874 [CONTRIBUTION No. 2026
DOUGLAS E. XPPLEQUISTAND J O l i K D. FROM THE
ROBERTS
Vol. 7 8
GATESAND CRELLINLABORATORIES OF CHEMISTRY, CALIFORNIA INSTITUTE OF TECHNOLOGY]
Small-Ring Compounds. XII. The Bromine Addition Products of Methylenecyclobutane' BY DOUGLAS E. APPLEQUIST~ JOHN D. ROBERTS RECEIVED AUGUST22, 1955
The dibromide(s) (I) from addition of bromine to methyl-group infrared absorptions. 111 was likemethylenecyclobutane has been studied by several wise saturated to permanganate and showed no worker^^-^ but the structural evidence is confusing double-bond or methyl absorption in the infrared. and inconclusive. Compound I has been reported I11 reacted with sodium iodide in acetone on to be reconverted to methylenecyclobutane by warming to give only a white precipitate. I11 treatment with z i n ~ and ~ . ~hydrolyzed to cyclo- reacted instantly with alcoholic silver nitrate a t pentanone in the presence of lead d i ~ x i d e . ~ JI t room temperature. I V rapidly reduced permanreacts with potassium cyanide in aqueous ethanol ganate in acetone, reacted with sodium iodide in to give an unidentified dinitrile16-10 and with acetone a t room temperature to give a white predimethylamine or trimethylamine to give both cipitate, and showed a strong C=C stretching band saturated (permanganate) and unsaturated prod- a t 6.05 p . The infrared spectrum showed no ucts in which both bromines are d i ~ p l a c e d . ~ ~methyl * group absorption. Dibromides prepared from methylenecyclobutane The assignment of structures 11-IV is in agreeeither by direct bromination or by Ziegler bromina- ment with mechanistic expectations if the bromine tion have been found to contain unsaturated addition goes by way of a non-classical carboniummaterial by permanganate test and possess boiling ion intermediate] which might be represented as ranges of 4-7°.7 V.I1 That the three dibromides are actually Slobodin and Shokhorgconsidered the dibromide, prepared in ice-cold ether solution, to be l-bromomethyl-1-bromocyclobutane (11) and reported that i t was converted to pure 1-(hydroxymethy1)I1 + 111 + I\' v cyclobutanol diacetate by treatment with silver acetate in acetic acid, but no satisfactory proofs of anionotropically related was proved by treatment structure were presented. of mixtures of variable compositions with zinc We have found that addition of bromine to bromide a t room temperature] which caused intermethylenecyclobutane in carbon tetrachloride gives conversions of the isomers. The equilibrium cona mixture of three dibromides 11-IV, containing stant between I1 and I11 was found to be 5.7 approximately 617, of 11, 327, of 111, and 6% of 2.1 in favor of 11 a t room temperature, as expected IV. Partial separation of the mixture was effected from strain considerations. The concentration of I V increased in every run, so its relative free energy was not determined. The structural assignments of I1 and I11 were confirmed also by their nuclear-magnetic-resonance I1 111 IV spectra.I2 by fractional distillation] which afforded I1 in Experimental' about 93% purity as judged by infrared analysis. Methylenecyclobutane Dibromides.-To a stirred soluPure I11 was obtained by treatment of the crude tion of 33 g. ( 0 . 5 mole) of methylenecyclob~tane'~ (b.p. bromination mixture with powdered potassium 41.0-42.0", 739 mm.) in 100 ml. of carbon tetrachloride was hydroxide] which selectively destroyed I1 and IV. added 73.6 g. (0.46 mole) of bromine in 50 ml. of carbon The properties of I V were inferred by comparison tetrachloride. The reaction temperature was kept between of the properties of fractions containing only I11 -5' and -2". Simple distillation of the reaction mixture gave a forerun, b.p. 28-48" (3.4 mm.), 6.2 g.; the dibromide and I V with those of pure 111. mixture, b.p. 48-59' (3.4 mm.), 72.3 g.; and a dark residue, I1 was saturated to permanganate, reacted with 12.3 g. sodium iodide in acetone to give a red color and a Part of the dibromide mixture was distilled through a 61white precipitate, and showed no double-bond or cm. modified-Podbielniak (spiral-wire) column with a total
*
(1) Supported i n p a r t b y t h e Petroleum Research F u n d of t h e American Chemical Society. ( 2 ) National Science Foundation Predoctoral Fellow, 1953-1955. (3) G. Gustavson, J . p r a k l . C h e m . , [2] 54, 97 (1896). (4) G. Gustavson a n d H . Bulatoff, ibid., 66, 9 3 (1897). ( 5 ) H. F e c h t , Ber., 40, 3883 (1907). (6) N. J. Demjanow a n d M. Dojarenko, ibid., 65, 2730 (1922). (7) E. R. Buchman a n d D. R. Howton, THISJ O U R N A L , 70, 2517 (1948). (8) E. R. B u c h m a n a n d D . R . Howton, unpublished results. (9) Ya. M . Slobodin a n d I. N. Shokhor, Sbornik Stafei Obschei Khim.,2 , 846 (1953); C. A , , 49, 6813 (1955). (10) (a) K. Llereshkovski, J . Russ. P h y s . C h e m Sac , 46,517(1911): .I Chem. S o r . , 106, 0 3 3 ( 1 9 1 4 ) ; ( h ) R C i i r t i s a n d T. K r n n e r . ibid., 105, 282 (1014).
reflux head, and collected in 14 fractions over the range 53-
(11) (a) J. D. Roberts a n d R . H. Mazur, THISJ O U R N A L , 73, 25GQ (1951); (b) J . D . Roberts a n d R . H. Mazur, ibid., 73, 3642 (1951); ( c ) J . D. Roberts a n d E. P. C o x , unpublished work. (12) Determined b y Dr. C. A. Reilly of t h e Shell Development Corp., to whom t h e a u t h o r s are indebted. I t was found possible t,i distinguish anddetermine quantitatively b y graphical integration methylene hydrogens of t y p e s C-CHz-C a n d C-CHI-Br. T h e ratio of t h e former to t h e l a t t e r was 0.985 (theoretical 1.000) in TI1 a n d 2.94 (theoretical 3.00) in 11. (13) Boiling points are uncorrected. I n f r a r e d spectra r e r e ohtained with a Model 21 Perkin-Elmer spectrophotometer. Microanalyses were performed b y D r . A. Elek. (14) J. n. Roberts and C. W. Sauer, 'Trirs J O T I K N A I . , 71, 3 ! ) 2 3 (1949).
Feb. 20. 1956
BROMINE ADDITION PRODUCTS OF METHYLENECYCLOBUTANE
65' (5 mm.). Indices of refraction were not useful in following composition changes during the distillation, but the infrared spectra of the various fractions revealed that the mixture probably contained three main constituents and one minor one (VI) which appeared only in the lowest-boiling fractions. From these spectra and from the spectrum OF pure 111, extinction coefficients were deduced for useful analytical peaks for each of the four components of the mixture. The fractions were analyzed by the base-line method.i6 The compositions were plotted against total distillate and the curves integrated graphically to obtain the over-all composition: 11, 61%; 111, 32%; IV, 6%; and the fourth constituent VI, 1%. The fist five fractions, b.p. 53-58', contained primarily 11, so were combined and redistilled through the same column. A fraction was obtained of b.p. 56.156.6" (5.1 mm.), 12% 1.5370, which had an infrared spectrum (Fig. 1) indicating 93% of 11, about 5% of 111, and about 4% of VI. The last four fractions of the f i s t distillation, b.p. 65.0-65.0' (5.3 mm.), contained primarily I11 and IV, and neither of these could be separated conveniently by distillation. Chemical tests on the various fractions from the first fractional distillation showed that fraction 4 was nenrly inert to permanganate in acetone, while fraction 13reacted rapidly with that reagent. Fraction 4 (containing 11, I11 and VI) reacted with sodium iodide in acetone to give a white precipitate and a red color, both a t a moderate rate, while fraction 7 (68% 11, 29% 111, 3% IV, and no trace of VI) reacted with the same reagent to give a white precipitate readily and a red solution slowly. Fraction 13 (containing only I11 and IV) gave a white precipitate rapidly but no color. A carbon-hydrogen analysis was obtained on fraction 13. Anal. Calcd. for CrHaBrz: C, 26.34; H , 3.54. Found: C, 26.41; H , 3.58. Isolation of 1,l-Bis-(bromomethy1)-cyclopropane (III).A mixture of 8.74 g. of mixed methylenecyclobutane dibromides (over 50% I11 from infrared analysis) and 13.1 g. of powdered potassium hydroxide was heated a t 100' for 2.5 hours. The reaction mixture was diluted with ether and filtered. Distillation of the filtrate through a semi-micro column gave 2.06 g. of 111, b.p. 63.7-64.0' (5.1 mm.), lt%D 1.5346. There was a viscous residue (0.5 g.) and a substantial forerun of lower-boiling, unsaturated material, which was not further investigated. The infrared spectrum of I11 is shown in Fig. 1. I11 was unreactive to permanganate in acetone. It reacted with sodium iodide in acetone a t 50' and with alcoholic silver nitrate a t room temperature t o give precipitates of sodium and silver bromides, respectively. Anal. Calcd. for CsHsBrz: C, 26.34; H , 3.54. Found: C, 26.48; H, 3.55. Isomerization of Methylenecyclobutane Dibromides.In each experiment, a few drops of dibromide mixture was treated with a small amount of powdered zinc bromide a t room temperature for 17 hr. The volatile material was removed in vacuo with gentle warming and analyzed by its infrared spectrum. A very small amount of undistillable tar was formed in each reaction. The results are shown in Table I. (15) J. J. Heigl, M . F. Bell and J. U. White, Anal. Chcm., 19, 293 (1947).
87 5
Fig. 1.-Infrared spectra taken in 0.054-mm. cells (no solvent) with a Perkin-Elmer model 21 spectrophotometer and NaCl prism: A, 1,l-bis-( bromomethy1)-cyclopropane (111); B, I-bromomethyl-1-bromocyclobutane(11), about 93% containing about 5% of 111 and 4% of VI.
TABLE I ISOMERIZATION OF METHYLENECYCLOBUTANE DIBROMIDES WITH ZINCBROMIDE AT ROOM TEMPERATURE
I1 I11 IV VI Total
7.7 71.3 14.2 0.0 93.3
36.6 19.2 38.2 0.0 94.0
89.0 4.0 0.0 5.1 98.1
80.9 8.4 20.3 5.0 94.6
VI is tentatively considered to be 1,2-dibromo-2-methylbutane, since i t is saturated to permanganate, shows methyl group absorption in the infrared (7.23 p ) , and is not interconvertible with 11-IV by zinc bromide (Table I ) . 2Methyl-I-butene, from which VI could have been formed with bromine, is known to contaminate methylenecyclobutane prepared by the zinc debromination of pentaerythrityl tetrabromide.lS If the interconversions of I1 and 111 are assumed to be first-order reactions and fast relative to isomerization of either to the IV,ila then the equilibrium constant between I1 and I11 is calculated from the data in Table I to be 5.7 f 2.1 in favor of 11. The error was estimated from maximum errors in the infrared analyses and does not include errors inherent in the foregoing assumptions.
PASADENA 4, CALIFORNIA (16) F.C. Whitmore and P. H. Williams, private communication to S. H. Bauer and J. Y. Beach, THISJOURNAL, 64, 1142 (1942).
