Intramolecular Fe(II)-Catalyzed N-O or N-N bond formation from aryl azides

Article abstract of DOI:10.1021/ol101040p

(Figure presented) Iron(II) bromide catalyzes the transformation of aryl and vinyl azides with ketone or methyl oxime substituents into 2,1-benzisoxazoles, indazoles, or pyrazoles through the formation of an N-O or N-N bond. This transformation tolerates a variety of different functional groups to facilitate access to a range of benzisoxazoles or indazoles. The unreactivity of the Z-methyloxime indicates that N-heterocycle formation occurs through a nucleophilic attack of the ketone or oxime onto an activated planar iron azide complex.

Full text of DOI:10.1021/ol101040p

ORGANIC
LETTERS
2010
Vol. 12, No. 12
2884-2887
Intramolecular Fe(II)-Catalyzed N-O or
N-N Bond Formation from Aryl Azides
Benjamin J. Stokes, Carl V. Vogel, Linda K. Urnezis, Minjie Pan, and
Tom G. Driver*
Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois, 60607-7061
tgd@uic.edu
Received May 6, 2010
ABSTRACT
Iron(II) bromide catalyzes the transformation of aryl and vinyl azides with ketone or methyl oxime substituents into 2,1-benzisoxazoles, indazoles,
or pyrazoles through the formation of an N-O or N-N bond. This transformation tolerates a variety of different functional groups to facilitate
access to a range of benzisoxazoles or indazoles. The unreactivity of the Z-methyloxime indicates that N-heterocycle formation occurs through
a nucleophilic attack of the ketone or oxime onto an activated planar iron azide complex.
Despite the prevalence of nitrogen-heteroatom bonds in
biologically active N-heterocycles, methods that form these
bonds remain rare. 2,1-Benzisoxazoles1,2 or indazoles3,4
typically originate from starting materials with pre-existing
N-O or N-N bonds such as oximes, nitrile oxides, or
hydrazines. Direct construction of the N-O5 or N-N6 bond
provides an attractive alternative approach to these N-
heterocycles. Surprisingly, there are no prior examples of
transition-metal-catalyzed routes to these bonds even though
such complexes promote the oxidation7,8 or imination9 of
sulfides. Since we established that benzimidazoles could be
produced from N-aryl imines with ortho-azido substituents
upon exposure to FeBr2 (Scheme 1),10 we were curious if
nitrogen-heteroatom bond formation could be achieved by
transposing the ortho-heteroatom substituent from the R-po-
sition to the ꢀ-position. While thermolysis of 3 affords the
new nitrogen-heteroatom bond,5a,b,6a-c the high temperature
(6) (a) Krbechek, L.; Takimoto, H. J. Org. Chem. 1964, 29, 1150. (b)
Cadogan, J. I. G.; Cameron-Wood, M.; Mackie, R. K.; Searle, R. J. G.
J. Chem. Soc. 1965, 4831. (c) Takada, K.; Thoe, K. W.; Boulton, A. J. J.
Org. Chem. 1982, 47, 4323. (d) Counceller, C. M.; Eichman, C. C.; Wray,
B. C.; Stambuli, J. P. Org. Lett. 2008, 10, 1021.
(1) Recent reviews, see: (a) Elguero, J. ComprehensiVe Heterocyclic
Chemistry; Katrizky, A. R., Rees, C. W., Eds.; Pergamon Press: New York,
1984; Vol. 4, p 167. (b) Stadlbauer, W. Sci. Synth. 2002, 12, 243. (c)
Smalley, R. K. Sci. Synth. 2002, 11, 337
.
(2) (a) Davis, R. B.; Pizzini, L. C. J. Org. Chem. 1960, 25, 1884. (b)
Musso, H.; Schro¨der, H. Chem. Ber. 1965, 98, 1562. (c) Wro´bel, Z. Synthesis
1997, 753. (d) Boduszek, B.; Halama, A.; Zon, J. Tetrahedron 1997, 53,
(7) Reviews on asymmetric sulfoxidation: (a) Bolm, C.; Mun˜iz, K.;
Hildebrand, J. P. ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N.,
Pfaltz, A., Yamamoto, H., Eds.; Springer: Berlin, 1999; p 697. (b) Kagan,
H. B. Catalytic Asymmetric Synthesis; Ogima, I., Ed.; Wiley-VCH: New
York, 2000; p 327. (c) Bolm, C. Coord. Chem. ReV. 2003, 237, 245.
(8) For recent reports, see: (a) (Fe) Legros, J.; Bolm, C. Angew. Chem.,
Int. Ed. 2004, 43, 4225. (b) (V) Drago, C.; Caggiano, L.; Jackson, R. F. W.
Angew. Chem., Int. Ed. 2005, 44, 7221. (c) (Al) Yamaguchi, T.; Matsumoto,
K.; Saito, B.; Katsuki, T. Angew. Chem., Int. Ed. 2007, 46, 4729. (d) (Fe)
Egami, H.; Katsuki, T. J. Am. Chem. Soc. 2007, 129, 8940.
11399
(3) (a) Jacobsen, P.; Huber, L. Ber. Chem. Ges. 1908, 41, 660. (b)
Ru¨chardt, C.; Hassmann, V. Liebigs Ann. Chem. 1980, 908
.
.
(4) For recent examples, see: (a) Song, J. J.; Yee, Y. K. Org. Lett. 2000,
2, 519. (b) Lukin, K.; Hsu, M. C.; Fernando, D.; Leanna, M. R. J. Org.
Chem. 2006, 71, 8166. (c) Vina, D.; del Olmo, E.; Lopez-Perez, J. L.; San
Feliciano, A. Org. Lett. 2007, 9, 525. (d) Jin, T.; Yamamoto, Y. Angew.
Chem., Int. Ed. 2007, 46, 3323. (e) Wu, C.; Fang, Y.; Larock, R. C.; Shi,
(9) (a) (Cu) Takada, H.; Ohe, K.; Uemura, S. Angew. Chem., Int. Ed.
1999, 38, 1288. (b) (Fe) Bach, T.; Ko¨rber, C. J. Org. Chem. 2000, 65,
2358. (c) (Ru) Murakami, M.; Uchida, T.; Katsuki, T. Tetrahedron Lett.
2001, 42, 7071. (d) (Ru) Tamura, Y.; Uchida, T.; Katsuki, T. Tetrahedron
Lett. 2003, 44, 3301. (e) (Rh) Okamura, H.; Bolm, C. Org. Lett. 2004, 6,
1305. (f) Manchen˜o, O. G.; Bolm, C. Chem.sEur. J. 2007, 13, 6674.
(10) Shen, M.; Driver, T. G. Org. Lett. 2008, 10, 3367.
F. Org. Lett. 2010, 12, 2234
.
(5) (a) Smith, P. A. S.; Brown, B. B.; Putney, R. K.; Reinisch, P. F.
J. Am. Chem. Soc. 1953, 75, 6335. (b) Boulton, A. J.; Ghosh, P. B.;
Katritzky, A. R. Angew. Chem., Int. Ed. Engl. 1964, 3, 693. (c) Prakash,
O.; Saini, R. K.; Singh, S. P.; Varma, R. S. Tetrahedron Lett. 1997, 38,
3147.
10.1021/ol101040p  2010 American Chemical Society
Published on Web 05/27/2010

significantly reduced temperatures. Screening other metal
complexes known to decompose azidessincluding Cu(I) and
Cu(II) saltssdid not result in nitrogen-oxygen bond forma-
tion.13 Our data show that the oxidation state of iron affected
the efficiency of the reaction because lower conversions were
observed when more Lewis acidic iron(III) salts were used.
If the reaction temperature was increased to 40 °C, the
catalyst loading could be reduced to 5 mol %. The low
catalyst loading contrasted with benzimidazole formation,
which required 30 mol % FeBr2 for satisfactory conver-
sions.10 Molecular sieves, while not required, were added
to eliminate any adventitious water.
Scheme 1. Iron(II) Bromide-Catalyzed N-Heterocycle Formation
With these optimized conditions, the scope and limitations
of iron(II)-catalyzed N-O bond formation were examined
(Table 2).14 2,1-Benzisoxazole formation depended on the
limits the functional group tolerance of the reaction. Herein,
we report that iron(II) bromide catalyzes the formation of
N-O or N-N bonds to transform azides 3 into 2,1-
benzisoxazoles 4 or pyrazoles 5 under markedly benign
conditions.
Table 2. Scope of Fe(II)-Catalyzed N-O Bond Formation
To achieve transition-metal-catalyzed nitrogen-heteroatom
bond formation, the reactivity of 2-azidobenzophenone
toward a variety of transition metal complexes was inves-
tigated (Table 1). This aryl azide is available in one step
entry
6
R1
R2
R3
yield (%)a
98
89
57
1
2
3
4
5
6
7
8
a
b
c
d
e
f
g
h
i
j
k
l
m
n
H
H
H
H
H
Cl
NO2
H
Ph
Ph
Ph
Ph
Ph
Ph
Me
OMe
CF3
H
H
H
H
H
H
H
Table 1. Optimization of Metal-Catalyzed N-O Bond
Formation
trace
90
dec.b,c
78
4-ClC6H4
4-BrC6H4
2-FC6H4
2-ClC6H4
Me
i-Pr
c-C6H11
H
H
86
91
97
9
Cl
Cl
H
H
H
entry
catalyst
none
Rh2(O2CC3F7)4
CuCl
Cu(OTf)2
RuCl3·H2O
ZnI2
Fe(acac)2
FeCl3
FeCl2
FeBr2
FeBr2
FeBr2
mol %
temp (°C)
yield (%)a
92
trace
no reaction
trace
trace
no reaction
8
20
97
97
99
90
10
11
12
13
14
1b
2
3
n.a.
5
120
25
25
25
25
25
25
25
25
25
40
40
83
H
H
H
72b,d
74b
10
10
10
10
10
10
10
10
5
H
no reaction
4
a After SiO2 chromatography. b 10 mol % of FeBr2 used. c 10% aniline
formed. d Isomerization to the 1H-benzisoxazole tautomer occurred upon
SiO2 purification.
5b
6
7
8
9
electronic nature of the azide. While the reaction tolerated a
range of R1- or R2-substituents, strong electron-withdrawing
groups are incompatible with the reaction conditions (entries
1-7). The effect of the R3-substituent on the reaction was
briefly surveyed. 2,1-Benzisoxazoles were made from sub-
strates bearing aryl or alkyl R3-groups (entries 8-14). Our
data suggest that a carbon R3-substituent is necessary to
achieve N-O bond formation because aldehydes, such as
6n, did not react. Varying the electronic nature of o-
azidobenzaldehyde 6n (R1 ) OMe or CF3) did not result in
benzisoxazole formation.
10
11
12b
a
5
As determined using H NMR spectroscopy. b No molecular sieves
1
were added.
from commercially available 2-aminobenzophenone.11 While
a plethora of methods have been reported to generate an
electrophilic nitrogen atom from aryl azides,12 we found that
only iron salts promoted 2,1-benzisoxazole 7a formation at
(11) Many 2-amino-substituted aryl ketones are commercially available.
Alternatively, the substrates were synthesized from 2-aminobenzaldehyde
derivatives. See the Supporting Information for the details of their syn-
thesis.
(12) For reviews, see: (a) Katsuki, T. Chem. Lett. 2005, 34, 1304. (b)
Bra¨se, S.; Gil, C.; Knepper, K.; Zimmermann, V. Angew. Chem., Int. Ed.
2005, 44, 5188. (c) Cenini, S.; Gallo, E.; Caselli, A.; Ragaini, F.; Fantauzzi,
S.; Piangiolino, C. Coord. Chem. ReV. 2006, 250, 1234.
(13) The use of other transition metal complexes resulted in attenuated
conversions. Please refer to the Supporting Information for more details.
(14) To a mixture of aryl azide 6 (0.1 mmol), 100% w/w of crushed 4
Å molecular sieves, and metal salt (5-10 mol %) in a conical vial with
Teflon septum was added 0.250 mL of solvent. The resulting mixture was
heated, and after 16 h, the heterogenous mixture was filtered through SiO2.
The filtrate was concentrated in vacuo. Purification by MPLC afforded the
desired product.
Org. Lett., Vol. 12, No. 12, 2010
2885

To determine if our method could be used to form N-N
bonds, aryl azide 8, bearing a ꢀ-nitrogen atom in the tether,
was screened (eq 1). In addition to testing the possibility of
N-N bond formation, this substrate would also allow an
examination of the stereochemical requirements of the
reaction. When a mixture of 8 (25:75, E:Z) was exposed to
FeBr2, the E-isomer reacted quantitatively to form 2H-
indazole 9, while Z-8 was entirely recovered from the
reaction mixture.15 Isomerization of Z-8 was never observed,
even when the reaction conditions were substantially al-
tered.16
bearing either electron-donating or electron-withdrawing R1-
and R2-substituents were cleanly transformed into 11 (entries
1-6). The reaction also tolerated aryl-, alkyl-, and hydrogen
R3-substituents (entries 7-11). As with 7m, indazole 11h,
which contained a secondary alkyl substituent, isomerized
to the 1H-aromatic product upon exposure to SiO2 (entry
9). The identity of the imine nitrogen substituent affected
the reaction outcome. While methyl ketoximes were com-
petent substrates, no reaction was observed for ketoximes.
The reaction was not limited to aryl azide substrates:
trisubstituted pyrazoles 11m-11p could be accessed from
vinyl azides upon exposure to 5 mol % of FeBr2 (entries
12-16). Vinyl azides 10m-10p are readily synthesized from
the cycloalkanone via a three-step procedure (eq 2).18
Iron(II)-catalyzed N-N bond formation from E-methyl
oximes 10 exhibited a broader substrate scope than our
benzisoxazole study (Table 3).17 Biaryl ketoximes (10a-10f)
The reactivity of diazides 10q and 10r showcases the
mildness of our Fe(II) process (Scheme 2). Whereas ther-
Table 3. Scope of Fe(II)-Catalyzed N-N Bond Formation
Scheme 2. Comparison of the Reactivity of Diazides 10
molysis of either diazide produced a thick tar containing trace
amounts of indazole (<5%) and aldoximine, exposure of
diazide 10q or 10r to FeBr2 cleanly afforded the 4- or
6-azido-substituted indazole. In contrast to the thermal
process, the second azide group remained intact in our Fe(II)-
catalyzed process.
N-Heterocycles 7 or 11 can be elaborated into other
synthetically valuable products (Scheme 3). SmI2-mediated
reduction of the N-O bond in benzisoxazole 7a in the
presence of acetophenone or acetone formed quinolines 12a
and 12b.19 Palladium-catalyzed hydrogenolysis of the N-OMe
bond present in 11q and 11p produced 1H-indazole 13 and
(15) Z-Oximes were reported to be unreactive in the metal-free thermal
reaction. See ref 6c.
(16) For a complete description of the isomerization conditions screened,
refer to the Supporting Information.
(17) E-10 (or a mixture of E-, Z-10), 100% w/w of crushed 4 Å
molecular sieves, and metal salt (5-10 mol %) were added to a conical
vial, which was sealed with a PTFE septum, followed by the addition of
CH2Cl2 as solvent. The resulting mixture was heated to 40 °C for 16 h.
The heterogeneous mixture was purified by MPLC to afford the product.
(18) Tabyaoui, B.; Aubert, T.; Farnier, M.; Guilard, R. Synth. Commun.
1988, 18, 1475. Refer to the Supporting Information for the synthesis of
10m-10p.
a After SiO2 chromatography. b Reaction performed using an E-/Z-
mixture of 10. c The remainder of the reaction mixture was unreacted 10.
d 10 mol % of FeBr2 used. e Isomerization to the 1H-benzisoxazole tautomer
occurred upon SiO2 purification.
(19) For a related SmI2-mediated pyridine synthesis, see: Fan, X.; Zhang,
X.; Zhang, Y. Heteroatom Chem. 2005, 16, 637.
2886
Org. Lett., Vol. 12, No. 12, 2010

Scheme 3
.
Functionalization of N-Heterocyclic Products
Scheme 4. Potential Mechanism for 2H-Indazole Formation
the oxime on the activated azide to form the N-N bond.22,25
Extrusion of N2 from 16 followed by dissociation of the iron
catalyst would provide the 2H-indazole product. The origin
of the Z-isomer’s inertia can be inferred from examination
of the iron azide complex 18: in addition to the destabilizing
steric interactions produced upon planarization of 18, the lone
pair of electrons on the oxime are oriented in the wrong
direction for N-N bond formation.
In conclusion, we have demonstrated that iron(II) bromide
readily forms N-O or N-N bonds from a range of ortho-
azide-substituted aryl ketones or E-methyl oximes. Future
experiments will determine if a contiguous π-system is
required for Fe(II)-catalyzed azide decomposition and if any
of the reactive intermediates can be trapped by dipolarophiles
to produce additional C-C or C-N bonds.
2H-pyrazole 14.20 The vanilloid receptor (TRPV1) antago-
nist, ABT-102,21 contains 1H-indazole 13. Abbott Pharma-
ceuticals developed ABT-102 as a clinical candidate for the
treatment of chronic pain.
While several mechanisms could account for N-heterocycle
formation,22 the reactivity differences between the E- and
Z-oxime isomers of 10 suggest that N-X bond formation
occurs through a planar reactive intermediate (Scheme 4).23
Because the corresponding aniline was unreactive, we assign
this reactive intermediate to be iron(II)-azide complex 15.
This intermediate could be produced from E-10 by coordina-
tion of the iron(II) catalyst to the terminal N-atom of the
azide.24 Planarization then triggers a nucleophilic attack of
(20) Lower yields of 1H-indazoles were obtained when alternative
reduction protocols were attempted.
(21) (a) Drizin, I.; Gomtsyan, A.; Bayburt, E. K.; Schmidt, R. G.; Zheng,
G. Z.; Perner, R. J.; DiDomenico, S.; Koenig, J. R.; Turner, S. C.; Jinkerson,
T. K.; Brown, B. S.; Keddy, R. G.; McDonald, H. A.; Honore, P.; Wismer,
C. T.; Marsh, K. C.; Wetter, J. M.; Polakowski, J. S.; Segreti, J. A.; Jarvis,
M. F.; Faltynek, C. R.; Lee, C.-H. Bioorg. Med. Chem. 2006, 14, 4740. (b)
Lukin, K.; Hsu, M. C.; Chambournier, G.; Kotecki, B.; Venkatramani, C. J.;
Leanna, M. R. Org. Process Res. DeV. 2007, 11, 578.
Acknowledgment. We are grateful to the National Insti-
tutes of Health NIGMS (R01GM084945), Petroleum Re-
search Fund administered by the American Chemical Society
(46850-G), and the University of Illinois at Chicago for their
generous support. We also thank Prof. L. Anderson (UIC)
for insightful discussions and Mr. Furong Sun (UIUC) for
mass spectrometry data.
(22) (a) The lack of aniline byproducts suggests that an alternative
mechanism involving the formation of iron nitrenoid intermediates is not
occurring. For leading references on iron-mediated nitrene transfer, see:
(a) ref 9b. (b) Bacci, J. P.; Greenman, K. L.; Van Vranken, D. L. J. Org.
Chem. 2003, 68, 4955. (c) Cowley, R. E.; Eckert, N. A.; Elhaik, J.; Holland,
P. L. Chem. Commun. 2009, 1760.
Supporting Information Available: Complete experi-
mental procedures and spectroscopic and analytical data for
the products. This material is available free of charge via
the Internet at http://pubs.acs.org.
(23) For leading mechanistic studies of the thermolysis reaction, see:
(a) Smith, P. A. S.; Budde, G. F.; Chou, S.-S. P. J. Org. Chem. 1985, 50,
2062. (b) Dyall, L. K.; Karpa, G. J. Aust. J. Chem. 1988, 41, 1231. (c)
Dyall, L. K.; Holmes, A. L. Aust. J. Chem. 1988, 41, 1677.
(24) For crystal structures of metal azide complexes, see: (a) (Nγ-
coordination) Proulx, G. R.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117,
6382. (b) (Nγ-coordination) Fickes, M. G.; Davis, W. M.; Cummins, C. C.
J. Am. Chem. Soc. 1995, 117, 6384. (c) (NR- and Nγ-coordination) Dias,
H. V. R.; Polach, S. A.; Goh, S.-K.; Archibong, E. F.; Marynick, D. S.
Inorg. Chem. 2000, 39, 3894. (d) (η2-coordination) Waterman, R.; Hillhouse,
G. L. J. Am. Chem. Soc. 2008, 130, 12628.
OL101040P
(25) For an alternative mechanism of N-N bond formation through an
intramolecular 1,3-dipolar cyclization, see: Hall, J. H.; Behr, F. E.; Reed,
R. L. J. Am. Chem. Soc. 1972, 94, 4952
.
Org. Lett., Vol. 12, No. 12, 2010
2887