Aug. 5 , 1954
3925
ACID-CATALYZED REARRANGEMENT OF BENZOPINACOL
ancy recorded in one of the parameters. Moreover, the fact that carbonato-apatite differs from hydroxy-apatite because of the replacement of hydroxyls by carbonate ions is evidenced by the information obtained by means of infrared spectrosCOPY
Acknowledgment.-The author wishes to express his thanks to Dr. Rustum Roy for reading the manuscript. DIvIsIoN OF THEpENNsYLVANIA sTATE uNIVERSITY STATECOLLEGE, PENNSYLVANIA
[CONTRIBUTIONFROM THE DEPARTMENT OF CHEMISTRY, ST. LOUISUNIVERSITY]
Kinetics and Mechanism of the Pinacol Rearrangement. I. The Perchloric Acidcatalyzed Rearrangement of Benzopinacol and of Tetraphenylethylene Oxide in Acetic Acid Solution’ BY HERBERT J. GEBHART, J R . , ~ AND KENNETH H. ADAMS RECEIVED FEBRUARY 17, 1954 The perchloric acid-catalyzed rearrangement of benzopinacol (P) to p-benzopinacolone ( K ) in acetic acid occurs by two kinetically distinguishable routes, one of which involves direct rearrangement and the other involves intermediate formation of tetraphenylethylene oxide (E). At 75” approximately SO% of the ketone is formed by the latter route. Correlation of Ho values of the medium with the rate constants of the reaction system P -t E + K shows that for the three reactions a
I
t
reversible proton transfer t o P or E is followed by the rate-determining ionization of the oxonium complex. The large positive values of the entropy of activation for the rearrangement of E and for the composite reaction of P suggest that the classical carbonium ion is an intermediate in these processes.
Although the molecular rearrangements of tetraaryl pinacols to pinacolones have been studied extensively with respect to the relative migratory aptitudes of the aryl g ~ o u p sthe , ~ kinetic characteristics of these reactions have received comparatively little attention. The rearrangement of 4,4’,4””’’’tetrachlorobenzopinacol and of the corresponding epoxide in acetyl chloride was studied kinetically by Meerburg4 with the aid of an analytical method which involved measurement of the alcoholic alkali consumed by cleavage of the pinacolone. At 30’ he observed fluctuations in the concentration, or apparent concentration, of the ketone for which he was unable to account. He concluded, however, that the epoxide could not be an intermediate in the rearrangement of the pinacol. The kinetics of the rearrangement of cis- and trans-1,2-diphenylacenaphthene-1,2-diol in acetic acid has been the subject of three investigations5in all of which the progress of reaction was followed by analysis for unchanged pinacol by oxidation with lead tetraacetate. The results of these investigations are not completely conclusive, particularly with respect to the ability of the trans isomer to undergo rearrangement. In its gross aspects the mechanism of the pinacol rearrangement is firmly established.6 However, there are numerous detailed features of these reac(1) Presented before the Division of Organic Chemistry, American Chemical Society, Atlantic City, N. J., September, 1952; Abstracts of Papers 24 M . (2) Shell Oil Co., Wood River, Ill. This paper is abstracted from the Pb.D. Thesis of Herbert J. Gebhart, Jr., St. Louis University, June,
1952. (3) For an excellent review of the pinacol rearrangement see G. W. Wheland, “Advanced Organic Chemistry,” John Wiley and Sons, Inc., New York. N. Y., 1949, pp. 451-534. (4) P. A. Meerburg, Rec. Iraw. chim., ‘24, 131 (1905); 18, 267. 270
(1909). ( 5 ) P. D. Bartlett and R . F. Brown, THISJOURNAL, 6’2,2927 (1940);
R.Criegee and K. H. Plate, Bcr., 71B,178 (1939): H.Sello, University Microfilms, Ann Arbor, Mich., Pub. No. 1286 (1949). (R) C. K. Ingold, “Structure and Mechanism in Organic Chemistry,” Cornell University Press, Ithaca, N. Y., 1953. Chapt. IX.
tions which have been inadequately investigated, particularly in the tetraaryl pinacols in which the shift is to a tertiary carbon atom. These include the stereochemical changes at the migration terminus, the effect of substituent groups upon energy and entropy of activation, and the relative importance in the transition state of the carbonium ion (IV) and the phenonium ion7 (V). The tetraarylethylene oxides (11) have received only scant attention and there is inadequate experimental support for the commonly held opinion that the rearrangements of the epoxides and of the corresponding pinacols are completely analogous. Some of these problems are under investigation in this Laboratory. The present communication records the results of the kinetic study of the rearrangement of benzopinacol (I) and of tetraphenylethylene oxide (11) to 4-benzopinacolone (111) in acetic acid with perchloric acid as the catalyst.
0 Ill ( K )
OH
+
IV
+
, .Ar..
Ar-C-LC-Ar
/
Ar
\
v
OH
Experimental Part Materials.-Glacial acetic acid (Mallinckrodt Chemical Works, bichromate test grade) was further purified by the method of Orton and Bradfield* modified by omission of acetic anhydride. The acid was distilled through an 18(7) D. J. Cram, Tms JOURNAL, 71, 38133,3875 (1949). (8) IC. J. P.Orton and A. E. Bradfield, J . Chcm. SOC.,983 (1927).
3926
HERBERTJ. GEBHART, JR., AND KENNETH H. ADAMS
inch Widmer column in which the spiral packing had been replaced by glass helices. The first 15% of the distillate was rejected. The main fraction was redistilled and the fraction of boiling range 117.0-118.1" was collected. The melting point of the product was consistently within the range of 16.20-16.40". The water content was calculated on the basis of a freezing point depression of 0.2" per 0.1% by weight of water.Q I n the later stages of the investigation more nearly anhydrous acid was obtained by carrying out the second distillation as an azeotropic distiIlation with benzene.'O The small residual water content was determined by titration with Karl Fischer reagent." The purified acid was collected and stored in all-glass equipment and was protected from contact with atmospheric moisture. Benzopinacol and benzopinacolone were prepared by the method of Bachmann.lz The pinacol was recrystallized from a benzene-petroleum ether mixture until a 0.02 M solution in benzene was optically clear when examined a t a wave length of 3300 8.in a 1-cm. quartz cell in a Beckman model DU spectrophotometer. The product melted sharply a t 195-196" on a pre-heated hot-stage. The pinacolone was recrystallized from benzene-petroleum ether until a constant molar extinction coefficient was obtained a t 3300 A.; m.p. 180-181". Tetraphenylethylene was prepared from benzopinacolone by the method of Bachmanni3 and was converted to tetraphenylethylene oxide by the action of perbenzoic acid14 in chloroform a t room temperature for four days. The oxide was recrystallized from a 10-1 mixture of anhydrous ethanol and chloroform until a 0.02 Af solution in benzene was optically clear a t 3300 b.; m.p. 2Otj-206". The standard stock solution of perchloric acid was prepared by the dilution (1: 100) of 70% aqueous perchloric acid with glacial acetic acid. This solution, approximately 0.12 M , was standardized by titration with a standard solution of sodium acetate in acetic acid with brom phenol blue.15 The standard sodium acetate was prepared by dissolving primary standard grade of sodium carbonate in glacial acetic acid. Analytical Methods.-Benzopinacol was estimated by oxidation with lead tetraacetate.'B A 5.00-ml. sample of the solution t o be analyzed was mixed in a 250-ml. glassstoppered flask with an equal volume of an acetic acid solution which was approximately 0.06' \ A in lead tetraacetate and 1.0 N in potassium acetate. A blank was prepared in the same manner by replacing with acetic acid the solution of the pinacol. The solutions were allowed t o stand a t room temperature for two hours and were then analyzed for lead tetraacetate" b y the addition of potassium iodide (0.5 g.), sodium acetate (6.0 g.) and 100 ml. of water. The liberated iodine was titrated with 0.04 3 sodium thiosulfate. Benzopinacolone was estimated by measuring the optical density in a 1.00-cm. quartz cell a t a wave length of 3300 A. of the solution obtained by the dilution of 5.00 ml. of the acetic acid solution of the ketone t o 25.00 nil. or 50.00 ml. with &butyl alcohol. The molar extinction coefficient of pinacolone was measured in the same solvent mixture. The mean value of 16 independent measurements of the molar extinction coefficient was 259 =k 2 I./mole-cm. Reaction Rate Measurements.-An oil-bath, the temperature of which could be controlled to &0.05O, was used. T h e rcactior. flask was of 2 k g >> k--4. The requirement that R + reacts predominantly by the internal processes of ring closure and rearrangement and is not sensitive to external attack by a solvent molecule seems not unreasonable for a molecule in which steric factors must be of major importance. The high susceptibility to acid-catalyzed hydration to the glycol which is characteristic of the phenylsubstituted ethylene oxides, including triphenylethylene oxide,za and the essentially complete failure of this reaction in tetraphenylethylene oxide suggests serious steric retardation in the latter reaction. Mechanisms I and I1 represent limiting types. In view of the generalization that neighboring group participation tends t9 decrease with increasing phenyl substitution of the carbon atom a t which heterolysis occurs24and in view of the opportunities for resonance stabilization of R + in the tetraaryl pinacols it may be expected that these substances react in a polar medium by a mechanism which approaches the liniitirig carbonium ion type. I n the case of benzopinacol the stability of R+ appears to be inadequate to allow it to be highly discriminating in its subsequent reactions, with the result that it reacts exclusively by one of the two internal processes available to it.’” The predominance of ring ( 2 8 ) J . I,. Lane and D. K . Walters, THIS J O U R K A L , 73,4234 (1951). (24) S . Winstein and E Grunwald, ibid., 70, 828 (1948); S. Win-
btein, B. K . Morae, E. Grunwald, K. C . Schreiher and J . Corse, i b i d . , 74, 1113 (1952) ( 2 5 ) With pinacols u-hich can produce a more stable carbonium ion, external reactions with solveiit molecule.; could compete mare success-
closure relative to rearrangement probably reflects the rotational configuration of the pinacol molecule a t the moment of reaction. If the pinacol assumes a favored rotational conformation it probably would be that having a trans arrangement of the hydroxyl groups since this would minimize the C-0 dipole and phenyl-phenyl steric interactions. The molecule would be thus in a favorable state for epoxide formation. The minority of the molecules (ca. 20%) which undergo direct rearrangement would be those having the somewhat less stable cis configuration. The observed difference of approximately 1000 cal. in AFT for epoxide forination and rearrangement is consistent with this interpretation if it may be assumed that the two reactions have energetically similar transition states. Consistent with mechanism I1 are the large positive values of the entropy of activation, which are in marked contrast to the values usually found ( - 5 to + 2 e.u.) for solvolytic reactions in acetic acid which are known to be subject to neighboring group effects.26 X firmer decision regarding the relative importance of the two preceding mechanisms could be made with the aid of information concerning the extent to which the rate-determining ionization of PHT is accelerated by participation by Co-OH or by CB-phenyl and that of EH+ bv C3-phenyl participation. Recause experimental work in progress is designed to provide this information discussion of that subject will be resen-ed for a subsequent communication. ST. LOUIS,MISSOURI fully with these internal reactions. Thus. ethcrificati
