April 20, 1964
COMMUNICATIONS TO THE EDITOR
0
0
IV V One of the most interesting properties of p-ketosulfoxonium ylides is their susceptibility to photochemical fission of the dipolar S-C bond by ultraviolet irradiation. Theylides 11, R = C ~ Hor S cyclo-C~H1l,upon excitation with 233 mp radiation (low-pressure mercury arc) are transformed smoothly according to eq. 3, presumably via keto carbene and ketene intermediates. The combination of (1) and (3) is a process of chain extension
RCOCHS+O(CH~)~
hv
--+ RCH~COOR'+ ( C H ~ ) ~ S O (3) R'OH
I1
R'
H, CHI
which is analogous to the Arndt-Eistert sequence. Cyclic 6-ketosulfoxonium ylides decompose by a pathway which parallels (3). Irradiation of the cyclic ylide I11 in methanol a t 253 mp produces the sulfoxide ester VI, infrared maximum 5.74. 9.5 p , in 80% yield; in water the corresponding carboxylic acid is formed. This reaction opens a new route to carbocycles: oxidation of VI with peracetic acid affords the corresponding sulfone which undergoes cyclization with potassium tbutoxide t o give the keto sulfone VII, R = S02CH3, and this in turn is converted by treatment with aluminum amalgam4 to trans-hydrindan-2-one (VII. R = H, 60% over-all yield from 111), shown to be the pure
1641
from I and ethyl acrylate; m.p. 181-185' dec., ultraviolet maximum 241 mp ( e 15,800). The transformations described above are easily carried out and are of practical value in synthesis. The photochemical reactions are especially interesting, and these are being studied further. (8) National Institutes of Health Postdoctoral Fellow 1962-1964.
E. J . COREY DEPARTMEXT O F CHEMISTRY HARVARD UNIVERSITY MICHAEL CHAYKOVSKY~ CAMBRIDGE 38, MASSACHUSETTS RECEIVED FEBRCARY 4, 1964
A New Synthetic Approach to Medium-Size Carbocyclic Systems
Sir : In the first stage of an investigation directed a t the total synthesis of the naturally occurring substance humulene (I),a study has been made of the reaction of a doubly allylic a,w-dihalide with nickel carbonyl as a model cyclization process. This communication describes the results of these experiments' and the successful synthesis of 1,6-dimethylcyclododeca-l,5,9-triene (TI); the new knowledge which has been obtained suggests simple synthetic routes to humulene and related structures and thus provides a key to these hitherto unsolved problems.
I &H
hv
+SO(CH3)CH30H I11
~
Dec-5-ene-2,g-dione (111)* was synthesized by alkylation of ethyl 6-dimethylaminocrotonate with trans-
&,oocH3 CH2SOCH3
trans
._ I
H VI
C H 3 C O C H z C ~ z ~ H = ~ ~ ~ ~ 2 ~ ~ 2 ~ I11
H
trans
1.101 2. 1-BuO'
RCH=CCHzCHzCH=CHCHzCHzC=CHR I
R VI1
trans isomer by vapor phase chromatography and conversion to known solid derivatives5 The stereospecific formation of trans-hydrindan-2-one is of special interest. Preliminary studies indicate that the reaction of I with ethyl esters of A'-cycloalkenyl carboxylic acids produces trans-fused cyclic ylides in preference to the cis isomers. Starting from ethyl cinnamate and I , 3-phenylcyclopentanone6 was synthesized via the cyclic P-ketosulfoxonium ylide' by a sequence corresponding to I11 + VI1 in 30Yc over-all yield. This new carbocycle synthesis can in some cases be modified t o permit the synthesis of cis-fused ring systems. For example, the ylide IT, R = A'-cyclopentyl, undergoes cyclization upon heating in methanol a t reflux (15 hr.) to give a cyclic product (55%) analogous to I11 but largely cis-fused since it is converted to cispentalan-%one by a sequence of the type I11 -+ VII. The parent member of the new class of cyclic pketosulfoxonium ylides VI11 has also been prepared J . Corey a n d hl. Chaykovsky, J . A m . Chem S o c . , 86, 1639 (1964). ( 5 ) W . Huckel a n d H. Friedrich, A n n . , 461, 132 (1926). (6) W . Borsche a n d W Menz, Chem. Ber., 41, 190 (1908) (7) E . J Corey a n d M Chaykovsky, S'efrahedron L e l l w s , No. 4, 169 (1963) (4) E
CHI
IT
CH3
1,4-dibromo-2-butene8 followed by hydrolysis and decarboxylation ; m.p. 41-42', infrared maximum 5.81 1.1, n.m.r. peaks at 5.20 6 (2 olefinic H ) and 1.98 6 ( 2 CH,C0).4 The diketone I11 was converted t o diethyl 3,10-dimethyldodeca-2,6,1O-triene-l, 12-dioate (IV, R = COOCzHs) by the procedure of U'adsworth and Emmons,6 and this was reduced by lithium aluminum hydride to 1,12-dihydroxy-3,10-dimethy1dodeca-2,6,10triene (IV, R = CHzOH). The n.m.r. spectrum of the diol IV, R = CHzOH, so obtained indicated i t to be a mixture of SOYc all-trans isomer and 20Yc cis, trans, trans isomer0; as expected the XTittig process leading to (1) This approach to cyclization seemed worthy of s t u d y because of t h e known coupling of simple allylic halides by nickel carbonyl to give 1 . 5 dienes [I. D . Webb and G . T. Borcherdt, J A m . Chem S o c . , 73, 2654 (195131 a n d the catalytic action of low-valent nickel compounds i n t h e trimerization of 1,3-butadiene to 1,.5,9-cyclododecatrienes[G. Wilke, Angew Chem , 76 10 (1963)l. M o r e recently t h e coupling of allylic acetates has been reported [ S .L. Bauld, Telrahedron Lellers, No. 19, 859 (1962)l. (2) Satisfactory analytical and spectroscopic d a t a were obtained f o r a I1 new compounds reported. (3) See ( a ) S. A . Glickman and A . C C o p e , J . A m . Chem S o c . , 67, 1017 , (19.53) (1945); ( b ) G Eglinton and M. C . Whiting, J . Chem. S O L 3052 (4) l'.m,r. chemical shifts expressed as p p . m . downfield displacement from tetramethylsilane a s internal standard and measured a t 60 A4c ( 5 ) W . S. Wadsworth and W . D E m m o n s , J . A m . C h e m . SOL.,83, 1733 (1961) (6) See R. B . Bates and D. M . Gale, ibid.,82, 5749 f l 9 6 0 ) , for the n . m . r methodology
1642
COMMCNICATIONS TO THE EDITOR
Vol. 80
Dicarbaundecaborane( 13) and Derivatives IV, K = COOC~HS,was not stereospecific. The alltrans isomer of the diol IV, R = C H 2 0 H ,was obtained Sir : in pure condition (as a colorless oil) via the crystalline Carborane [ 1,2-dicarbaclovododecaborane( 12) ] and ;3,3-dinitrobenzoate derivative, m.p. 136- 137'. The its derivatives, prepared from the reaction of BI0Hl2pure diol IV, K = C H 2 0 H , and the 4 : 1 mixture of (ligand)* compounds and acetylenes, l 2 have been stereoisomeric diols were transformed into 1,12-dibromoselectively degraded by rnethanolic potassium hydroxide ~3,lO-dimethyldodeca-2,6,10-triene (IV, R = CH2Br)by a t 40'. The degradation reaction is described by the reaction with phosphorus tribromide and pyridine in equation ethereal solution a t -15'; the two products had esB I o C ~ H+I ~CH30- + BCHIOH + sentially identical infrared and n.m.r. spectra. Slow H? B(OCHs)3 B s C ~ H I ~ addition of the dibromide IV, R = CH2Br, to a stirred solution of nickel carbonyl (excess) in 1,2-dimethoxyOne mole each of hydrogen and boric ester were proethane a t 78' followed by continued heating for 12 hr., duced and 1 equiv. of base was consumed per mole of removal of solVent under reduced pressure! partitioning carborane employed, The BgC2H12anion was precipibetween ether and water, and distillation of the product tated as its tetramethylammonium salt I (98% yield) from the organic layer a t 100' bath temperature (6 upon the addition of aqueous tetramethylammonium mm.) gave 03-G8% yield of I1 as a colorless oil, appearchloride. :1nal. Calcd. for B9C2H24N:B, 4G.90; C , ing homogeneous in vapor phase chromatographv 34.70; H , 11.65; N , 6.74. Found: B, 46.90; C , using fluorosilicone, tris(P-cyanoethoxy) propane, and 35.61; H , 12.30; N, 6.60. The corresponding salts diethylene glycol succinate columns. The product were prepared from C-methyl- and C-phenylcarboranes. showed a parent peak in the mass spectrum a t m/e = The infrared spectra of the B9C2H12anion exhibited a 190 and absorbed 3.16 molar equivalents of hydrogen sharp C-H stretching band a t 3.35 p and B-H stretchbased on molecular weight 190 using prereduced ing centered a t 4.0 p . No absorptions which could be platinum oxide catalvst in acetic acid. Ozonolysis7 of attributed to B-H-B bridge stretching were observed. the triene followed by reduction of the total mixture Acidification of a diethyl ether slurry of I with with sodium borohydride and acetylation gave a liquid anhydrous hydrogen chloride followed by sublimation product which was shown by v.p.c. and spectral analy(40') of the product retained in the ethereal solvent sis to be a lG:i..? mixture of pentane-l,l-diol diacetate produced dicarbaundecaborane( 13) (11) in 89% yield. and butane-l,l-diol diacetate (91Yc of the theoretical Elemental analysis of BgCzHla was thwarted by its yield). These facts clearly establish that the product low stability and hydroscopic nature, m.p. 110' dec. from the reaction of dibrornide IV, R = CH?Br, with However. I1 was titrated as a monoprotic acid (equiv. nickel carbonyl must be 1 ,G-dimethylcyclododeca-l,5.9-wt. 1 3 7 ; ' theory 131) with an apparent pK,\ of 2.95 triene. in 33% (volume) methanol-water. X cryoscopic molecAnalysis of the triene so produced by thin layer chroular weight of 137 was observed in benzene solution. matography (on silica gel-silver nitrate) and n.ni.r. The infrared spectrum of I1 displayed a C-H stretching analysis indicated t h a t the material was, in fact, a mixband a t :?.30 p and a B-H stretching band a t 3.85 p , ture of two isomers in a ratio of ca. 2 : 1. The same rnixX weak absorption a t 5.10 p could be reasonably attribture was obtained starting from the pure all-trans diol uted to a B-H-B bridge absorption although the H ' I V , I< = C H 2 0 H , or the 4 : 1 mixture described above. n.m.r. spectrum of I1 did not reveal clean evidence for .Uthough it is clear that one or both of the last two steps such an interaction. i n the synthesis of the cyclic triene may be nonstereoThe degradation reaction was examined kinetically specific, we are inclined to believe that most of the isomwith hydroxide ion in 507c volume aqueous ethanol erization occurs a t the last stage. Further work is in a t 73" and with C-phenylcarborane as the substrate. progress on this point. The major isomer, which can Hydrogen evolution was employed to monitor the rebe separated from the mixture by column chromatogaction. The reaction was first order in both hydroxide raphy using 10% silver nitrate on silica gel, is probably ion and C-phenylcarborane with a second-order rate the all-trans structure 11; the n.m.r. spectrum shows constant of 1.48 f 0.03 X l o p 4 1. mole-' set.-'. two sharp methyl peaks (GH) a t 1.39 and 1.42 6,8 a peak The similar reaction of C-phenylcarborane labeled in the a t 1.9 6 (12H), and broad olefinic proton absorption a t 9, 10, and 12 positions? with deuterium produced ail 4,:3-4,9 6 (4H). The minor isomer, presumably the cis, identical rate constant and 97yo pure hydrogen was trans, irans form of 11, shows methyl absorption a t 1 .5S evolved. These results clearly indicate that the 8, 9, 10, p.p.111,.closer to a normal value for C=C-CH3. ( I ) C C Clark, L ' S P a t e n t 3,0li2.756 ( N o v fi, 1962), T 1,. H r y i n g . J Further work on the reactions of nickel carbonyl with \Y Ager, J r , S I. Clark, I) J hlangold, H L. Goldstein, hI H i l l m a n . I< J Polak, and J \V Szymanski. l n o r g C h e m , 2 , 1089 (19631, H Schl-oeder-. dihalides and related substances is in progress with the T 1. H e y i n g , and J K K e i n e r , i b i d , 2 , 1092 (19631, T I- Heying. J W object of investigating the scope and mechanism of this Ager. J r . . S I, C l a r k , I< P Alexander. S P a p e t t i , J . A Reid, and S 1 process and applying the ring closure to the synthesis of T r o t z . i b i d , 2 , 1097 ( I g M ) , S Papetti and T 1. Heying, i b i d , 2 , 1105 ( 1 9 6 3 ) , R P Alexander and H Schroeder, i b i L 2 , 1107 (19fiB); hf 51 natural products.q
+
C h e m . S O L ,4086 (1963). nd A. G . H o r t m a n n , J . A m . C h e m Soc
,
his work w a s pat tially suppoi ted by t h e National Science F o u n d a -
, 10)
Elmer Petel- Kohlel- Fellow, Harvard University. 1963-1964
I ~ E P A R T M LOF N TCHEMISTRY HARVARD I-SIVERSIT~. CAXBRIDGE
E. J . COREY ERSESTH A M A S A K A ' ~
:38, hfhSS.4CHL'SETTS RECEIVED FERRCARY
4, 1964
+
F e i n , J Bobinski, N Mayes. N . Schwa!-tz, a n d Rf. S. C o h e n , i b i i l , 2 , 1111 (1OC,8); RI 11 Fein, I 1 Grafstein, J . B Paustian, J Bobinski, B. Irl I.ichstein. Z I I a y e s , S S Schmartz, and h l . S Cohen, ibid , Z, 1115 ( 1 9 6 8 ) , 11 O r a f s t e i n , J . Bobinski. J l l v o r a k , H S m i t h , S S c h n a r t z , hI S C o h e n . and I\I h l , F e i n , i b i d , 2 , 1120,1R(j,3), I 1 Grafstein, J . Bobinski, J Llvorak, J . E P a u s t i a n , H . P S m i t h , S Karlan. C Vogel, a n d hl M. F e i n , i b r d , 2 , 112.5 (1SGR); I3 Chafitein a n d J I 3 v o r a k . ibid , 2 , 1128 ( 1 Y A 3 ) ( 2 ) Cai-borane(l2) I S xeasonahly assumed t o have a n icosahedral s t r u c t u r e with t w o near-neighbur- catbon a t o m s placed a t positions 1 a n d 2 The numhering system employed was adopted b y t h e S o m e n c l a t u r e C o m m i t t e e of t h e Inor-ganic 13ivisiiin of t h e .4met-ican Chemical Society See R A d a m % iilOlE rizp,,l,, a , 1087 8 I O I K
