Proc. NoL Acad, Sci. USA Vol. 79, pp. 4495-4500, July 1982 Review

Dodecahedrane-The chemical transliteration of Plato's universe (A Review) LEO A. PAQUETTE Evans Chemical Laboratories, The Ohio State University, Columbus, Ohio 43210 Communicated by Daniel E. Koshlond, Jr., April 26, 1982

ABSTRACT The development ofchemical interest in three of deviate the greatest from normalcy, this should pos- Plato's five convex polyhedra is described from an historical per- sess the greatest per . In contrast, spective. The successful synthesis of 1,16-dimethyldodecahedrane is characterized by minimal angle strain, al- and its structural characteristics are outlined. Finally, an account though unparalleled high levels oftorsional strain exist because ofrecent work leading to the still more aesthetically appealing and ofthe precise eclipsing ofall the C-H bonds. The chemist's fas- ultrasymmetric parent dodecahedrane is given. cination with 1, 2, and 3 is fueled further by the aesthetically delightful topologies ofthese structures, the presence ofcavities Men oflearningin ancient Greece exhibited particular fondness ofincreasing size (perhaps reaching a practical lower limit in 3), for, and interest in, five exquisitely symmetric convex poly- the spectral consequences of 4, 8, or 20 symmetry-equivalent hedra, each composed of totally equivalent vertices and edges. methine units, the absence of structurally allied substances in In Plato's thinking, constituted the smallest com- nature, and much more. ponents of the element fire and pentagonal dodecahedra com- Despite dire theoretical predictions about the intrinsic insta- prised the building blocks ofheavenly matter; in between were bility of , there has been no shortage of attempts cubes for earth, octahedra for air, and icosahedra for water. to synthesize this fascinating molecule. Although the parent Much later, in the 17th century, the astronomer Johannes Kep- remains unknown, Maier et aL succeeded in 1978 ler made extraordinary use of these five regular bodies in his in preparing the tetra-tert-butyl derivative 5(1, 2). The critical model of the universe. steps in the synthesis (Scheme I), which depends very heavily That the heritage of this wonder should inevitably pass into on light-induced reactions, are the photoisomerization and de- the realm of synthetic organic comes as no surprise. carbonylation of 4. The colorless crystalline 5 is amazingly re- However, the requisite transliteration into hydrocarbon equiv- sistant to oxygen, moisture, and heat (mp 1350( dec). At the alents necessarily excludes octahedrane and icosahedrane for melting , isomerization to tetra-tert-butylcyclo- practical reasons. Ofthe remaining three, 1 belongs to Td point occurs. The extraordinary stability of 5 has been group with its 24 identity operations, 2 has full Oh symmetry, and 3, with its 'h character and 120 identity operations, is the hvi -hv&-I. II + 0 /a -c

hVi hv *0

I 2 5 4 a, Br2; b, KOH; c, tert-BuLi. Scheme I attributed to the fact that rupture of any framework bond is opposed by increased steric interaction between the pendant tert-butyl groups. The most notable chemical property of5 un- covered to date is its penchant for oxidation to the isomerized radical cation in the presence of many reagents. The tactical elaboration of was masterfully achieved by Eaton and Cole (3) at the University of Chicago as early as 3 1964. Their protocol (Scheme H) took good advantage of the facility with which 2-bromocyclopentadienone dimerizes to 6. most highly symmetric hydrocarbon possible. All three mole- Following photocyclization of the dimer, 2-fold Favorskii ring cules clearly are structurally unusual, and their properties are contraction was achieved with hot . Once not predictable with certainty a priori from current theories of dicarboxylic acid 7 was in hand, it was a simple matter to remove chemical bonding. Because the facial angles of tetrahedrane the carboxyl groups by conventional methods. 4495 Downloaded by guest on October 3, 2021 4496 Review: Paquette Proc. NatL Acad, Sci. USA 79 (1982)

Br f-t Y0 6%A Lto0 ].

6 lhV cOc )H _ 0 d IO-OC O= Br I 4-d NW Br HOOC

2 7

a, NBS; b, Br2; c, (C2H5)3N; d, KOH, H20, A, and then H30+; e, SOC12; f, tert-BuOOH; g, triisopropylbenzene, A. Scheme II Presumably because oforbital symmetry constraints, 2 is also carbon (26-28). From the outset, it was that impressively stable to heat. Only when cubane is kept at 2000C projected for 23 should derive from the cyclopentadienide anion (10), a sim- 2 weeks does one obtain cyclooctatetraene via cuneane (8) ple and inexpensive commodity. Following oxidative and semibullvalene (4). Silver(I) ion greatly accelerates this par- of 10 to coupling ticular isomerization (5); rhodium(I) leads 9,10-dihydrofulvalene, a domino Diels-Alder reaction alternatively to the (29-31) is carried out with dimethyl to syn dimer of (9) (6, 7). acetylenedicarboxylate give 11, a C2,,-symmetric molecule whose role it is to serve as the cornerstone of the spherical target molecule (Scheme Controlled reduction in the Ill). Ag (I) Rh(I) symmetry level ofthis intermediate -.Vl to diketo diester provide 12 which possesses only a simple C2 molecular axis is entirely possible by a hydroboration-oxidation sequence (32). In practice, however, a somewhat longer, more 8 2 efficient catercorner oxygenation sequence is preferred (26-28) because the unwanted The challenge offered by the synthesis ofthe pentagonal do- C. diketo diester is not produced decahedrane is of large proportions. The ultimate objective is concomitantly. to interlink 20 Bisspiroannulation of 12 with diphenyleyclopropylsulfonium methine units with 30 carbon-carbon bonds in ylide followed oxidation delivers the such a way that 12 multiply fused five-membered rings result. by peroxide ax- ially symmetric spirolactone 13 which can be isomerized under Additionally, the topology of 3 is such that all ring fusions are strongly and cis-syn, a feature which causes acidic conditions subsequently fashioned into di- many partially constructed ore- lactone 14. It must be recognized that the continued mainte- cursors to suffer excessively high levels of nonbonded steric nance of C2 symmetry now reduces our analysis of the subse- strain and consequently to exhibit a pronounced tendency for quent chemistry of14 to the manipulation ofa pair ofsymmetry- transannular bonding, framework isomerization, and nonevelo- related carbonyl pentanoid cyclization (8, 9). It is groups. Furthermore, this intermediate chiefly for these reasons and already contains all requisite 20 carbon atoms and has ofthe other related factors that the several convergent approaches to 12 3 all-important cis-locked stereocenters properly in- which have been attempted by various groups have not suc- stalled. methine Although the chemical reactivity of 14 far less ceeded as yet. These include the formal combination oftwo proved C1o controllable than originally the con- units (10-18), capping of a C15 unit with a anticipated (33, 34), pivotal cyclopentane ring version to 15 can be accomplished quite (19-22), amalgamation of C,6-hexaquinacene with a C4 segment efficiently. (23, 24), and controlled Having arrived at 15 in a minimum of nine laboratory steps, Lewis acid-catalyzed isomerization of we an number to a more readily required equal craft the monoseco 20 accessible (CH)20 (ref. 25; P. von R. Under the influence of in Schleyer, personal communication). (SchemeIV). liquid , 15 experiences a splendid reduction-alkylation sequence to gen- These serious complications prompted consideration of a se- erate a dienolate which is directly methylated to yield 16 (35). rial synthetic approach to 3. Importantly, we hoped to overcome In this way, the endo orientation of the pair of groups the tactical disadvantages of nonconvergency by on carbonyl capitalizing is guaranteed and the methyl substituents serve as the symmetry ofthe target molecule throughout the sequence. 1H NMR invaluable Early in beacons. Light-induced ring closure at the site 1981, my colleagues and I were pleased to be able to (ester are report our groups generally photoinactivatable), dehydration of successful acquisition ofthe 1, 16-dimethyl derivative the resulting tertiary alcohol, and diimide reduction delivers (23) in a number ofsteps fewer than the total count offramework the triseco ester 17. While this intermediate clearly possesses Downloaded by guest on October 3, 2021 Review: Paquette Proc. Natd. Acad. Sci. USA 79 (1982) 4497

H 3O0%IV!:: 0

No+ a b, c CH3.9-

10 11 12

d, e

CH30 C

|, .,% f-h '-VI I s o CH3

15 14 13

a, 12, then CH30OC=CCOOCH3; b, [(CH3)2CHCH]2-BH; c, Cr03; d, tA-SPh2, then H+; CH3 e, H22; f, P205, CH3SO3H; g, H0, Pd/C; h, NaCNBH3; i, HCI, CH30H. Scheme III

many of the desirable structural features being sought, its op- efficiencyofthis process was appreciably lower (29%) than here- posed methylene groups remain unfunctionalized. After some tofore because of the well-established tendency of tertiary al- preliminary studies, the decision was made to address this po- dehydes to experience decarbonylation from their excited tentially complex issue at a later stage. Following reduction of states. Nevertheless, the important point here is that a five-ring 17 to the level, 18 was photocyclized as before. The ketone (19) is produced upon which light may be impinged for CH30 CH30 CH34, 0

a b-d CI co3

is2 POP 3 CH3 15 16 17 e ,f OH 13 CH3H 0

b b, f

OH3 CH3 20 19 18

a, Li/NH3, then CH31; b, hv; c, TsOH, C6H6, d, HN NH, H3OH; e, (i-Bu)2AIH; f, PCC CH2CI2. Scheme IV Downloaded by guest on October 3, 2021 AA.98 Review: Paquette Proc. Nad Acad. Set. USA 79 (1982)

a b

OH3 20 21 22

c lc CH3 io

23 a, TsOH, CAH6; b, H2NNH2, H202; c, CF3SO3H, 0H2C12. Scheme V

the third time. Closure to the monosecododecahedral alcohol the disubstituted D-rsymmetric dodecahedrane 23. The un- 20 proved to be encouragingly simple and efficient, despite the expected occurrence of a 1,2-methyl migration during this in- extreme steric congestion which develops in the gap and exerts stallation of the final framework bond was substantiated by 1H major geometric alterations in the superstructure as a whole and 13C NMR (merely four carbon resonances) as well as x-ray (26-28). crystallography. Some of the more unique physical properties With arrival at 20, it proved an easy matter to proceed to the of23 include its reluctance to melt (discoloration begins above beautifully crystalline hydrocarbon 22 which possesses an eight- 350C) its density (1.412 g/em3), its relatively small transcavity line 13C NMR spectrum due to a return to C2, symmetry. Both distance (0.9 A), and its marked resistance to fragmentation in 20 and 21 could be cyclized under extreme acid conditions to the mass spectrometer.

CH3O 0 Cl Na,NH3 9 steps --b-, "49CHI CH3I I

15W 24 25

HN NH2, 4CF3SH, H202 1 H2C12

27 26 28 Scheme VI Downloaded by guest on October 3, 2021 Review: Paquette Proc. Natl. Acad. Sci. USA 79 (1982) 4499 The situation at this point was such as to make pursuit of the parent (C)20 hydrocarbon irresistable. Utilization of 23 as a serviceable precursor was not viewed favorably because of the inoperability of Schleyer's catalyst systems (36-37) under the high-vacuum conditions necessary for its volatilization (D. XV. Balogh, personal communication) and the unlikely workability ofchemospecific methyl oxidation and ensuing decarboxylation in this instance (38). Furthermore, the entire synthetic se- quence would be protracted by the inclusion ofthese additional proposed transformations. As a result, we sought to adapt the 29 30 synthesis just described in as efficient a manner as possible. Highlv pertinent to this end was our eventual discovery of At about the same time, we succeeded in preparing aldehyde a method for transforming dichloro diester 15 into monoalkyl- 29. To our disappointment, this substance has proven to be too ated keto ester 24 by strict control of the amount of methyl io- sensitive to deal with on a preparative scale, being remarkably dide. Equally encouragingwas the expeditious mannerin which prone to oxidative and conversion to 30, pre- methylsecododecahedrane (26) was fashioned from this inter- sumably as a direct consequence ofits high content. Suit- mediate (Scheme VI). However, it was noted in the course of able resolution ofthis difficulty has mandated the introduction this study that treatment of 25 with trifluoromethanesulfonic of a blocking group whose appropriateness wvas to be dictated acid in dichloromethane as before resulted in isomerization to by its inertness toward several reagents (including light) and its a complex mixture ofseveral among which 27 was ready removal at a later stage. clearly not dominant (R. J. Ternansky, personal communica- As matters worked out, our first attempt in this direction was tion). By careful repeated recrystallization, the major constit- successful, use ofchloromethyl phenyl etheras the electrophile uent was obtained in a pure state and identified as the "iso- providing convenient access to 31. As shown in Scheme VII, dodecahedrane" 28 by x-ray crystal structure analysis (G. C. the phenoxymethyl side chain causes no interference during the Christoph, personal communication). Evidently, the substantial three-step "stitching together" of the other side of the sphere. steric compression present in the protonated form of25 is more Nor does it become entangled with proximate reactive func- than adequate to foster a most remarkable transannular elec- tional groups as 32 is modified to the aldehyde level and sub- trophilic substitution with inversion of configuration! To elim- jected to photochemical cyclization. Sequential Birch reduc- inate this reaction channel, the fully saturated seco hydrocarbon tion, acid hydrolysis, and pyridinium chlorochromate oxidation was prepared and subjected to dehydrogenation over palladium of33 proceed to make keto aldehyde 34 cleanly available. This on charcoal. The best conditions uncovered to this time appear intermediate, which is seen to be a /-keto aldehyde, conforms to produce 27 admixed with other equally nonpolar materials to the susceptibility of this class of compounds to retroaldol (R. J. Ternansky, personal communication). At present, we are cleavage in alkaline solution. Since the monoseco olefin derived actively engaged in enhancing the proportion of 27 in these from 35, like 25, is prone to uncontrolled carbocationic rear- mixtures and in isolating this substance in pure form. rangement, its diimide reduction was effected. When recourse Cl HO PhOCH.-

a-c d,e,o

31 32 33

Iffgqe

0

h

I~~~~~~~~~~~~~~~~~

36 35 34 a, hiv; b, TsOH, CAH6; c, NH2NHO, H202; d, (i-Bu)2AH; e? PCC, CH2CI2; f, Li, NH3, C2H5OH; g, HCI, H20, THF; h, KOH, H20, C2H5OH; i, 10% Pd-C, 2500C. Scheme V-II Downloaded by guest on October 3, 2021 4500 Review: Paquette Proc. Natil. Acad. Sci. USA 79 (1982) is made to catalytic dehydrogenation of the resulting secodo- 11. Repic, O. (1976) Dissertation (Harvard University, Cambridge, decahedrane, a mixture of compounds containing the desired MA). 12. Roberts, W. P. & Shoham, G. (1981) Lett., 4895. 36 is produced. Although this drawback makes isolation of do- 13. Jacobson, I. T. (1967) Acta Chem. Scand. 21, 2235. decahedrane difficult, this accomplishment has recently been 14. Paquette, L. A., Farnham, W. B. & Ley, S. V. (1975) J. Am. achieved (R. J. Ternansky, personal communication). Due to Chem. Soc. 97, 7273. severely limited quantities of 36 at this time, our study of its 15. Paquette, L. A., Itoh, I. & Farnham, W. B. (1975)J. Am. Chen. properties has not yet progressed far. Nonetheless, the expec- Soc. 97, 7280. tation that its 'H (8 3.35 in CDC13) and N3C (66.93 ppm in 16. Paquette, L. A., Itoh, I. & Lipkowitz, K. (1976)J. Org. Chern. has 41, 3524. CDC13) spectra should be characterized by lone singlets 17. Clardy, J., Solheim, B. A., Springer, J. P., Itoh, I. & Paquette, been borne out. L. A. (1979)J. Chem. Soc. Perkin Trans. 2, 296. Predictions concerning dodecahedrane abound (39-46). As 18. Deslongehamps, P. & Soucy, P. (1981) Tetrahedron 37, 4385. our synthetic approach becomes more streamlined and reason- 19. Eaton, P. & Mueller, R. H. (1972)]. Am. Chem. Soc. 94, 1014. able quantities of this hydrocarbon become available, we shall 20. Eaton, P. E., Mueller, R. H., Carlson, C. R., Cullison, D. A., explore the preparation of its amino derivative (a potential an- Cooper, C. F., Chou, T.-C. & Krebs, E.-P. (1977)J. Am. C/hern. tiviral and anti-parkinsonism agent) and carbonation (a stable, Soc. 99, 2751. 21. Eaton, P. E., Andrews, C. D., Krebs, E.-P. & Kunai, A. (1970) highly fluxional species?), as well as its conversion to dodeca- J. Org. Chem. 44, 3824. hedrene (a highly distorted cyclic olefin). In addition, it is an- 22. Eaton, P. E. (1979) Tetrahedron 35, 2189. ticipated that several intermediates in our synthetic scheme 23. Paquette, L. A., Snow, R. A., Muthard, J. L. & Cynkowski, T. might be successfully deployed to arrive at azadodecahedrane (1978) J. Am. Chern. Soc. 100, 1600. and the homododecahedryl cation [a totally degenerate (CH)21 24. Paquette, L. A., Snow, R. A., Muthard, J. L. & Cynkowski, T. species?]. Attention shall also be given to the possibility of en- (1979)J. Am. Chem. Soc. 101, 6991. 25. Jones, N. J., Deadman, W. D. & LeGoff, E. (1973) Tetrahedron larging the cavity by supplanting select symmetrically dis- Lett., 2087. 26. Paquette, L. A., Balogh, D. WV., Usha, R., Kountz, D. & Chris- toph, G. G. (1981) Science 211, 575. 27. Paquette, L. A. & Balogh, D. W. (1982)J. Am. Chemr. Soc. 104, 774. 28. Christoph, G. G., Engel, P., Usha, R., Balogh, D. W. & Pa- quette, L. A. (1982) J. Am. Chern. Soc. 104, 784. 29. Paquette, L. A. & Wyvratt, M. J. (1974)]. Am. Chern. Soc. 96, 4671. 30. McNeil, D., Vogt, B. R., Sudol, J. J., Theodoropulos, S. & He- daya, E. (1974) J. Am. Cher. Soc. 96, 4673. 31. Paquette, L. A., Wyvratt, M. J., Berk, H. C. & Moerck, R. E. 37 38 (1978)J. Am. C/er. Soc 100, 5845. 32. Paquette, L. A., Wyvratt, M. J., Schallner, O., Schneider, D. posed cyclopentane rings with larger six-membered ones as in F., Begley, W. J. & Blankenship, R. M. (1976)J. Am. Chemr. Soc. pentakaidecahedrane (37) and hexakaidecahedrane (38). XWith 98, 6744. these hydrocarbons in the possibility of sealing an atom 33. Paquette, L. A., Wyvratt, M. J., Schailner, O., Muthard, J. L., hand, Begley, W. J., Blankenship, R. M. & Balogh, D. WV. (1979) . such as in the interior becomes more realistic. Org. C/em. 44, 3616. Our accomplishments described herein are a reflection of the hard 34. Balogh, D. WV. & Paquette, L. A. (1980) J. Org. Chem. 45, 3038. work produced by several inspired students whose names are given in 35. Paquette, L. A., Balogh, D. W. & Blount, J. F. (1981) ]. Am. the references. Financial support for this program was generously pro- Chern. Soc. 103, 228. vided by the National Institutes of Health (Grant AI-11490). 36. Grubmflller, P., Schleyer, P. von R. & McKervey, M. A. (1979) Tetrahedron Lett., 181. 1. Maier, G., Pfriem, S., Schafer, U. & Matusch, L. (1978) Angew. 37. Grubmflller, P., Maier, W. F., Schleyer, P. von R. & McKeney, Chem. 90, 552-000. M. A. (1980) Chem. Ber. 113, 1989. 2. Maier, G., Pfriem, S., Schafer, U., Malsch, K.-D. & Matusch, 38. Paquette, L. A., Begley, XV. J., Balogh, D. WV., Wyvratt, M. J. R. (1981) Chem. Ber. 114, 3965-000. & Bremner, D. (1979) J. Org. Chem. 44, 3630. 3. Eaton, P. E. & Cole, T. W., Jr. (1964) J. Am. C/tem. Soc. 86, 39. Engler, E. M., Andose, J. D. & Schleyer, P. von R. (1973)J. Am. 3157. Chem. Soc. 95, 8005. 4. Grahn, W. (1981) Chem. Unserer Zeit 15, 52. 40. Ermer, 0. (19711) Angew. C/er. Int. Ed. EngI 16, 411. 5. -Cassar, L., Eaton, P. E. & Halpern, J. (1970)J. Am. Chem. Soc. 41. Schulman, J. M., Venanzl, T. & Disch, R. L. (1975) J. Am. 92, 6366. Chem. Soc. 97, 5335.

6. Byrd, J. E., Cassar, L., Eaton, P. B. & Halpern, J. (1970) J. 42. Schulman, J. M. & Disch, R. L. (1978) J. Am. Chem. Soc. 100, Chem. Soc. Chem. Commun. 40. 5677. 7. Cassar, L., Eaton, P. E. & Halpern, J. (1970)]. Am. C/er. Soc 43. Clark, T., Knox, T. McO., Mackde, H. & McKervey, M- A. (1975) 92, 3515. J. Chem. Soc. Cten. Commun., 666. 8. Paquette, L. A. (1978) Pure App/ Chem. 50, 1291. 44. Clark, T., Knox, T. McO., McKervey, M. A., Mackle, H. & Roo- 9. Paquette, L. A. (1981) in -Today and Tomor- ney, J. J. (1979)J. Am. Chem. Soc. 101, 2404. row, eds. Trost, B. M. & Hutchinson, C. R. (Pergamon, New 45. Dixon, D. A., Deerfield, D. & Graham, C. D. (1981) C/ten. York), p. 335. Phys. Let. 78, 161. 10. Woodward, R. B., Fukunaga, T. & Kelly, R. C. (1964) J. Am. 46. Disch, R. L. & Schulman, J. M. (1982) J. Am. Chem. Soc. 103, Chem. Soc. 86, 3162. 3297. Downloaded by guest on October 3, 2021