Synthesis and reactions of highly strained 2,3-bridged 2H-azirines

Article abstract of DOI:10.1002/anie.200600483

Take the strain: Despite extreme ring strain and reactivity, the heterocycles 2, which are easily obtained from azides 1 and detectable by NMR spectroscopy, can be converted stereoselectively by novel addition and cycloaddition reactions. (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200600483

Angewandte
Chemie
Strained Molecules
DOI: 10.1002/anie.200600483
Synthesis and Reactions of Highly Strained
2,3-Bridged 2H-Azirines**
Klaus Banert* and Barbara Meier
Strained compounds are of special interest because of their
increased energy content and the enhanced reactivity that
frequently results from this. The already considerable ring
strain of simple cyclopropenes (ca. 54 kcalmolꢀ1) is signifi-
cantly increased if the three-membered ring is incorporated
into the bicyclic framework 1 (Scheme 1).[1] It should be
possible to transfer this concept from the hydrocarbons 1 to
the heterocycles 3.[2,3] The 2,3-bridged azirine 3d, which is
easily obtainable from azide 2d by photolysis or thermolysis,
can be distilled in vacuo,[4] whereas the more unstable
compound 4 could be characterized only in solution.[5] Until
now, the even more strained azirine 3b and many similar
bridged azirines bearing a six-membered ring can be gener-
ated only in situ and detected by trapping reactions.[4a,6] Thus,
[*] Prof. Dr. K. Banert, Dr. B. Meier
Lehrstuhl für Organische Chemie
Technische Universität Chemnitz
Strasse der Nationen 62, 09111 Chemnitz (Germany)
Fax: (+49)371-531-1839
E-mail: klaus.banert@chemie.tu-chemnitz.de
[**] Reactions of Unsaturated Azides, Part 19. This work was supported
by the Fonds der Chemischen Industrie. We thank Dr. M. Hagedorn,
Dr. F. Köhler, and Dr. J. Lehmann for support with spectroscopic
investigations. Part 18: J. R. Fotsing, M. Hagedorn, K. Banert,
Tetrahedron 2005, 61, 8904–8909.
Supporting information for thisarticle isavailable on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 4015 –4019
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4015

Communications
Scheme 1.
the prospect of spectroscopically characterizing 3b in solution
1
appears to be gloomy, especially as the H NMR data of the
analogous short-lived hydrocarbon 1b were obtained even at
ꢀ908C only partially.[7]
We nowreport the heterocycles 3b, 3c, 3e, 3 f, and 3h–j
(Table 2), which have been analyzed spectroscopically for the
first time, as well as some further 2,3-bridged azirines. The
highly strained title compounds undergo several addition and
cycloaddition reactions that are not possible for simple 2H-
azirines.[2]
To produce the corresponding azirines, we prepared first
the necessary vinyl azides[8] 2c,[9] 2d,[4b,c] and 2e[9] by using
Hassnerꢀs method,[10] that is, by electrophilic addition of
iodine azide, generated in situ, to cycloalkenes followed by
base-induced elimination of hydrogen iodide. In the case of
five- and six-membered rings, it is well-known[4b,c,11–13] that this
method does not lead to vinyl azides, but rather to allyl azides.
Thus, we switched over to Zbiralꢀs[11b] sequence for the
synthesis of 2a and 2b. Neither route was successful in the
synthesis of 1-azido-2-methylcycloalkenes such as 2 f, 2g, and
2h (Scheme 2). The dehydration with thionyl chloride of a
mixture of the regioisomeric azidoalcohols 6[14] and 7,[14]
which could be easily synthesized from the inexpensive
limonene oxide 5, was problematic and led only to a very
lowyield of 2 f. However, the formal addition of iodine azide
to 1-methylcycloalkenes 8g,h assisted by cerium(IV) ammo-
nium nitrate (CAN)[15] and the subsequent treatment of 9g,h
with potassium tert-butylate gave somewhat better yields of
the desired vinyl azides 2g,h.
When a solution of the azide 2b in anhydrous CDCl3 or
[D8]toluene was irradiated at ꢀ508C with a mercury high-
pressure lamp, surprisingly the bridgehead azirine 3b could
be detected even at room temperature by its IR as well as
1H NMR and 13C NMR data (Table 1).[16] However, a rapid
subsequent reaction led to the formation of the dimer 11b,
and this product was quickly oxidized to 12b when oxygen
was not excluded rigorously (Table 2).[17] Addition of tetra-
chloro-1,4-benzoquinone caused instantaneous transforma-
tion of 11b to 12b. If photolysis of 2b is performed in CDCl3
or [D8]toluene saturated with water, the dimerization 3b!
11b is accelerated so strongly that the intermediate 3b can no
longer be observed by NMR spectroscopy. We assume that
even traces of water act as a catalyst and initiate the reaction
sequence shown in Scheme 3. The less strained bridgehead
azirine 3c dimerizes considerably slower than 3b (3 months,
Scheme 2.
Table 1: Selected physical data of 2,3-bridged 2H-azirines 3b and 3c.[a]
3b: IR (CDCl3): n˜ =1743 cmꢀ1 (C N); H-NMR (CDCl3, ꢀ408C): d=1.01
1
=
(ddddd, 2J=14.0, 3J=9.9, 8.9, 4.7, 2.6 Hz, 1H, 4-Hendo), 1.28 (ddddd,
2J=14.7, 3J=5.5, 4.7, J=1.1, J=0.6 Hz, 1H, 5-Hendo), 1.41 (ddddd,
2J=14.0, 3J=7.3, 5.5, 5.0, 2.6 Hz, 1H, 4-Hexo), 1.54 (dddddd, 2J=12.6,
7.3, 6.9, 5.0, 2.6, 4J=1.1 Hz, 1H, 3-Hendo), 1.64 (dddd, 2J=14.7, 3J=8.9,
5.0, 3.8, 1H, 5-Hexo), 1.82 (ddddd, 2J=12.6, 3J=9.9, 7.6, 7.1, 2.6 Hz, 1H,
3-Hexo), 2.28 (dd, 3J=3.8, 0.6 Hz, 1H, 6-H), 2.84 (ddd, 2J=12.7, 3J=7.1,
6.9 Hz, 1H, 2-Hexo), 3.24 ppm (ddd, 2J=12.7, 3J=7.6, 5.0 Hz, 1H, 2-
4
3
H
endo); 13C-NMR (CDCl3, ꢀ408C): d=21.39 (t, C-4), 23.85 (t, C-3), 25.96
(d, 1JC,H ꢁ 192 Hz, C-6), 27.27 (t, C-2), 28.33 (t, C-5), 179.04 ppm (s, C-
1).
3c: Colorless liquid; IR (CDCl3): n˜ =1764 cmꢀ1 (C N); H NMR (CDCl3):
1
=
d=0.87 (m, 1H), 1.27 (m, 1H), 1.41 (m, 1H), 1.54 (m, 1H), 1.66–1.76
(m, 3H), 2.07 (m, 1H), 2.13 (m, 1H), 2.79 (m, 1H, 2-H), 3.00 ppm (m,
1H, 2-H); 13C NMR (CDCl3): d=26.37 (t), 27.46 (t), 27.82 (t), 28.02 (t),
28.96 (d, 1JC,H =183 Hz C-7), 31.53 (t), 175.35 ppm (s, C-1).
[a] The data of the other bridgehead azirinesand those of other new
compoundsare summarized in the Supporting Information. 1H NMR:
300 MHz, 13C NMR: 75 MHz. The assignment of NMR signals and the
measurement of coupling constants were performed with the help of
1H,1H double-resonance experiments, 1H NMR NOE difference spectra,
1H,1H COSY experiments, 13C,1H shift correlations, DEPT135, GATED,
and 2D J-resolved experiments, as well as spectrum simulation in the
case of 3b. The vicinal coupling constants found for 3b correlate with the
H-C-C-H torsion angles, which were calculated semiempirically
(MOPAC), with the Karplusrelationship.
4016 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4015 –4019

Angewandte
Chemie
Table 2: Generation and successive reactions of 2,3-bridged 2H-azirines 3.
11g, respectively (Table 2), and we
were able to isolate the nitrile 15k
as well as the azo compound 16k
after photolysis of 2k and workup
by chromatography (Scheme 4).
The diradical intermediates shown
in Scheme 4 could possibly play a
role in the formation of 15k and
16k; thus, the difficulties in detect-
ing azirine 3k by NMR spectros-
copy may result not only from its
short lifetime.
As a result of their increased
ring strain, 2,3-bridged azirines
such as 3b, 3c, and 3h can undergo
addition and cycloaddition reac-
tions that are not possible for
Substrate 2
R1
Yield[a] of 3 [%]
after photolysis
Yield[a] of 3 [%]
after thermolysis
Yield
of 11 [%] of 12 [%]
Yield[a]
n
R2
X
2a
2b
2c
5
6
7
H
H
H
H
H
H
H
H
CH2
CH2
CH2
CH2
0 (CDCl2F, ꢀ1208C)
83 (CDCl3, ꢀ508C)
91 (CDCl3, ꢀ508C)
100 (CDCl3, ꢀ508C)
94[f] (CDCl3, ꢀ508C)
0 (CD2Cl2, ꢀ808C)
85 (CDCl3, ꢀ508C)
n.i.[b] n.d.[b]
45[c]
49[d]
54[d]
0
0 (CDCl3, 808C)
85 (CDCl3, 808C)
96 (CDCl3, 808C)
n.m.[b]
n.d.
0
2e[e] 12
2 f
2g
2h
2i
6
5
6
6
6
CMe=CH2 Me CH2
n.i.
n.i.
–
H
H
H
H
Me CH2
Me CH2
H
n.i. 65[c,g,h]
n.i. 52[a,d,i]
–
–
C=O 22[j] (CD2Cl2, ꢀ808C)
n.i.
n.i.
0
0
0
–
2j
Me C=O 65 (CDCl3, ꢀ508C)
[a] Yield based on 1H NMR standard. [b] n.i.=reaction not investigated, n.d.=compound not detected
because of rapid oxidation to 12, n.m.=yield not measured. [c] Yield based on azide 2. [d] Yield based
on azirine 3. [e] Mixture of (Z)-2e/(E)-2e=2.5:1. [f] 1:1 mixture of diastereomers. [g] Yield based on
isolated products. [h] Mixture of diastereomers (60:40). [i] Mixture of diastereomers (ca. 5:1). [j] 47% 2i
wasrecovered; prolonged irradiation led to destruction of 3i.
simple
2H-azirines.
Whereas
alkyl- or aryl-substituted azirines
do not react with cyclopentadie-
ne,[2b] the corresponding Diels–
Alder reactions of 3b and 3c
Scheme 3.
208C, 54% yield of 12c). Moreover, we were not able to
observe the corresponding reaction of 3e. The azirines 3i,j
generated from azides 2i,j[6f,g] did not form pyrazines
analogous to 11 and 12, but decomposed by unknown
subsequent reactions.
The significantly increased ring strain for 3b relative to 3c
is demonstrated by the 1J(13C,1H) NMR coupling constants of
ꢀ
the C H unit at the bridgehead (Scheme 1). Consequently, 3b
is distinctly more reactive than the azirine 3c, which can also
be observed, similarly to 3e, after thermolysis in solution
(Table 2) and which can be isolated after flash vacuum
pyrolysis (3008C, 2 ” 10ꢀ4 Torr) in 78% yield. In contrast, it
Scheme 4.
was not possible to remove the solvent from a solution of
photochemically generated 3b without complete destruction.
Furthermore, no trace of 3b could be detected by monitoring
the thermolysis of a solution of 2b in chloroform or
cyclohexane by NMR spectroscopy,[16] whereas flash vacuum
pyrolysis led to lowyields of 3b. Comparison of the properties
of 3b with those of 3h and of 3i with those of 3j shows that an
additional methyl group at the bridgehead increases the
kinetical stability of the azirines significantly.
Our attempts to detect azirines bridged by a five-
membered ring, such as 3a, 3g, or 3k, by spectroscopy
directly led at best to strongly broadened NMR signals of the
starting materials even after irradiation and measurement at
lowtemperatures (Table 2 and Scheme 4). Nevertheless, the
photolyzed solutions yielded definite products after thawing.
The azides 2a and 2g gave the pyrazine derivatives 12a and
were successful under mild conditions (1 h, ꢀ508C for 17b
and 4 days, 208C for 17c, Scheme 5). Of the four imaginable
diastereomeric [4+2] cycloadducts (exo,exo; endo,exo;
exo,endo; endo,endo), we isolated always the exo,endo
stereoisomer 17, which has the cyclopentadiene added to
the exo side of the bridgehead azirine to form a 5-azanorbor-
nene with an attached three-membered ring in the endo
position. The addition of hydrogen cyanide to the highly
strained azirines takes place already at ꢀ508C and leads to
good yields of the aziridines 18b,c,h. The less strained 2,3-
bridged heterocycle 3e does not undergo addition of hydro-
gen cyanide even after 7 days at 908C.[18] However, the
photolysis of the azide 2l[19] in the presence of hydrogen
cyanide can be used to gain evidence for the highly strained
Angew. Chem. Int. Ed. 2006, 45, 4015 –4019
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4017

Communications
(Table 3). These tricyclic compounds, as well as 19c, could be
1
characterized by their H and 13C NMR data. The analogous
transformations of 3c or the less reactive azirine 3d required
1 h at ꢀ408C and 4 days at ꢀ228C, respectively. The
intermediates 19 decomposed already at ꢀ208C to yield the
allyl azides[22] 20 and 21, hence 19d especially was always
present only in small proportions. We observed that at low
temperatures, at which the rearrangement of allyl azides
(equilibration of 20 and 21) had not yet begun, 20b–d as well
as 21b–d formed from 19b–d. Consequently, 19 reacts not
ꢀ
only by cleavage of the two C N bonds of the three-
membered ring, but also by cycloreversion of the five-
membered ring.
Highly strained 2,3-bridged 2H-azirines can also be
converted in the presence of ketenes, acyl isocyanates, or
nitrile oxides. We will report on these remarkable reactions
elsewhere.
Received: February 6, 2006
Revised: March 10, 2006
Published online: May 9, 2006
Scheme 5.
Keywords: azides· cycloaddition · dimerization ·
.
nitrogen heterocycles· strained molecules
bridgehead azirine 3l, which is not detectable by NMR
spectroscopy. The trapping product 18l proves the partic-
ularly short-lived intermediate 3l, in which the azirine unit is
further destabilized by the electron-withdrawing carbonyl
group.
The reaction of simple 2H-azirines with diazo compounds
to give allyl azides via 1,2,3-triazabicyclo[3.1.0]hex-2-enes
(compared with 19 in Table 3) has been known for more than
[1] W. E. Billups, M. M. Haley, G.-A. Lee, Chem. Rev. 1989, 89,
1147 – 1159.
[2] Reviews on 2H-azirines: a) J. Backes in Houben-Weyl, Vol. E16c
(Ed.: D. Klamann), Thieme, Stuttgart, 1992, pp. 321 – 369; b) V.
Nair in The Chemistry of Heterocyclic Compounds, Small-Ring
Heterocycles, Vol. 42, Part 1 (Ed.: A. Hassner), Wiley, NewYork,
1983, pp. 215 – 332; c) W. H. Pearson, B. W.
Lian, S. C. Bergmeier in Comprehensive
Heterocyclic Chemistry II, Vol. 1A (Ed.: A.
Table 3: Treatment of 2,3-bridged 2H-azirineswith diazomethane.
Padwa), Pergamon, New York, 1996, pp. 1 –
60; d) F. Palacios, A. N. Ochoa de Retana,
E. Martinez de Marigorta, J. M. de los San-
tos, Eur. J. Org. Chem. 2001, 2401 – 2414;
e) T. L. Gilchrist, Aldrichimica Acta 2001,
Substrate 3
Max. yield[a]
of 19 [%]
Yield[a,b] of 20+21 [%]
Initial ratio
20/21 after
equilibration
34, 51 – 55; f) K. M. L. Rai, A. Hassner in
Advances in Strained and Interesting Organic
Molecules, Vol. 8 (Ed.: B. Halton), Jai,
Greenwich, 2000, pp. 187 – 257.
20/21[c]
n
R
X
3b
3c
3d
3j
6
7
8
6
H
H
H
Me
CH2
CH2
CH2
89
97
<8
86
89
4:1 (ꢀ208C)
1:1 (ꢀ208C)
2:1 (08C)
1:0 (ꢀ208C)
4:1
2:1
5:1
1:0
[3] For the ring strain in 2H-azirines, see: E.-U.
Würthwein, T. Hergenröther, H. Quast, Eur.
J. Org. Chem. 2002, 1750 – 1755.
[4] a) A. Hassner, F. W. Fowler, J. Am. Chem.
Soc. 1968, 90, 2869 – 2875; b) A. Hassner,
F. W. Fowler, Tetrahedron Lett. 1967, 8,
1545 – 1548; c) B. Rose, D. Schollmeyer, H.
Meier, Liebigs Ann. 1997, 409 – 412.
84 (62)[d]
96 (78)[d]
85 (37)[d]
=
C O
[a] Yield based on 1H NMR standard. [b] Yield based on 3. [c] At low temperature. [d] Valuesin brackets
correspond to the yields of isolated products 20+21 based on the precursor 2.
[5] R. S. Atkinson, M. J. Grimshire, J. Chem.
Soc. Perkin Trans. 1 1986, 1215 – 1224.
forty years.[20] The 1,3-dipolar cycloaddition of the starting
materials proceeds relatively slowly at room temperature
relative to the fragmentation of the postulated intermediate
to yield the final products. Hence, short-lived 1,2,3-
triazabicyclo[3.1.0]hex-2-enes, which were also discussed as
intermediates of the rearrangement of allyl azides,[21] could
not be proved spectroscopically until now.
At ꢀ508C, the highly strained bridgehead azirines 3b and
3j react instantaneously with diazomethane by exo addition
to give the tricyclic compounds 19b and 19j, respectively
[6] a) K. M. L. Rai, A. Hassner in Comprehensive Heterocyclic
Chemistry II, Vol. 1A (Ed.: A. Padwa), Elsevier, Oxford, 1996,
pp. 61 – 96, and references therein; b) J. Laue, G. Seitz, Liebigs
Ann. 1996, 645 – 648; c) M. Drögemüller, R. Jautelat, E. Winter-
feldt, Angew. Chem. 1996, 108, 1669 – 1671; Angew. Chem. Int.
Ed. Engl. 1996, 35, 1572 – 1574; d) M. Drögemüller, T. Flessner,
R. Jautelat, U. Scholz, E. Winterfeldt, Eur. J. Org. Chem. 1998,
2811 – 2831; e) E. Haak, E. Winterfeldt, Synlett 2004, 1414 –
1418; f) Y. Tamura, Y. Yoshimura, T. Nishimura, S. Kato, Y.
Kita, Tetrahedron Lett. 1973, 14, 351 – 354; g) Y. Tamura, S. Kato,
4018 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4015 –4019

Angewandte
Chemie
Y. Yoshimura, T. Nishimura, Y. Kita, Chem. Pharm. Bull. 1974,
22, 1291 – 1296; h) S. Senda, K. Hirota, T. Asao, K. Maruhashi, J.
Am. Chem. Soc. 1978, 100, 7661 – 7664; i) D. S. Pearce, M. J.
Locke, H. W. Moore, J. Am. Chem. Soc. 1975, 97, 6181 – 6186;
j) M. Zaidlewicz, I. G. Uzarewicz, Heteroat. Chem. 1993, 4, 73 –
77.
[7] W. E. Billups, G.-A. Lee, B. E. Arny, K. H. Whitmire, J. Am.
Chem. Soc. 1991, 113, 7980 – 7984.
[8] Reviews on vinyl azides: a) K. Banert in Houben-Weyl, Vol. E15
(Eds.: H. Kropf, E. Schaumann), Thieme, Stuttgart, 1993,
pp. 818 – 875; b) A. Hassner in Azides and Nitrenes (Ed.:
E. F. V. Scriven), Academic Press, Orlando, 1984, pp. 35 – 76.
[9] Treatment of 1-azido-2-iodocycloheptane with KOtBu followed
by separation by chromatography gave 2c (31%) as well as 3-
azidocycloheptene (25%). Hassnerꢀs route afforded 2e (73%
overall yield from cyclododecene) as a yellowoil.
[10] A. Hassner, Acc. Chem. Res. 1971, 4, 9 – 16.
[11] a) A. Hassner, F. W. Fowler, J. Org. Chem. 1968, 33, 2686 – 2691;
b) J. Schweng, E. Zbiral, Justus Liebigs Ann. Chem. 1978, 1089 –
1095.
[12] From cyclohexene, we obtained by using Hassnerꢀs route 3-
azidocyclohexene along with a portion of 2b ꢂ 1%; see the
contrast to reference [13].
[13] a) F. Müller, Dissertation, Universität Münster, 1992; b) F.
Müller, J. Mattay, Angew. Chem. 1992, 104, 207 – 208; Angew.
Chem. Int. Ed. Engl. 1992, 31, 209 – 210; c) F. Müller, J. Mattay,
Chem. Ber. 1993, 126, 543 – 549.
[14] C. E. Davis, J. L. Bailey, J. W. Lockner, R. M. Coates, J. Org.
Chem. 2003, 68, 75 – 82.
[15] V. Nair, T. G. George, V. Sheeba, A. Augustine, L. Balagopal,
L. G. Nair, Synlett 2000, 1597 – 1598.
[16] Our spectroscopic data of the four azirines 3b–e (Table 1 and
Supporting Information) and the properties found for 3b differ
very clearly from the statements in reference [13] (e.g., see the
data of 3c and 3e in reference [13c] for correct data of 3d, see
also reference [4a,c]).
[17] Alternative syntheses for 11b,h and 12a,b,c: a) R. K. Grasselli,
D. D. Suresh, D. R. Bridgeman, US Patent US4379925A, 1983
[Chem. Abstr. 1983, 99, 88228]; b) D. G. Farnum, G. R. Carlson,
Synthesis 1972, 191 – 192; c) T. Kobayashi, S.-Y. Yamamoto, H.
Kato, Bull. Chem. Soc. Jpn. 1997, 70, 1193 – 1197; d) C. Herzig, J.
Gasteiger, Chem. Ber. 1981, 114, 2348 – 2354; e) T. Kobayashi, S.-
Y. Yamamoto, H. Kato, Bull. Chem. Soc. Jpn. 1997, 70, 1193 –
1197; f) H. E. Baumgarten, F. A. Bower, J. Am. Chem. Soc. 1954,
76, 4561 – 4564; g) J.-X. Chen, J.-P. Jiang, W.-X. Chen, T.-Y. Kao,
Heterocycles 1991, 32, 2339 – 2342.
[18] For the addition of exactly one equivalent of HCN to 2-
methylene-2H-azirines, see: K. Banert, M. Hagedorn, E. Knö-
zinger, A. Becker, E.-U. Würthwein, J. Am. Chem. Soc. 1994,
116, 60 – 62.
[19] M. Mizuno, T. Shioiri, Tetrahedron Lett. 1999, 40, 7105 – 7108.
[20] a) A. L. Logothetis, J. Org. Chem. 1964, 29, 3049 – 3052; b) V.
Nair, J. Org. Chem. 1968, 33, 2121 – 2123 und V. Nair, J. Org.
Chem. 1968, 33, 4316; c) T. C. Gallagher, R. C. Storr, Tetrahe-
dron Lett. 1981, 22, 2909 – 2912; d) K. Banert, M. Hagedorn, C.
Liedtke, A. Melzer, C. Schöffler, Eur. J. Org. Chem. 2000, 257 –
267; e) T. M. V. D. Pinho e Melo, A. L. Cardoso, C. S. B. Gomes,
A. M. dꢀA. Rocha Gonsalves, Tetrahedron Lett. 2003, 44, 6313 –
6315.
[21] A. Gagneux, S. Winstein, W. G. Young, J. Am. Chem. Soc. 1960,
82, 5956 – 5957.
[22] Alternative syntheses for 20b,d and 21b: a) E. Maxa, E. Zbiral,
G. Schulz, E. Haslinger, Justus Liebigs Ann. Chem. 1975, 1705 –
1720; b) R. C. Hayward, G. H. Whitham, J. Chem. Soc. Perkin
Trans. 1 1975, 2267 – 2270.
Angew. Chem. Int. Ed. 2006, 45, 4015 –4019
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4019