Hexaradialenes by successive ring openings of tris(alkoxytricyclobutabenzenes): Synthesis and characterization

Article abstract of DOI:10.1002/anie.200907305

(Figure Presented) You rang m'lord? Thermally induced successive ring openings of tris (alkoxytricyclobutabenzene) to hexaradialene are described. The isomerization reaction was highly stereoselective to afford a stereodefined hexaradialene 1, given that the hydroxy groups on the four-membered rings were protected by bulky substituents.

Full text of DOI:10.1002/anie.200907305

Communications
DOI: 10.1002/anie.200907305
Ring Opening Reactions
Hexaradialenes by Successive Ring Openings of Tris(alkoxy-
tricyclobutabenzenes): Synthesis and Characterization**
Shinya Shinozaki, Toshiyuki Hamura, Yousuke Ibusuki, Kotaro Fujii, Hidehiro Uekusa, and
Keisuke Suzuki*
Hexaradialenes are alicyclic cross-conjugated hydrocarbons,
in which all of the carbon atoms in the six-membered ring are
sp2 hybridized and have as many exocyclic double bonds as
possible. Interesting physical properties and reactivities are
expected from their unique structures, which also have
potential utility for constructing polycyclic molecules.[1,2]
The parent system (I),[3–5] however, is unstable and polymer-
izes immediately on formation; only a few substituted
hexaradialenes have been structurally characterized, includ-
ing hexamethyl,[6] dodecamethyl,[7] and hexabromo[8] deriva-
tives II–IV.
their tricyclobutabenzene analogues have remained unstud-
ied, presumably owing to their poor availability.
We previously developed a flexible synthetic route to
tricyclobutabenzenes that contain various oxygenated motifs
by repeated [2+2] cycloadditions of benzyne and ketene silyl
acetal (KSA),[10] which allowed the study of the thermal
behavior of oxygenated triannelated systems, such as VI.
Herein, we report that the thermal isomerization of VI indeed
offers a highly stereoselective access to stereodefined hexar-
adialenes VII, given that the hydroxy groups on the four-
membered rings were protected by bulky substituents.
One possible synthetic route to such systems is from the
valence isomer, tricyclobutabenzene V; this route is partic-
ularly interesting as it considers the question of the relative
thermodynamic preference of the two isomers. Although the
corresponding interconversion between monocyclobutaben-
zene and quinodimethane has been thoroughly studied,[9]
Scheme 1 shows the preparation of tritosylate 7, the key
intermediate for introducing three alkoxy substituents using
SN2 reactions. Starting from iodide 1, bromotosylate 2[11] was
selectively prepared by the repeated [2+2] cycloaddition
reactions of benzyne in a similar manner to that previously
reported.[10b] The third cycloaddition reaction occurred regio-
selectively[12] upon treatment of bromotosylate 2 with nBuLi
in the presence of KSA 8 to give a single cycloadduct, which
was converted into ketone 3 by selective hydrolysis. The
carbonyl group in 3 was converted into a methylene group by
a four-step sequence. Acid hydrolysis of triacetal 5 and
subsequent reduction of the resulting C3-symmetric triketone
gave triol 6. Tosylation of 6 afforded tritosylate 7 as a mixture
of diastereomers, which were separated by preparative TLC
to give the syn–anti isomer 7a and the syn–syn isomer 7b in
53% and 32% yields, respectively. The relative stereochem-
istry of 7a and 7b was easily assigned by NMR analysis with
symmetry considerations.
Tritosylate 7 served as a versatile platform to introduce
various oxygen nucleophiles onto the tricyclobutabenzene
framework (Scheme 2). Upon treatment of 7a with five
equivalents of sodium 4-methoxyphenoxide in N,N-dimethyl-
formamide at 458C, the three-fold SN2 reactions occurred
smoothly to give syn–anti ether 9a in 75% yield. Similarly, the
reaction of syn–syn isomer 7b proceeded cleanly and
stereospecifically, affording syn–syn ether 9b in 73% yield.
The corresponding sterically hindered 2,6-xylyloxy diastereo-
mers 10a and 10b were also prepared in high yields.
[*] S. Shinozaki, Dr. T. Hamura,[+] Dr. Y. Ibusuki, Dr. K. Fujii,
Dr. H. Uekusa, Prof. Dr. K. Suzuki
SORST, JST Agency
and
Department of Chemistry, Tokyo Institute of Technology
2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan)
Fax: (+81)3-5734-2788
E-mail: ksuzuki@chem.titech.ac.jp
and
Present address: Department of Chemistry
School of Science and Technology, Kwansei Gakuin University
2-1 Gakuen, Sanda, Hyogo 669-1337 (Japan)
[+] PRESTO, JST Agency
[**] This research was supported by the G-COE program and Grant-in-
Aid for Young Scientists (B). We thank Y. Tanaka (IMCE, Kyusyu
University) for HRMS measurements.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200907305.
3026
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3026 –3029

Angewandte
Chemie
Scheme 1. Preparation of tris(ester) 7. Reagents and conditions: a) 8,
nBuLi, Et2O, ꢀ788C, 5 min. b) PPTS, MeOH, 258C, 2 h (3: 62% yield,
2 steps). c) NaBH4, MeOH, 08C, 15 min. d) Tf2O, pyridine, CH2Cl2,
08C, 15 min. e) nBu4NI, toluene, 958C, 4.5 h. f) LiEt3BH, Et2O, 258C,
42 h. (5: 74% yield, 4 steps). g) aq HF, CH3CN, 258C, 2 h. h) LiEt3BH,
THF, ꢀ100!ꢀ788C (6: 71% yield, 2 steps). i) TsCl, pyridine, CH2Cl2,
258C, 18 h (7a: 53% yield, 7b: 32% yield). PPTS=pyridinium para-
toluenesulfonate, Tf=trifluoromethanesulfonyl, Ts=para-toluenesul-
fonyl.
Upon heating of syn–anti ether 9a in toluene at 1108C, the
expected hexaradialene 11 was not obtained: only a complex
mixture of products was produced, and the starting material
was completely consumed (Scheme 2).[13] On the other hand,
when the bulkier 2,6-xylyloxy substrate 10a (syn–anti isomer)
was heated in toluene at 1008C for 3 hours, the ring-opened
product 12 was obtained in 74% yield.[14] Importantly, the
product was obtained as a stable white solid after purification
by preparative TLC. It is interesting that the isomerization
was rigorously stereoselective, irrespective of the stereo-
chemistry of the starting material: the same C3-symmetric
product 12 was also cleanly obtained from the syn–syn isomer
10b (Scheme 2).
Scheme 2. SN2 reaction of 7 and thermal isomerization of aryl ethers
9a and 10a. Reagents and conditions: a) NaH, 4-methoxyphenol,
DMF, 458C, 36 h (9a: 75% yield from 7a; 9b: 73% yield from 7b).
b) NaH, 2,6-dimethylphenol, DMF, 458C, 28 h (10a: 77% yield from
7a; 10b: 73% yield from 7b). c) Toluene, 1108C, (11: 0% yield from
9a and 9b). d) Toluene, 1008C, 3 h. (12: 74% yield from 10a, 73%
yield from 10b). DMF=N,N-dimethylformamide.
1
Figure 1 shows the H NMR spectroscopic monitoring of
the thermal reaction of 10a ([D8]toluene, 1008C). The
reaction proceeded cleanly, and the peaks of the starting
material were almost completely converted into those of the
corresponding hexaradialene 12 after 3 hours (spectrum (d);
Figure 1). During the reaction, neither quinodimethane (one-
ring opening) nor bis(quinodimethane) (two ring openings)
was observed (see below).
Slow crystallization of hexaradialene 12 gave single
crystals suitable for X-ray analysis (hexane, ꢀ158C) as
shown in Figure 2.[15,16] The central six-membered ring
adopts a non-planar twist-boat conformation[17] so as to
avoid the non-bonding interactions of the exocyclic double
bonds.
Trisilyl ethers 14 and 15, available from the silylation of
triol 6, were also subjected to thermal isomerization con-
ditions (Scheme 3). Upon heating a mixture of syn–syn
and syn–anti isomers of TBDMS-protected 14 (TBDMS =
tBuMe2Si; toluene, 1058C), the ring opening reaction oc-
curred cleanly and stereoselectively to give the C3-symmetric
Figure 1. Thermal reaction of 10a monitored by 1H NMR spectroscopy
(400 MHz, [D8]toluene, 1008C). a) 10a (RT), b) 15 min (1008C), c) 1 h
(1008C), d) 3 h (1008C).
hexaradialene 16 as a single stereoisomer. Although clean
1
formation of 16 was evident by the H NMR analysis, it was
highly prone to hydrolysis during chromatography on silica-
gel, presumably owing to the lability of the enol silyl ether
Angew. Chem. Int. Ed. 2010, 49, 3026 –3029
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3027

Communications
Scheme 4. Thermal isomerization of non-C3-symmetric aryl ether 18.
Figure 2. X-ray structure of 12. Thermal ellipsoids drawn at 50%
probability. The hexane molecule and the minor part of the disordered
molecule are omitted for clarity, see reference [16].
Upon heating of 10a with dimethyl-2-butynedioate at 1008C
for 24 hours, three-fold [4+2] cycloadditions occurred
smoothly, and subsequent treatment with acid effected
three-fold eliminations of the tris(annelated) product 20[21]
to give triphenylene 21 in 60% yield (Scheme 5).[22] TLC
analysis suggested that the initially observed products, which
corresponded to the mono and the bis(cycloaddition) prod-
ucts were gradually converted into the cycloadduct 20.
Indeed, gentle heating of 10a (toluene, 908C, 4 h) gave
monoannelated product 22[23] and diannelated product 23 in
30% and 15% yields, respectively, after acid treatment of the
reaction mixture.[24]
It is interesting that the three-fold [4+2] cycloadditions
occurred smoothly to give triphenylene 21 when phenyl ether
9a was heated in the presence of the trapping agent. In
contrast, simple heating of 9a gave a complex mixture of
products (see above).
Scheme 3. Thermal isomerization of silyl ethers 14 and 15. a) Toluene,
1058C, 2.5 h. b) Chlorobenzene, 1258C, 1 h (17: 75% yield).
moieties. On the other hand, the more sterically congested
TBDPS (tBuPh2Si) derivative 15 was cleanly converted
(chlorobenzene, 1258C) into the C3-symmetric hexaradialene
17 as a stable, colorless oil after purification on preparative
TLC.
The observed stereoselectivities in these reactions can be
explained by the torquoselectivity of the substituted cyclo-
butene derivatives.[18] The outward rotation of the electron-
donating oxy groups on the four-membered rings accounts for
the ring-opened product. Also consistent with the formation
of 17 would be the, albeit less likely, three-fold inward
rotations of three alkoxy groups.
These results suggest that the three ring opening reactions
occur independently and the intermediary dienes and/or
tetraenes undergo [4+2] cycloaddition to give the corre-
sponding mono and/or diannelated cycloadducts.
In summary, tricyclobutabenzenes, which have bulky
alkoxy groups on the four-membered rings, cleanly rearrange
To address this point, the regioisomeric non-C3-symmetric
substrate 18,[19] consisting of a mixture of four stereoisomers,
was subjected to the thermal reaction (Scheme 4). Here again,
a high stereoselectivity was observed, irrespective of the
stereochemistry of the starting material, to afford 19a as a
single stereoisomer. The formation of 19b, by the three-fold
inward rotations was not observed. Thus, the energetically
preferred outward rotation of the electron-donating three
aryloxy groups accounts for the formation of the sterically
congested product. The stereochemistry of 19a was deter-
mined by NMR spectroscopic analysis (NOE, HMQC,
HMBC).[20]
As for the reaction mechanism, we addressed the
interesting question of whether the three ring openings
proceed in a concerted or stepwise manner. To gain insight
into these possibilities, trapping experiments were carried out.
Scheme 5. Three-fold [4+2] cycloaddition reactions of 12. Reagents
and conditions: a) Toluene, 1008C. b) Conc H2SO4, toluene, RT (20:
67% yield from 9a, 60% yield from 10a).
3028 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3026 –3029

Angewandte
Chemie
[13] The reaction has a threshold temperature for the ring opening:
into the corresponding hexaradialenes with high stereoselec-
tivity. Further studies on the reactivity and physical properties
of such oxygenated alicyclic cross-conjugated hydrocarbons
are underway in our laboratory.
upon heating of 9a at 1058C, the starting material was gradually
consumed, although the reaction did not occur in toluene at
1008C.
[14] All new compounds were fully characterized by spectroscopic
methods, combustion analysis, or high-resolution mass spectral
analysis. For details, see the Supporting Information.
[15] Crystallographic data for 12: C39H43O3, Mr = 559.7669, colorless
Experimental Section
ꢀ
Synthesis of hexaradialene 12: A solution of tricyclobutabenzene 10a
(10.3 mg, 0.0199 mmol) was heated in toluene at 1008C. After 3 h, the
crude product was purified by column chromatography on silica-gel
(hexane/EtOAc = 95:5) to give 12 (7.6 mg, 74% yield) as a white
solid. Recrystallization from hexane gave 12 as colorless prisms:
m.p. = 118.3–120.48C.
crystal, 0.40 ꢂ 0.30 ꢂ 0.20 mm, monoclinic, space group P1, Z = 2,
T= 123(2) K, a = 7.608(3), b = 12.346(3), c = 18.172(7) ꢃ, a =
100.71(2), b = 100.99(2), g = 104.92(2)8, V= 1568.6(9) ꢃ3, l-
(MoKa) = 0.71075 ꢃ, m = 0.073 mmꢀ1. Intensity data were col-
lected on a Bruker SMART 1000 diffractometer. The structure
was solved by direct methods (SHELXS97) and refined with full-
matrix least-squares on F2 (SHELXL97). A total of 13951
reflections were measured and 6654 were independent. Final
R1 = 0.1153, wR2 = 0.2801 (4411 refs; I > 2s(I)), and GOF =
1.095 (for all data, R1 = 0.1595, wR2 = 0.3101). CCDC 759632
(12) contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.
uk/data_request/cif.
Received: December 28, 2009
Published online: March 23, 2010
Keywords: diastereoselectivity · isomerization · radialene ·
.
ring opening reactions · tricyclobutabenzene
[16] The high R factors are mainly due to the disorder of the hexane
crystalline solvent molecules incorporated in the crystal. The
hexane molecule is in a narrow channel structure along the a axis
and is extremely disordered on two inversion centers (0, 0, 0 and
1/2, 0, 0), which results in the continuous carbon chain model
without hydrogen atoms. The 2,6-xylyloxy groups of 12 are also
disordered. The disorder consists of the flipping of the 2,6-
xylyloxy groups and the occupancy for the minor part of the
disorder is 12.6(4)%. In spite of the disorder, the conformation
of the central six-membered ring has been clearly determined
from the electron density map and refined.
[17] For the preparation of planar hexaradialenes, see: A. Patra,
Y. H. Wijsboom, L. J. W. Shimon, M. Bendikov, Angew. Chem.
2007, 119, 8970 – 8974; Angew. Chem. Int. Ed. 2007, 46, 8814 –
8818.
[18] a) N. G. Rondan, K. N. Houk, J. Am. Chem. Soc. 1985, 107,
2099 – 2111; b) K. N. Houk, D. C. Spellmeyer, C. W. Jefford,
C. G. Rimbault, Y. Wang, R. D. Miller, J. Org. Chem. 1988, 53,
2125 – 2127; c) K. Nakamura, K. N. Houk, J. Org. Chem. 1995,
60, 686 – 691.
[19] For the preparation of 18, see the Supporting Information.
[20] For details, see the Supporting Information.
[21] Tris(annelated) product 20 was highly prone to aromatization
during column chromatography on silica gel, although its
formation was confirmed by 1H NMR analysis.
[22] For related reactions, see: a) H. Hopff, G. Kormany, Helv. Chim.
Acta 1963, 46, 2533 – 2538; b) H. Hopff, G. Kormany, Helv. Chim.
Acta 1965, 48, 437 – 443; c) W. E. Billups, D. J. McCord, B. R.
Maughon, J. Am. Chem. Soc. 1994, 116, 8831 – 8832. See also
Refs [4] and [5].
[1] For reviews of [n]radialenes, see; a) G. Maas, H. Hopf in The
Chemistry of Dienes and Polyenes, Vol. 1 (Ed.: Z. Rappoport),
Wiley, Chichester, 1997, Chapter 21, pp. 927 – 977; b) H. Hopf in
Classics in Hydrocarbon Chemistry, Wiley-VCH, Weinheim,
2000, Chapter 11.5, pp. 290 – 300.
[2] a) H. Hopf, G. Maas, Angew. Chem. 1992, 104, 953 – 977; Angew.
Chem. Int. Ed. Engl. 1992, 31, 931 – 954; b) T. Hꢀpfner, P. G.
Jones, B. Ahrens, I. Dix, L. Ernst, H. Hopf, Eur. J. Org. Chem.
2003, 2596 – 2611; c) M. Gholami, R. R. Tykwinski, Chem. Rev.
2006, 106, 4997 – 5027.
[3] L. G. Harruff, M. Brown, V. Boekelheide, J. Am. Chem. Soc.
1978, 100, 2893 – 2894.
[4] a) A. J. Barkovich, K. P. C. Vollhardt, J. Am. Chem. Soc. 1976,
98, 2667 – 2668; b) A. J. Barkovich, E. S. Strauss, K. P. C. Voll-
hardt, J. Am. Chem. Soc. 1977, 99, 8321 – 8322.
[5] P. Schiess, M. Heitzmann, Helv. Chim. Acta 1978, 61, 844 – 847.
[6] a) H. Hopff, A. K. Wick, Helv. Chim. Acta 1961, 44, 19 – 24;
b) H. Hopff, A. K. Wick, Helv. Chim. Acta 1961, 44, 380 – 386;
c) M. Trætteberg, P. Bakken, H. Hopf, T. Hꢀpfner, J. Mol. Struct.
1998, 445, 99 – 105.
[7] M. Iyoda, S. Tanaka, H. Otani, M. Nose, M. Oda, J. Am. Chem.
Soc. 1988, 110, 8494 – 8500.
[8] A. Stanger, N. Ashkenazi, R. Boese, D. Blꢁser, P. Stellberg,
Chem. Eur. J. 1997, 3, 208 – 211.
[9] a) J. L. Segura, N. Martin, Chem. Rev. 1999, 99, 3199 – 3246;
b) A. K. Sadana, R. K. Saini, W. E. Billups, Chem. Rev. 2003,
103, 1539 – 1602.
[10] a) T. Hamura, Y. Ibusuki, H. Uekusa, T. Matsumoto, K. Suzuki,
J. Am. Chem. Soc. 2006, 128, 3534 – 3535; b) T. Hamura, Y.
Ibusuki, H. Uekusa, T. Matsumoto, J. S. Siegel, K. K. Baldridge,
K. Suzuki, J. Am. Chem. Soc. 2006, 128, 10032 – 10033.
[11] For the preparation of 2, see the Supporting Information. See
also Ref. [10b].
[23] Monoannelated product 22 consisted of two stereoisomers,
which were not separable.
[24] Triphenylene 21 was also produced in 5% yield, accompanied by
the starting material 10a (37%).
[12] The regiochemistry of the [2+2] cycloadduct was determined by
NMR spectroscopic analysis with symmetric consideration after
converting into the C3-symmetric triacetal 5.
Angew. Chem. Int. Ed. 2010, 49, 3026 –3029
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3029