BEN M. BENJAMIN,HOWARD J. SCHAEFFER AND CLAIRJ. COLLINS
6160
Vol. 79
perchlorate ion is not nucleophilic, this imine prob- transition state the carbonium carbon atom is ably reacts almost entirely by S N 1 in dilute per- shielded by the departing nitrogen atom from atchloric acid. 1,2-Iminoethane probably reacts tack by a nucleophilic amine molecule on the same only by S N ~ .The other imines would react by a side of the carbon atom. Thus inversion of concombination of S N 1 and SN2. The relative rates figuration is not incompatible with s N 1 mechanism, of these acid-catalyzed hydrations of imines can be but on the other hand retention of configuration accounted for on the basis of a transition from cannot occur in an SN2 reaction. Actually S N 1 to S N 2 mechanism as in the case of epoxidesz4 L( +)-trans-N-ethyl-2,3-iminobutane with aqueous and the assumption that reaction is predominantly ethylamineI2 gave 99.5% rneso-2,3-bis-(ethylS N a~t a primary carbon atom, a mixture of S N ~amino)-butane and 0.5y0 L( +)-2,3-bis-(ethylaand S N 2 a t a secondary carbon atom and pre- mino)-butane. The active product mustresultfrom dominantly S N 1 a t a tertiary carbon atom. an S N 1 reaction, and a frontal attack on a carbonThe S N rate ~ a t a primary carbon atom could ium ion would seem to be the only way in which it well be approximately equal to the combined S N 1 could be formed. Here the conditions are quite and Siv2 rates a t a secondary carbon atom and unfavorable for frontal attack on the carbonium definitely slower than the SN1 rate a t a tertiary ion, owing t o the presence of the two ethyl groups, carbon atom. Thus 1,2-imino-2-methylpropane one on the imine nitrogen, the other on the attackwould be the fastest reacting and the others would ing amine nitrogen atom. The fact that some vary, depending on the different contributions of frontal attack took place points to a significant s N 1 and sN2 reactions a t secondary carbon atoms. contribution of S N 1 in the weakly basic medium. Although the steric results of the opening of the This would be enhanced under the much more rings of L ( +)-N-ethyl-2,3-iminobutaneby am- favorable conditions of acid-catalyzed hydration, as in the case of epoxides.25 Therefore it is reasonmonia and ethylamine and of L( -)-2,3-iminobutane by ethylamine are in all cases inversions of able to expect that in acid solution there would be configurations,12 an S N 1 mechanism even under a mixture of S N 1 and S N 2 reactions a t the two basic conditions is not ruled out because in the secondary carbon atoms of this imine. (24) C. K Ingold, “Structure and Mechanism in Organic Chemlst r y , ’ Cornel1 University Press, Ithaca, N. Y . , 1953, p. 344.
[CONTRIBUTION FROM
THE
(25) Reference 23, p. 38.
PASADENA. CALIFORNIA
CHEMISTRY DIVISIONO F
OAK
RIDGEXATIONAL LABORATORY]
Molecular Rearrangements. XI. The Deamination of 1,l-Diphenyl-2-amino1-propanol1 BY BEN M. BENJAMIN,HOWARD J. SCHAEFFER AND CLAIR J. COLLINS RECEIVED MAY11, 1987 I t has been shown that the addition of phenylmagnesium bromide and of P-tolylmagnesium bromide to a-aminopropiophenone are 98-99y0 sterospecific. The rearrangement, in the presence of nitrous acid, of (+)- or (-)-l,l-diphenyl-2aminopropanol (IV) stereospecifically labeled in only one of the two phenyls, has been shown to produce a-phenylpropiophenone (VII) which has undergone approximately 88% inversion (with 12% retention) a t the migration terminus. These results are in agreement with those of McKenzie, Roger and Wills.@ The product from the deamination of (+)-IV was resolved into a (-)-fraction [ [ m l z 4 -210’1 ~ and a very nearly racemic [ [ a I z-1.6’1 4~ fraction. Degradation of these two fractions followed by radioactivity assay of the degradation products disclosed that migration of the labeled phenyl resulted in inversion to produce (-)-VII, whereas migration of the unlabeled phenyl resulted in retention to produce (+)-VII. The deamination of labeled 2-amino-1,1,2-triphenylethanol also has been studied. The results of the present research are explained in terms of open carbonium ion intermediates whose rotation about the C-C+ bond is restricted and whose phenyl groups migrate solely through trans-transition states.
I n the rearrangement with nitrous acid of amino alcohols of general structure I, i t often has been assumed: (1) that the ion I1 is first formed, followed
I
I1
111
by a one-stage migration of R1 or Rf;(2) that the rate of rotation about the bond connecting the two central carbon atoms is very fast compared to re(1) (a) This paper is based upon work performed a t Oak Ridge Sational Laboratory, which is operated by Union Carbide Nuclear Co for the Atomic Energy Commission. Paper X, W. A. Bonner and C. J. Collins, THIS JOURNAL, 78, 5587 (1956); (b) portions of this research were presented a t the International Conference on Radioimtopes in Scientific Research, Paris, France, Sept. 13, 1967,
arrangement; and (3) that the steric factor, called the “cis-effect” by Curtin,2 “is a function only of the difference in free energy between the two transition states for the rearrangement step and is independent of the relative populations of the conformations of the initial m o l e c ~ l e . ” ~An alternate possibility involves the formation, from 11, of ion 111. The relative proportions, then, of products formed will depend upon the populations of the conformations of ion I11 which will allow migration of R1or Rz; this, in turn, will depend upon the rate ( 2 ) P . I. Pollak and D . Y. Curtin, THISJ O U R N A L , 71, 961 ( I Q N ) ; D. Y . Curtin and P. I. Pollak, i b i d . , 73, 9 0 2 (1951): D. 1’. Curtin, E . E. Harris and P . I. Pollak, i b i d . , 73, 3453 (1931). (3) (a) Quoted from D. Y . Curtin and hl. C . Crew, ibid., 77, 365 (1955); (b) W. G. Dauben and K , S. Pitzer, Chapter 1, “Steric Effects in Organic Chemistry,” edited by hl. S. Kewman. John Wilry and Sons, Ins.,New York, N. P., 19613, pp. 11, 44.
Dec. 5 . 1957
THEDEAMINATION OF 1,1-DIPHENYL-2-AMINO-1-PROPANOL
6161
rotation about the central carbon-carbon bond chloride complex salt of 2-aminopropiophenonecompared to the rates a t which R1 and Rz shift phen~1-C'~.Resolution of IV was accomplished the migration terminus. I n the rearrangements through the d-10-camphorsulfonate. The optically the erythro (I, R1 = anisyl, Rz = phenyl, Ra pure amino alcohols were deaminated by the proce= methyl) and threo (I, R1 = phenyl, Rz = anisyl, dure of McKenzie, Roger and Wills.4a I n experiRB = methyl) isomers of l-p-anisyl-l-phenyl-2- ments 1 and 2 the ketonic products produced amino-1-propanol, studied by Curtin and quantitatively were cleaved with isoamyl nitrite the ratios of the two ketones produced (p-anisyl in ethanolic alkali to yield benzoic acid and acetophenylethyl ketone t o phenyl a-p-anisylethyl ke- phenone oxime, which were assayed for radiotone) were, respectively, 88:12 and 6:94. It ap- activity content. In experiment 3 the product was peared to us that there is a striking parallel be- resolved by repeated crystallization from ethanol tween these r e s u l t ~ and , ~ ~ those obtained by Mc- into a (-)-ketone and into an almost completely Kenzie and his co-workers4 in the deamination6 of racemic fraction. Both of these fractions were optically active 1,l-diphenyl-2-amino-l-propanolcleaved as before to yield the degradation products, (IV), since i t was clearly shown4 by McKenzie that benzoic acid and acetophenone oxime which likethe product consisted of a mixture of the enantio- wise were assayed for radioactivity. The results of meric ketones (VII) in the ratio of approximately all of these experiments are given in Table I. 88.12. TABLE I The stereochemical relationships of (3.)-and (-)-IV and (+)- and (-)-VI1 have been worked SUMMARY OF DEAMINATION EXPERIMENTS UPON(+)- AND o u t by Bernstein and W h i t m ~ r e . * ~If the mech( - )-IV anism2j3generally accepted is the correct one for [ a I z 4 D Of RetenYoC1;in Reactant Product tion, * the deamination of IV, then both phenyls must % PhCOOHa PhCOCHa Expt. IV VI1 migrate in such a way that the configuration of the 11.1 89.9 +158.5' 12.5 1 -59.3' migration terminus is inverted. The considerable 13.3 89.9 +158.3 12.5 2 -59.3 amount of racemization (about %yo)which at11.5 44.3b 53.0b 3 $59.3 -161.6 tends this reaction would then be required to arise 4.40 98.0C through instability of the reactant (IV) or product a The molar radioactivity of the a ~ e t o p h e n o n e - C oxime ~~ (VII) to the reaction conditions. Since both prod- is more reliable than that of the ben~0ic-C'~acid, since uct and reactant are presumed"a to be unrace- during the cleavage of VII, some of the acetophenone fragmized under the conditions of the deamination re- ment is oxidized to benzoic acid. This accounts for the the sum of radioactivity content of the two fracaction, however, the approximately 12y0of prod- fact that is usually greater than 100%. The radioactivity uct VI1 with retained configuration4" must arise tions assay (expt. 3) of the acetophenone oxime from (-)-VI1 is from the conversion VI (+)-VII-presuming the proof that the initial Grignard addition to a-aminopropioreacting amino alcohol I V to have the (+)-con- phenone was a t least 98y0 stereospecific. This high degree of stereospecificity is consistent with the observations of figuration as shown in the formulas of as to of
-
.c&
Zb
c . i , a
F===
-Ti!
Cti2&H
P h*
HO
b.
31-
(+,
CQxGde:
Curtin and Crew (ref. 3 ) upon similar Grignard addition reactions. From nearly racemic fraction of product, [CY]% -1.6'. From completely resolved fraction of product whose [cY]*~D-210'.
?t
'>Y2
Ph*
P
I
I H ,
PhCCCLCH3 I ?hf
(-1 -
m
Ph I ?h*CO C --- H CH,
\
(+)
- rn
This hypothesis has now been tested and found correct by the synthesis of, and subsequent deamination experiments, with, the enantiomorphs of IV labeled with carbon-14 in only one o i the two phenyls. Racemic IV was prepared by the addition of phenylmagnesium bromide to the stannic (1) (a) A. McKenzie. R. Roger and G. D. Wills, J . Chcm. Soc., 779 (1926); (b) H. I. Bernstein and F. C . Whitmore, THISJOURNAL, 61, 1324 (1939). ( 5 ) This is a classic reaction which has been widely quoted a9 evidence for inversion at the migration terminus: (a) C . K. Ingold, "Structure and Mechanism in Organic Chemistry," CorneU University Press, Ithaca, N. Y., 1953, pp. 504-505; (b) D. J. Cram, Chapter 5, "Steric Effects in Organic Chemistry," edited by M. S. Newman, John Wiley and Sons, I n c . , New York, N. Y., 1956, p. 253; (c) P. D. Bartlett, "Organic Chemistry," Vol. 111, edited by H. Gilman, John Wiley and Sons, Inc,, N@w YQ&, N. Y.,l Q Z , Chapter 1. pp. 69-60.
I n order to test the generality of the Pxtremely high degree of stereospecificity (see later discussion) noted in the addition of phenylmagnesium pbromide to a-aminopr~piophenone-phenyl-C~~, tolylmagnesium bromide was added to the stannic chloride complex of a-amin~-(propio-l-C~~)-phenone, and the yields of erythro- and threo-l-phenyl-lp-tolyl-2-amino-1-propanol-l-C14 were determined by the isotope dilution method. I n this way it was shown that the product of this reaction consisted almost exclusively of erythro-l-phenyl-lp-tolyl-2-amino-1-propanolsince there was not more than 1% of the threo isomer produced. Finally, the chain- and ring-labeled isotopeposition isomers of 2-amino-l,1,2-triphenylethanol (VJH) were prepared. OH
"2
PhbC!!HPh
+ PhMgBr
---f
NHz PhCObPh
-
+ $hMgBr
* I
.--)
I*
?SHz
I
PhZC-CHPh VIIIa OH NHz
I
PhzC-CHPh VIIIC
(6) The tetffls crythvo and threo have been employed for these compounds In the same a w s e as that used by Curtin and Crew (ref. 3a. footaate 4),
6162
BEN M. BENJAMIN,HOWARD J. SCHAEFFER AND CLAIR J. COLLINS
The amino alcohols VIIIa and VIIIc were then subjected to nitrous acid deamination; and the benzhydryl phenyl ketone products were subjected to degradation procedures. H NOS KMn04 LXIa ---+ PhsOCHPh2 ___) PhEOOH
+
PhzC=O 0.9%
IXa
Vol. 79
involving the rigid geometry of XI i t has been postulated' that the reagent attacks the molecule (e.g., XI) preferentially from such a direction as to M
A
R'
S
be flanked by the medium (M) and small (S) groups. Because the conformation of the molecule (XI) being attacked is not rigidly fixed, how* IXcd PhzC=O 9.3% ever, as it is in intermediate X, the carbonj.1 group The percentages of the original molar radioactivities may not always be exposed between the M and 9 of VIIIa and VIIIc are shown under the appro- flanks and thus can suffer attack sometimes bepriate benzophenone fractions. tween the L (large) and S, or even between the L and M groups. The difference in degree of stercoDiscussion specificity, then, exhibited by X and X I may be From the foregoing data we conclude: (1) The summed up by the statement that in the case of the addition of ptolylmagnesium bromide to the stan- intermediate X, the environment adjusts to the nic chloride salt of a-amin~-(propio-l-C~~)-phenone intermediate, whereas in the case of molecdes such is a t least 99% stereospecific. This follows from as XI the molecule must adjust to the ermslronthe observation that not more than 0.S% of threo-l- merit.* phenyl-1-p-tolyl-2-amino-1-propanol6 was formed (4) During the deamination of (+)-IV, I'h during this reaction, whereas the erythro isomer was (labeled) migrates nearly exclusively to yield produced in 70% yield. product of inverted (-) configuration whereas Ph (2) The addition of phenylmagnesium bromide (unlabeled) migrates nearly exclusively to yield to the stannic chloride salt of a-aminopropio- product of retained (+) configuration. The forphen~ne-phenyZ-C'~ is a t least 98y0 stereospecific. mation of (+)-VII thus is not the result of the deThis follows from our results upon the deamination composition of intermediate 11 with a simultaneous of (+)-IV to (-)-VI1 and (+)-VI1 (see Table I). migration of Ph since Ph (unlabeled) approaches the The (-)-a-phenylpropiophenone (VII) from this migration terminusf r o m the same side as that occupied reaction was shown to have 98% of its radioactivity by the amino group and not, as has been assumed,3 in the a-phenyl group, thus demonstrating stereo- through a transition state in which methyl a77d phenyl specificity during the Grignard reaction to the ex- are in a cis configuration. tent of 98%. It follows also from this observa( 5 ) Because of the high degree of stereospecificity tion that neither the reactant (+)-VI, nor the prod- associated with the additions of phenylniagnesium ucts (+)- or (-)-VI1 is racemized significantly bromide and fi-tolylmagnesium bromide to cyunder the conditions of the deamination reaction, aminopropiophenone, it is almost certain that the for to the extent that such racemization takes addition of phenyl-C 14-magnesium bromide to aplace, there should be an interchange of the labeled aminodesoxybenzoin hydrochloride to yield VIIIc and unlabeled phenyls: such an exchange would proceeds also with 98-1 00% stereospecificity. I n result in an equal distribution of radioactivity in the deamination of VIIIc, therefore, we propose the two phenyls of (+)- or (-)-VII. that Ph (unlabeled) migrates to the opposite side (3) The high degree of stereospecificity exhibited of the migration terminus originally bonded t o during these two Grignard additions undoubtedly nitrogen, whereas Ph* (labeled) migrates to the arises through an intermediate such as X, in which same side of the migration terminus originally bonded to nitrogen. .T( I (6) Although these data do not rule out the interM g vention of ions of general structure I1 during the / \ deamination of IV, we believe the simplest general HN 0 X explanation of the data of Table I t o be one in which I I1 CH,-----C-C (a) ions of type 11 are unimportant, (b) the open 1 carbonium ions V and VI are formed and ( e ) the H >ll equilibrium V V I involving rotation about the the preferred side of attack is that by which the central carbon-carbon bond is not extremely fast Grignard reagent is flanked by hydrogen and compared with the rate of phenyl migration. Alphenyl, rather than by methyl and phenyl.' The Other work which seems to support stereospecificity of 5O:L intermediate X, possessing a five-membered ring, to (8) 1OO:l when intermediates similar t o X can intervene are: (a) A . might be expected to provide a much higher degree McKenzie and H. Wren, J . Chcm. SOC.,97, 473 (1910), in which it of asymmetric induction than is found in most is reported that only one product was isolated from the addition Of methylmagnesium iodide to 1-benzoin; this experience was repeated stereospecific syntheses; for example, in cases not when racemic benzoin was employed. These authors state ". . . i t was VIIIc
€IKO* * ---+ PhCOCHPh2
RMnOl
hhCOOH
I__)
+
(7) For previous discussions of similar intermediates see H. S.Moser and E. La Combre, TEISJOURNAL, 73,3994,4891 (1950); W. E. Doering and T. C. Aschner, ibid., 71, 838 (1949); W. E. Doering and R. W. Young, ibid., 72, 631 (1950). D.J. Cram and F. A. Abd Elhafez, ibid., 74, 5833 (1952). have recognized the specific intermediates X and XI. Sea al-o V. Frelox, 0. Cedar and M. Wilhelm. Rclr. Ciiim. Acto, 88, 303 /i9,53)
obvious from the yield that the other isomeride could have been present only in small amount or not at all."; (b) the work of Curtin and co-workers (ref. 2, 3a), in which a variety of Grignard additions t o a-eminoketones result in the isolation of only one racemic modification of product; and (c) L. H. Welsh, THISJOURNAL, 71, 3500 (1949). io which is mported a 70-fold difference in the rates of rearrangement of N-beuzo,ylephrdrine and N-beuzoyl-~-ephedrine.
Dec. 5 , 1957
THEDEAMINATION OF 1,1-DIPHENYL-%AMINO-1-PROPANOL
6163
though the presence of open carbonium ions has been demonstrated previously in the deamination of secondary amine^,^ we believe the present work is the first clear-cut experimental evidence which demands the presence of open carbonium-ion intermediates in the amino alcohol deamination, a reaction which previously had generally been held2va to be concerted. Acknowledgment.-The authors are pleased to acknowledge the stimulation derived from a discussion of the contents of this manuscript with Professor Cram. Experimental
of alcohol and was treated with 9.3 g. of &tartaric acid in 25 ml. of water. When the &tartrate salt of IV had precipitated, it was removed by filtration and crystallized twice from alcohol; [ c r I a 4 ~-54.4' (water). The free (+)-amine was recovered as described for the (-)-isomer; m.p. 101.5102.5', [CY]~'D4-59.3' (ethanol), 2.691 mc. of carbon-14/ mole. A total of 4.5 g. of (-)-isomer and 4 g . of (+)-isomer of IV was recovered after working up the mother liquors. Non-radioactive optically pure I V was prepared by an alternate procedure4*starting with pure Z-alanine purchased from California Foundation for Biochemical Research. The amino alcohol product was purified by repeated crystallization from hexane. The highest rotation obtained was [ U I ~ ~ D -59.6' (ethanol). Reaction of 1,l-Diphenyl-2-aminopropanol-C~4 with Nitrous Acid.-The Drocedure of M c K e n ~ i eel, ~al.. ~ was used. Pure (-)-IV (1 g.) was dissolved in 37 ml. of 25% acetic acid, and the solution was cooled to 0'. A solution of 1 g. Preparation of l,l-Diphenyl-2-aminopropanol-C~~ (IV) Phenyl-labeled propiophenone-C14, prepared by treating of sodium nitrite in 5 ml. of water was added dropwise during 15 minutes, and the solution was stirred a t 0' for 5 hr. phenyl-labeled benzoyl-C14 chloride with diethylcadmium, The oily ketone which separated was removed from the was converted t o the isonitroso derivative.Io The product was reduced with stannous chloride to the stannic chloride solution, after dilution with water, by ether extraction (three 50-ml. portions). The ether extracts were washed complex salt of 2-aminopr0piophenone.~~The dry powdered salt (22.5 g.) was added in small portions to the Grignard with water, sodium carbonate solution, water, and then the ether was evaporated in a current of air. The yield of dried reagent prepared from 83 g. of bromobenzene and 12.7 g. of magnesium. After the addition was complete, the mixture ketone VI1 was quantitative; [ c Y ] ~ ~ D 158.5' (ethanol). In a second experiment (-)-IV also gave a quantitative was heated a t reflux temperature for 1 hr. The Grignard yield of ketone, [ c L ] ~ ~+158.3' D (ethanol). In a third exaddition product was then hydrolyzed with a solution of ammonium chloride and ammonium hydroxide. The periment (+)-IV was deaminated and gave a quantitative D This material was reether was removed by evaporation in a current of air and the yield of ketone, [ c L ] ~ ~ -161.6'. remaining solid material was collected on a filter. The or- solved by repeated crystallization from alcohol; m.p. 36", [aIz4D -210", 2.693 mc. of carbon-I4/mole. A small ganic material was separated from inorganic salts by washing the mass thoroughly with ether. Crude racemic amino sample (cu. 0.3 g.) of almost racemic ketone I11 was recov~ alcohol IV was extracted from the ether solution with dilute ered from the mother liquors: m.p. 50-51', [ a I z 4-1.6". Cleavage of 2-Phenylpropiophenone-C14(VII).-To the hydrochloric acid. It was precipitated from the acid solution by careful addition of dilute sodium hydroxide. The total product from the deamination of (- )-IV described crystals (11.5 g.) were collected on a filter, washed with above was added 6 ml. of absolute alcohol containing 0.2 g. of sodium hydroxide. When the solution was well mixed, water and dried in vacuum, m.p. 101-102". 0.6 g. of amyl nitrite was added and the solution was placed Resolution of I,l-Diphenyl-2-aminopropanol-C".-The racemic phenyl-labeled amino alcohol IV (11.2 g.) was in the refrigerator (approximately 0 " ) . After seven days niised with 5 g. of non-radioactive pure (-)-isomer of IV the solution was diluted, neutralized with hydrochloric acid and the mixture was dissolved in 60 ml. of ethanol. To and extracted with ether. The ether was evaporated and this was added a solution of 17 g . of d-10-camphorsulfonic the remaining solid material was treated with two 10-ml. acid in 60 ml. of water. After several hours the d-10-cam- portions of sodium bicarbonate solution followed by 5 ml. of phorsulfonate of the amine precipitated. This was removed water. A gummy material remained in the flask. The by filtration and crystallized twice from alcohol; [ c x ] ~ ~ D combined aqueous solutions were treated with Norite, +37.8' (ethanol). The free (-)-amine was obtained by filtered and neutralized. The precipitated benzoic acid was removed and sublimed; m.p. 122', 0.259 mc. of carbontreating the above salt with dilute sodium hydroxide. was crystallized twice from hexane; m.p. 101.5-102.5 I,t, 14/mole. Acetophenone oxime was extracted from the gummy material, with five 10-ml. portions of hot water. [ c L ] ~ ~-59.3' D (ethanol), 2.601 mc. of carbon-I4/mole. The free amine (9 g.) was recovered from the salts in the The hot water was poured through a filter and was cooled crystallization liquors from above and mixed with 5 g. of in an ice-bath. The precipitate was collected and sublimed non-radioactive pure (+)-IV. This was dissolved in 50 ml. a t 60', m.p. 60', 2.338 mc. of carbon-14/mole. The ketone from the second deamination experiment was cleaved (9) (a) W. A. Bonner and C. J. Collins, THIS JOURNAL, 78, 5590 to give benzoic acid, 2.337 mc. of carbon-I4/mole; aceto(1956), have demonstrated open carbonium ion intermediates in the phenone oxime, 0.347 mc. of carbon-I4/mole. deamination of 1,2,2-triphenylethylamine,a system whose solvolytic The resolved ketone ( [ c x ] ~-210') ~D obtained from the reactions also give rise to open carbonium ions. (b) D. J. Cram and product of the third deamination experiment was cleaved J. E. McCarty, ibid., 79, 2866 (1957), have demonstrated open car- t o give benzoic acid, 0.1158 mc. of carbon-I4/mole; and bonium ion intermediates in the deamination of 3-phenyl-2-butylacetophenone oxime, 2.636 mc. of carbon-I4/mole. The amine, a system whose solvolytic reactions are best explained through racemic ketone was cleaved t o give benzoic acid, 1.192 mc. the intervention of bridged ions. However, the claim of Cram and of carbon-l4/mole; and acetophenone oxime, 1.425 mc. of McCarty that "the neighboring @-substituentsmust stiil approach the carbon-I4/mole. face of the carbonium Ion opposite from that which the nitrogen Preparation of erythro- and threo-l-Phenyl-l-p-tolyl-2is not supported by our data. ( c ) A. Streitwieser and W. D. left.. amhopropanol .-Procedures very similar t o those already Schaeffer, ibid., 79, 2888 (1957), have postulated open carbonium ion describedz were used for preparing the diastereomeric forms intermediates, among others, in the deamination of I-aminobutane-14. of this amino alcohol. The main difference is that the stanThese authors conceive that the migrating group in such deaminations nic chloride complex salt of the appropriate amino ketone, is engaged in a modified participation and thus undergoes a simulinstead of the usual amino ketone hydrochloride, was added taneous shift as the leaving group is lost from the molecule. Our own to a Grignard reagent. The general method of preparation data and conclusions are not in agreement with this concept. See is demonstrated by the procedure for synthesis of the carbonparticularly A. Streitwieser, J. Org. Chcn., 22, 866 (1957), who assumes 14 labeled erythro compound given in the following section. a concerted rearrangement during deaminatlon of 1,l-diphenyl-ZBoth diastereomers were purified by crystallization from aminopropanol (IV). (d) The general possibility that the stereohexane followed by sublimation. The erythro isomer had chemistry of the products from ions such as 111, V and VI might be a m.p. of 86-87' instead of 73-75' as reported.z No hemicontrolled by restricted rotation about the central carbon-carbon bond hvdrate was found. has been mentioned, without experimental evidence, by S. Winstein Anal. Calcd. for CleHlgNO: C, 79.63; H, 7.94. Found: and E. Grunwald. THISTOURNAL. 70. 835!f19481: S. Winstein and . ., C,79.56; H, 8.01. B. K. Morse, ibid., 74, 1184 (1952); S: Winstdn and L. L. Ingraham. The diacEty1 derivative of the erythro isomer had a m.p. i b i d . . 77.. 1739 (1955). . . of 192-193 (10) W. H. Hartung and J. G. Munch, TEIa JOUXNAL, 51, 2262 (1929). Anal. Calcd. for C ~ O H ~ ~ C, NO 76.29; ~ : H, 7.47. Found: C, 76.43; H, 7.46. (11) L. Behr-Bregowski, Ber., 30, 1515 (1897). ~~
.-
+
.I'
~~
.
6164
ALFREDR. BADER
Vol. 79
The erythro-hydrochloride had a m.D. of 255-256 '. phenylaminoacetate-l-Cl4 hydrochloride12; m.p. 153.5Anal. Calcd. for CloH&lPu'O: C, 69.17; H , 7.26. 155",'3 2.158 mc. of carbon-l4/mole. Phenyl-labeled amino alcohol VIIIc was synthesized by the action of phenylFound: C, 69.30; H , 7.25. The m.p. 100-101" was found for the threo isomer after C14-magnesium bromide upon aminodesoxybenzoin hydrochloride; m.p. 153-154", 0.8883 mc. of carbon-l4/mole. purification by sublimation. Deamination of VIIIa and VII1c.-The amino alcohol Determination of Yields of erythro and threo Isomers in (250 mg.) was dissolved in 20 ml. of 50% aqueous acetic the Synthesis of I-Phenyl-1-p-tolyl-2-aminopropanol .-Caracid, and the solution was cooled in an ice-salt-bath. To bonyl-labeled propiophenone-C14, prepared by the FriedelCrafts synthesis from pr0pionic-1-C~~ acid and benzene, was this cold solution was added a solution of 179 mg. of sodium converted through the isonitroso derivative to the stannic nitrite in 3 ml. of water dropwise while stirring over a period chloride complex of 2-aminopropiophenone as described of 1 hr. The mixture was allowed t o warm to room temabove. After the salt (15.1 g., 0.0502 mole) was dried and perature and stirring was continued overnight. The solid finely powdered, it was added in small portions to the phenyldesoxybenzoin which formed was removed by filtration and was then crystallized from ethanol (68% yield after Grignard reagent prepared from 85.5 g. of p-bromobenzene and 12.2 g. of magnesium. The mixture was heated a t re- purification); m.p. 135-136'.14 Radiochemical Structure Determination of VIIIa, VIIIc flux temperature for 1 hr. and then hydrolyzed with ammonium chloride solution. The ether layer was removed and and Their Deamination Products, IXa and 1Xcd.-The pothe aqueous layer was washed three times with 100-ml. por- sition of carbon-14 in the compounds was demonstrated-by tions of ether. The combined ether extracts were filtered oxidation of each compound to benzophenone as described and the ether was removed by evaporation. The organic previo~sly.1~In each case the benzophenone was converted material was quantitatively transferred to a 250-ml. volu- to its 2,4-dinitrophenylhydrazonewhich was purified by metric flask and diluted to the mark with benzene. Two crystallization from dioxane before assaying for carbon-14
