June20, 11360
322 1
COMMUNICATIONS TO THE EDITOR
sponding data previously published for 2,2-5,5tetramethyl-THFD.2 With the latter compounds the E of the tetraalkyldioxolane, analogous to VII, indicates alcoholysis of the dioxolane ring in methanol. Grateful acknowledgment is given to Dr. H. Stange of F.M.C., Princeton, N. J . , for infrared spectrograms and to Prof. E. C. Taylor of Princeton University for assistance both practical and theoretical. This work was made possible through grants from the Research Corporation of New York City, Merck and Company, Kahway, N. J., the Victoria Foundation of New York City and the Pew Foundation of Philadelphia, Pa.
C H (CH3)2
Na -and- p orientat ion iri
tr,-Amxa
K
,,-AmK AmCl K
m-and p orientation
+
Ka o-and p orient at ion
(2) C . Sandris and G. Ourisson, Bull. SOC.Chrm. France, 338 (1958).
PRINCETON UNIVERSITY EDWARD C. KENUALI, DEPARTMENT O F CHEMISTRY ZOLTANG. HAIOS PRINCETON, NEW JERSEY RECEIVED MAY 12, 1960 FACTORS GOVERNING ORIENTATION I N HETEROGENEOUS METALATION REACTIONS
Sir:
metalations (which in all likelihood are surface reactions) might take different courses. We are hereby reporting (see Table I) that namylsodium, prepared either from n-amyl chloride or di-n-amylmercury, metalates cumene almost entirely in the ring. Practically identical isomer distributions result in each case. Clearly the metal enters the ring in such fashion as to be as far removed from the isopropyl group as possible, which is contrary to findings in earlier reports.6*6 We can only conclude that the meta isomer was not recognized as a major product by these earlier workers. Table I also discloses that when n-amyl-
Compelling evidence has been advanced t h a t homogeneous metalation reactions involve a nucleophilic attack by the anion of the metalating agent on hydrogen.' This view never has been generally accepted for heterogeneous metalations, particularly those involving organoalkali compounds, although similarities between the two systems have been pointed out recently. 3 , 4 Typical of the controversies which have prevailed TABLE I in the area of heterogeneous metalations are the METALATION O F CUMENE BY n-AMYLSODIUM AND %-AMYLresults which have been reported for the metalation POTASSIUM of cumene by alkylalkali reagents. One labora% Isomer distrihution'*d, * n-AmylI'repn. Solvent Alpha hletr Pard tory6t6 reports that n-amylsodium metalates cumene predominantly para and some ortho to the iso1 Na a Curnene 2 6 55.6 41.8 propyl group, while another3 claims nearly a 1:1 2 Na b Cumene 1.2 LB.0 42.8 metalpara ratio. Very little side chain metalation 3 K a Cumene 88.0 4.5 7.5 was observed with n-amylsodium by either Labora4 K b Cumene 91.6 .. 8.4 tory. n-Amylpotassium is reported to give almost 5 K a Heptane' 100 exclusive side chain metalation, when the potas6 K a Heptane',@ 42.0 39.0 19.0 sium reagent is prepared from n-amyl chloride and n-AmC1 metal. n-Am2Hg + metal. e All reacpotassium,? but only 13% side chain product when tions were stirred a t high speed at room temperature. All analyses were of t h e methyl esters obtained by treating it is prepared from n-amyllithium and a sodiumt h e carbonation products with excess diazomethane. potassium alloy.3 Vapor phase chromatography was used in every case. e T h e An attempt? was made recently to explain these reaction time was 20 hr. in every case except for entry 6. puzzling discrepancies on the basis that each LabContained a two-mole excess of cumene. Y The reaction oratory had prepared their organoalkali reagents by time was 3 hr. different methods. It was reasoned' that different modes of preparation would cause variations in the potassium was prepared from both the chloride and surfaces of these insoluble reagents and hence the di-n-amylmercury (entries 3 and 4) almost exclusive alpha metalation resulted. While the isomer ratios (1) G. E. Hall. R. Piccolini a n d J. D. Roberts, THIS JOURNAL, 1 7 , in the two cases were not identical they were ex4540 (1955). tremely close. Further, if the metalation of cu(2) A. A. M o r t o n , ibid., 69, 969 (1947); see also A. A. Morton, Chem. Reas., 36, 1 (1944); A. A. Morton, C. E. Claff, Jr., a n d F. W. mene by n-amylpotassium in heptane was allowed Collins, J . Ovz. Chem., SO, 428 (1955). to proceed for 20 hours, almost exclusive alpha (3) D . Bryce-Smith, J. Chcm. SOL.,1079 (1954). metalation occurred (entry 5). If the reaction was (4) D. Bryce-Smith, V. Gold a n d D. P. N. Satchell, ibid., 2743 interrupted a t the end of 3 hours (entry 6), however, (1954). (5) A. A. M o r t o n , J. T. hlassengale a n d M. L. Brown, THIS JOURconsiderable quantities of meta and para isomer were NAL, 67, 1020 (1945). found. Clearly the latter two isomers disappear with (6) A. A. hfortou a n d C. E. Claff, Jr., Abstracts of t h e 119th hfeettime. This is graphically illustrated by the results ing of t h e American Chemical Society, Boston, Mass., 1951, p. 11-M. shown in Table 11,wherein highly purified p-chloro(7) A. A. Morton a n d E. J. Lanpher, J . Org. Chem., 23, 1030 (1958). cumene was treated with potassium in both excess
+
f
cumene and heptane as a solvent h t t!ie end of 20 hours, only the alpha product remained (entry 1). On the other hand, when the reaction was stopped after 4 hours, not only alpha product, but rnrta a$ TABLE11 REAC I O~ Y
OF
HIGHLYPURIFIED
P-CHLOROCUMETF I\ ITH
POTASSIL \I 5 ,l\ent
1 2
"
Curneiie Curnene
Reartioo time hr
