DONALD J. CRAM,H. LEROY NYQUIST AND F.ATHY AHMEDXRD ELHAFEZ
28i6
Vol.
[ ~ O S T R I H L T I o XFROM T H E DEPARTXENT O F CHEMISTRY O F THE USIVERSITY OF C A L I F O R N I A A T T,OS :\XGELES]
Studies in Stereochemistry. XXVIII. Reactivity Differences between Diastereomers in the Wagner-Meerwein Rearrangement' BY DONALD J. CRAM,H. LEROYNYQUISTAND FATIIY AHMED ABDELHAFEZ RECEIVED DECEMBER 7, 1956 The rates of formolysis and acetolysis of the diastereomers of 3-phenyl-%butyl tosylate, 4-phenyl-&hexyl tosylate and of 2,5-dimethyl-4-phenyl-3-hexyl brosylate have been measured. The kerrthro/ktbreo(titrimetric)values for acetolysis are 1.2, 1.8and 6.1, respectively, and 1.2, 2.6 and 8.2 for formolysis, respectively. T h a t these ratios do not change much in the two solvents suggests t h a t in each system, the division of phenonium tosylate (or hrosylate) ion-pairs between collapse and exchange reactions with solvent is quantitatively similar for the threo and erythro diastereomers. The values of the ratios, k,/kt (polarimetric and titrimetric rate constants), decrease in value for t h e threo isomers in passing from the less to the more sterically hindered systems. As the bulk of the eclipsed groups increases, the ratio of exchange (with solvent) t o collapse of these symmetrical cis-phenonium tosylate (or brosylate) ion-pairs increases. The products of solvolysis of 2.5dimethyl-4-phenyl-3-hexyl brosylate have been examined in detail. Olefin or the products of hydrogen migration (mostly from the 2-position) dominate over those resulting from phenyl migration, Only traces of products of simple soivolysis were found. The yields of these various products indicate t h a t t h e diastereumeric rate factor of 8.2 for the formolysis in this system is due not t o differences in rate for the part of the reaction leading to phenonium ion, but t o differences in rate for the part leading t o olefin, and the products of hydrogen migration. These results suggest t h a t in this system the differences in rate for solvolysis for the threo and erythro isomers reflect more the differences in energy of the diastereomeric starting states rxther than differences in eclipsing rffects for the diastereomeric transition states.
Previous studies of the Wagner-Meerwein rearrangement in systems represented by I have uncovered marked differences in behavior of I with R = CHs2 and R = C2H5.3 hlthough OTs I k RCHCHR
I
C6H6
OTs-
k,
RCH-CHR
_____f
...+..
exchange (+SOH-HX)
CsHj
1
i
k , collapse
os I
R-CH-CH-R
OTs 'R-CII--&H
1 C6H5
--R
OTs
I + R-CH--CH-R ~
C6H5
CsH5
anism for solvolysis for both systems can be encompassed by the formulation, the ratios kp/k, and k c / k e vary enough to affect the balance of products. ,The rates of solvolysis (k, is the titrimetric rate constant) and of change of optical activity (k, is the polarimetric rate constant) have been reported for threo-I with R = CH, in both formic and acetic acids,4 and k t was reported for erythro-I with R = CHa in acetic acid.4a The rates kt and k, also were obtained in acetic acid for threo-I with R = CzHj, and K , (but not k t ) was also determined in formic acid. This paper is concerned with reactivity differences between diastereomers of I with R = CH8, C2Hs and CH(CH3)2. The new kinetic data on particularly the erythro isomers of the first two systems coupled with product and kinetic data on the third system allow an evaluation of conformational and eclipsing effects in a series of homologous (1) This work was sponsored by the Offce of Ordnance Research, U. S. Army. (2) (a) D. J. Cram, TRXS J O W R N . ~ L ,71, 3863 (1949), (b) 74, 2129 (1952); (c) 74, 2137 (1952). (3) D. J. Cram and F. A . Abd Elhafez, ibid., 75, 3189 (1953). (4) (a) S. Winatein, B. K. Morse. E. Grunwald. K. C. Schreiber and J . Corse. ihid., 74, 1114 (1952); (b) S. Winstein and K. C. Schreiber, i b i d . . 74, 2164 11952); (c) S. Winstein and H. Marshall, i b i d . , 74, 1120 (19'52).
compounds of increasing steric constraints. XI1 three of these systenis possess symmetry properties which allow a maximum amount of information to be extracted concerning the role that phenoniurii ions play in the Wagner-hleerwein rearrangement. Kinetics of Solvolyses The kinetics of acetolysis and formolysis were followed utilizing methods previously described, 4b " and in all cases were found to be cleanly first order. A11 four stereomers of 3-phenyl-2-butanol (11), 2 4-phenyl-3-hexanol (111)5 and 2,5-dimethyl4-phenyl-3-hexanol (IV) have been previously prepared in an optically pure state, and the absolute configurations of I1 and I11 have been cornpletely assigned. -4s a result of the present work, all of the relative configurations of all four isomers of I V can be definitely assigned. None of the absolute configurations of the isomers of IV are known, and therefore a completely arbitrary assignment is made for convenience (see formulatiorl).
(+)-II, R '= CH1 ( + ) - I I , R = CH? (+)-III, R = C2H5 (+)-III, R = C2H.j (+)-I\., R = (CH3)iCH ( + ) - I V , R = (CHZ)lC€I
Table I reports the rate constants needed for comparison of the three systems and of the two diastereomers within each system. The rates were not all run under completely identical conditions. For instance, runs 1-6 were conducted in the absence of a salt for neutralization of the acid liberated, whereas runs 7-21 were made in the presence of such a salt. Enough data were obtained on system IT4:v4b to indicate that the rate ratios reported in this paper were only slightly sensitive to the presence of salt. In the formolysis of systems I11 and IV, about 20% by volume of chloroform had ( 5 ) D. J . Cram, F. A . Ahd Elhafez and H. Weingartner, i b i d . , 7 6 , 2293 (1953). (6) D. J. Cram, F A Abd Elhafez and H . LeRoy Nyquist. zDid., 76, 22 (1954).
2877
DIASTEREOMERS IN IIIE U'AGNER-~IEERWEIN RCARRAKGEXENT
June 5 , 195i
TABLE I RATECOKSTASTS FOR SOLVOLYSES OF SCLFONATE ESTERS OF DIASTEREOMERS OF ALcorioLs 11. 111 A N D 11' Run
1" 2b 3b 4" jb
6* 7 8
Compound
Solvent
threo-11-Ts threo-11-Ts threo-11-Ts evythro-11-Ts threo-11-Ts threo-11-Ts erythro-11-Ts erythro-11-Ts threo-111-Ts threo-111-Ts threo-111-Ts erythro-111-Ts threo-111-Ts threo-111-Ts erythro-111-Ts threo-IV-Bs threo-IV-Bs erythro-IV-Bs erythro-IV-Bs
AcOH AcOH AcOH -4cOH HCOOH HCOOH HCOOH HCOOH AcOH AcOH XcOH AcOH HCOOH' HCOOH' HCOOH' AcOH AcOH AcOH AcOH HCOOH' HCOOH"
lod Ild 12 13 14d 15 16 17 18 19 20 threo-IV-Bs 21 erythro-IV-Bs
Concn ester, mole/
0.030 ,116 ,116 ,032
.OiO ,116 ,096 ,098 ,0882 ,0882 ,0900 ,030 ,091 ,103 ,030 ,079 ,079 ,079 ,079 .035 .034
Concn. salt, mole/l.
Salt added
.. ..
. ...
iYaOi\c Na0.h
0.0957 0.0957
..,. . .. .. .
. .. . .. ...,
XaOOCH SaOOCH KOAc KOXc KOAc KOXc NaOOCH KaOOCH NaOOCH KOAc KO.-lc K0.k KOAc NaOOCH NaOOCH
0.119 ,119 ,105 ,105 ,120 ,035 ,123 ,130 ,034 ,105 ,109 ,105 ,109 ,041 ,041
to be added to make the system homogeneous. The solvent was always the same for any two rates appearing as a ratio in this paper. The polarimetric and titrimetric rates for erytlzro-I1 tosylate formolysis are almost the same (runs 7 and 8) and serve as a check on the method. Unfortunately the polarimetric rate of acetolysis could not be determined because some of the optically active products resulting initially underwent reaction a t a rate comparable to that a t which the solvolytic reaction occurred (see Discussion). An analysis of the meaning of these kinetic data depends upon a knowledge of the reaction products, a subject taken up in the next section.
T, Proc.
OC.
Tit. Tit. Pol.' Tit. Tit. Pol.' Tit. PoLC Tit. Pol.' Tit." Tit. Tit. Pol.' Tit. Tit. Pol.' Tit. Pol.' Tit. Tit.
74.71 74.91 74.91 74.72 24.98 25.12 25.00 25.00 74.61 74.64 50.08 50.00 25.00 25.2 25.00 49.74 49.74 49.74 49.74 25.00 25.00
Rotations a0 Init. Final
.. ..
k, sec. -1
(4.95 f 0.17) X , . . . (6.72 f .16) X 0.53 -0.01 (3.33 f . l o ) X , . . . . . . . ( 5 . 7 7 i .26) x , . . , .. ( 2 . 2 8 f .03) X 3 . 9 3 - 0 . 0 4 ( 2 . 6 5 i .12) X .. , , . . . . ( 2 . 8 0 f ,021 x -2.53 - 1 . 6 0 ( 2 . 4 9 f .18) X .... . . , . ( 1 . 6 8 zk .03) X -1.47 - 0 . 0 6 ( 5 . 4 8 i .19) X .. . . ..,. (1.01f .02) x (1.80 =t . o i ) x ... . . . . (2.90 f .03) X -1.44 0.06 (3.90 f . l o ) X , . . . , . . . ( 7 . 6 5 f .09) X . .. . .. (4.97 + .09) x -4.64 +0.84 ( 6 . 0 5 i .05) X .... . . . . ( 3 . 1 8 i .03) X -0.83 $ 1 . 6 5 ( 3 . 2 8 =I= .07) X , , , , .... (3.10 f ,011 x (2.53 i . O I ) x , , . ,
, ,
.
,
,
,
,
10-5
loT4 10-4
loF4 10-5 10-5
lop4
10-5
lo-' 10-4 10-3
been previously reported.?^^ The pattern of products obtained by the solvolyses of 2,5-dimethyl-4phenyl-3-hexyl brosylate' (IV-Bs) is somewhat different from that found for the two simpler homologs. These products (alcohols V and VII) as well as the potential product VI were all prepared by conventional means since their detection in the alcohol mixtures from the hydrolysis of the solvolysis mixtures demanded a knowledge of their individual physical properties. Alcohol V was prepared in an essentially optically pure form since the analytical scheme necessitated a knowledge of its rotation. Alcohol VI was only prepared as its racemate. The third alcohol (product of two
.1
1, (CH3)KHLi 2, H30+ CHARTI :
OH
I
purified through its CH3-CH-C-CH2-CH-CH3 3,5-dinitrobenzoate I 1 I CHI C6H5 CH3 VI NaOEt ( C S H ~ O O C ) ~ C H C G H (CH3)2CHCH2Br S -+1, O H ( CzHa0OC)nC-CHnCHCHa CH300CCH-CHzCHCHa I I 2, CHaO'H2 I I CcHj CHa CsH5 CHI
+
Products of Solvolysis in the 2,5-Dimethyl-4phenyl-3-hexyl System.-The products of formolysis and Of the tosylates Of the 3PhenYl-2-butyl and 4-phenyl-3-hexyl Systems have
CHI 1, CH3MgI
2, .&OF
I
CH3C-CHCH2CHCH3
I
1
OH C6H5 VI1
I
CH3
1,2-H shifts) was also synthesized (see formulas) as a racemate, (7) The tosylates of this system, although less reactive kinetically, are much harder to obtain in a pure state than the brosylates.
28713
DONALD J. CRAM,H. LEROY NYQUIST.WD FATIIT ai^^^^^^ Xnn ELII~ F I - L
OH
OH (CHdzCHCHhHCH(CH3)2 I CsHi IV, Secondary alcohol fraction r
Products Sec. alc. fraction Yield, yo
(+)-thrco-IV-Bs AcOH ( r u n 1 )
OH
I
+ ( C H ~ ) Z C H ~ H C H Z C ( C4-H J( C) ~H ~ )I~ C C H C H Z C H ( C-tH ~ole511 )~ I
I
CBHS
CaHs YII
v
i
Tertiary alcohol fraction
-
Startinr materiai-
(+)-prythro-IV Bs AcOH (run 2)
( - ) -thrPo-IV-ns HCOOIi (run 31
(--)-~yth?o-IY-13~
IlCOOH
(rut1 I J
[alz5D
Composition (fraction = 100%) Tert. alc. fraction Yield, % a% (1 = 1 dm., neat)
n% Composition Olefin fraction Yield, % fl"D
(1 = 1 dm., neat) (232 (c, 2.1, CI1C13).
aybD t
IZ
-23.00 1592
-72.1' 474
After recrystnllization (c 2, CHCla).
The brosylates of all four of the optically pure isomers of IV6 were prepared. The sulfonate esters of the (+)-threo- and (+)-erythro-alcohols (these isomers differ only in configuration a t the carbon carrying the oxygen) were acetolyzed a t 50' for 8 titrimetric half-lives. The brosylates of the (--)-threo- and (-)-erythro-alcohols were formolyzed a t 25' for 5 titrimetric half-lives. The olefinester products were converted to olefin-alcohol mixtures with lithium aluminum hydride. This mixture was split into an alcohol and olefin coiiiponent by chromatography. The alcohol fraction was then itself split by careful chromatography into a secondary and tertiary alcohol component, each of which was then examined in detail. Table 11 records the important data. Control experiments established that the separations were almost quantitative and that the olefins and tertiary formates interconverted somewhat in the formolyses. That tertiary and secondary acetates can be reduced with lithium aluminum hydride without producing olefin has been established previously.2c Secondary Alcohols Obtained from Solvolyses.Infrared analysis6 of the alcohol fraction obtained in the two acetolysis runs established that from threo-brosylate (run 1), the final secondary alcohol product was about 97y0 threo-IV and 3y0 erythroIV. From erythro-brosylate (run Z ) , the final secondary alcohol product was about 95yo erythroVI and 5% threo-IV. The rotations of the secondary alcohols obtained in these two runs indicate that the threo-alcohol obtained from threo-brosylate (run 1) is essentially racemic, whereas the erythroalcohol obtained from erythro-brosylate (run 2) is essentially optically pure. Although the infrared :in,rlysis wa5 iiot performed on the alcohols from
+(I2 7 3 i2 1
+ 5 4 80 1788
I = 1 dm., neat
Cycloliiu,inc
AcOH 25 HCOOHd 50 AcOH 2 5 HCOOH'
-
,I6 .71 ..58
,5f
-
.83 .87 .42 ,3g
.I3
-
.OQ .67 ,770
-
.04 .04 .01
.OQ
+ kl, + ko + k.. *Temperature difference of 0.14" between runs 5 and 6, ka being the higher. OTemk, = k,
perature difference of 0.20" between runs 13 and 14, ka being t h e higher. d Actually a 3.33/1 misture by volume of HCOOH-CHCla. e Actually a 2.86/1 mixture by volume of HCOOH-CHC13. f This ratio was calculated assuming that the same fractions of products were produced via racemized brosylate ill formolysis and acetolysis, a n assumption supported by the rotations of the olefins. 0 Calculations based on same assumptions as in (f).
Table I V compares the kinetics of the two diastereomers of each system in terms of the partial rate factors kp (phenyl participation in ionization) ko (a rough measure of hydrogen particiand k h pation). The virtue of comparing reactivity differences between diastereomers lies in the fact that any differences must derive fundamentally from geometry. Unfortunately kp cannot be calculated directly for the eryfhro runs, since the ratio ke,l(k, ke) cannot be determined. The best that can be done, therefore, is to calculate k, (erythro) by kc) do not assuming that the values of k e / ( k e vary much between the diastereomers. This assumption is supported by the fact that although the values of k,/kt (Table IV) vary by factors of 4 and 2.5 in passing from acetic to formic acids (for systems I1 and 111), the values of kt (erythro)/kt
+
+
+
I
I
H
H
J
I
C&
. . .H
\
CH3
K these ratios in these two systems is 2.8, a"fact which indicates both that eclipsing effects are small and that the differences in energy between the starting states are small. In IV, the ratio of kp's remains near unity, but ( k h ko) e r y t h r o / ( k h ko) threo amounts to about 9. In this more hindered system, k h ko measures mostly involvement of C61-H. Since Cgf carries two methyls, eclipsing effects in the transition state K should be about the
+
+
+
(11) (a) D. J . Cram and F. A. Abd Elhafez, THIS JOURNAL, 75, 339 (1953); (b) D. J. Cram and F. D. Greene, ibid., 76, 6005 (1953). D. H. R. Barton and R. C. Cookson [Quarterly Rcuicws, 10, 48 (195811 have also recognized these principles and quote as an example the work of R. P. Linstead and M. Whalley [ J . Chcm. Sot., 3722 (1954)l on the relative stability of mcso- and DL-2,3-disubstituted succinic acids. In this and similar systems as well as those such as the stilbene dibromides l8.g.. R. E. Buckles, W .E. Steinmetz and N. G. Wheeler, TEE JOURNAL, 72, 2496 (1950)l. dipole-dipole interactions may play a much more important role than simple steric effects. In these types
of systems, the electronic and steric effects happen to reinforce one another. The only tests of the purely steric principles (of which the authors are aware) are those quoted in ref. 11s and llh. (12) I n the analysis of IV, phenyl is selected as being effectively more bulky than isopropyl, a choice supported hy the direction of asymmetric induction in the preparation of the system (ref. 6). (13) See D. Y . Curtin [Record Chcm. Progress, 16, 111 (1954)j for a general discussion.
DIASTEREOMERS IN THE WAGNER-MEERWEIN REARRANGEMENT
June 5, 1957
2883
same for the two diastereomers. Therefore, the factor of 9 is mainly attributable to differences in energy of the two diastereomeric starting states. The proximity of k , (erythro)/k, (threo) to unity therefore implies that the geometries of the transition states for phenyl participation resemble those of the starting states, and that phenyl provides only a small driving force for ionization in this particular system. The relative absence of eclipsing effects in IV recalls the data obtained in a study of the E2 reaction in the 1,2-diphenyl-1propyl-X system. The ratio X?E, (threO)/kE, (erythro) in this system could be varied from 1 to 57, depending on the strength of the base, the character of the solvent and of the leaving group.i4 I n system I V and probably in I1 and I11 as well, the transition states for phenyl migration still leave the substituents on C, and Cp well staggered, and the three-membered ring only slightly formed. Experimental
-
a m W
2
E
s P
2
E
w
9
s $E a 5 a
0 0
cn
F
2
O
m
2
Preparation of the Brosylates of the Stereoisomers of 2,s-Dimethyl - 4-phenyl - 3 - hexanol (IV) .-The ordinary methods of preparation of either tosylates or brosylates in this system failed. T h e following procedure was found applicable t o the preparation of all four stereomers. T h e starting four alcohols were reported previously in an optically pure state, as well a s in their two racemic forms.6 T o 25 ml. of benzene (distilled from potassium) were added 1.23 g. (1.42 ml.) of sodium-potassium alloy and 2.58 g. of ( -)-threo-IV. T h e mixture was stirred under dry, pure nitrogen a t 25" for one hour and a t reflux for 4 hours. Control experiments indicated this to be the minimum time for formation of t h e alcoholate. T h e reaction mixture was cooled t o Oo, the remaining alloy removed (under nitrogen), and a solution of 3.26 g. of pure p-bromobenzenesulfonyl chloride in 15 ml. of dry benzene was added rapidly. T h e resulting mixture was permitted to warm to 15' during the first hour and then stirred a t 25" for a n additional hour. A few drops of solution gave a neutral test with phenolphthalein solution. T h e reaction mixture was washed with water, dried (sodium carbonate), treated with a few drops of pyridine and evaporated a t 25' under diminished pressure. T h e resulting oil was crystallized a t -20" from an equal volume of pentane (containing a few drops of pyridine). T h e liquor was decanted quickly, and the crystals were dissolved a t 25' in a minimum amount of ether containing a few drops of pyridine and recrystallized by cooling to -20". After two recrystallizations, the sulfonate ester was obtained ( b y decantation), dried under vacuum, wt. 3.70 g . (70% yield), m.p. indeterminate because of decomposition. Anal. Calcd. for C20H2503SBr: C, 56.46; H , 5.90. Found: C, 56.64; H, 6.12. T h e other three stereomers were prepared in from 50707, yields, and showed similar instability, which prevented the determination of any meaningful m.p.'s or rotations. T h e ester of ( +)-threo-lV was not successfully analyzed because of its instability. T h e analyses of (+)-eryLitro-IV and ( - )-erythro-IV, respectively, are as follows. Anal. Calcd. for CZOH2503SBr: C, 56.46; H , 5.90. Found: C, 56.54, 56.33; H, 5.88, 5.79. Preparation of the Acetates of the Stereomers of 2,5-Dimethyl-4-phenyl-3-hexanol (IV) .-In the hope t h a t the acetates of the four stereomers of I V would have rotations t h a t would aid in the analysis of unknown mixtures of stereomers, these esters mere prepared by the usual pyridine method in Isomer ( f)-threo-IV gave Zz6D 1.4868, yields of 85-93%. CPD f104.3" (1 = 1 dm., neat). Anal. Calcd. for ClcHz40a: C, 77.37; €1, 9.74. Found: C, 77.59; H , 9.72. Isomer ( -)-tlzreo-IIr gave n 2 5 ~1.4868, 0 1 2 5 ~ - 104.4' ( I = 1dm., neat). A n d . Found: C, 77.31; 11, 9.47. (14) D. J. Cram, F. D. Greene and C. H. Depuy. THISJ O U R N A L , 78, 790 (1956).
Isoiiier (-t)-erylhro-IV gave tii.p. 45-46;', a% +85.7' = 1 din., iieat). Anal. Found: C, 77.15; H, 9.52. Isomer (-)-erythuo-IV gave m.p. 45-46', a25~ -85.7" ( 1 = 1 din., neat:. Anal. Found: C, 77.58; H, 9.64. Kinetics of Solvolysis of the Brosylates of the Stereomers of 1V.-The acetolysis solutions were prepared from C.P. 'icetic acid, anhydrous potassium carbonate arid sufficient acetic anhydride to emsunie t h e water produced and leave ; L I();, excess. The ampoule technique w a 5 emIilo>-edin both tlic titrimetric aiid polarimetric runs. T h e titratiiin reagents ,i~itlprncetlures Iiave been reported previously. l o T h e polarimetric readiiigs \yere taken in a therino,t;ttetl 2-dni. tube a t 25'. In a11 cases from 8 t o I:! points \yere taken, :mil the reactioiis followed t o 85--95c,b coinpletioti. A fixed infinity reading was olitaiiicd in t h e pol:irinietric runs. T h e lorriiic :icitl uwd in forniolysis \vas purified as before4c (100.01 '," by titratirm with &drl I'isclier reagent), and the requisite :tniounts of anli>-drous sodium formate and r i f etliiiniil-free anli)-tirou:, chloroform were added. T h e , i n ~ o u n tof elili~rofi~rni needed was determined by trial experiments, liornrigeiicity throughout t h e run being t h e criterioii. T h e titrimetric runs were followed by withdrawing .i-rnI. aliquots of solution from t h e reaction and quenching tliis in 50 ml. o f purified dioxane, and titrating this solution with standardized perchloric acid-dioxane solution (method ( i f \Vinsteiti :-l iodide and 8.0 g. of magnesium turning\, was added a solution of 3.i) g. of meth\-l 4-nietliyl-3-pheiiylpentanoate iii 50 nil. of ether. T h e reaction mixture w a < stirred for 10 hours and tlien refluxed vigorously for one boutbefore decomposing tlie reaction mixture with 200 ml. o f ice-cold saturated ; L I ~ I I I ~ O I I ~ ~chloride III solution. Tlic phases were separated, aiitl t h e aquerms phase \vas acidifieti with sulfuric acid and then estracted with ether. T h e ether extract was washrtl, tlrictl, ev;ipiir;ited. and t h e remaining oil was absorbetl 011 200 g. of hasic aluniin:i of :ICtivity I . I 9 -1fter elutiim of the column with pentane h u l produced i i o residue, the prodnet v a s eluted wit11 50' , methanol iii ether. Tlie eluent was evaporated, : ~ n dthe liquid residue ab diwilvcti iu pentane, washed with water and dried. Tlie pentarie w;is evaporated, and the liquitl residue was distilled a t 100" pot temperature (1.4 IIIITI.) t(1 yield 2.76 g. (92[ yicltl I of 2,:i-tlin1etl1yl-4-phen~-1-2-l~ex:r11~~l ~ ~ D ~ D(I::, ~ z 1.j021, A n u l . Caieti. f i r ClrH&: c , 81.30; 11, 10.75. Foui~tl: C , 81.53; 13, 10.60. T h e 3,5-(iitiitrciherizo3tc of L7 was 1)reparecl from 0.10 g. of Y,0.3 ml. o f reagent pyridine and 0.12 g. of freshly prepared 3,5-diiiitrobenzoyI chloride a t 90" ( 2 hours'i. The product \vas isolated in t h e usual way and recrystallized twice from ether-pentane t o gire 0.15 g. (i7e;) yield) of light yellor needles, m.p. 59.5-90.5". a l n u i . Calcd. for C ? 1 1 ~ 2 4 0 6 SC, ~ : 62.99; 11, 0.0-1. Found: C, 63.06; 11, 6.13. Hydroll-sis of the :rbo\-e clei-i\-;itivcit1 ;I n i i x t u ~ ~of! 0.05 g. of potassium h>-droxidciti one 1n1. o f \vater and 4 1111. of cthm o l a t 95' fiir 8 hours gave linck the original alcohol Y in almost quantitative yield. T h e acetate of V proved tliliicult to prepare, tlie folloiviilg inethod giving the beit result. r\ mixture of 0.109 g. of I., 0.5 mi. of pure acetic anli!dride and 0.4 nil. of pure pyridine was held a t 50" for 12 ( l ~ y s . T h e product T Y ~ Sisolated ill t h e usual way, absorbetl 011 a column of 50 g. i i f neutral alumina of activity 1 1 9 made up in pure pentane. T h e product was eluted with 350 ml. of 125; ether in pure pentane, the solvent first was evaporated at 30 mm., and then at 1 cquiv., 192. Found: C , 75.16; 1-1, 8.37; iieut. equiv., 192. nini. T h e acetate was not distilled due t o t h e possibility of Tlic anilide \vas prepared iii the usual way, ni.p. 113.5slight decomposition h i t was analyzed directly, 1 z Z 5 1.4864. ~ 116" (ethl-l acetate). z l n n l . Calcrl. fur C , o ~ 1 2 4 0 2C,: 77.37; H,'3.74. F o U l l d : C, 77.38; H, 9.80. . I d . C.ilccl. for ClblLISO: C,S0.Xti; 11, 7.92. I~CILIIK!: c, 80.80; 11, 7.92. :\ttetnpts t o form tlic fnriiiate o f 1; g a v e ;I inixturc of olcliil :lilt1 formate. I!tilizing tlic ; i h i v ~Iiriicrcliirc, ~ iiljticdly pure l-ii1etl1yl-~~~~lieii~-ipeiitatir,ic acid (0.875 g . ) was cotiverterl to its rnetliyl ester (0.870 g . ) , YL% t.3931, a% +33.47 ( I = 1 titn., lieat). . l n d . Calcd. for C13111~02: C, 75.(1 \ \ C I Ctliswlvctl i i i iieiitaiic, tlic solution was washed Ivitli sotliutn sulfite soluticm,
water, dried and evapuratcd. Distillation of the oil a t a p o t tempcrature of 102-103" (2 nim.) gave 0.47 g. of tertiary alcohol fraction. See Table I1 for properties. T h e infrarctl spectrum of this material proved to be identical to t h a t of 2,5-di1nethyl-4-phenyi-2-he~anol atid was demonstrated through comparisons of this spectrum with tliat of it syiithetic mixture of V containing 270 of IV to contain less than 2 % of this secondary alcohol. Unfortunately this material was lost before a derivative could be prepared. Acetolysis of the Brosylate of ( +)-thrco-2,5-Dimethyl-4phenyl-3-hexanol (Run 1 of Table 11).-The brosylate ester (6.123 g.) of optically pure alcoliol was acetolyzed as in run 2 in 150 ml. of stock solution for 31 hours (8 titrimetric halflives). T h e products were isolated as in run 2 t o give 3.03 g. of olefin-ester mixture, which gave 1.43 g. of olefin and 1.20 g. of a mixture of secondary and tertiary alcohol (see Table I1 for physical properties). This mixture was submitted t o infrared analysis (see later section). T h e secoiidary and tertiary alcohol fractions were separated by chromatography, 1750 ml. of solvent developer coniing through the column without any dissolved alcohol in between the secondary and tertiary alcohol fractions. T h e secondary alcohol fraction amounted t o 0.69 g., its infrared spectrum being identical to t h a t of racemic threo-IV. This material (0.161 g.) was converted t o its p-nitrobenzoate derivative,B yield) being obtained after being crystallized 0.197 g. (71yU once from ethanol, m.p. l15.5-118° (undepressed by admixture with an authentic sample),6 [ o ~ ] ~+0.45' ~ D (c 3.3, CHCla). T h e same derivative o f (+)-threo-IV has m.p. 145-116', [Q]"D +lO5" ( C 3.3, CHCl.i).6 Mixed melting points of the folloFvirig p-nitrobenzoatcs were obtained: (1) 5y0 derivative of (-)-threo-IV +05yo racemic derivative, 118.5-12:;"; (2) 0.5yoderivative of ( - ) lh7eO-IV 09.5% racemic derivative, 11?.5-11g.5°; 1';;) derivative of ( +)-erytizuo-IY $- OO',-h racemic derivative of threo-IV, m.m.p. 110.5-14O".
+
0 461 1010
1.0.5
7ti8
0.254 0.232 0.620 ,530 . '763 ,061 1.00 ,061 .293 0.819 ,219 .226 ,840 . 174 ,242 . l7ii ,490 .404 .385 ,166 .'$79 The readings ivere taken succcsiivcly iii
was accounted for. This mixture was submitted to infrared analysis (see later section). T h e alcohol fraction \vas chroniatographed on 171 g. ( 3 cm. by 35 cm.) of neutral alumina of activity 1'9 in pentane, tlic clironiatograph column being wrapped in heavy paper to exclude light. T h e column was developed with 32% ether in pentane (ether was washed free of peroxides just prior to use and carefully distilled). T h e results are tabulated (tlie chromatograph was run continuously for 27 hours). T h e residues from cuts 1-19 were combined with pentane and dried to give 0.0426 g. of solid material, m.p. 66-68', [ a I z 5+17.3' ~ (c 4.3, CHC13). Recrystallization of the material from pentane gave 0.0243 g., m.p. 68.469.2", [ a l Z z$20.7' ~ (c 1.7, CHCls), undepressed by adniixture with an authentic sample of optically pure (+)erylhro-IV. El tien t
1 s t set of equations
0 420 1049
2ud set of equations 0 BOO 1.20 838:l 766.3
1.16 .209 .014
.os2 ,041 .;33 ,799 ;I O.O:(-iiiui.
0.012 ,271 ,002 ,1131 ,019 ,012
0 . 109 .3:;3 . '335
,800 , s39 32.'
.U0(i
.:;12 cell, homogeneous
T h e tertiary alcohol fractioii ainouritcd t o (1.24 g., and i l i infrared spectrum was almost identical t o t h a t of authentic alcoliol V. This fractiou (0.160 9.) was coiiverted to its 3,5-dinitrobenzoate derivative (see previous section) to give u p m crystallizatiou froiii ethaiiol 0.1 554 g. (>O',L vield) of orange derivative. This inaterial \vas crystallized three times from ethanol (charcoal treatmeut) t u give 0.048 g. of derivative, m.p. 82.5-84' (decomposition, 162.5163.5'), [ o ~ ] ~-8.37 ~ D ( c 3 . 3 , CHCla). The same derivative of optically pure V (n1.p. 89.5-90.5') gave [.I]"D + l 5 . 0 l o (c 3.3 CHC11). The mised melting poitit of this derivative uf active and racemic V gave 80.5-82'. I;ractiimal cry\tallizatiori of the filtrates from the above purification g a v e no material t h a t melted above the melting point of the derivative of optically pure 11 (89.5-90.5O). Infrared Analysis of Secondary-Tertiary Alcohol Mixtures from Runs 1 and 2.-The infrared spectra o f pure threo-IV, erythvo-IV and V were determined, and the most atlvantngeous wave lcngtlis for analysis selected. T h e ai,sorbance of the two unknown mixtures a t these ~ ~ a lengths v e \\-asdetermined. Two independent sets of three simultaneous equations were solved t o give estimates of thc thrcc components. Syiitlietic mixtures of approximately the s a n ~ e composition were prepared and similarly aiialyzctl, a1111the iinknown mixtures were corvxted to the k ~ i c l ~ vfor n deviatims frliiri Beer's law. T a l ~ l eV records the pertinent data. For 2 , the first set of equations guvc lLj0 threo-IV, 17 ,,o-IV and 88% V,ant1 the second set gave If% thu~.o-I erythio-IV a ~ i dbrO,;' V. F(Jr ruii 1, tile first s c i . ~ t h , ~ o - iari(1 V timis gave TO',;of thico-l e of secondary and tci ticci-y alcohol obtained froin tlie chromatograms. Preparation of Formolysis Stock Solution.-Formic iiciil was purified by the method of JVinstein and and the chloroform w i i s purified inirnediately before use. ;2 inisture of 105 ml. (1.55 g.) of chloroform in ;i00 mi. of formic acid was prepared, aiid the solution made 0.040 iV in sodium formate with anhydrous, mialytical grade reagelit. Formolysis of the Brosylate of ( - )-erythro-Z,5-Dimethyl-4phenyl-3-hexanol (Run 4 of Table 11).--4 mixture of 4.7% g. of brosylate in 320 ml. of stock formolysis solutioii \va$ shaken vigorously until solution K ~ Scoiiiplete, and the resulting solution was held a t 25.00 zk 0.01' for 23 minutcs ( 3 titrimetric half-lives). The products v-ere isolated as in run 2 t o give 0.293 g. of ulefin and 1.552 g. of a mixture of the secondary and tertiary alcohols (814.;yield of products a t this point). The alcohol mixture was submitted to chromatographic separation to give a secondary and tertiary alcohol fraction with a large blank fraction intervening. T h e secondary alcohol fraction came to 0.137 g. of ITlatCrial, m.p. 68.4-69", m.m.p. with authentic (--)-evythuo-lV, 69..5-70.5" (see Table I1 for other properties). A satiiple of this fraction (0.0795 9.) was converted to its p-tiitrobenzoa t e derivative6 t o give 0.0913 g. (G770 yield) of material, m.p. 168.5-169.5' (undepressed by admixture thentic material), [ a ] * %-46.1' (c 3.3, CHC13). material6 gave m.p. 168.5-170.O0, [a] *51) -47. CHClj). The tertiary alcohol fraction caiue to I .IS9 g . I I fur propcrties). I t s iufrared spec'tiulri iuclicatcd tlic PI v i C I I C C o f less than 256 of evythrothis niaterial \vas coiiverted to previous section) to gi\se 0.257 g
June 5, 1057
DIBSTEREOMERS I N THE WAGNER-Lf EEKWEIN REARFdNGEMENT
2887
product, m.p. 87-89', [ a j Z 5 D +12.91' (c 3.3, CHC13). ane); a t the end, n z 61.5007, ~ a Z 7 D -47.24" (1 = 1 dm., One recrystallization of this material from ethanol (char- neat), A232 gave E 2040 (cyclohexane). coal) gave0.145 g. of purematerial, m.p. 89.3-90.2", m.m.p. Controls on the Stability of the Tertiary Esters Formed in the Solvolyses.-The formate of V could not be prepared, with a n authentic sample, 89.5-90.5", [ O ~ ] ' ~ D +14.2" (c 3.3, CHC13). Fractional crystallization of t h e filtrates so t h e more stable acetate of V was used in the control experiment. A solution of 0.201 g. of material was dissolved gave crops melting above t h e melting point of t h e 3,5-dinitrobenzoate of V, e.g., m.p. 87-104'. This material was in 40 ml. of formolysis stock solution and maintained a t 25' for 190 minutes. T h e product was isolated in the usual probably contaminated ester of VI1 (m.p. 133.5-134.5'). Formolysis of Brosylate of ( -)-threo-Z,S-Dimethyl-4- way, treated with lithium aluminum hydride and chromatographed to give 0.0347 g. of olefin arid 0.0749 g. of an alcohol phenyl-3-hexanol (Run3 of Table II).--A mixture of 5.0 g. of t h e above ester and 335 ml. of formolysis stock solution fraction. Clearly t h e formate of V would have been unstable under the conditions of its formation in the formolywas shaken until homogeneous, and the resulting solution ses. (0.035 molar in ester and 0.041 in sodium formate) was held ,4 solution of 0.0346 g. of the acetate of 2,5-dimethyl-3a t 25.00 0.01' for 190 minutes ( 5 titrimetric half-lives). phenyl-2-hexanol (VII) in 5 ml. of acetolysis stock solution The products were isolated as in run 2 to give 0.960 g. of was held a t 50" for 31 hours (8 titrimetric half-lives of the olefin (see Table 11) and 1.141 g. of alcohol fraction (91% yield of products a t this point). T h e alcohol mixture was acetolysis of brosylate of threo-IV). The product was isochromatographically separated into secondary and tertiary lated, treated with lithium aluminum hydride, and the alcohol and olefin separated by chromatography. The olefin alcohol components with a large blank fraction intervening. The secondary alcohol fraction came t o 0.226 g. (see fraction amounted to 0.0013 g. and the alcohol fraction to Table 11), and possessed an infrared spectrum identical t o 0.0225 g. (69% recovery). This acetate appeared to det h a t of threo-IV. Its p-nitrobenzoate0 was prepared from compose only slightly under the conditions of its possible formation. T h e similarities in structure between X r and 0.129 g. of alcohol t o give 0.158 g. (71y0 yield), m.p. 119122' (undepressed by admixture with an authentic sample VI1 indicate the acetate of V should be equally stable under the acetolysis conditions. of derivative of racemic threo-IV), [ a l Z s-2.13' ~ (c 3.3, A solution of 0.0989 g. of the acetate of VI1 in 30 ml. of forCHC13). T h e tertiary alcohol fraction came t o 0.536 g., and its molysis stock solution was held a t 25' for 190 minutes. infrared spectrum showed the presence of not more than 370 Work-up of t h e mixture gave 0.0066 g. of alcohol ( 3 7 5 rethveo-IV. This fraction (0.201 g.) was converted t o its 3,5- covery) and 0.0214 g. of olefin. This experiment indicates t h a t the formate of VI1 would have been unstable under the dinitrobenzoate derivative to give 0.270 g. (69% yield) of conditions of the formolyses, probably equilibrating with light orange crystals, m.p. 72-83', [a]"D +9.38' (c 3.3, CHC13). Three recrystallizations of this material from the conjugated olefins. Control on the Dilution Effect of Racemic 2,5-Dimethyl-3ethanol (charcoal treatment) gave 0.0348 g. of material, m.p. 83-87", [ a I z 6+13.1 ~ (c 3.3, CHCla). This material phenyl-2-hexanol (VII) in Optically Pure 2,5-Dimethyl-4solution of 0.0392 g. of pure was partially racemized derivative of V. Fractional re- phenyl-2-hexanol (V).-A racemic VI1 in 0.0832 g. of optically pure V ( ~ Y * ~+1.81", D crystallization of the second crops gave 0.001 g. of the p (2 = 1 dm., neat) was prepared and gave nZ5D 1.5013, nitrobenzoate of 2,5-dimethyl-3-phenyl-2-hexanol(\TI), C Y ~ ~ .+0.82 ~D f 0.06" (1 = 1 dm , neat). The infrared m.p. 131.5-132.5', m.m.p. with authentic material, 132spectrum of this mixture was essentially identical with t h a t 133.5'. of an authentic sample of V. Controls on the Olefin Products of Acetolysis and ForControl on the Extent of Resolution During the Fractional mo1ysis.--A solution of 0.208 g. of olefin from acetolysis Crystallization of the 3,5-Dinitrobenzoate of Partially Acrun 1 in 20 ml. of acetolysis stock solution was held a t 50" opfor 31 hours (8 titrimetric half-lives for acetolysis of the tive 2,5-Dimethyl-4-phenyl-Z-hexanol(V) .-Partially tically active 4-methyl-3-phenylpentanoic acid (1.06 g., brosylate of threo-IV) and reisolated (a chromatographic m.p. 46-47.5') was esterified to give 1.05 g. of methyl ester, step was included). This treatment changed t h e rotation of n*5~ 1.4931, 0 1 2 5 ~-17.77' ( I = 1 dm., neat), 53% optically t h e olefin only slightly, from a 2 5-42.91' ~ ( 2 = 1 dm., neat) pure. This material gave 0.90 g. of V, n 2 5 1.5021, ~ 0.305 to -43.94" (2 = 1 dm., neat). A similar treatment of the g. of which was converted to its 3,5-dinitrobenzoate. This olefin fraction from t h e formolysis run 3 only changed its ro; derivative was recrystallized from ethanol. T h e first tation from a Z 5f54.83' ~ (2 = 1 dm., neat) t o a Z 5 D +55.43 crop of 0.33 g., m.p. 81.5-86", [ c u ] ' ~ D-7.33" ( ~ 3 . 3 CHCL), , ( I = 1 dm., neat). was recrystallized t o yield 0.119 g., m.p. 81-87', [ C Y Y ] ~ ~ D A solution of 0.207 g. of olefin from acetolysis run 1 in -7.25" (c 3.3, CHCl3). A second crop of 0.090 g., m.p. 60 ml. of formolysis stock solution was held a t 25" for 190 81-84, [ a ] " ~-8.58' (c 3.3, CHC13), was recrystallized to minutes ( 5 titrimetric half-lives for t h e formolysis of the yield 0.0483 g., m.p. 81.5-83.5", [ C Y ] ~ ; D-9.37" (c 3.3, brosylate of threo-IV). T h e olefin was reisolated, but CHC13). T h e mother liquor after removal of the second showed a strong band in t h e infrared corresponding t o forD ( c 2.5, CHCl1). crop gave material, 0.0251 g., [ C U ] ~ -7.97" mate ester a t 1715 cm.-'. This material was treated with The percentage increase in resolution with each crystallizalithium aluminum hydride, and the product chromatotion appears t o be about +5 5 12570. graphed t o give 0.0027 g. of an alcohol fraction. A similar experiment with the olefin from formolysis run 3 gave similar Volume of results. Eluent eluent, ml. cut Characteristics A solution of 0.207 g. of olefin from acetolysis run 2 in 65 ml. of formolysis stock solution was held a t 25" for 190 32% ether in 1000 1- 52 Blank minutes ( 5 titrimetric half-lives for t h e formolysis of the 525 53- 75 Liquid pentane brosylate of threo-IV). T h e product was isolated, treated 1800 76-147 Solid with lithium aluminum hydride and t h e olefin fraction isolated by chromatography as before to give 0.16 g. of material. 75 148- 150 Blank During this treatment the properties of the olefin changed as 600 151-153 Blank follows: a t the start, n Z 51.4994, ~ a Z 5-72.09' ~ (2 = 1 dm., 100% ether 800 154-155 Liquid neat), A232 gave E 474 (cyclohexane); a t t h e end, n z s ~ 400 156 Blank 1.5016, a Z 5-84.49' ~ ( 2 = 1 dm., neat), A232 gave.€ 599.;' (cyclohexane). A second experiment was carried out identical t o t h a t above except the reaction was permitted to proControl on the Chromatographic Separation of 2,5-Diceed for 14 hours (20 titrimetric half-lives). T h e recovered methyl-4-phenyl-3-hexanol (IV) and 2,5-Dimethy1-4-phenylolefin fraction gave n z 61.5020, ~ -88.12' ( 1 = 1 dm., 2-hexanol (V).-A mixture of 0.225 g. of (-)-erylhyo-Il', neat), A 2 3 2 gave E 575 (cyclohexane). [ a I z 5-19.02' ~ ( c 3.3, CHC13) and 0.232 g. of (-)-ihreo-IV, A solution of 0.210 g. of olefin from acetolysis run 1 in 70 [ C Y ] ~ ~-7.25" D ( c 5 , CHCIs) was melted and well mixed, ml. of formolysis stock solution was held a t 50" for 190 [ a ]2 5 ~ 12.69' (c 5 , CHCla). A portion (0.3083 g.) of this minutes. T h e product was isolated, treated with lithium mixture was diluted with 0.2465 g. of racemic V, and the aluminum hydride, chromatographed, and the olefin frac- resulting sample was chromatographed on 166 g. of neutral tion was reisolated. During this treatment t h e properties of made up alumina ( 3 cm. by 25 cm. column) of activity the olefin changed as follows: a t the start, n 2 5 1.4992, ~ in pentane. The secondary alcohol was eluted from the C Y ~ ~-4'2.91" D ( 2 = 1 dm., neat), AB, gave e 15% (cyclohex- column with 325% ether, and the tertiary with 100% ether
2888
L ~ N D R E WS T R E I T W I E S E R ,
1701. i ! )
JH.,. W D \ ~ I I , L I . 3 h l L). S C I L i E F F E K
( t h c d a t a are suininarizcd below). Cuts 53 through 147 were to give 0.25 g. o f tertiarj alcohol, combined t o give 0.23 g. of secondary alcohol, [ a ] " ~ ( 6 5 , CHC11). - 12.00" (c .5, CHCls). Cuts 154 and 155 were combined LOSAKGEIES 21, CALIFOR\IA
1zZ5u 1.3026, [u]*'D-0.21"
~ C O S l R I B U T I O XFROM THE DEPARTMENT O F CHEMISTRY A B D CHEMICAL E N G I S E E K I X C , U N I V E R S I T Y O F C A L I F O R N I A , BERKELEY]
Stereochemistry of the Primary Carbon. VI. The Reaction of Optically Active 1-Aminobutane-1-d with Nitrous Acid. Mechanism of the Amine-Nitrous Acid Reaction1 BY ANDREWSTREITWIESER, JR.,
AND
WILLIAM
RECEIVED OCTOBER 2, 1956 The reaction of n-butylamine with sodium nitrite in acetic acid gives a mixture of esters in which the ratio of n-butJl acetate t o sec-butyl acetate is 2 : 1 a n d a small amount of alkyl nitrates. The 1-butyl-1-d-acetate obtained from optically active 1-aminobutane-1-d is 69 i: 770 inverted, 31 zk 7% racemized. +Imass spectral investigation of the product from I-aminobutane-1,142 shows that 0% ethyl rearrangement occurs. These results in conjunction with other d a t a lead to a complete theory for the product-determining steps of the amine-nitrous acid reaction. T h e fundamental tenet of this hypothesis is that the alkyldiazonium ion, the last common intermediate in the reaction, Can undergo a number of competing reactions because of its great instability.
Introduction Since the first example of Piria4 and the later extended investigations of Linnemann5the reaction of aliphatic primary amines with nitrous acid has been used extensively in preparative and theoretical organic chemistry. I n aqueous solution, the reaction generally produces alcohols and olefins, frequently with rearrangement. Although the kinetics of the reaction has been investigated6 it pertains only t o the initial phase of the reaction of the amine with a nitrous acid moiety; the productdetermining steps occur later in the reaction sequence and are not amenable to direct kinetic approaches. Analogy to the aromatic series strongly suggests that diazonium ions are intermediates in the reaction.' The nature of the reaction products and particularly the similarity of the many rearrangement products to products of typical carbonium ion reactions have led to the interpretation of the product-determining steps of the aminenitrous acid reaction in terms of solvolytic displacement reactions. The commonly accepted mechanism of the reaction is a competing direct displacement ( S N ~on ) the alkyldiazoniuin ion by solvent (reaction 1 ) and a unimolecular fission of the diazonium ion (Sxl) to a carbonium ion which subsequently may react with solvent to form prod(1) Paper V , A. Streitwieser, Jr., and J. R. Wolfe, Jr., 'PHIS JOUR19, 908 (1'337). (2) Taken in part from the dissertation submitted by W.D.S. in partial fulfillment for the degree of Doctor of Philosophy, University of California, June, 1 9 5 6 (:1) General Electric Fellow, 1955-1956. ( 4 ) R. Piria, A N F Zchim. . P h y s . , [ 3 ] 2 2 , 173 (1848). ( 5 ) (a) E. 1-innemann, A n n . , 144, 129 (1867); (b) A . Siersch, ibid., 144, 137 (1867); ( c ) E, Linnemann and 1 ' . v. Zotta, ibid., 161, 49 (1871); (d) 162, 3 (1872); (e) E. Linnemann, ibid., 162, 12 (1872); ( f ) He?., 10, 1111 (1877). (0) T. W . J. Taylor, J. Chein. Soc., 1099, 1897 (1928); T. W. J. Taylor and L. S. Price, ibid., 2032 (1929); J. H. Dusenbury and R. E. T'owell, THISJ O U K X A L , 13, 3269 (1951); A. T.Austin, E. D. Hughes, C. K . Ingold and J. H. Kidd, ibid.,74, 555 (1952); R. H. Sahasrabu(ley, R. Singh and Vasundhara, J . I n d i a n Chem. SOC.,31, 69 (1954); the subject has been reviewed recently by J. C. Earl, Reseavch, 3, 120 NAL,
uct, eliminate a proton to yield olefin or rearrange a hydrogen or a carbon function to a new carboiiium ion which ultiniately results in products of such rearrangement (reaction 2 ) .*-I1 HzO
-
N
I
