Diastereomeric cyclic tris-allenes

Article abstract of DOI:10.1039/c3cc40354h

Both diastereomers of the tris-allene, cyclododeca-1,2,5,6,9,10-hexaene have been obtained using a triple cyclopropylidene-allene rearrangement. On the NMR timescale, one has D₃ symmetry, and is the smallest hydrocarbon synthesised to have this symmetry, and the second has C₂ symmetry.

Full text of DOI:10.1039/c3cc40354h

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Diastereomeric cyclic tris-allenes†
a
a
a
Hussein H. Mustafa, Mark S. Baird,* Juma’a R. Al Dulayymi and
Cite this: Chem. Commun., 2013,
b
4
9, 2497
Viacheslav V. Tverezovskiy
Received 27th November 2012,
Accepted 7th February 2013
DOI: 10.1039/c3cc40354h
www.rsc.org/chemcomm
Both diastereomers of the tris-allene, cyclododeca-1,2,5,6,9,10-
hexaene have been obtained using a triple cyclopropylidene–allene
rearrangement. On the NMR timescale, one has D
the smallest hydrocarbon synthesised to have this symmetry, and
the second has C symmetry.
3
symmetry, and is
2
Scheme 1
3
Molecules with C symmetry have been of considerable recent
1
interest; among the most common of these are tri-aryl substituted
benzenes, where the molecule is twisted from planarity by the steric
of a cyclopropylidene (3), or a related carbenoid, to the allene in a
reaction which in many cases occurs rapidly and efficiently even at
hindrance between the substituents. Those with D symmetry are
3
ꢀ
90 1C (Scheme 1). This reaction has been used to produce cyclic
2
rarer, for example, trans-hexa-substituted trindanes, while the
7–11
mono- and bis-allenes,
and indeed it is 50 years since Moore
3
smallest reported hydrocarbon of this type is [4.4.4]propellahexaene.
12–14
and Ward and Skattebøl reported the first such products.
As part of a study of the applications of cis,cis,cis-cyclonona-1,4,7-
triene 1, we became interested in the corresponding allene, cyclo-
dodeca-1,2,5,6,9,10-hexaene 2. Because of the inherent chirality of
However, where the strain generated in producing the allene
is high, alternative reactions of the carbene, such as insertion into
6,15
C–H bonds, may occur. We were intrigued by the possibility that
the tris-dibromocarbene adduct of cyclonona-1,4,7-triene (i.e. 4)
would undergo three sequential carbenoid formations and rear-
rangements to give the tris-allene 2, and, if it did, whether this
would produce one or both diastereoisomers of the allene. Indeed
it is known that the mono-dibromocarbene adduct of 1 does
react with methyllithium to give the corresponding allene,
1
,3-disubstituted allenes, this molecule can in principle exist either
as R,R,R- or R,R,S-forms, or as their enantiomers. Little is known
about these molecules except for some calculations of their shape
and energy which predicted a C symmetry for the former and C or,
3
2
4,5
1
more probably, C symmetry for the latter (Fig. 1).
The reaction of a 1,1-dibromocyclopropane with methyl-
6
lithium provides one of the most efficient routes to allenes.
This process proceeds by initial lithium–bromine exchange,
followed by formal loss of lithium bromide and rearrangement
1
9
cyclodeca-1,2,5,8-tetraene in 78% yield.
1
6–21
cis,cis,cis-Cyclonona-1,4,7-triene (1)
was reacted with an
excess of dibromocarbene generated from bromoform and base
under phase transfer conditions. This gave the bis-adduct 5 and
the tris-adduct 4 (Fig. 2). The tris-adduct has been prepared
before using bromoform and potassium t-butoxide and converted
Fig. 1 Cyclonona-1,4,7-triene and the related tris-allene.
a
School of Chemistry, Bangor University, Bangor, Gwynedd, Wales, UK LL57 2UW.
E-mail: chs028@bangor.ac.uk; Fax: +44 (0)1248370528; Tel: +44 (0)1248382374
b
Biocomposites Centre, Bangor University, Bangor, Gwynedd, Wales, UK LL57 2UW
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3cc40354h
Fig. 2 Tris- and bis-dibromocarbene adducts of cyclonona-1,4,7-triene.
Chem. Commun., 2013, 49, 2497--2499 2497
This journal is c The Royal Society of Chemistry 2013

View Article Online
Communication
ChemComm
Scheme 2
22
into the corresponding hydrocarbon. The stereochemistry of the
latter had been established as being all cis since only one three
hydrogen doublet was seen, corresponding to the trans-protons of
13
the methylene group between the cyclopropanes. The C NMR
23
spectrum of 4 showed just three lines, confirming this assignment.
The bis-adduct 5 would most reasonably also be cis- since it is a likely
intermediate in the formation of 4; this was confirmed by the
presence of a sharp doublet for one hydrogen of the methylene
group between the cyclopropanes at d 2.45, coupled only to the
geminal proton, while the latter proton appeared at d 1.0 as a
complex single-hydrogen multiplet, in this case coupled to the
Fig. 4 (a) A model of R,R,R-7 with D
3
symmetry; the C
3
axis is vertical through
the centre of the molecule. The central allene carbons are shown in red; (b)
showing one of three C axes through the front central allenic carbon and the
back) opposite methylene group. The symmetry axes are shown on the structure
in the ESI.† (c) R,R,S-8 with the R-allene segments in red and the S-allene segment
axis between the methylene group and the centre
carbon of the S-allene (black).
2
(
24
adjacent cyclopropane hydrogens.
Reaction of 4 with methyllithium in ether at 0 to ꢀ40 1C led in blue and showing the C
2
to a mixture of 7 and 8 (81%) in a ratio of ca. 1 : 3 and the NMR
spectrum after work-up showed the presence of no other
significant products (Scheme 2). Hydrogenation of the mixture
using 5% Pd/C as catalyst gave a single product, cyclododecane and irradiation of the multiplet at d 5.22 caused the signal at
90%). The isomeric allenes could be partially separated by 2.62 to become a single line, and vice versa (see ESI†). Moreover,
selective crystallisation of the minor isomer. They were not there was no change when the spectrum was run at ꢀ45 1C in
separated by column chromatography over silica gel but were CDCl . The spectra were consistent with those expected for a
separated by chromatography on silica gel impregnated with molecule 7 with D symmetry (at least on the NMR timescale) in
% silver nitrate. Silver nitrate complexes of a number of cyclic which there are three allenes of the same absolute stereochemistry
allenes, including cyclodeca-1,2,6,7-tetraene have been (Fig. 4a), showing six chemically equivalent but magnetically
(
3
3
5
2
5
reported. The first fraction, consisting of 7 (8% from 4), inequivalent hydrogens and in which the six hydrogens of the
1
3
showed just three peaks in the C NMR spectrum in CDCl3, methylene groups are also chemically equivalent but magnetically
at d 27.8 (CH ), 89.8 (CH) and 206.5 (central allenic carbon). inequivalent – each one having the opposite combination of a pair
2
1
The H NMR spectrum showed just two six-hydrogen multiplets of dihedral angles to the two adjacent allenic hydrogens. The
centred at d 5.22 and 2.62 (Fig. 3). The multiplets at first molecule has a three-fold axis of symmetry as seen in Fig. 4a,
26
sight approximated to complex pentuplets or triplets of triplets and three C
J ca. 3.8, 4.0 Hz), though additional lines were also present component (37% from 4 after silica gel/AgNO
indicative of non-first-order multiplets. There was no difference showed d (CDCl ) 28.2, 27.7 (both CH ), 91.7, 90.2 and 89.8
in the shapes and line separations of the multiplets when the (all CH) and 206.7 and 205.8 for the central carbons of the allenes.
solvent was changed from CDCl to C D (see ESI†), or if the This is consistent with a molecule 8 having C symmetry on the
spectrum was run at 500 MHz rather than 400 MHz. Each NMR timescale (Fig. 4c). The H NMR spectrum in CDCl (see ESI†)
2
axes, one of which is shown in Fig. 4b. The second
(
3
chromatography)
C
3
2
3
6 6
2
1
3
multiplet had a width of 14.8 Hz between the major outer lines showed three two hydrogen multiplets in the range d 2.5–2.8, a two
hydrogen multiplet at 5.2 and a four hydrogen multiplet at 5.3; in
C D solution, the signals in the allenic region were split into three
6 6
complex two hydrogen multiplets corresponding to the three types
3
of allenic hydrogen in Fig. 4c. The spectra showed JH,H coupling
between each different allenic hydrogen and both the adjacent
5
hydrogens of the methylene groups, and JH,H to the two protons of
each methylene group at the other end of the allene. The multiplets
show non-first order effects due to the constrained geometry of
the molecule and the presence of chemically equivalent, but
magnetically inequivalent, protons.
Preliminary calculations at a semi-empirical level representing a
3
D structure of 7 reveal a molecule having a central cavity
Fig. 3 The 400 MHz HNMR signals for 7 (a) the methylene groups centred at d
2
.62; (b) the allenic hydrogens centred at d 5.22.
with a HOMO providing a triple co-ordination site on each face
2
498 Chem. Commun., 2013, 49, 2497--2499
This journal is c The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
Notes and references
1
2
3
4
5
For example, S. E. Gibson and M. P. Castaldi, Chem. Commun., 2006,
045–3062.
G. Borsato, M. Crisma, O. De Lucchi, V. Lucchini and A. Zambo,
Angew. Chem., Int. Ed., 2005, 44, 7435–7439.
L. A. Paquette, S. Liang, L. Waykole, G. DeLucca, H. Jendralla,
R. D. Rogers, D. Kratz and R. Gleiter, J. Org. Chem., 1990, 55, 1598–1611.
I. Yavari, H. Norouzi-Arasi, D. Nori-Shargh and H. Fallah-Bagher-
Shaldaei, J. Chem. Res., Synop., 1999, 154–155.
3
3
The orbitals of the D symmetric form of the smallest possible tris-
allene, cyclonona-1,2,4,5,7,8-hexaene, have been reported to show
Mobius character (a) H. S. Rzepa, Chem. Rev., 2005, 105, 3697–3715;
see also: (b) J. R. Hutchison and H. S. Rzepa, J. Am. Chem. Soc., 2004,
Fig. 5 (a) and (b) The calculated HOMO orbital of a D
3
model of 7 seen from
27
opposite faces.
1
26, 14865–14870.
6
J. Backes and U. H. Brinker, Carbene(oide) Part 1, in Methoden der
Organische Chimie, ed. M. Regitz, 1989, vol. 19b, pp. 391–510.
2
6,27
(
Fig. 5a and b).
In each fragment QCH
x
–CH
a
H
a
0
–CH
x
0
Q,
7 (a) L. Skattebøl, Tetrahedron Lett., 1967, 1961; (b) , Acta Chem. Scand.,
963, 17, 1683–1693; (c) K. Kleveland and L. Skattebøl, Acta Chem.
1
the H has a dihedral angle of approximately 231 to H and 921
a
x
Scand. Ser. B, 1975, 29, 191–196.
M. S. Baird and C. B. Reese, Tetrahedron, 1976, 32, 2153–2156.
to Hx
0
, and Ha
0
has a dihedral angle of approximately 231 to Hx
and 921 to Hx.
Compounds 7 and 8 are the first reported carbocyclic tris-
allenes, though macrolide tris- and tetra-allenes, and more
recently enantiopure, shape-persistent alleno-acetylenic macro- 11 (a) K.-L. Noble, H. Hopf and L. Ernst, Chem. Ber., 1984, 117, 474–488;
cycles, have been reported. The ring-opening of a cyclopropylidene
to an allene requires an overall monorotation of the substituents on
one ring carbon. The reactions of bis-dibromocarbene adducts of
0
8
9 R. W. Thies, J. L. Boop, M. Schiedler, D. C. Zimmerman and
T. H. LaPage, J. Org. Chem., 1983, 48, 2021–2024.
10 H. Nozaki, T. Aratani, T. Toraya and R. Noyori, Tetrahedron, 1971,
2
8
27, 905–913.
29
(b) S. N. Moorthy and D. Devaprabhakara, Synthesis, 1972, 612.
2 W. R. Moore and H. R. Ward, J. Org. Chem., 1960, 38, 2073–2073.
3 W. R. Moore and H. R. Ward, J. Org. Chem., 1962, 27, 4179–4181.
4 L. Skattebøl, Tetrahedron Lett., 1961, 167–172.
1
1
1
cis,cis-cyclo-octa-1,5-diene or cyclotetradeca-1,9-diene with methyl 15 X. Creary, Z. Jiang, M. Butchko and K. McLean, Tetrahedron Lett.,
1
996, 37, 579–582.
lithium are known to give the corresponding bis-allene exclusively
as one diastereoisomer, in the case of the smaller ring, the
meso-form.
1
1
6 K. G. Untch and D. J. Martin, J. Org. Chem., 1964, 29, 1903–1904.
7 K. G. Untch, J. Am. Chem. Soc., 1963, 85, 345–346.
3
0,31
Dehmlow showed that either syn- or anti-bis- 18 W. R. Roth, Liebigs Ann. Chem., 1964, 671, 10–25.
1
2
9 F. A. L. Anet and M. Ghiaci, J. Am. Chem. Soc., 1980, 102, 2528–2533.
0 (a) R. W. Thies, P.-K. Hong, R. Buswell and J. L. Boop, J. Org. Chem.,
975, 40, 585–590; (b) R. W. Thies, P. K. Hong and R. Buswell,
J. Chem. Soc., Chem. Commun., 1972, 1091–1092.
21 M. R. Detty and L. A. Paquette, Tetrahedron Lett., 1977, 18, 347–350.
dibromocarbene adducts of cyclo-1,5-octadiene give the same
stereoisomer of allene, and proposed that when the initial
alkene contains a relatively small ring, the reaction will be
1
3
1
controlled by strain, and be highly stereoselective. Moreover it
2
2 C. D. Poulter, R. S. Boikes, J. I. Brauman and S. Winstein, J. Am.
Chem. Soc., 1972, 94, 2291–2296.
is known that either syn- or anti-isomers of the tetrabromo-
C 3 2
dicyclopropanes derived by addition of two molecules of dibromo- 23 This showed d (125 MHz, CDCl ): 32.4 (CH ), 31.7 (C), 24.6 (CH).
2
4 The J value between the trans-methylene hydrogen and the adjacent
cyclopropane for these compounds is essentially zero. The signal at
d 2.45 became a singlet on selective decoupling of the signal at 1.0.
carbene to 1,1,6,6-tetramethoxy-cis,cis-cyclodeca-3,8-diene react
with methyllithium to give a 1 : 2 mixture of the racemic and
meso forms of the diallene, 1,1,7,7-tetramethoxycyclododeca- 25 G. Nagendrappa, G. C. Joshi and D. Devaprabhakara, J. Organomet.
3
2,33
Chem., 1971, 27, 421–426.
6 Initial semi-empirical and ab initio calculations predict that the D
3
form is not an energy minimum, but a transition state between two
minima (R. A. Davies, unpublished results) (see ESI†).
7 Visualised using Chem3D Pro 12.0.
8 M. Yoshida, N. Harada, H. Nakamura and K. Kanematsu,
Tetrahedron Lett., 1988, 29, 6129–6132.
3,4,9,10-tetraene.
The ring opening of cyclopropylidenes
2
formed by reaction of dibromocyclopropanes with methyl-
lithium is an extremely fast process. The formation of 7 and 8
would therefore appear most likely to occur by three sequential
cyclopropylidene–allene rearrangements. The first of these
6
2
2
would produce a 1 : 1 mixture of R and S-monoallenes; if there 29 P. Rivera-Fuentes, J. L. Alonso-Gomez, A. G. Petrovic, P. Seiler,
F. Santoro, N. Harada, N. Berova, H. S. Rzepa and F. Diederich,
Chem.–Eur. J., 2010, 16, 9796–9807.
0 M. Irngartinger and H.-U. Jager, Tetrahedron Lett., 1970, 4047–4050.
is no control by the stereochemistry of the second process, this
would statistically produce a 1 : 1 mixture of RR(SS) and RS(SR)
3
bis-allenes, and the third a 1 : 3 mixture of RRR(SSS) and 31 E. V. Dehmlow and T. Stiehm, Tetrahedron Lett., 1990, 1841–1844.
3
2 P. J. Garratt, K. C. Nicolau and F. Sondheimer, J. Am. Chem. Soc.,
973, 95, 4582–4592.
RRS(SSR) forms. This is relatively close to the observed ratio.
The detailed structures of the allenes, and in particular
allene 7, are currently being examined, together with an
ab initio analysis of its structure. The allenes 7 and 8 may both
provide interesting opportunities as ligands; this is also being
investigated.
1
3
3 Reaction of 5 with methyllithium also led efficiently to both dia-
stereoisomers of the corresponding allene, cycloundeca-1,2,5,6,9-
pentaene. One of these rearranges over a period of hours at ambient
3
temperature in CDCl solution to a major and a minor product. The
major product rearranges further over a period of 18 h. This will be
described elsewhere.
This journal is c The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 2497--2499 2499