4358
Vnl. 84
COVRIGNICATIOSS TO THE EDITOR
pared readily from VIa without protection of the 3-keto group by ketalization. Reaction of VIa with ethylene glycol and p-toluenesulfonic acid furnished mainly the A6-3-ketal VIIa, m.p. 143144'; [ a ] ~ 152'; 76,7H 4.34 (two.proton singlet) and the A5-3-keta1, m.p. 150-151'; [ a ] ~ 25'; 76H 4.18 (single proton, triplet J = 3.5 c.P.s.). The 5-hydrogen in VIIa is p since its hydrogenation product IIIhI3 (Pd;'C in ethanol), m.p. 189190' and 208-209'; [ a ] ~ 67'; on hydrolysis with SOYo acetic acid a t room temperature, conditions which do not isomerize the trans-ketone IIIb, yielded the &-ketone IIIc. Hydrolysis of VIIa with 80y0 acetic acid gave the A6-3-ketone VIIb, m.p. 171-174'; [ a ]f~ 300'; no absorption a t 240 mp; 76HJH 4.13, 4.31; which with N HCl in methanol a t room temperature isomerized to VIa. Hydrolysis of VIIa with I\- KOH in methanol under helium gave the Afi-ketalacid VIIc,14 m.p. 176-177'; [ a ] ~ 150'; 76,iH 4.33 ( 2 proton singlet); The acid chloride, m.p. 140-150'; prepared from VIIc with oxalyl chloride in benzene was converted to the diazoketone and the latter with excess H I into the 21-methyl ketone VIe, m.p. 28'; GI:, 240 mp (E = 8,100); 203-204'; [ a ] D and with HCI in chloroform to the 21-chloro ke[ a ] ~ 1'; 241 tone VIf, m.p. 218-219'; mp (E = 8,700). Treatment of VIf with potassium acetate and iodide in acetone gave the 21-acetoxy ketone VIg, m.p. 160-161'; [a111 0'; X;', 240 mp ( E = 9,300).
+
constant and the exchange of one hydrogen bonded carbanion for another. (1)
s-H+g+K
+
+
+
(13) This product was identical with t h e product of direct ketalization of IIIb or IIIc. (14) Similar hydrolysis of t h e Aj-ketal methyl ester V I b gave t h e [ a ] D +ZOO; 7 6 4.18 ~ (single pro&ketal acid VIc, m.p. 237-238'; ton). Deketalization of VIc and VIIc with boiling 0.05 A' H2S04 in methanol for 2 hours gave t h e S - 3 - k e t o acid VId, m.p. 296-299'; m~.l 240 mp ( 6 = 9400).
$-. . . H B . . .K+R-.
bB k-i
+
+
+ kz *
ki
( 2 ) keoi,J (3)
kaoba
=
k-
(ky i
. . .R-. .
+ k3)
= ki ke
k-
. . D B . . . .E=.
DB
J
k 3 4 DB R-D t- BD. Racemization
HB :+
1
=
.DB
K(k2
R-D Retentioil
+ k:)
Kk3
A clear implication of k W 1> kp or ks is that in a system such as IV with a driving force for rearrangement, an intramolecular base-catalyzed proton transfer is a distinct possibility. Such a process has been observed in tert-butyl alcohol-0-D (98% deuterated) with potassium tert-butoxide as catalyst. CsHs ( 4 ) CHI*C---CH=CH?
+ KOIZ IhK
H
k, -,l
( +)-1\..2-h
CHs
CsHi
\* / C.
/ I * -'.HO.. . I-IC pJ I CH! R
+
.
R i .OD
C6H5
k,
+ CHaC=CH-CHa
Hydrogen bonded carbanion
c-is-V-4-h
THESQUIBB INSTITUTE FOR MEDICAL RESEARCH YEW BRUNSWICK, XEWJERSEY
JOSEF FRIED EMILY F. SABO RECEIVED SEPTEMBER 4, 1962
INTRAMOLECULAR PROTON TRANSFER IN A BASECATALYZED ALLYLIC REARRANGEMENT
.Tip In earlier work. li!.drogen-deuteriu~n isotope effects that ranged from 1j.3 to 3.fJ were obserl-ed ior the base-catalyzed raceniization and isotope exchange reactions of 1 , l aI I l a and III.Ib CI1, C?HixG-H(13j C:IT: I
CH
CH.
C H , O * C - H : I ~ ~ C,H.SOivC-H(D:
c
11:
I1
C f lI : - - i t Ill
The Ion \ d u e s and certain aspects of thc stercochemical course oi the reactions led to postulation of a xnechanisiii embodied in equation ( 1 ) . Ii-ith solvents in which kH, k" < 1, then k k l > k2 or ka. If kobs.e is the observed rate constant for isotopic exchange and kobsa for rxciiiizatioii, then equations (2) and ( 3 ) can be \vritten, which provide an equilibrium isotop effects associated o n l ~ with . 11,
..J
C J . C r a m . C. A. K i n g s b u r y and B. Rickborn.
J. AIII
Chew.
5m..83, 3688 :l981); i b ) 1). J ( ' r a m , I) .4. Scott and \V D. S i e l s e n . cbid., 83, 3696 (1961).
k,. CsHa I
CH?*CI CH=CHz
(+)-IN-2-d
Optically active ( +)-IV-2-h2 was dissolved in tert-butyl alcohol-0-D (0.386 M) which was 0.409 M in potassium tert-butoxide, and the solution was heated a t 75' for 257 min. The olefin was recovered (75%) through pentane extraction, olefins IV and cts-V2 were separated from one another by preparative vapor phase chromatography, and analyzed for deuterium and optical activity. Compound IV was completely free of deuterium," and exhibited the same rotation as the starting material, a Z 5 ~3.25' (1 1 dm., neat). Compound V contained only 0.46 of one atom of deuterium per molecule. The nuclear magnetic resonance spectrum of cis-V indicated that most of this deuterium was a t C-4. Clearly a minimum of 54% of the re-
+
(2) D. J. Cram, i b i d . , 74, 2141 (1952). (3) Comhustion and falling drop method.
Nov. 20, 1962
COMMUNICATIONS TO THE EDITOR
,4359
arrangement involved intramolecular proton trans- (96% sulfuric acid)? Xmax 268 (sh), 274 (4310), fer from C-2 to C-4. The possibility exists that 280 mp. A n d . Calcd. for C7H8Br2: C7H7+, some of the 0.46 atom of deuterium was incorpo- 36.17; Br, 63.42. Found: C7H7+, 35.9; Br, rated in cis-V after it was formed. These obser- 63.45, 63.21. Both salts dissolve instantly in water to give colorless solutions which neutralize vations indicate that in scheme (41, kc k d , and that if any deuterium was incorporated into czs-V two moles of base per mole of salt; qualitative during the rearrangement (as distinct from after), ultraviolet spectra of such solutions showed the presence of tropenium ion (Amax 273, 280 (sh) then k f > > keand k c > > k-@. Although in recent years a number of intra- mp)4 and no covalent species. Infrared spectraS molecular rearrangements involving carbonium ion- confirm the presence of tropenium ion in these anion ion pairs have been reported, t o the author's salts. The hydrogen dichloride does not show a knowledge this is the first example of an intra- discrete melting point, but decomposes slowly molecular carbanionic process. This phenomenon over 100'; the decomposition temperature is a may be rather general, and the scope and intimate function of rate and temperature of initial heating. mechanistic details are under active i n ~ e s t i g a t i o n . ~The hydrogen dibromide is converted to the brilliant yellow bromide on warming to 100'. (4) This research was sponsored by t h e U. S. Army Research Office (Durham). Both compounds are reasonably hygroscopic and DEPARTMENT OF CHEMISTRY were handled in the glove box a t all times. The DONALD J. CRAM stability of these compounds is demonstrated by UNIVERSITY OF CALIFORNIA ROYT. UYEDA AT Los ANGELES storage i n ~tacuounder continuous pumping a t 1 Los ANGELES24, CALIF. mm. for 18 hours which gave a 1% increase in RECEIVED SEPTEMBER 5, 1962 tropenium ion concentration in the hydrogen dichloride and a 4Yc increase in the case of the diCARBONIUM ION SALTS. VI. TROPENIUM HYDROGEN DIHALIDES AND AN IMPROVED ROUTE bromide. The vapor pressure of hydrogen halide over primary and secondary ammonium hydrogen TO TROPENIUM CHLORIDE' dihalides is quite high3" and cesium hydrogen diSir : Recent theoretical studies of the hydrogen di- chloride is reported to be unstable a t temperatures chloride anion2 indicate considerable interest in greater than 0°.3d Even tetramethylammonium this novel type of hydrogen-bonded species. We hydrogen dichloridezchas a vapor pressure of 2 mm. wish to report the preparation of tropenium hydro- a t room temperature, which suggests that this gen dichloride and hydrogen dibromide whose ease salt would be destroyed under these conditions. of preparation and marked stability may render Addition of hydrogen chloride gas to the brilliant them more amenable to physical chemical studies yellow solution of tropenium chloride in methylene than previously reported3 salts of these anions, chloride gives a colorless solution in which the and to describe a synthesis of tropenium chloride broad tailing of the tropenium peak from chloride which in yield and quality of product is superior charge-transferg is lost and a new C-T band (Xmax to any yet reported. 314 nip ( 1 i 0 0 ) ) appears. IX'hen hydrogen bromide Tropenyl methyl ether4was added to a saturated is passed into the deep orange solution of tropenium solution of hydrogen chloride in ether under a bromide in methylene chloride the color becomes stream of dry hydrogen chloride gas5; an immediate light yellow and the bromide C-T band (hxnax precipitate formed which when dried in vacuo gave 40% mp (1380))gis replaced by a new band (Amax 86.1% tropenium hydrogen dichloride as white 356 mp (1300)). This apparent shift in the C-T microcrystals, ultraviolet spectrum (96% sulfuric spectral" to shorter wave lengths shows the presence acid)' Xmax 268 (sh), 274 (4350), 280 mp. A n a l . of anions with higher ionization potentials than the Calcd. for C7H8C12: C7H7+, 55.89; C1, 43.49. corresponding halidesll as would be expected if Found: C7H7+, 55.8; C1, 43.23, 43.32. Use of the hydrogen dihalides were formed. hydrogen bromide in the above reaction gave a Tropeiiium chloride is a difficult compound; white precipitate containing excess hydrogen it is implacably hygroscopic, and we find---once bromide which upon adiabatic removal of solvent having overcome the moisture problem enough to in vacuo gave 78.0% tropenium hydrogen dibromide tell-that it is also extremely sensitive to light. as small light yellow needles, ultraviolet spectrum Diffuse sunlight rapidly darkens the chloride through Pyrex glass, and a sunlamp destroys it (1) Supported b y t h e Petroleum Research Fund and t h e National Science Foundation. completely in minutes to give several as yet un-
7
(2) (a) T. C. Waddington, J . Chem. Soc., 1708 (1958); (b) D. W. A . Sharp, ibid., 2538 (1958); (c) S. Chang a n d E. F . \\'estrum, Jr., J . Chem. P h y s . , S6, 2671 (1962). (3) (a) F. Kauffler and E. Kunz, Ber., 43, 38.5, 2482 (1909); (b) R . West, J . A m Chem SOL..79, 4568 (1967); (c) H. F . Herbrandson. R. T. Dickerson, Jr., and 1. Weinstein, ibid., 76, 4046 (1954); (d) R . E. ValleC and D. H. McDaniel, ibid., 84, 3412 (1962). (4) W. von E. Doering and L. H Knox, ibid., 76, 3203 (1934). ( 5 ) This method was suggested by our observation t h a t when tropenium chloride is prepared by passing hydrogen chloride over a n ethereal solution of tropenyl methyl e t h e 6 a a excess of hydrogen chloride gives products with low tropenium ion content. (6) Method suggested by H. J. Dauben, Jr., and L. R. Honnen, private communication, ( 7 ) H. J. Dauben, J r . , F. A. Gadecki, K . M. Harmon and D. L. Pearson, J . A m . Chem. Soc., 79, 4557 (1957), report Amax 268 (sh), 273.5 (4350), 280 mr for tropenium ion in this solvent.
(8) We wish t o thank Ilennis J. Dieitler who determined t h e infrared spectra of t h e salts by t h e potassium bromide disk technique ( 8 ) K . M. Harmon, F. E. Cummings, I ) . .4.I,a\,isand I ) . J. Iliestler, J . A m . Chem. Soc., 84, 120, 3 x 4 9 (1962). (10) Our placement of the chloride C-T hand a t 299 mgQis hased o n t h e assumption t h a t subtraction of t h e spectrum of tropenium fluoroborate from t h a t of trnpenium chloride of equivalent concentration would give a measure of the chloride C-T absorption, t h e fact t h a t t h e chloride t o hydrogen dichloride shift produces a peak a t 311 mr a n d also a loss of color makes os question t h e validity uf this subtraction. Although a value of 299 mg (33,500 an.-') gives a reasonably straight line when t h e energies of t h e tropenium halide C-T bands are plotted against halide ion ionization potentials. a true straight plot calls for a value of 31,200 ern.-', which lies t o longer wave lengths t h a n t h e hydrogen dichloride C-T band. (11) S. P . McGlynn, Chem. Re%, 68, 1113 (1958).
