960
COMMUNICATIONS TO THE EDITOR
Vol. 86
CH3CHZC1'H* +CHaCH=C"H, it has become possible to apply a double tracer tech(2) nique to the study of atoms produced by nuclear reThe resulting radical then abstracts a hydrogen atom coil.8 We report here preliminary results of such from C2H6 (or C Z D ~ ) . I - ~ studies on carbon atom reaction mechanisms. Ethylene from C Atom Reaction with Ethane.Technique.-C1l atoms are produced by C"(y,n)C" One of the more important products from the C atom reaction induced by 40 MeV, Bremsstrahlung irradiareaction in saturated hydrocarbons is ethylene.'-4 tion of the appropriate isotopically mixed system. When produced in an equimolar CzH,&zD6 mixture (Either a mixture of protonated and deuterated we find approximately equal amounts of CCl1H4, molecules or a single partially deuterated compound CC11H3D,CC"HD3, and CC"D4. The ratio of CCllis used.) Nuclear techniques for production of the HzDz t o CC''HD3 is less than 1:5. This finding recoil atoms under conditions of low radiation damage excludes a number of possible mechanisms, but is conAnalysis of products are described elsewhere. 1-4 sistent with C atom insertion into a C-H bond. The involves gas chromatography followed by flow countintermediate formed by reaction 1 de-excites by C-C ing.lP4 Isotopic separation is effected by a 100-ft. rupture (a favorable mode) ethylene glycol-silver nitrate column (25' and 0'). CH3CHzC"H --+ CH3. + CHzC"=H. (3) This column was calibrated for isotopic ethylenes using The vinyl radical then abstracts hydrogen from C2Hs CzD4, CzH4, C2H3D,CH2CD2,and a mixture prepared or C2D6. The data are equally consistent with a by reducing C2Hz with Dz on a platinum catalyst. postulated alternative mechanism involving insertion The five equally spaced compounds found are assumed into the C-H bond by a C'lH intermediate.lo.ll to correspond t o C2H,D4-. (n = 0-4). No information on possible separation of the three CzHzDz isomers was C" + CHICHI --+ C"H. + CHaCH2. (4) obtained. Allene calibrations were based on C3H4 C"H. + CHsCH3 +CH3CHzC"Hz. + and C3D4. Retention volumes of partially deuterated CH3. CHz=C"Hz (5) isomers were interpolated. These were checked by irAcknowledgments.-We are indebted to the operatradiation with C1' of a C2H6-CzD6-CeH2mixture. The ing staff of the Yale linear electron accelerator. This predicted unsymmetrical distribution of allenes, corwork was supported by the U. S . Atomic Energy Comresponding to the expected productiong of CaH4, mission. C3H3D, and C3H2D2, was found. The effectiveness of the column for separating isotopic propylenes was (10) Insertion b y C H has been recognized as a possible alternative t o insertion b y C.2 I t was suggested as a specific precursor of ethylene b y A. P . checked using a propylene mixture prepared by Wolf (Hot-Atom Conference, Amsterdam, M a y , 1963) catalytic reduction of CH3C=CH with Dp. (11) I t may be relevant t o point o u t t h a t t h e conclusions drawn from this Allene from C Atom Reactions with Ethylene.-Alwork regarding t h e existence of radical precursors could not be derived from lene is a principal product of the reaction of C with work with scavengers. T h u s t h e yields of neither allene from ethylene nor ethylene from ethane are markedly oxygen-sensitive. Yet t h e double tracer ethylene. ' When produced from an equimolar mixture experiments show t h a t t h e former does not generally involve a radical preof CzH4 and C2D4 the following isotopic yield while t h e latter does This limitation of t h e scavenger technique ratios were found : C2C"H4 : C2C"H3D : CzC"H2Dp : CZ- cursor, in atomic carbon systems is probably d u e t o t h e very high thermal reactiviC11HD3:CpC11D, = 7 : l + 0 . 3 : < 0 . 2 : 1 =t 0.3:7 i 1. ties of t h e radicals involved (Scavengers a r e only effective when radicals have a relatively long lifetime in t h e given system.) Thus allene is produced predominantly by interaction of a carbon atom with a single ethylene molecule. DEPARTMEKTS OF CHEMISTRY JAMESDUBRIN YALEUNIVERSITY COLINMACKAY This result is consistent with the two previously XEWHAVEN,CONNECTICUT RICHARD WOLFGANG postulated mechanisms of insertion into the double HAVERFORD COLLEGE bond,' and into the C-H bond,l,? of ethylene. I t HAVERFORD, PENXSYLVANIA seems plausible that the isotopic mixing that does RECEIVED OCTOBER 21, 1963 occur is associated with radical production following insertion into the C-H bond. Propylene from C Atom Reaction with Ethane.A Reinvestigation of the Pyrolysis of The formation of propylene from ethane has a formal Tetramethylammonium Hydroxide similarity t o the formation of allene from ethylene. Sir: However, i t has a much lower This was atThe thermal decomposition of tetramethylammonium tributed to the different nature of the intermediates hydroxide has been found t o yield trimethylamine and formed by the C-C and C-H insertions.' This difdimethyl ether as the major decomposition products ference in predominant mechanisms is reflected in the rather than the reported products, trimethylamine results on propylene formation from an equimolar and methanol.' The ether is formed in >90% yield mixture of c2H6 and CzDG. Four propylene peaks are based on the recovered trimethylamine. Minor prodfound, probably C3H6, C3H5D, C3HD6, and C3D6. The peak ratios are C3H6:C3H5D = C ~ D O : C ~ H GD ~ 2[(CH3)3G-CH3]0H- --+ 1.3:1. Regardless of peak identities it is obvious that there is very extensive, though probably incomplete, isotopic mixing involving one hydrogen atom. ucts are methanol (ca. 5%), and apparently small Thus, unlike allene from ethylene, most propylene quantities of higher alcohols and polyethylene. from ethane seems to have a radical precursor. This The pyrolysis was carried out in a vacuum line result is consistent with the previously postulated' usirg a stock solution of 10yotetramethylammonium carbon insertion into a C-H bond hydroxide in water. Most of the water was removed C" + CHsCHs + CHJCHzC"H* (1) by heating the solution a t 65' to give a white amorphous solid. The solid decomposed a t 135-140' and the The intermediate can de-excite by several modes and products were collected continuously during the course is thus a precursor for a number of One of the reaction in a series of cold traps; water and such mode is H atom elimination methanol were removed a t - 78' (COz-isopropyl alcohol bath), and the amine and ether a t -196' (8) T h e use o f stable isotopes as tracers for recoil a t o m reactions should not be confused with their use to measure isotope effects. (liquid Nz bath).
+
(9) T h e allene produced in this system appears to result from addition to acetylene of methylene arising from reaction of C atoms with ethane.
( I ) A U' H o f m a n n , B e y , 14, 494 (1881)
March 5, 1964
COMMUNICATIONS TO THE EDITOR
The methanol was determined by gas-liquid chromatography with a flame ionization detector. The amine and ether mixture was condensed into a tared reaction tube and treated with a known amount of hydrogen chloride. The ether and excess hydrogen chloride were distilled and the amine content was determined by the isolation of the hydrochloride salt.2 In a partial trap-to-trap fractionation of the original mixture, both the amine and ether fractions were analyzed by infrared spectroscopy. The spectrum of the ether fraction revealed traces of amine as the only impurity. A typical decomposition using 2 ml. of 10% tetramethylammonium hydroxide (2.20 mmoles) gave the following products. Millimoles
2.13" 0.97' 0 11' Trace a Recovered as the hydrochloride. Recovered ether and HC1 mixture less the HCl not used in forming the amine hydrochloride (see ref. 2 ) . < Methanol and water mixture analyzed by gas-liquid chromatography with flame ionization detector.
A series of pyrolyses was carried out under a variety of conditions (a-e) without significant changes in the yield of products: (a) in an atmosphere of nitrogen, (b) in a rapid nitrogen flow system with a head pressure of 10 mm., (c) in a n atmosphere of 25 mm. of water vapor, (d) in the presence of Ag+ ion, and (e) in an initially evacuated system without continuous collection of products. Trace amounts (0.09%) of dimethyl ether have been reported3 as a product in the pyrolysis of cyclohexylmethyl-P-d-trimethylammonium hydroxide and its formation was explained by a three-step mechanism. An analogous mechanism for the decomposition of tetramethylammonium hydroxide would account for the observed products.
+
/CHa)4fiOH--+( C H 3 ) 3 S CHIOH CHa0H O H - $ H20 CHaOCHaO-
+ + (CH3),k --+
+
CHaOCHa
+ (CH3)3N
However, if methanol is formed under the present conditions it would be expected to leave the reaction zone rapidly. The formation of the methoxide ion and its subsequent reaction to form the ether in high yield would involve retaining the methanol completely but allowing the water to escape. Consequently, i t is suggested t h a t the hydroxide ion behaves as a protonabstracting agent to give a nitrogen ylide and the observed products result from the decomposition of the ylide in various ways. Nitrogen ylides have been shown to exhibit carbene character, and carbenes have been reported5 to react with alcohols to give ethers. .Thus, the ether may be formed by the reaction of ylide with the small amount of water remaining a t the decomposition temperature. Methanol may be formed either by the decomposition of the ylide in the presence of a large amount of water, or by a displacement reaction, whereas polyethylene is formed only in the latter stages of reaction when water is no longer present. The decomposition of this ylide in ether to form polyethylene and trimethylamine has been reported.6 (2) Dimethyloxonium chloride decomposes a t -2' and can b e removed completely from the amine salt. ( 3 ) A C Cope, N . A LeBel, P . T Moore. and W R . Moore, J . A m Chem S o c . , 83, 3861 (1961) (4) V . Franzen and G Wittig, A n p e w C h e m , 72, 417 (1960). ( 5 ) W. Kirmse, A n n . , 666, 9 (1963). (6) G Wittig and R Polster, ibid., 699, 1 (1956):
96 1
From this preliminary study, i t appears t h a t in the absence of hydroxylic solvent, or after the removal of hydroxylic solvent from the reaction site, a-hydrogen abstraction by strong bases' (OH-, NHZ-, phenyl-) is the favored reaction. With weaker bases (Cl-, Br-, etc.)* displacement may be the predominant reaction. I t has been suggested3 that the mechanism of the Hofmann degradation may depend on the manner in which i t is carried out. However, the mechanism may change during the course of reaction. In the presence of hydroxylic solventg,10or during the removal of solvent at elevated temperature, P-elimination appears to be ' the primary mode of decomposition. It is known that a-hydrogen abstraction occurs in the presence of these protonic but apparently reprotonation occurs before the ylide can decompose by any competitive path and alternative mechanisms are As the solvent is removed, a-hydrogen abstraction becomes irreversible and the course of reaction depends on the mode of decomposition of the ylide. Acknowledgment.-The author wishes to thank the National Science Foundation (GP-1704) and The University of California for supporting this research. (7) A . C. Cope and E . K Trumbull, Org. Reaclions, 11, 317 ( 1 9 6 0 ) . (8) A . T. Lawson and S Collie, J Chem Soc , 63, 624 (1888) (9) G. Ayrey, E Buncel, and A . N Bourns, Proc Chem Soc , 458 (1961) (10) V J . Shiner and M . I, Smith, J A m Chem S o c . , 80, 409.5 (1938). (11) W von E Iloering and A K . Hoffmann, i b i d . , 77, 521 (1955)
UXIVERSITY OF CALIFORSIA W. KENNETH MUSKER OF CHEMISTRY DEPARTMENT DAVIS,CALIFORNIA RECEIVED NOVEMBER 25, 1963
Chelation a s a Driving Force in Organic Reactions. A Generic Synthesis of Amino Acids by the Carboxylation and Alkylation of 3-Phenylhydantoin
Sir: In the course of our work on the carboxylation of compounds containing activated methylene positions, we became interested in 3-phenylhydantoin. Wheeler and Hoffman2 showed that the methylene position of hydantoin is sufficiently activated to condense with aromatic aldehydes, and thus hydantoin might be a member of the group of compounds which react with magnesium methylcarbonate. When 3 - p h e n y l h y d a n t ~ i nwas ~ heated a t 80' with magnesium methyl carbonate and the reaction mixture sampled periodically by diluting I-ml. portions with methanol, a new absorption peak in the ultraviolet was observed a t 278 m p . A basic solution of %phenylhydantoin does not absorb in this region. This suggested t h a t the 5-position of 3-phenylhydantoin has been carboxylated and the chelate formed. The new absorption peak immediately disappeared on addition of a trace of aqueous acid. The identity of the carboxylation product was established by isolation of 3-phenyl-5carbomethoxyhydantoin. The product was precipitated by pouring the reaction mixture into ether. After decanting the liquid phase, methanolic hydrogen chloride, a t - 30', was added. After 2 days a t room temperature most of the methanol was removed and the remainder poured into water. The infrared and nuclear magnetic reso(1) Previous related papers: H . L. Finkbeiner and M Stiles, J . A m . C h r m . S o c . , 86, 616 (1963); H . I,. Finkbeiner and G. W. Wagner. J . 01.g C h P m . , 28, 21.5 (1463). and references cited. (2) H . I,. Wheeler and C. Hoffman, A m . C h e m J , 4 6 , 368 (1411) Subsequent work on hydantuins, their synthesis, and utility has been reviewed b y E . Ware, Chem Ria.. 408 (19.50). ( 3 ) This rompound is prepared from the direct reactinn oi glycine with phenylisocyanate [J. R. Bailey and C. P. Randolph, Be?'.,41, 2499 (1908)l.
