2 h
72:28
11
12
13
14
15
16
Co(NO3)2‚6H2O/37
Co(BF4)2‚6H2O/37
20
Co(acac)2
Mn(dpm)3 (23)
Mn(OAc)2‚4H2O/37
>98%; 18 h
>98%; 2 h
78%; 2.5 h
50%; 12 h
<10%; 24 h
61%; 18 h
70:30
77:23
58:42
ndc
t-BuOOH; 30%
t-BuOOH; 30%
t-BuOOH; 30%
t-BuOOH; 30%
t-BuOOH; 30%
ndc
72:28
a Standard conditions: 0.10 mmol 4-phenylbutene 1, 0.16 mmol phe-
nylsilane, 0.30 mmol ethanesulfonyl azide, and 6 mol % catalyst in 0.50
mL of ethanol at 23 °C under argon. b Determined by gas chromatography.
c Not determined.
product (eq 8). The ratio of azidation versus reduction was
determined to be 74:26, while the conversion was higher than
95%. Combined with the isolated yield of 70% for azide 27,
this means that alkane 38 is the only side product of the reaction
present in significant amounts, and improving the azide/alkane
ratio was paramount to improving the efficiency of the process.
We began our studies toward this goal by careful examination
of the conditions for catalyst formation and the effect of the
tert-butyl hydroperoxide additive in the hydroazidation reaction
of 4-phenylbutene 1 with phenylsilane and ethanesulfonyl azide
(Table 7). From these studies, it was concluded that: (1) tert-
butyl hydroperoxide (30 mol %) was necessary to obtain useful
reaction rates (runs 1-3, 8, and 11), (2) in situ formation of
the catalyst allowed suppression of the batch dependency
observed with the isolated catalyst; the most convenient Co salt
for complex formation was Co(BF4)2‚6H2O, as it allowed nearly
instantaneous catalyst formation (runs 4-6 and 12), (3) the ratio
of ligand to Co had nearly no influence on the reaction (runs 9
and 10), and (4) other catalysts tested to assess the generality
of the activating effect of tert-butyl hydroperoxide gave inferior
results (runs 13-16).
The examination of the catalytic system allowed us to develop
a more practical and reproducible procedure for the hydroazi-
dation reaction, but we could not solve the problem of
chemoselectivity (olefin azidation vs reduction). To solve this
problem, we further examined the variation of the solvent and
sulfonyl azide (Table 8). In these studies, no solvents proved
to be superior to ethanol (runs 1-4). In contrast to these results,
modification of the sulfonyl azide was crucial to favor the
desired hydroazidation reaction against double-bond reduction.
Phenyl and toluenesulfonyl aides were optimal (runs 5-9).
To further optimize the reaction conditions, we then turned
to gas chromatography for analysis of the reaction mixture. This
allowed us to identify 4-phenylbutane (38) as the main side
(47) As 28b is chiral, the enantiomeric excess of the obtained azide 27 was
determined. However, 27 was racemic.
(48) Sterically hindered complex 36 could not be synthesized following the same
procedure as 20, as the in situ condensation to form the Schiff-base ligand
did not occur. However, preformation of the Schiff-base ligand 37, followed
by complex formation with Co(II) salts and subsequent oxidation with
hydrogen peroxide, afforded 36.
9
11702 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 10. Hydroazidation of Monosubstituted Olefins
Table 8. Influence of Solvents and Sulfonyl Azide on the
Hydroazidation of 4-Phenylbutene 1 (Eq 8)
run
azide solvent
catalyst ligand
conversion;a time
27:38b
1
EtSO2N3
EtOH
Co(BF4)2‚6H2O
37
>98%; 2 h
77:23
2
3
EtSO2N3
Co(NO3)2‚6H2O
>98%; 2 h
>98%; 18 h
>98%; 2 h
>98%; 2 h
>98%; 4 h
>98%; 18 h
<10%; 24 h
>98%; 4 h
97%; 16 h
71:29
50:50
72:28
77:23
89:11
89:11
ndd
EtOHc
37
EtSO2N3
MeOH
EtSO2N3
i-PrOH
MeSO2N3
EtOH
TolSO2N3
EtOH
MesSO2N3
EtOH
NsN3
EtOH
PhSO2N3
EtOH
TolSO2N3
EtOH
Co(NO3)2‚6H2O
37
Co(NO3)2‚6H2O
37
Co(NO3)2‚6H2O
37
Co(NO3)2‚6H2O
37
Co(NO3)2‚6H2O
37
Co(NO3)2‚6H2O
37
Co(BF4)2‚6H2O
37
20
4
5
6
7
8
9
90:10
90:10
10
a Standard conditions: 0.10 mmol 4-phenylbutene 1, 0.16 mmol phe-
nylsilane, 0.30 mmol sulfonyl azide, 30 mol % t-BuOOH, and 6 mol %
catalyst in 0.50 mL of solvent at 23 °C under argon. b Determined by gas
chromatography. c Ethanol was distilled over CaH2 and degassed before
use. d Not determined.
Table 9. Influence of Silane on the Hydroazidation of
4-Phenylbutene 1 (Eq 8)
run
sulfonyl azide
silane
conversion;a time
27:38b
1
2
3
4
5
EtSO2N3
TolSO2N3
EtSO2N3
EtSO2N3
EtSO2N3
PhSiH3, 1.6 equiv
PhSiH3, 1.6 equiv
TMDSO, 2 equiv
PMHS, 4H equiv
PMHS, 4H equiv
PhSiH3, 0.2 equiv
Et3SiH, 4 equiv
>98%; 1.5 h
>98%; 2 h
>98%; 2 h
20%; 24 h
81%; 18 h
81%; 18 h
<10%; 24 h
<10%; 24 h
>98%; 3 h
77:23
89:11
84:16
ndc
a General procedure A: 0.5 mmol alkene, 0.8 mmol PhSiH3, 1.5 mmol
TsN3, 30 mol % t-BuOOH, 6 mol % ligand 37, 6 mol % Co(BF4)2‚6H2O,
and 2.5 mL of ethanol at 23 °C under argon. b General procedure B: 1.0
mmol TMDSO was used instead of PhSiH3. c 2.0 mmol TMDSO was used.
d Starting material was partially recovered (see Supporting Information for
further details).
90:10
6
7
8
EtSO2N3
EtSO2N3
TolSO2N3
ndc
(EtO)3SiH, 4 equiv
TMDSO, 2 equiv
ndc
96:4
yields. Surprisingly, styrene derivatives (not shown) were not
reactive, although they had proven to be excellent substrates
for the hydrohydrazination reaction. The functionalization of
safrole (entry 3) is of special interest, as safrole is widely
available in bulk quantities and the amines derived from the
obtained azide are a class of biologically active compounds well-
known for their psychopharmacological activity.49 The hy-
droazidation of unprotected allylic and homoallylic alcohols was
not successful, but the introduction of a bulky silyl protecting
group allowed the conversion of these substrates (entries 4 and
5). A benzyl protecting group was less successful (entry 6). An
ester (entry 7) and a ketone (entry 8) were tolerated with
excellent chemoselectivity, as no CdO reduction was observed.
Di- and trisubstituted olefins also undergo reaction in the
hydroazidation reaction (Table 11). R,R-Disubstituted olefins
(entries 1-3) give good yields of the tertiary azides, which are
difficult to synthesize via substitution of halides. For these
substrates, competitive reduction of the alkene was less pro-
nounced and, thus, the more reactive phenylsilane gives higher
yields. Cyclooctene also reacts, but the yield is moderate (entry
4). Finally, trisubstituted olefins (entries 5 and 6) were examined.
In these cases, full conversion of the starting material could
not be achieved. Nevertheless, useful yields could be obtained
with phenylsilane as reductant.
a Standard conditions: 0.10 mmol 4-phenylbutene 1, 0.30 mmol sulfonyl
azide, 30 mol % t-BuOOH, 6 mol % Co(BF4)2‚6H2O, and 6 mol % ligand
37 in 0.50 mL of ethanol at 23 °C under argon. b Determined by gas
chromatography. c Not determined.
Interestingly, the hydrohydrazination catalyst 20 could also be
used with tosyl azide (run 10) with good selectivity, although
the reaction was slower.
Finally, we examined the effect of varying the silane structure
on reaction rate and selectivity for the hydroazidation of
4-phenylbutene with Co(BF4)2‚6H2O and ligand 37 as catalyst
(Table 9). It was found that TMDSO offers the best compromise
between selectivity and reactivity, and combining tosyl azide
and TMDSO gave full conversion of 4-phenylbutene 1 in 3 h,
with an improved azide/alkane ratio of 96:4 (run 8). Using the
conditions of runs 2 and 8 for preparative scale (0.50 mmol)
hydroazidation of 4-phenylbutene 1, the desired azide 27 was
obtained in 90 and 86% isolated yield, respectively.
Next, we proceeded to examine the scope of the hydroazi-
dation reaction for monosubstituted olefins, both with phenyl-
silane and TMDSO (general procedure A and B, Table 10). The
reaction displayed very good Markovnikov selectivity, and only
the secondary azides were isolated. The yields with phenylsilane
or TMDSO are comparable, except in the case for which the
competitive reduction is more pronounced (entries 4 and 5).
Apart from 4-phenylbutene 1 (entry 1), 4-naphthylbutene (entry
2) and safrole (entry 3) also give the desired azide in useful
(49) Nichols, D. E.; Hoffman, A. J.; Oberlender, R. A.; Jacob, P.; Shulgin, A.
T. J. Med. Chem. 1986, 29, 2009.
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11703

A R T I C L E S
Waser et al.
Table 11. Hydroazidation of Di- and Trisubstituted Olefins
Figure 3. Azides tested as nitrogen sources for the hydroazidation of
4-phenylbutene 1.51
the Co catalyst, we examined arylsulfonyl azides functionalized
at the ortho position (Figure 3, compounds 44-47). The first
result obtained with azide 44 bearing a free OH group was
encouraging, as full conversion was observed after 2 h. More
importantly, no precipitation of the catalyst occurred. The
synthesis of 44 was not optimal, however, mainly because the
intermediate sulfonyl chloride is very unstable and could not
be purified. Unfortunately, these impurities could not be
removed from azide 44. Azide 45 could be synthesized with
high purity, but it gives only low conversion in the hydroazi-
dation reaction of 4-phenylbutene 1. We then turned to azide
46 bearing a methoxy group instead of the free OH. We were
pleased to see that this azide is also a good azide source, giving
full conversion in 4 h without precipitation of the catalyst. Azide
46 is an easy to handle solid and is obtained in two steps from
4-methoxy toluene without the need of chromatography.50 To
rule out a simple electron-donating effect of the methoxy group,
azide 47 bearing an electron-withdrawing ester group was next
examined. Azide 47 gives nearly identical results to 46, although
the addition of methylene chloride as cosolvent for the hy-
droazidation reaction was necessary due to the low solubility
of this reagent in ethanol. We have no experimental-based
explanation for the success of 46 and 47 in stabilizing the
catalytic system, and further studies are needed to better
understand this effect. Interestingly, the reaction mixture in the
hydroazidation reaction of 4-phenylbutene 1 with TMDSO and
azides 46 and 47 had a different color than that observed for
tosyl azide (from brown-green to dark green). That could suggest
that indeed an interaction between the sulfonyl azide and the
Co catalyst is occurring during the reaction.
Substrates that could be functionalized only in moderate yields
with tosyl azide were examined next (Table 12). For 4-phenyl-
butene 1, it was possible to use only 1.5 equiv of azide 46 or
47 to obtain more than 90% yield of the desired azide 27 (entry
1). When the reaction was performed with only 1.5 equiv of
tosyl azide, 27 was obtained in 70% yield only, and the reaction
failed to go to full conversion. The use of 46 or 47 gave no
improvement for the hydroazidation of monosubstituted ethers
(entries 2 and 3), but the yields were significantly improved
for R-methyl disubstituted allylic ether (entries 4 and 5), leading
to nearly complete conversion with half as much sulfonyl azide.
Finally, a trisubstituted olefin could also be functionalized in
good yield (entry 6).
a General procedure A: 0.5 mmol alkene, 0.8 mmol PhSiH3, 1.5 mmol
TsN3, 30 mol % t-BuOOH, 6 mol % ligand 37, 6 mol % Co(BF4)2‚6H2O,
and 2.5 mL of ethanol at 23 °C under argon. b General procedure B: 1.0
mmol TMDSO was used instead of PhSiH3. c 2.0 mmol TMDSO was used.
d Starting material was partially recovered (see Supporting Information for
further details).
When considering the scope of the hydroazidation reaction,
two major limitations are apparent: substrates bearing a
stabilizing group (ester, phenyl, alkyne) in conjugation with the
olefin do not react, and full conversion could not be obtained
with sterically more-hindered substrates (di- and trisubstituted
olefins). Even for monosubstituted olefins, a large excess (3
equiv) of sulfonyl azides is needed to achieve full conversion.
Careful monitoring of the reaction with tosyl azide showed that
in the case of sterically hindered substrates the problem does
not reside in activity but in the stability of the catalyst. After
about a 2-5 h reaction, precipitation of the catalyst occurs and
no further conversion is observed. Unfortunately, the exact
identity of the precipitate could not be assigned, but one
explanation would be the formation of insoluble oligomeric Co-
sulfinate salts.
The reaction conditions are similar to those for the hydro-
hydrazination reaction, and we hypothesize that the same active
Co-alkyl intermediates should be generated in the two reac-
tions. Consequently, the solution to the limitations of the
hydroazidation reaction resides in the modification of the azide
source to obtain a more active nitrogen-transfer reagent, less
prone to deactivate the catalyst. For this reason, several other
potential azide sources were examined for the hydroazidation
of 4-phenylbutene 1 (Figure 3).
Phosphorus-based azide reagents 39 and 40 could not be used
as azide sources. The chiral sulfonyl azide 41 also showed no
conversion. Electron-rich sulfonyl azide 42 and commercially
available sulfonyl azide 43 react in the hydroazidation reaction
of 4-phenylbutene 1, but full conversion was not observed, as
precipitation of the catalyst occurs already after 1 h. To study
potential favorable interactions between the sulfonyl azide and
(50) Chlorosulfonylation, followed by reaction with sodium azide, see Supporting
Information for further details.
9
11704 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 12. Comparison of TsN3, Azide 46, and Azide 47 in the
Hydroazidation Reaction
a Standard conditions: 0.50 mmol 4-phenylbutene 1, 1.5 mmol TsN3,
1.0 mmol TMDSO, 30 mol % t-BuOOH, 6 mol % Co(BF4)2‚6H2O, and 6
mol % ligand 37 in 2.5 mL of ethanol at 23 °C under argon. b Sulfonyl
azide 46 (0.75 mmol) and 0.75 mmol TMDSO were used. c Sulfonyl 47
(0.75 mmol) and 0.75 mmol TMDSO were used and 1.0 mL of methylene
chloride was added as cosolvent.
Substrates with the C-C double bond in conjugation with
stabilizing groups, such as an ester, a phenyl, or an alkyne, still
showed no conversion to the desired azides under the improved
reaction conditions. This shows the possibility and limitations
of the newly introduced azide sources 46 and 47: these reagents
give much better yield for substrates that react slowly but cleanly
in the hydroazidation reaction (especially geminally disubstituted
and trisubstituted alkenes), but they are not well-suited for
substrates prone to a side reaction (such as monosubstituted
allylic ethers) or olefins that have shown no conversion at all
when using tosyl azide.
Mechanistic Investigations. To further develop the hydro-
hydrazination and hydroazidation reactions, a better understand-
ing of the reaction mechanism is needed. When compared with
the classical hydroamination reaction, the mechanistic studies
of these reactions are complicated by the presence of both an
oxidizing nitrogen source and a reducing hydride source. In the
case of the hydroazidation, the tert-butyl hydroperoxide additive
adds a level of complexity to the problem. Consequently, the
primary goal of our investigations was a qualitative understand-
ing of the catalytic cycle, focusing mostly on the hydrohydrazi-
nation reaction. This will facilitate the design of further
experiments.
Figure 4. 1H NMR monitoring of the hydrohydrazination reaction using a
stoichiometric amount of Co catalyst 20. A: Only catalyst 20. B:
Immediately after addition of PhSiH3. C: After warming to 40 °C over 20
min. D: Back to 23 °C, 2 min after addition of alkene 1. E: 30 min after
addition of 2.
visible changes in the signals of 20 when the reaction was
1
monitored by H NMR spectroscopy. However, when a sto-
ichiometric amount of phenylsilane was added to 20, the
1
observed H NMR spectrum changed completely and a new
complex was detected (B, Figure 4). Unfortunately, this was
accompanied by a broadening of all signals, precluding any
precise structure assignment. This is likely due to a partial
reduction of Co(III) to Co(II). Interestingly, a new broad signal
appeared between -0.5 and 1 ppm (X in Figure 4). This signal
was first thought to indicate the presence of a Co hydride
complex. With the hope of obtaining better-resolved signals,
the reaction mixture was heated slowly to 40 °C (C, Figure 4).
Under these conditions, signal X was shifted to about -0.5 ppm
and was sharper, but this was accompanied by a broadening of
all other signals. Furthermore, a control experiment using
phenylsilane-d3 showed no change in shape or intensity for
signal X, and this peak most probably results from a paramag-
netic Co(II) complex and not from a hydride signal. No changes
were observed when 4-phenylbutene 1 was added (D, Figure
4). The addition of azodicarboxylate 2 resulted in a fast
conversion of 2 into hydrohydrazination product 3 and hydrazine
4. In contrast to the reaction with catalytic amount of 20,
Our first attempt to get further information about potential
intermediates in the reaction was to perform the hydrohydrazi-
nation reaction with a stoichiometric amount of Co complex
20, with the hope of detecting some intermediate complexes
during the course of the reaction (Figure 4). Addition of a
stoichiometric amount of 4-phenylbutene 1 or azodicarboxylate
2 to a solution of complex 20 in methanol-d4 resulted in no
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11705

A R T I C L E S
Waser et al.
Scheme 3. Possible Pathways for the Hydrohydrazination of
Scheme 2. Hydrohydrazination of 4-Phenylbutene 1 and Indene
49 with PhSiD3 and Azodicarboxylate 2
Cyclooctadiene (52)
hydrazine 4 was the major product (4:3 ) 4:1). The reaction
was accompanied by a shift of the signals corresponding to the
Co complex, and after 30 min, several sets of sharp peaks similar
to the starting complex 20 were observed, showing again the
presence of Co(III) complexes (E, Figure 4).
As we were not successful in identifying reaction intermedi-
ates via 1H NMR spectroscopy, we turned to deuterium-labeling
experiments. As a first experiment, the hydrohydrazination of
4-phenylbutene 1 was run under standard conditions, but using
PhSiD3 instead of PhSiH3 (Scheme 2, A). The reaction displayed
clean and complete deuterium incorporation at the primary
1
position, as shown by H and 13C NMR spectroscopy. This
suggests that the silane acts as a hydride source in these
reactions, such that a pathway involving protonolysis of a Co-
alkyl complex with ethanol is unlikely. Finally, a kinetic isotope
effect on the reaction rate of 2.2(3) ( 0.5 was observed. This
moderate primary isotope effect suggests that the formation of
the C-H bond in the product is rate-determining. Next, the
hydrohydrazination reaction of indene (49) using PhSiD3 was
examined to see if the formation of the product was stereospe-
cific (Scheme 2, B). This proved not to be the case, as the
product was obtained as a 1:1 mixture of diastereoisomers, as
shown by 1H NMR spectroscopy. This result is expected if the
formed Co-alkyl complex is not stable and undergoes epimer-
ization prior to reaction with azodicarboxylate 2, if the reaction
proceeds from a dissociated free radical, or if the amination is
not stereospecific.
As in the case of the hydrohydrazination reaction, the
hydroazidation of 4-phenylbutene with phenylsilane-d3 resulted
in the exclusive formation of azide 51 bearing a deuterium atom
at the primary position (eq 9). Interestingly, the kinetic isotope
effect displayed by the hydroazidation reaction (1.65(15) ( 0.32)
is smaller than that for the hydrohydrazination reaction (2.2(3)
( 0.5).
mechanism, products 53a, 53b, 53c, or a mixture thereof are
expected. In path A, a Co-D complex is first formed, and
insertion of one of the double bonds of cyclooctadiene (52) gives
the regioisomeric Co-allyl intermediates I or III. From I, direct
addition to azodicarboxylate 2 gives product 53a, whereas
reaction via an ene pathway gives product 53b.52 Alternatively,
complex I could be in fast equilibrium with the isomeric
complex II. Intermediate II could also react directly to give
53b or via ene reaction to give 53a. Co intermediate III would
be expected to lead to the exclusive formation of homoallylic
hydrazine 53c. In path B, a Co-hydrazido complex formed from
azodicarboxylate 2, and PhSiD3 reacts via insertion in the C-N
bond to give regioisomeric intermediates IV or VI. Subsequent
reaction of complex IV with PhSiD3 is expected to lead to
homoallylic hydrazine 53c or to allylic hydrazine 53b. Once
again, a fast equilibrium between complex IV and V can be
envisaged, V can then further react to give 53b or 53c.
Intermediate VI is anticipated to lead to 53a through reaction
with PhSiD3, and no other products are expected.
When the hydrohydrazination reaction of 52 was run under
standard conditions with PhSiD3, a 1:1 mixture of products 53a
and 53b was obtained in 52% combined yield. Homoallylic
hydrazine 53c was never observed so that the pathways leading
to this compound are not operative. Path B is unlikely, as in
this case formation of hydrazines 53a, 53b, or 53c only, or a
mixture of 53b and 53c, are the most likely results.53 However,
a 1:1 mixture of 53a and 53b is consistent with the operation
of path A, in the case of a fast equilibrium of the two allyl
From these first labeling experiments, it was clear that an
insertion of the double bond in a Co-D complex with the Co
atom ending at the primary or homobenzylic position cannot
account for the reaction outcome, whereas an insertion placing
the Co atom at the secondary or benzylic position and
subsequent reaction with 2 leads to the observed products.
However, another alternative cannot be excluded: insertion of
the double bond in the C-N bond of a Co-hydrazido complex,
placing the Co atom at the primary position, followed by
reductive elimination. To gain more information about this step
of the reaction, we decided to examine the hydrohydrazination
of cyclooctadiene using PhSiD3 (Scheme 3). Depending on the
(51) The conversions using 2 equiv of azide are given.
(52) Alternatively, dissociation of a free radical from the Co complex prior to
amination can be envisaged.
(53) In fact, the only possibility to obtain the observed mixture via pathway B
would be a nonregioselective insertion of 30 in the Co-N bond and the
exclusive formation of 31b from either IV or V.
9
11706 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Table 13. Radical Clocks Examined in the Hydrohydrazination
Reaction
the major product was the nonrearranged hydrohydrazination
product for both Co and Mn catalysts. Ring-opening products
were also obtained, but, due to multiamination, several products
were formed, and they could not be separated. Nevertheless,
the ratio of rearranged/unrearranged products can be estimated
roughly to be 1:2-1:3. This leads to a reaction rate of about
1.5-2.5 × 108 s-1 mol-1 for the amination step, with no
significant difference between the Co and the Mn catalysts.
Next, we decided to examine another type of radical clock
based on the 5-exo-trig cyclization of dienes (entries 3 to 5).55c
The hexenyl radical derived from heptadiene cyclizes quite
slowly (1 × 105 s-1 to the cis product and 3 × 104 s-1 to the
trans product).57 In accord with these reported data, the main
products resulting from the hydrohydrazination reaction of
heptadiene were not cyclized monoaminated and diaminated
products, but small amounts of the cyclization product could
be isolated. The rate for the amination step can be estimated
again to be 1.5-2.0 × 108 s-1 mol-1 for the Co catalyst and
2.0-2.5 × 108 s-1 mol-1 for the Mn catalyst. Although the
values obtained are only rough estimations,58 the fact that two
systems based on different mechanisms gave the same value is
noteworthy. This enhances the probability that the reaction coud
proceed via free radicals or at least Co-alkyl radical pairs. If
another reactive intermediate was involved, the reaction rates
would have not been expected to display exactly the same
behavior as known for free radical reactions in two structurally
distinct substrates. However, a rearrangement or cyclization “on
Co” could still be envisaged.
Finally, in the case of dienes more prone to cyclization,
preparatively useful yields of the tandem cyclization/hydrohy-
drazination reaction can be obtained (entries 4 and 5). For diallyl
ether,59 the two diastereisomers could be separated, and the
major product was found to be the cis-isomer by NOE 1H NMR
(entry 4). Again, this is in accordance with the well-known
preference of free radicals for the formation of cis products in
the 5-exo-trig cyclization.60 In these reactions, the Mn catalyst
23 usually displays higher yield and selectivity, especially in
the case of diallyl malonate (entry 5), for which the cyclization/
hydrohydrazination product was obtained in 94% yield and 9:1
diastereoselectivity.
a Standard conditions: 0.50 mmol alkene, 0.50 mmol PhSiH3, 0.75 mmol
di-tert-butyl azodicarboxylate (2), 5 mol % catalyst 20, and 2.5 mL of
ethanol at 23 °C under argon. b Standard conditions: 0.50 mmol alkene,
0.50 mmol PhSiH3, 0.75 mmol di-tert-butyl azodicarboxylate (2), 2 mol %
catalyst 23, and 2.5 mL of i-PrOH at 0 °C under argon.
complexes I and II, followed by reaction via either ene or direct
pathway. The same result would be obtained if I reacts via ene
and direct pathways at the same rate.
Deuterium-labeling experiments have allowed us to better
understand the transfer of the H atom to the double bond, but
nearly no information on the crucial amination step was
obtained. One commonly proposed mechanism for the Co-
catalyzed Mukaiyama hydration reaction involves carbon radi-
cals.54 To test the presence of radical intermediates, we made
use of radical clocks based on cyclopropane ring opening or
cyclization.55
Next, we proceeded to examine the kinetics of the hydrohy-
drazination and hydroazidation reactions. We decided to limit
our kinetic investigations to the Co-catalyzed hydrohydrazination
and hydroazidation of 4-phenylbutene 1. The hydrohydrazination
reaction could be monitored by 1H NMR spectroscopy, as
running the reaction in methanol-d4 gave only minor differences
in rate when compared with the standard conditions (ethanol).
The hydroazidation reaction was monitored by gas chromatog-
raphy.
We first examined the dependence of the reaction rate on
the initial concentration of 4-phenylbutene 1. When the con-
centration of 1 was systematically increased, the rate of the
reaction also increased for both reactions. An approximation
of the reaction order was obtained from a van’t Hoff plot on
the initial rate. In the case of the hydrohydrazination reaction,
We first tested vinylcyclopropane itself (entry 1, Table 13).
The secondary carbinyl cyclopropyl radical is known to undergo
rearrangement with a rate of 4 × 107 s-1 55c
.
Usually, competition
experiments are run with a large excess of the radical trap, but
it was difficult in our case due to the limited solubility of
azodicarboxylate 2. For these reasons, we preferred to use our
standard reaction conditions, as the obtained rates can be
mathematically corrected.55c Vinylcyclopropane gave hydrazine
derivative resulting from the opening of the cyclopropane ring
as the only isolated product and, as such, set a higher limit for
the reaction rate of the intermolecular amination reaction (entry
1). We then turned to R-phenyl vinylcyclopropane, as the
stabilizing effect of the phenyl group slows down the rear-
rangement for this compound (4 × 105 s-1; entry 2).56 In fact,
(57) Lusztyk, J.; Maillard, B.; Deycard, S.; Lindsay, D. A.; Ingold, K. U. J.
Org. Chem. 1987, 52, 3509.
(54) Tokuyasu, T.; Kunikawa, S.; Masuyama, A.; Nojima, M. Org. Lett. 2002,
4, 3595.
(58) For more exact values, more data would be needed, for example through
a series of reactions at different concentrations.
(55) (a) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317. (b) Nonhebel,
D. C. Chem. Soc. ReV. 1993, 22, 347. (c) Newcomb, M. Tetrahedron 1993,
49, 1151.
(59) Burkhard, P.; Roduner, E.; Hochmann, J.; Fischer, H. J. Phys. Chem. 1984,
88, 773.
(56) Bullock, R. M.; Samsel, E. G. J. Am. Chem. Soc. 1987, 109, 6542.
(60) Spellmeyer, D. C.; Houk, K. N. J. Org. Chem. 1987, 52, 959.
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11707

A R T I C L E S
Waser et al.
this gives an order of 0.98(16) ( 0.23, which means that the
reaction is first order in the substrate and suggests that the alkene
is involved in the rate-determining step of the reaction in
accordance with the results of the deuterium-labeling studies.
For the hydroazidation reaction, the calculated reaction order
was only 0.45(3) ( 0.07, suggesting that several steps contribute
equally to the reaction rate.
In the case of the nitrogen source, azodicarboxylate 2 or tosyl
azide, the rate of the reactions was completely independent of
the initial concentrations. The van’t Hoff plot gives an order of
-0.18(8) ( 0.1261 and 0.06(4) ( 0.26, respectively, which
correspond to a zero order for 2 or tosyl azide. This suggests
that the amination step is faster than the rate-determining step
and, as such, does not influence the apparent reaction rate.
For phenylsilane and the Co catalyst 20, the situation is not
so clear, and we obtained mixed-order results. The reaction rate
shows only a very weak dependence from the initial concentra-
tion of silane when enough silane (>0.8 equiv) is present.
However, at low concentration, the silane begins to play a more
pronounced role on the reaction rate. As the high concentrations
are more relevant to our standard conditions, the order of the
reaction in phenylsilane was determined only for concentrations
higher than 0.14 M (0.8 equiv), and we obtained an order of
0.22(11) ( 0.16 for the hydrohydrazination reaction and 0.33-
(3) ( 0.1 for the hydroazidation reaction. In the case of the
hydrohydrazination with Co catalyst 20, the data obtained were
of lower quality and the order of 0.54(10) ( 0.22 is difficult to
interpret. The influence of the catalyst was stronger in the case
of the hydroazidation reaction (order of 0.82(5) ( 0.11 with a
1:1 ligand-to-Co ratio). Changing the ligand-to-Co ratio gives
nearly no change in rate. In the case of the peroxide, we
observed a good linearity for the van’t Hoff plot in a concentra-
tion range of 0.014 to 0.2 M (7 to 100%) with a reaction order
of 0.643(14) ( 0.020. This indicates that the peroxide is not
only important as an initiator, but also during the reaction. Up
to now, we have still no explanation for this very strong effect.
The fact that the hydrohydrazination reaction needed no additive
to attain a useful rate suggests that the effect of the peroxide
occurs after hydrocobaltation of the double bond. This effect
could be due to activation of the Co species such that it is more
prone to react with the azide or by helping the regeneration of
a Co-hydride complex at the end of the catalytic cycle. A
(partially) rate-determining catalyst regeneration step would
further be consistent with the lower kinetic isotope effect and
the mixed order in alkene observed for the hydroazidation
reaction.
Figure 5. Arrhenius plot for the hydrohydrazination reaction.
Figure 6. Arrhenius plot for the hydroazidation reaction.
Scheme 4. Working Model for the Mechanism of the
Hydrohydrazination Reaction
reactions: 76(7) ( 15 kJ‚mol-1 for the hydrohydrazination and
60(3) ( 6 kJ‚mol-1 for the hydroazidation reaction. Interest-
ingly, these values are lower than the activation barrier
calculated by Togni and co-workers for the direct hydroami-
nation reaction via olefin activation using Ni catalysts (108
kJ‚mol-1), which was predicted to be the lowest of all the metals
examined.62
The results of our preliminary mechanistic investigations led
us to propose the working model shown in Scheme 4 for the
mechanism of the hydrohydrazination reaction. The first step
of the reaction is the initiation of the catalytic cycle via
formation of active Co-hydride complex I. Already at this
stage, there are a number of questions about the identity of I.
As we were not successful in isolating or detecting a clean Co-
(III)-hydride complex, its presence in the catalytic cycle
As a result of the complexity of the reactions, including
probably an initiation step for the generation of a Co-hydride
complex, mixed orders are not unexpected. Furthermore, the
kinetics are complicated by the presence of the direct reduction
of azodicarboxylate in the case of the hydroydrazination reaction
and the strong effect of the peroxide in the hydroazidation
reaction. For these reasons, although the data obtained in these
studies were statistically significant, their interpretation in terms
of mechanistic relevance is difficult.
The temperature dependence of the reactions was examined
next (Figure 5 and Figure 6). The Arrhenius plot for the reaction
allowed us to determine the energy of activation for the
(61) The low correlation (R ) -0.412) observed for the data further confirms
that the reaction rate is indeed independent of the concentration of 2.
(62) Senn, H. M.; Blochl, P. E.; Togni, A. J. Am. Chem. Soc. 2000, 122, 4098.
9
11708 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Scheme 5. Working Model for the Mechanism of the Hydroazidation Reaction
remains speculative. In the case of the Co-catalyzed Mukaiyama
The scarce mechanistic information obtained so far for the
peroxidation reaction, the generation of a hydride complex is
thought to proceed via prior activation of Co(II) with oxygen.54
In our case, the silane seems able to reduce at least partially
the catalyst, as shown by our NMR experiments, but it is
difficult to know if the catalyst is further activated by the
azodicarboxylate. It would be interesting in the future to
synthesize Co-hydrazido complexes to see if they are also
active catalysts.
From I, hydrocobaltation of the double bond led to Co-alkyl
complex II. The insertion of the double bond in the Co-H bond
placing the Co atom at the secondary center is in accordance
with our deuterium-labeling experiments. Furthermore, both the
reaction order in alkene substrate and the significant kinetic
isotope effect observed for the reaction suggest that this step is
rate-determining, whereas the following amination step is fast.
This explains why we were never able to observe Co-alkyl
complex II. As adding the alkene to a stoichiometric mixture
of Co complex 20 and PhSiH3 leads to no reaction, two
possibilities can be envisaged: either the insertion of the double
bond is reversible and the equilibrium lies in the favor of I, or
the azodicarboxylate is indeed needed to generate an active
catalyst. However, we have never observed an isomerization
of the C-C double bond in the olefins tested, making reversible
olefin insertion less probable.
hydroazidation reaction does not allow the proposal of a precise
mechanism for the reaction. However, a speculative working
model, which is in accordance with our observations, is
presented in Scheme 5. We hypothesize a similar entry in the
catalytic cycle via a Co-hydride complex I. For this step, the
tert-butyl hydroperoxide additive could already play an impor-
tant role. Hydrocobaltation would then give a Co-alkyl complex
II. At this point, we can only speculate at what happens next.
Again, two pathways can be envisaged: a free radical pathway
(A) or direct reaction of Co-akyl complex II with sulfonyl azide
(B). Free radicals have been reported to react with sulfonyl azide
by Renaud and co-workers.18b,c However, this reaction was run
at elevated temperature. The radical adduct formed in the
addition to the sulfonyl azide (as a result of addition on the
terminal or internal NdN bond of the sulfonyl azide) could
either be recaptured by a Co(II) complex to give a new Co
complex III, in analogy to the Co-hydrazido complex III
proposed for the hydrohydrazination reaction, or collapse
directly to the azide and a sulfonyl radical. A direct reaction of
the Co-alkyl complex II with the sulfonyl azide would also
give III. In contrast to the hydrohydrazination reaction, the
amination step is much more substrate-dependent: if the R group
on the alkene has a stabilizing effect (for example with phenyl,
ester, or alkyne), no reaction is observed.
From II, the crucial amination step is difficult to examine,
as it is very fast. The reaction of radical clocks, both via
rearrangement and via cyclization strongly suggest the presence
of free radicals. However, this is not a proof that the generated
radical also adds on azodicarboxylate 2 (path A). One could
envision that radical generation is an unproductive pathway,
and the reaction in fact proceeds from the Co-bound intermediate
(path B). In this aspect, it will be interesting in the future to
generate free radicals via an independent method and test their
addition on azodicarboxylate 2.
After the reaction with 2, the generated Co-hydrazido
complex could react with the silane to regenerate an active Co-
hydride complex I. We did not observe silylated hydrazine
derivatives, but instead, product 54, resulting from the etha-
nolysis of phenylsilane, was isolated. The strong effect of protic
solvents on the reaction rate is probably effective in this step,
and ethanol must play a role in the regeneration of complex I
and not be simply limited to the alcoholysis of silylated
hydrazines.
The fate of the catalyst after the azidation step is uncertain:
A Co-sulfonato complex IV could be formed directly from
III via elimination or after recombination with a Co(II) complex,
but the formation of IV is speculative and Co-hydride complex
I could also be regenerated from III directly. It is also possible
that IV could be the inactive precipitate observed in the reaction
(vide supra). As the order of the reaction in the sulfonyl azide
was zero, a possible explanation for the strong additive effect
of the peroxide is the acceleration of the conversion of III or
IV to Co-hydride complex I, allowing useful turnovers. The
end product resulting from the silane has yet to be identified,
but silane derivatives of polymeric nature were often detected
in the crude mixture. A possible structure would be a mixed
silane/sulfonate such as 55. Obviously, further efforts are needed
to identify the inactive precipitate as well as the end products
resulting from the silane and the sulfonyl azides.
The fact that Mn catalyst 23 was more active for the
hydrohydrazination reaction but failed completely in the case
of the hydroazidation reaction is also intriguing. It could be
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11709

A R T I C L E S
Waser et al.
envisaged that Mn-alkyl complexes similar to II are unable to
react with the sulfonyl azide, but in this case, a common free-
radical manifold for the two catalysts would be excluded.
Another possibility is that catalyst 23 is too unstable under the
reaction conditions, and deactivation occurs immediately after
one turnover. In fact, we observed that Mn catalysts bearing
the more robust tridentate ligand 37 indeed display some activity
in the hydroazidation reaction.
However, the initially developed procedure using TsN3 is not
convenient for reactions conducted on a larger scale because a
large excess of reagents (3 equiv of TsN3 and 2 equiv of
TMDSO) is needed to achieve full conversion. To limit the
amount of waste, we made use of azide 46 for the reaction of
4-phenylbutene 1 on a 5 mmol scale (eq 11). We were pleased
to see that in this case the use of only 1.2 equiv of 46 and 1
equiv of TMDSO were sufficient to afford the desired amine
57 in 68% overall yield.
Transformation of Azide and Hydrazine Products. Hy-
drazines and azides are useful building blocks by themselves,
but the final products in pharmaceutical and natural products
seldom contain these functionalities.63 A fast and convenient
access to the free amine is highly desirable. At this stage of
development, we focused on the transformation of a few chosen
substrates to test and optimize known literature procedures.
There are several reports for the conversion of Boc-protected
hydrazines to the free amines.64 The most frequently used
methods are hydrogenation reactions over metal catalysts. The
procedure involving Raney Ni is inconvenient because it requires
high pressure hydrogen (>20 bar).64a Pt-based catalysts are more
active,64c-e but they also promoted the hydrogenation of
aromatic rings of the substrates in our experience. For these
reasons, we turned to the reduction of the N-N bond of the
hydrazines using Zn dust.65 This method is very mild but
requires prior removal of the Boc protecting groups. One
possibility is to perform this deprotection in trifluoroacetic acid
(TFA), but as the Zn reduction did not work in TFA, removal
of the highly corrosive TFA was needed. However, it was
possible to devise a one-pot procedure for the transformation
of hydrazine 3, the product of the hydrohydrazination of
4-phenylbutene 1, to the free amine hydrochloride 56 (eq 10).
Deprotection occurred in acetic acid when refluxing for 12 h.
Addition of acetone resulted in the formation of the hydrazone,
which was subsequently reduced with Zn dust.66 The crude
amine was treated with aqueous HCl and water was removed,
and pure amine hydrochloride 56 could be isolated in 51% yield
after recrystallization. This represents a 43% overall yield for
the formal addition of ammonia on the double bond of
4-phenylbutene 1.
Small, geometrically constrained amines such as methylcy-
clobutylamine and methylcyclopropylamine are of interest as
building blocks in medicinal chemistry in the optimization of
the activity of pharmaceuticals.67a-c The influence of the ring
structure on the basicity of the amine is also of theoretical
interest.67d However, these compounds are not easy to prepare,
and published procedures made use of multistep sequences based
on the Ritter reaction68 or the Hofmann rearrangement.69 Our
methods offer the possibility to access these compounds directly
from commercially available methylenecyclobutane (58) and
methylenecyclopropane (59). The amination of methylene-
cyclopropanes mediated by lanthanide catalysts has been
reported, but opening of the cyclopropyl ring was observed in
this reaction.12b We decided to further demonstrate the strengths
and limitations of the hydrohydrazination and hydroazidation
reactions in the synthesis of these interesting amines.
We first examined the reaction of methylenecyclobutane (58;
Scheme 6). The Co-catalyzed hydrohydrazination reaction of
58 gives 90% yield of the desired tertiary hydrazine derivative
60 after column chromatography (path A, Scheme 6). The Mn-
catalyzed hydrohydrazination reaction could also be used, but
in this case, a small amount of the primary hydrazine derivative
was also obtained and could not be separated from the main
product. At this point, removal of the Boc group of 60 in boiling
acetic acid did not proceed in good yield. Clean removal of the
Boc groups was achieved in a 1:1 mixture of 8 M HCl and
THF in 1 h at room temperature. After subsequent reduction
with Zn dust in acetone/acetic acid, formation of the amine
hydrochloride salt and recrystallization, 62 was obtained in 84%
yield together with a small amount of 63 resulting from a
reductive amination of 62 with acetone. Unfortunately, 62 and
63 were difficult to separate via recrystallization, and higher
purity of 62 could be obtained only with great substance loss.
Nevertheless, this reaction sequence represents a formal addition
of ammonia on methylenecyclobutane 58 with 75% overall
yield.
In the case of azides, the reduction to the free amine is
possible under much milder conditions.40 Furthermore, the
cycloaddition reaction of azides with terminal alkynes increases
the versatility of these compounds.41 The mild conditions of
the olefin hydroazidation reaction had allowed us to develop a
one-pot conversion of olefins to amines using in situ reduction40a
and to 1,4-disubstituted triazoles using Sharpless’ procedure41b
for 4-phenylbutene 1 in useful yields.25
The hydroazidation reaction of 58 was examined next (path
B, Scheme 6). Hydroazidation of 58 gives the volatile and
(63) The wide-spread use of AZT in the treatment of AIDS constitutes a
noteworthy exception.
(64) (a) Evans, D. A.; Britton, T. C.; Dorow, R. L.; Dellaria, J. F. J. Am. Chem.
Soc. 1986, 108, 6395. (b) Evans, D. A.; Britton, T. C.; Dorow, R. L.;
Dellaria, J. F. Tetrahedron 1988, 44, 5525. (c) Gennari, C.; Colombo, L.;
Bertolini, G. J. Am. Chem. Soc. 1986, 108, 6394. (d) Oppolzer, W.; Moretti,
R. HelV. Chim. Acta 1986, 69, 1923. (e) Oppolzer, W.; Moretti, R.
Tetrahedron 1988, 44, 5541. (f) Trimble, L. A.; Vederas, J. C. J. Am. Chem.
Soc. 1986, 108, 6397.
(67) (a) Castagnoli, N.; Rimoldi, J. M.; Bloomquist, J.; Castagnoli, K. P. Chem.
Res. Toxicol. 1997, 10, 924. (b) Burgess, K.; Ke, C. Y. J. Org. Chem.
1996, 61, 8627. (c) Jimenez, A. I.; Cativiela, C.; Aubry, A.; Marraud, M.
J. Am. Chem. Soc. 1998, 120, 9452. (d) McDonald, R. N.; Reitz, R. R. J.
Am. Chem. Soc. 1976, 98, 8144.
(68) Cox, E. F.; Silver, M. S.; Roberts, J. D.; Caserio, M. C. J. Am. Chem. Soc.
1961, 83, 2719.
(69) Vaidyanathan, G.; Wilson, J. W. J. Org. Chem. 1989, 54, 1815.
(65) Leblanc, Y.; Fitzsimmons, B. J. Tetrahedron Lett. 1989, 30, 2889.
(66) It is important to use a good quality of freshly activated (over diluted HCl)
Zn dust to obtain good yields.
9
11710 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006

Hydrohydrazination and Hydroazidation of Olefins
A R T I C L E S
Scheme 6. Comparison of the Hydrohydrazination and Hydroazidation Reaction for the Amination of Methylenecyclobutane (58)
Scheme 7. Hydrohydrazination Reaction for the Amination of Methylenecyclopropane (59)
potentially explosive azide 61, which was directly reduced in
situ. When we used the standard reduction conditions with
sodium borohydride and a Cu catalyst, a significant amount of
ring-opening product 64 was also obtained (path B1). Fortu-
nately, this problem could be circumvented by changing to the
Lindlar catalyst using hydrogen as reductant (path B2).40b In
this case, the methylcyclobutylamine hydrochloride 62 was
obtained in 60% overall yield with high purity after recrystal-
lization. The overall yield is slightly lower for the hydroazidation
reaction when compared with the hydrohydrazination reaction,
but amine hydrochloride 62 was obtained in higher purity, and
all reactions of the sequence were run in one pot without the
need of column chromatography for product purification.
Finally, the combination of the hydroazidation reaction with
Sharpless’ click chemistry was also successful and gives the
triazole 65 in 53% overall yield after column chromatography
(path B3).
In the case of methylenecyclopropane (59), the compound is
too volatile to run the reaction at room temperature (bp ) 9
°C). In this case, the Mn-catalyzed hydrohydrazination reaction
proved to be the best procedure, as this reaction readily proceeds
at 0 °C to give the desired hydrazine product 66 in 77% yield
after column chromatography (Scheme 7). Deprotection with
HCl followed by reduction with Zn dust and isolation as amine
hydrochloride proceeded smoothly to give 67 in 72% yield after
recrystallization. In this case, the reductive amination with
acetone occurred in 2% yield only, and pure 67 could be
obtained after a second recrystallization. Overall, the formal
Markovnikov amination of methylenecyclopropane with am-
monia has been achieved in 55% yield. Importantly, no products
resulting from ring opening have been detected.
corresponds to a direct hydroamination of C-C double bonds
with protected hydrazines but is based on a fundamentally
different concept in which the H and the N atoms come from
two different reagents, a silane and an azodicarboxylate. The
reaction is characterized by its ease of use: commercially
available silanes and azodicarboxylates as reagents, easily
synthesized and bench-stable Co and Mn catalysts 20 and 23,
alcohol solvents of commercially available purity, and opera-
tionaly simple procedures. Moreover, in contrast to many
classical hydroamination methods, the reaction readily proceeds
at room temperature under neutral conditions. The scope of the
reaction is very broad for an amination reaction, including
mono-, di-, tri-, and tetrasubstituted olefins, activated and
nonactivated olefins, dienes, enynes, and heterocycles.
Based on the same principles, we then described the develop-
ment of the Co-catalyzed hydroazidation of olefins. This reaction
allows a direct access to secondary and tertiary alkyl azides,
complementing existing methods using substitution or hydrazoic
acid addition. Alkyl azides are versatile intermediates, easy to
reduce to the free amine or to convert to triazoles and tetrazoles
via cycloaddition reactions. Key to the successful development
of the hydroazidation reaction was the use of sulfonyl azides
as nitrogen sources and the discovery of the activating effect
of tert-butyl hydroperoxide. After careful optimization, the
reaction was found to be efficient for the functionalization of
mono-, di-, and trisubstituted olefins. The scope is more limited
than that of the hydrohydrazination reaction, however, as several
functional groups were not well-tolerated and no azide products
were isolated for olefins in conjugation with a stabilizing group
(phenyl, ester, or alkyne).
The conversion of the hydrazine and azide products obtained
into the free amines was further demonstrated. In the case of
the hydrohydrazination reaction, this transformation was achieved
via acid-promoted deprotection and reduction with Zn dust. For
the hydroazidation reaction, it was possible to develop a one-
Conclusion
We have described the development of the Co- and Mn-
catalyzed hydrohydrazination reaction of olefins. This reaction
9
J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006 11711

A R T I C L E S
Waser et al.
pot hydroazidation/reduction protocol to obtain directly the free
amine. This corresponds formally to the direct hydroamination
of olefins with ammonia. Alternatively, a one-pot procedure
allowing direct access to triazoles was also disclosed.
Finally, our preliminary mechanistic investigations have been
presented. On the basis of NMR, deuterium-labeling studies,
radical clocks, and kinetic measurements, a first speculative
image of the catalytic cycle was proposed. However, many
important questions remain to be elucidated.
In summary, we have reported the first successful use of an
“oxido-reductive” approach for the hydroamination of unacti-
vated olefins. This allows a new direct access to hydrazines
and azides from olefins in high yields under operationally simple
reaction conditions. The methodology is still in its infancy, and
future work will be dedicated to overcome the limitations of
the scope observed for the hydroazidation reaction, to the
functionalization of more complex substrates en route to natural
products, bioconjugates, or pharmaceuticals, and to the identi-
fication of more convenient nitrogen sources, which do not have
the safety concern associated with azodicarboxylates and
sulfonyl azides. Finally, the development of asymmetric hy-
drohydrazination and hydroazidation reactions will be pursued,
as the access to chiral amines is highly desirable.
Acknowledgment. We thank the Swiss National Foundation
and INIT (ETH) for support of this research and the Roche
Research Foundation for the support of J.W. H.N. thanks the
Japan Society for the Promotion of Science (JSPS) for a
postdoctoral fellowship.
Supporting Information Available: Experimental procedures,
kinetics, and spectral data for all products (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA062355+
9
11712 J. AM. CHEM. SOC. VOL. 128, NO. 35, 2006