Aug. 20, 1964
COMMUSICATIONS TO THE EDITOR
Some examples of bromine insertion have also been observed. Photolysis of Ia (R = NO, X = H) in benzene containing excess CC13Br followed by oxidation of the product with chromium trioxide in acetone gave 18-bromoprednisone acetate (IIa, X = Br; 16(%), m.p. 172-175' dec., [ a ] D +179', 239 mp ( E 16,300). Similar photolysis and oxidation of 1lp-hydroxyandrosta-1,4-diene-3,17-dione nitrite6 gave 18-bromoandrosta- 1,4-diene-3,11,17-trione (18%), m.p. 183-186' dec., [ a ] D +212', XE:;H 239 m p ( e 17,200). The radical-exchange reaction has also been observed in the photolytic cleavage718of cyclic nitrites. Thus, photolysis of cyclopentanol nitrite gave 5-bromopentanal and 5-iodopentanal in the presence of CC13Br and of iodine, r e s p e ~ t i v e l y . ~ , ' ~ Acknowledgments.-We thank Dr. M. M. Pechet for his interest and encouragement and G. Berlinghieri, L. Salmansohn, and C. B. Pantuck for skillful technical assistance.
3395
P Y
I
Figure 1
(7) P. Kabasakalian and E. R . Townley, J . D v g , Chem., 27, 2918 (1962). (8) M. Akhtar and D. H. R . Barton, J . A m . Chem. Soc., 86, 1628 (1964). (9) Infrared and nuclear magnetic resonance spectra and microanalytical d a t a in accord with t h e structures assigned have been obtained for all new compounds. (10) For the photolytic rearrangement of iodohydrins formed i n situ see K . Heusler and J . Kalvoda, Helv. Chim. A d o . , 46, 2732 (1963), and earlier papers there cited.
collected with Mo Koc radiation with Weissenberg geometry employing a scintillation counter 'for a detector. The 875 observed intensities were transformed to structure amplitudes without any correction for absorption. The structure was solved using the iodine as a heavy atom and refined with a full-matrix T H E RESEARCH INSTITUTE FOR M. AKHTAR least-squares program.8 MEDICIXE A N D CHEMISTRY D. H. R. BARTON The final disagreement factor, R,was 11%. The CAMBRIDGE, MASSACHUSETTS P. G. SAMMES structure is shown in Fig. 1. The standard deviations RECEIVED J U N E13, 1964 of the bond distances are quite large (0.04 ; this is probably due to preferential absorption which was The Absolute Configuration of neglected in correcting the intensity data. The Menthyl Arenesulfinatesl average for the benzene bond lengths is 1.38 A.,and for Sir: the carbon-carbon saturated bond lengths it is 1.53 A. The absolute configuration around the sulfur atom The ready availability of optically active sulfoxides can be determined in this structure because the absoby the Grignard s y n t h e s i ~ ~has - ~ sparked new interest lute configuration of the (-)-menthyl group is known.g in the question of the absolute configuration of the preAs suggested by Mathieson, lo the absolute configuration cursor menthyl arenesulfinates. Andersen? assigned around an atom can be determined by ordinary Xthe (S)-configuration to the asymmetric sulfur in ( - ) ray techniques if an internal known asymmetric menthyl (-)-p-toluenesulfinate [(-)-I]; by comparing reference center is present in the molecule.ll In the the rotation of (-)-I with t h a t of (-)-menthyl (-)-$Ipresent case our internal reference is the (-)-menthyl iodobenzenesulfinate [( -)-I1l5 a t the D-line; Hergroup. Using the above method the absolute conjigurabrandson and Cusano6 had previously made a "tentation about the sulfur atom in (-)-I1i s ( S ) . tive" assignment of the (S)-configuration to the asymThe configurations of I and I1 were correlated as metric sulfur in (-)-I1 on the basis of kinetic and thermofollows. Reaction of p-toluenesulfinyl chloride and dynamic studies. Most recently, Andersen7 has rever(-)-menthol a t - 78' in ether-pyridine afforded a sed himself and has claimed the (R)-configuration for diastereomeric mixture of ( - )-menthyl p-toluenesul(-)-I and (-)-11. We now present unequivocal evifinates, [ a ] -38.5' ~ (Ia)," which was converted to dence which forces a choice between these two assignp-tolyl butyl sulfoxide (IIIa), [ a ] D -5l', by reaction ments. with an excess of n-butylmagnesium bromide. OptiCrystals of (-)-I1 are monoclinic with the space cally pure 111,prepared from n-butylmagnesium bromide group P21 and a density of 1.33 g . / ~ m . ~The . cell and diastereomerically pure (-)-I ( [ a ] D -206'), constants are a = 17.26 A., b = 7.656 f i . , c = 6.925 ~ Therefore the ratio of enantiomers A., and y = 105.1'. This requires two molecules per has [ a ] +187'. in I I I a is 36.4% (+)-111:63.6% (-)-HI. Th'IS acunit cell. Three-dimensional intensity data were curately reflects the ratio of diastereomers in precursor (1) We gratefully acknowledge support by t h e National Science FoundaIa (36.4% (-)-I : 63.6% ( + ) - I ) , as shown by an anciltion (grant GP-757) for the work conducted a t New York University and by the Public Health Service (grant AM-5785) for the work conducted a t lary experiment. A mixture ([oc]D - 119') of Ia (3.256 t h e University of Chicago g.) and (-)-I (3.000 g.) was subjected to the Grignard (2) K . K . Andersen, Tetvahedvon Letlevs, No, 3, 93 (1962).
A,)
(3) K Mislow, A L Ternay, J r , and J T. hlelillo, J . A m . Chem. Soc., 80, 2329 (1963). (4) C . J . hl Stirling, J . Chem. Soc., 5741 (1963). (5) T h e diastereomers of I and I1 prepared from (-)-menthol have opposite signs of rotation a t t h e D-line; t h e prefixes of I and 11 refer t o these signs. (6) H F Herbrandson and C . ?J. Cusano, J . A m . Chem. Soc., 83, 2124 (1961) (7) K . K . Andersen, J . O v g . Chem., 29, 1953 (1964).
(8) T h e observed structure amplitudes and the final coordinates for the structure are available upon request from E . B. F (9) V . Prelog, H e l v Chim Acla, 36, 308 (1953). (10) A M . Mathieson, Acla Cvyst , 9 , 317 (1956) (11) A pertinent example of this technique is t h e establishment of t h e absolute configuration a t sulfur in ( +)-methylcysteine sulfoxide (R.Hine and D. Rogers, Chem. I n d (London). 1428 (1956). (12) Except for the p-toluenesulfinates, whose rotations were measured in nitrobenzene, all rotations refer t o acetone solution.
3396
COMiMUSICATIOKS TO
reaction with n-butylmagnesium bromide. The product I I I b had [ a ] +63.2" ~ and thus was a mixture of 66.90/, (+)-I11 :33.17, ( - ) - I I I , in perfect accord with the figures predicted from the composition of the precursor. Reaction of p-iodobenzenesulfinyl chloride and ( - ) menthol under identical conditions gave a diastereomeric mixture of ( - )-menthyl p-iodobenzenesulfinates, [ a ] -20.9' ~ ( H a ) . From the rotations of the pure diastereoineric components6 the composition of I I a is 25.9% (-)-I1 : 74.17, ( + ) - I I . In the synthesis of I and 11, the (+) isomer is the more rapidly formed. Since steric factors unquestionably control the relative abundance of the diastereomeric product ratios, it follows that ( - ) - I and (-)-I1 have corresponding conjigiuations. This conclusion is buttressed by the following observation. Reaction of 1-butanesulfinyl chloride and (-)-menthol a t - T8' gave a diastereomeric mixture of (-)-menthyl 1butanesulfinates which afforded IIIc, [ a ] D +99', on treatment with p-tolylmagnesium bromide. Since the signs of I I I a and I I I c are opposite, i t follows t h a t the absolute configuration around sulfur in the predominant diastereomer leading to I I I a and I I I c must be the same. We shall discuss in a separate paper the relevance of the present findings to the resolution of the other discrepancy which we had earlier remarked upon.
THE
EDITOR
Vol. 86
The di-p-tolyl (Ib) and di-p-bromophenyl (IC) compounds were prepared from benzenediazonium-2: carboxylate and the appropriately substituted phenylacetylene. Similar generation of benzyne in the presence of methyl propiolate yielded the 3,B-dicarbomethoxy compound (Ie). .In equimolar mixture of phenylacetylene and p-bromophenylacetylene reacted with benzyne to furnish the monobromo compound (Id) in addition to I a and IC. Treatment of either ICor Id with butyllithiuni, followed by hydrolysis, succeeded in removing bromine to produce I a in yields of 33% and 337,. respectively. Each of the 5,F-disubstituted dibenzocyclooctatetraenes Ib-e underwent isomerization to the 5 , l l - isomer under conditions similar to those described for Ia. The yields of IIb-e were 937,, 8iyG,95yG,and 75%, respectively. Table I lists first-order rate constants and activation parameters for the rearrangement of Ia-d. TABLE I RATESOF REARRANGEMENT OF I Compd.
Solvent (temp., " C . )
105k, set.-'
TO
11"
A H * , kcal
A S * , e.u.
Decalin (164.5) 4.8 28.9 - 13 Triglyme (164.0) 4.6 29.1 - 13 7 1 29.1 - 12 Ib Decalin (164.9) 31.7 IC Decalin (164.8) 5.3 Id Decalin (164.8) 4.9 30 1 - 10 a The reactions were followed by spectrometric determination The rates were cleanly first EVERLY B. FLEISCHER of product appearance a t A,,. order over the measured course of the reaction, which was one t o three half-lives.
DEPARTMEM OF CHEMISTRY UNIVERSITY OF CHICAGO CHICAGO, ILLISOIS 60637 DEPARTMENT O F CHEMISTRY MICHAELAXELROD X E W E-ORKUSIVERSITY MARKGREEN NEXVYORK,A-EIV YORK 10453 KURTiMIsLon RECEIVED JUSE 18, 1964
Thermal and Photochemical Rearrangement of Substituted Dibenzo [a,e]cyclooctatetraenes
Sir: 3,6-Diphenyldibenzo [a,e]cyclooctatetraene (Ia) is available from the reaction of phenylacetylene with benzyne. Heating this hydrocarbon to temperatures of 14i)-2OUo, either as the pure melt or as a solution in decalin or triglyme, leads to the remarkable rearrangement shown in eq. 1.2
\
I
R2
TI a , It1 = It2 = C 6 H j ; b, R, = RS = p-CHBC6HI;c, R1 = R ? = P-BrC6HI; d , K I = C&; KS = p-BrCsH,; e, R I = Rz = COSCH3
Hydrocarbon I I a [A,, 260 mp ( E 34,700)] was formed in 967, yield. Hydrogenation (Pd-charcoal) furnished a tetrahydro derivative, m.p. 174-1745', the n.ni.r. spectrum of which exhibited a ratio of aromatic : methine :methylene hydrogens equal to 18.2: 2.0 : 3.9. Ozonization of I I a yielded o-benzoylbenzoic acid. (1) 51. Stiles, U Burckhardt, and 4 . Haag, J . Org. Chem., 27, 4715 (1962) ( 2 ) Satisfactory elemental analyses were obtained for all t h e compounds described klelting points obtained were: Ih, 168-16Q3, l c 222-222.5'; I d , 197 .5i1980, Ie, 176 5-177'; I I a , 183-184'. I I b , 181-183°, IIc, 201204.5O. 1 I d 17&1713, IIe. 187 5-188'.
Ia
-
Isomerization of an equimolar mixture of I a and IC in triglyme a t 193" produced a 99% yield of a mixture of I I a and IIc, with no contamination by I I d , as shown by thin-layer chromatography. Furthermore, the isomerization of Id yielded IId which contained no detectable I I a or IIc. These two experiments eliminate an obvious possibility for the mechanism, a dissociation-recombination such as t h a t of eq. 2 . R
R
R
R
A reaction path which would explain most simply the rearrangement of I to I1 includes the "twisted" intermediate 111. The structure of 111 is such t h a t reversion to the dibenzocyclooctatetraene structure could lead to either I or 11. The feasibility of structure I11 as an intermediate may be argued from the results of Srinivasan,3 who reported that photolysis of cis,cis-l,5cyclooctadiene yielded tricyclo [4.2.0.026]octane, which possesses the carbon skeleton central to 111.
&jfJ \
I11 Photolysis of a solution of hydrocarbon I a in methylcyclohexane (Hanovia lamp, Pyrex filter) a t 0' produced a 367, yield of isomer I I a The remainder was starting material, with no evidence for other isomers by thinlayer chromatography. (3) R Srini\aqan J A m Chem Soc
86,819 3048 (1963)
