Nov. 5 , 1957
ALLYL-AND PROPYLBENZENE WITH BUTYLLITHIUM
TABLE I DISPROPORTIONATION OF ETHYLBENZENE IN THE PRESENCE OF XYLENE AND 150 VOLUME % HF Feed composition, mole % 20 20 75 75 75 o-Xylene 0 0 0 40 0 0 0 40 67 33 75 m-Xylene 0 20 0 0 0 20 0 34 p-Xylene 25 25 20 33 33 25 25 20 Ethylbenzene 65 50 0 20 24 24 63 25 Temp., “C. Reaction time, min. 4160 240 60 90 60 30 30 30 1.1 1.1 1.1 2.0 1.1 Moles BFa per mole aromatic 1.6 1.5 1.0 Product distribution, mole % 24 Benzene 15 15.5 22 20 19 12.5 18 52 62 74 63 CSFraction 70 69 56 60 15 15.5 CIOFraction 24 22 20 19 13.5 19 Composition of CSfraction, % 0 0 4 1 1 o-Xylene 0 69 16 77 71 95 m-Xylene 98 15 73 95 83 18 0 0 2 1 0 3 7 p-Xylene 10 11 Trace 1 16 11 4 Ethylbenzene 6 Composition of Clo fraction, % 0 0 0 0 10 Trace 1,3-Diethylbenzene 85 90 10 100 Trace 20 1,3-Dimethyl-5-ethylbenzene 100 100 100 15 0 0 0 0 0 90 80 1,2-Dimethyl-4-ethylbenzene 0 0 0 0 0 0 0 1,4-Dimethyl-4-ethylbenzene 0 0 0 0 1.3-Dimethyl-4-ethylbenzene 0 0 0 0 0 0 Ethylbenzene converted, % 90 97 85 88 60 75 96 99
+
higher boiling o-xylene, then the lower-boiling fraction-ethylbenzene, m- and +-xylene-can be transformed under mild conditions into benzene, m- and p-xylene and 1,S-diethylbenzene. Also, under mild conditions, the o-xylene can be con-
[CONTRIBUTION FROM THE WALKER
5809
0 0 50 50 20 90 1.1 21 58 21 0 64 17 19 93
7 0 0 0 75
verted by transalkylation with ethylbenzene into 1,2-dimethyl-4-ethylbenzene,another pure tencarbon alkylbenzene not readily obtained by direct alkylation. WHITING,
INDIANA
LABORATORY OF THE RENSSELAER POLYTECHNIC
INSTITUTE]
The Metalation and Addition Reactions of Allylbenzene and Propenylbenzene with Butyllithium and Lithium Amide BY HARRYF. HERBRANDSON AND DAVID s. MOONEY RECEIVED MARCH 1, 1957 1-Phenylallyllithium has been prepared in liquid ammonia and in ether. Its reactions with proton donors and with carboii dioxide are very similar to those of 1-phenylallylsodium and cinnamylsodium leading to the conclusion that the compound is probably ionic in both solvents, albeit existing as ion pairs or higher aggregates in ether. A possible explanation is offered for the formation of propenylbenzene from the quenching of 1-phenylallyllithium in a kinetically controlled reaction. Propenylbenzene undergoes addition reactions with butyllithium and is quite unreactive to lithium amide.
Although a number of examples have appeared of allylic rearrangements of organolithium compounds,’ those which have been reported do not afford a good basis for comparison with allylic rearrangements of other organometallic compounds. The most extensive studies of allylic rearrangements of organometallic compounds have been concerned with the cinnamyl and butenyl Grignard reagents2 and c i n n a m y l ~ o d i u m as ~ ~well , ~ as 1-phenylallylsodium.3a,c Other reactions of allylic derivatives of so(1) H. Gilman and F. Breuer, TRISJOURNAL, 66, 1127 (1934); H. Gilman and C. W. Bradley, ibid., 60, 2333 (1938); K. Ziegler, N. Eimers, W. Hechplhammer and H. Wilms, Ann., (167, 43 (1950). (2) M. S. Kharasch and 0. Reinmuth, “Grignard Reactions of Nonmetallic Substances,” Prentice-Hall, Inc., New York, N . Y., 1954, pp. 1133 ff. (3) (a) T . W. Campbell and W. G. Young, THISJOURNAL, 69, 688 (1947); (b) A. A. Morton and E. Grovenstein, Jr., ibid., 74, 5437 (1952); ( c ) R. Y.Mixer and W. G. Young, rbid.. 78, 3379 (1956).
dium, 3b-c,4 potassium6 and cesium6also have been reported. That the inorganic alkali metal cation can have an effect on the behavior of the organic anion has been emphasized.’ In view of the marked differ(4) A. A. Morton, F. D. Marsh, R . D. Coombs, A. L. Lyons, S. E. Penner, H. E. Ramsden, V. B. Baker, E. L. Little and R. L. Letsinger, ibid.. 72, 3785 (1950). (5) A. A. Morton, M. L. Brown, M. E.T. Holden, R . L. Letsinger and E. E. Magat, ibid., 67, 2224 (1945). (6) J. de Postis, Comfit. rend., 324, 579 (1947). (7) J. B. Conant and G. W. Wheland, THIS JOURNAL, S4, 1212 (1932): C. G. Swain, ibid., 69, 2306 (1947); C. G. Swain and I,. Kent, ibid., 72, 518 (1950); 0. L. Brady and J. Jakobovits, J . Chcni Soc., 767 (1950); A. A. Morton and C. E.Claff, Jr., THIS JOURNAL, 76, 4935 (1954); H. Gilman and J. W, Morton, Jr., in “Organic Reactions,” R. Adams, editor, Vol. VIII, John Wiley and Sons, Inc., New York, h’. Y.,1954, p. 258; A. A. Morton and E. J. Lanpher, Abstracts of the 129th Meeting of the American Chemical Society, April 1956, p. 12-N.
5810
HARRY F. HERBRAKDSON AND DAVIDS. MOONEY
1701. 79
ences in behavior of cinnamylmagnesium chloride and cinnamylsodium and the marked difference in ionic character between organosodium and organolithium compounds,* the behavior of the lithium derivative of allylbenzene toward a number of electrophilic reagents was investigated. Some comparisons between the behavior of allylbenzene and propenylbenzene toward butyllithium and lithium amide were also made.
fractional vacuum sublimation and comparison of the m.p. of each fraction with a m.p.-composition curve1* determined from mixtures of authentic 2phenyl-2-butenoic acidI3 and 4-phenyl-3-butenoic acidI4 and in two instances by quantitative comparison of the ultraviolet spectra of the mixtures with those of authentic samples of the acids. In four preparations 4-phenyl-3-butenoic acid constituted 59-69y0 and 2-phenyl-2-butenoic acid 3141% of the total acid.lj Results Carbonation of the product of reaction of butylThe quenching of 1-phenylallyllithium with a lithium and propenylbenzene after 1 hr. a t 0” gave proton donor gave a mixture of allylbenzene and an 80% recovery of propenylbenzene, 5% of 3propenylbenzene whose compositioii depended on methyl-2-phenylheptanoic acid and 15% of hydrothe solvent. The complete reaction of allylben- carbon, mainly 2-methyl-1-phenylhexane. Refluxzene with lithium amide in liquid ammonia re- ing of the reactants in ether for 15 minutes before quired 45 minutes after which quenching with am- carbonation gave 5% of 3-methyl-2-phenylliepmonium chloride gave 80% propenylbenzene and tanoic acid, about S-lOCi-, of acid presumed to be acid, about 557; 2OrO allylbenzene. If the mixture was quenched 3,c5-dimethyl-2,4-diphenylnonanoic after only 30 minutes a greater proportion of allyl- of higher molecular weight acids and 30% of neubenzene was found in the product: 40-50% allyl- tral material composed mainly of 2-methyl-lbenzene and only 5 0 4 0 % propenylbenzene. In phenylhexane and 2,4-diniethyl-l,3-diphenyloctane. The carbonation of the products of competition ether the reaction of butyllithium with allylbenzene was also slow. As estimated from the carbonation between equivalent amounts of allylbenzene and experiments, less than 5% reaction was accom- propenylbenzene for butyllithium resulted in the plished after 2 hr. a t - l j O , whereas the reaction formation of 2-phenyl-3-butenoic acid (isolated as was better than 9Oy0 complete after 2 hr. of re- 2-phenyl-2-butenoic acid) atid 4-phenyl-3-butenoic fluxing. Added to a large excess of methanol, 1- acid as the only acids. The refluxing in ether for phenylallyllithium in ether gave approximately 15 minutes of equimolar quantities of allylbenzene, 54y0 allylbenzene and 469; propenylbenzene. propenylbenzene and butyllithium gave after carCorrection of these figures for unreacted allylben- bonation 4Oy0 of acid, 59-69y0 4-phenyl-3-butenoic zene (about 7yo)gives for the product of quench- acid arid 31-41% 2-phenyl-3-butenoic acid and 11% of recovered allylbenzene and propenylbenzene. ing a 50 :50 ratio of propenylbenzene :allylbenzene. Lithium amide in liquid ammonia did not react That these were the only acids was demonstrated to any appreciable extent with propenylbenzene dur- by the determination of the neutralization equivaing a period of 2.25 hr. The reaction of butyl- lent of the total acid recovered, the ul!r:iviolet lithium in ether with propenylbenzene was mare spectrum and the m.p. range of the acid. The rerapid than with allylbenzene; however, quenching mainder consisted largely of about equal anloL1lits after 15 minutes reflux gave only a 4oj, recovery of of 2-methyl-1-phenylhexane and 2 ,-l-dimethyl-1,3propenylbenzene. Of the original hydrocarbon, diphenyloctane and lesser amounts of higher molec22% was accounted for as 2-methyl-1-phenyl- ular weight hydrocarbons. A similar reaction hexane, 44% (25% purified) was presumed to be effected with 2 hr. of reflux gave only 2 i 7 0 of acid, 2,4-dimethyl-lJ3-diphenyloctane, and the remain- 63-64% 4-phenyl-.?-butenoic acid snd 3 6 4 7 % 2der appeared as higher molecular weight products. phenyl-3-butenoic acid and a greater amount of 2The carbonation of 1-phenylallyllithium in methyl-1-phenylhexane and higher molecular ether gave a 48-57y0 yield of acid which was re- weight hydrocarbons. The reaction of 1-phenylallyllithiuni with bcrizofluxed for 2 hr. with 10% sodium hydroxide to isomerize the 2-phenyl-3-butenoic acid to 2-phenyl-2- phenone gave 41y0 crude, or %yopure, l;l,’$-tributenoic acid.g That little of this acid was lost phenyl-3-butene-1-01. The alcohol was identified in the 2-hr. reflux is indicated by Fittig’s isolation through dehydration by vacuum distillation over a of SO-SO% of unchanged 4-phenyl-3-butenoic acid trace of iodine to give 1,1,4-triphenyl-l,3-butaafter 20-80 hr. reflux with 10% sodium hydrox- diene.” None of the isomeric alcohol, 1,1,2-triide,l0 although Linstead recovered only 5oY0 of phenyl-3-butene-1-01,could be isolated. (12) This m.p.-composition curve has been deposited 8s Docuthe acid from a more drastic 100-110 hour reflux ment number 8199 with t h e AD1 Auxiliary Publications Project, with 20y0 alkali.11 Photoduplication Service, Library of Congress, Washington 25, D . C. Quantitative estimations of the relative amounts A copy m a y be secured by citing t h e Document number and by reof 2-phenyl-2-butenoic acid and 4-phenyl-3-bute- mitting in advance 51.25 for photoprints or $1.25 for 35 mm. micronoic acid in the resulting product were made by film payable to: Chief, Photoduplication Service, Library ol Congress. (13) H. Rupe and E. Busolt, A n n . , 969, 332 (1909).
(8) (a) G. E. Coates, “Organo-Metallic Compounds,” John Wiley and Sons, Inc., New York, N. Y . , 1056, pp. 6 , 12; (b) M. T. Rogers and A. Young, THIS JOURNAL, 68, 2748 (1946); ( c ) L. Pauling, “ T h e h’atnre of t h e Chemical Bond,” 2nd Ed., Cornel1 University Press, Ithaca, N. Y . , p. 74. (9) H . Gilman and S. A. Harris, THISJOURNAL, 49, 1825 (1927); 53, 3541 (1931). (10) R. Fittig and A. Luib, A n n . , 283, 297 (1894). (11) R. P. Linstead a n d L. T. D. Williams, J . Chcm. Sac., 2735 (1926).
(14) R. Fittig and W. Jayne, ibid., 216, 98 (1883). (15) It is possible t h a t a small amount of o-propenylbenzoic acid
m a y have been presentsb as the ultraviolet spectra probably would not distinguish this acid from I-phenyl-3-butenoic acid (compare t h e spectra of sodium 4-phenyl-3-butenoate and fvuns-propenylbenzeneln; b u t from t h e sublimations i t must be concluded t h a t this amounted to less than 10% of the total acid. (16) R. Y . Mixer, R . F. Heck, S. Winstein atid W. G. Young, THIS JOURNAL. 75, 4094 (1953). (17) H . Staudinger, Ber., 41, 4249 (1909).
Nov. 5 , 1957
ALLYL-AND PROPYLBENZENE WITH BUTYLLITHIUM
5811
in ether which gives a greater proportion of the less stable allylbenzene on quenching as do also the sodium derivative in pentaneaC and the cinnamyl Grignard reagent in etherz1(Table I). One explanation is the possible existence in the organic solvents of the organometallic derivatives as ion-pairs or partially covalently bonded species which, with proton donors, would give the less stable olefin through a coordinated allylic displacement.3a.22 That a similarly bonded sodium salt of the dihydronaphthalene may exist in liquid ammonia a t -70" but not to as great an extent a t Discussion higher temperatures could result from the fact that 1-Phenylallyllithium in liquid ammonia behaves the sodium would be bonded to a more nucleophilic on quenching just as do l-phenylallylsodi~m~~~c secondary carbon than would be the case with the and cinnamylsodium, 3a giving about the same ratio corresponding allylbenzene anion which appears to of propenylbenzene to allylbenzene (Table I). exist as the free ion in liquid ammonia even a t TABLE I Dry Ice temperature. has 'Onsidered the PRODUCTS OF REACTION O F PROTON DONORS WITH METALLIC Of the dihydronaphthalene to exist as DERIVATIVES op ALLYLBENZENEA N D P R O P E ~ L B E N Z E'Odium NE free ions even a t low temperatures and has interPropenylbenzene in the preted the formation of 1,4-dihydronaphthalene as propenylbeing the result of electrostatic control of the posibenzeneOrganometallic compound Solvent ?$: %$trg tion of attack by the proton donor. Birchz4has applied this concept to other anions as well. The NHlCl 81" I-Phenylallyllithium NHs(1.) distribution of charge in the allylbenzene anionz6is CHIOH" 93 l-Phenylallylsodiumb NHa(1.) such that if it existed as the free anion and if interl-Phenylallylsodium.d NH8(l.) CHxOH, 85 action in the transition state were largely electroNHhCl, or phenylacetylene static, the preponderant product would be the ther85 Cinnamylsodiumd NHs(1.) CHIOH, modynamically less stable allylbenzene. It would NHICl, or phenylacetylene appear, if both exist as free anions in liquid am1-Phenylallyllithium Ether CHxOH 50" monia, that covalent bond formation in the transi26 1-Phenylallylsodium' Pentane' CH*OHC tion state must contribute more in the case of the Cinnamylmagnesium allylbenzene anion than in the case of the dihydroDil.HzSO, 25 chlorided.' Ether If the propenylbenzene contained as much as 20% of the phthalene anion. That the recombination of ions encis isomer,'* the amount of allylbenzene in the mixture from a pseudo-acid need not have a would be from 5-8% more than that reported here. b Ref. ergy of activation has been shown by Pearson.26 3c. Reverse addition. Ref. 3a. e Heterogeneous. From quenching in the less polar solvents one would Ref. 21. anticipate] if the organometallic compounds are It appears that each of these three compounds is ionic, a looser transition state than in liquid amessentially ionic and fhe reactive species in liquid monia and thus a greater influence of the charge ammonia, which can effectively solvate the cation distribution in the anion on the position which is attacked by the proton donor with formation of a and anion, is the free carbanion. The major product of quenching from liquid am- greater proportion of the less stable allylbenzene as mania is the thermodynamically more stable pro- is found to be the case. It is obvious that from penylbenzene, yet it is formed in a kinetically con- these observations alone on 1-phenylallyllithium trolled reaction. The best evidence for this is the it is impossible to conclude whether, in ether, the slow reaction of lithium amide with allylbenzene, lithium is covalently bonded or not. The intense which would seem to preclude rapid equilibration red color of the ether solution indicates easy deloduring the actual quenching and the recovery of a calization of the bonding electrons in any case. greater proportion of allylbenzene from reactions in The diminished reactivity of the allylic hydrogen which some unreacted allylbenzene remained at the of the more stable propenylbenzene is emphasized time of quenching. This result is in contrast to the by the fact that, in liquid m-"nia, propenylbenformation of the less stable isomer. 1.4-dihvdro- zene is quite unreactive to lithium amide and, in naphthalene rather than 1,2-dihydronaphthhene, (21) W. G. Young, G. Ballou and K. Nozaki, THISJ O U R N A L , 61, 12 from the anion of 1,4-dihydronaphthalene in liquid (1939). (22) J. E.Leffler, "The Reactive Intermediates of Organic Chemisammonia a t temperatures of -70 to - 7 5 O . I 9 At higher temperatures the sodium salt in liquid try," Interscience Publishers, Inc., New York, N. Y.,1956, p. 200. (23) G. S. Hammond, THIS JOURNAL, 77, 334 (1955). ammonia gives the more stable 1,2-dihydronaph(24) A. J. Birch, Quart. Revs., 4, 69 (1950); A. J. Birch, Disc. Farothalene, yet the lithium derivative in ether a t 0" d a y Soc., 2, 246 (1947); cf. also M. J. S. Dewar, "Electronic Theory of gives 1,4-dihydr0naphthalene.~OIn this respect i t Organic Chemistry," Oxford University Press, London, 1949, p. 103; is similar to the lithium derivative of allylbenzene M. J. S. Dewar, Disc. Faraday SOC.,2 , 261 (1947). 2-Methyl-1-phenylhexane, prepared from cymethyIcaprophenone18 by Wolff-Kishner reduction, proved to be identical with the product from the quenching of the butyllithium-propenylbenzene reaction mixtures. The structure of the 3methyl-2-phenylheptanoic acid follows from its manner of preparation, analysis] and neutralization equivalent and the assignments of structures to the 2,4-dimethyl-l,3-diphenyloctaneand the 3,5dimethyl-2,4-diphenylnonanoicacid follow from their analyses and by analogy.
f
(18) A. D. Campbell, C. L. Carter and S. N. Slater, J . Chcm. Soc., 1741 (1948). (19) W. Hiickel and H. Bretschneider, Ann., 640, 157 (1939). (20) W. Schlenk, E. Bergmann and J. Appenrodt, ibid., 463, 90 (1928).
(25) H. C. Longuet-Higgins, -7. Chcm. Phrs., 18, 265 (1950); M. J. S. Dewar, THIS JOURNAL, 74, 3345 (1952). (26) R. G. Pearson and R. L. Dillon,
