2086
HOW.ARD E. ZIMMERMAN ANI) LEOAIIRAMJIAN
Vol. 82
is approximately 1 X lo6 set.-'. If ka is assumed t o be of the same order of magnitude, then it can be shown using
tained when the relative intensities, I ( Y ) ,are plotted against 1/N where N is the number of carbon atoms. M o h o and co-workers16 have shown that the potential energy of both bonded and non-bonded kl = ( k r / h ) e A S * / R e - A r I l * / R T (8) atom pairs is related directly to the reciprocal that in order for k1 to be greater than k3, AIIl* of the reduced mass of the atoms. could be approximately 8 kcal., or less (assuming a The rates of hydrolysis of the aryl trifluororeasonable value for A S ) . I n the alkaline hy- acetates are qualitatively dependent upon the drolysis of ethyl benzoate2 which has an activa- basicity of the ArO-group. It was found, a t least qualitatively, that the rate tion energy of 14.56 kcal., there could be an energy barrier of approximately 8 kcal. for the unstable of hydrolysis of phenyl trifluoroacetate was very intermediate and kl could still be greater than k3. dependent upon the water concentration. This Thus it is very possible t h a t the life of the suggests that a possible mechanism for the nonunstable intermediate is too short for the two catalyzed hydrolysis of this ester might be oxygen atoms to become completely equivalent oa I by proton transfer. An explanation for equation 3 might be found in the effect of chain length upon the amplitude of the -C-OR bond vibrations or upon the probability that the bond will obtain the requisite energy for and thereby k-2 could be greater thail or a p the reaction. Theimer16 has shown, for examplc, proximately the same as K , ; thus, the effect of the that the relative intensities of the terminal modes ionization step, k , , could be felt in the ovc.r-:dI in straight chain n-paraffins decrease with increas- rate constant. ing chain length, ;.e., a linear relationship is ob(10) Y.Morino, el n l , ibid., 21, 1927 (19;:;) (15) 0 Theimer,
J. Chem.
Phys.. 27, 1041 (1057).
LAFAYIC'r-rE,
1x1).
[CONTRIBUTION FROM THE CHEMISTRY DEPARThlENT OF h'ORTlI\\'ESTERN
UNIVERSITY]
Overlap Control of Carbanionoid Reactions. 1I.I The Stereochemistry of the Perkin Reaction and Related Condensation Reactions B Y HOWARD E. ZIMMERMAN AND LEOl
Z
~
~
RECEIVED AUGUST29, 1958 The Perkin reaction of aromatic aldehydes with phenylacetic acids has been known for quite some time to afford prcferentially the a-phenylcinnamic acid stereoisomer with cis-phenyl groups but an unhindered carboxyl group. The nature o f the stereochemical driving force has now been elucidated. Evidence is presented that the initial contiensttion step is not reversible and that the reaction stereochemistry is eliminatively controlled. The role of overlap control in affortling condcnsation products with unhindered electron delocalizing groups is discussed.
I t has been known for some time that the Perkin condensation of aromatic aldehydes with phenylacetic acids affords the a-phenylcinnamic acid stereoisomer with cis-phenyl groups.
This knowledge derives mainly from the work of Stoeriner and co-workers2 and B a k ~ n i n . ~ Recently, Crawford and Moore4 have emphasized the generality of this reaction stereochemistry6 and commented on the enigma of a reaction (1) For paper I in this series see €1. E. Zimmerman, L. Singer and
I3. S. Thyagarajan, THISJ O U R N A L , 81, 10s (1959). (2) R. Stoermer, A n n . , 409, 15 (1!)15); I < . Stoermer and L. Priggc. ibid., 409, 20 (19IB); R. Stoermer and C,. Voht, ibid.. 409, RO ( I ! I l 5 ) . (3) M . nakunin, Gazr. chim. ilnl , 27. 11, 34 (1807); Chrm Z e n l r . , 68, 11, F62 (1897). (4) M. Crawford and G . W. Moore, J . Chem. Sac., 344.5 (1955). ( 5 ) T h e configurations of t h e various a-phenylcinnamic acids are tirmly established. T h e evidence includes t h e stereospecific decarboxylation t o the stilbene of the same configuration (ref. l ) , t h e much greater r a t e of esterification of a-phenyl-trans-cinnamic t h a n of aphenyl-cis-cinnamic ncid (ref. 0) a s would be expected from t h e less hindered carboxyl group of t h e former, t h e ready cyclization of only the a-phenyl-cis-cinnamic acld t o t h e corresponding indone (ref. 7),
which leads preferentially t o a product with cisphenyl groups. To account for the observed reaction coursc a mechanism was proposed which involved preferential formation of the ~rythro~-3-hydroxy-2,3-diphenylpropionkacid anhydride intermediatelo (Ib, R = Ac) followed by trans eliniination of acetic acid t h e similar cyclization of t h e related nitrile stereoisomer t o the i n d o n e contrasted with t h e non-reactivity of t h e a-phenyl-;rans-cinnam~~n;tr;~e (ref. 81, the Pwhorr reaction of a-phenyl-irons-a-amitiocinnamic acid compared with t h e formation of 3-phenylcoumarin from o-phenyl-cis-o. aminocinnamic acid on diazotization (ref. 2 ) . ( 6 ) J. J. Sudboroiigh and I,. I.. Loyd. J. Chrm. Soc., 81 (1898); J. J. Sudborough and D. J. Roberts, ibid.. 1851 (190.5) (7) M. Bakunin. Gasa. chim. i f d . , 30, 11, 310 (1900); cf. C h e m . Z r n f r . , 1 1 , 11, 1276 (1900). ( 8 ) P. Pfeiffer, H. Kiibler and H. Riiping, J. P r n f i f . Chrm., [ a ] 121, 85 (1929). (9) I t is convenient t o assign t h e errlhro a n d l h r e o designations t h e basis of similarity of relative size of groups rather than similarity of functionality. On this ba5is t h e w l h r o diastereomer will he t h e more stable one, since as has been Pointed o u t by both D. J. Cram, F. I). Greene and C. H . DePiiy, THISJ O U R N A L , 78, 700 ( l O j G ) , and D.H. R . Barton and R. C . Cookson, Quart. Rcr., 10, 4 8 (1956), it has available t o I t a low energy conformation in which not only the large groups (I, and L',) have an S-franS arrangement but also t h e medium groups (hI and M',). (10) T h e suggestion t h a t t h e stereochemistry of condensation reactions is controlled by preferential formation of t h e more stable cryihro
~
~
May 5, 1959
Ii
11
However, this interpretation sceiiied inconsistent with our finding^'^,^^ that the condensation of benzaldehyde with either the magnesium or the sodiumIb enolate of phenylacetic acid led in the Ivanov reaction preferentially to thrco-3-hydroxy-2,3diphenylpropionic acid (Ia, R = H) rather than
Experimental section, and calibration data and analyses of known compounds are given in Table 111. The second requirement preliminary to the main investigation was information bearing on the interconvertibility and relative stabilities of the aphenylcinnamic acid products. The early work of Stoermer2 indicated t h a t a-phenyl-trans-cinnamic acid (IIb) is the stable isomer. Moreover, recently Curtin's has reported the equilibrium constant for formation of I I b from I I a as 6. I n agreement with the literature indications i t was found in the present study that by refluxing for 22 hours in a dilute solution of acetic anhydride-triethylamine, a-phenyl-&-cinnamic acid (IIa) was converted to an equilibrium mixture containing 81% of a-phenyltrans-cinnamic acid (IIb) and 19yoIIa. With only 35 minutes of refluxing I I a was converted to the extent of only 21y0to IIb. CsHs, \
/
In
OR
the proposed intermediate I b (R = H ) . 1 6 With the intent of resolving this dificulty and elucidating the mechanism of the Perkin condensation, the present investigation was begun. Since both diastereomers (Ia and Ib) of the 3hydroxy-2,3-diphenylpropionic acid intermediate were available from the Ivanov reaction, a study of the behavior of each of these under Perkin conditions was undertaken. For this investigation a procedure for analyzing mixtures of the a-phenylcinnamic acid products (IIa and IIb) first had to be devised ; and secondly, knowledge of the conditions under which these stereoisomers interconverted was needed. I t was desirable to select reaction conditions for the mechanistic study mild enough such t h a t product equilibration would not be a serious complication and yet conditions severe enough that the Perkin reaction would proceed essentially to completion. The first requisite was fulfilled by using an infrared technique devised earlier, which allowed determination of the percentage of each isomer in a mixture with a maximum expected error of =k2 percentage units. The details are described in the diastereomer has been p u t forth by several other groups as well. Thus it has been postulated by N. H. Cromwell and R. A. Setterquist (ref. l l ) , H. Kwart and L. Kirt (ref. 12) and H. Dahn and L. Loewe (ref. 13) that t h e condensation step of t h e Darzens reaction leads preferentially t o the more stable crylhro diastereomer, (11) N. 11. Cromwell and R. A. Setterquist, Tim JOURNAL, 76, 5752 (1!64).
(12) I i . Kwart and L. Kirt. J . Org. Chcm., 82, 116 (1057). (1.3) I f . Dahn and L. Loewe, Chimio. 11, 98 (1957). (14) € I . E. Zimmerman and M. D. Traxler, Tars JOURNAL, 79, 1920 (1957). These findings were presented at the Sixth Reaction Mechanism Conference, held a t Swarthmore. Pa., September, 195G, (15) l i . E. Zimmerman, L. Ahramjian and M. D. Traxler. Abstracts of the Organic Division, ACS Meeting, Miami, Fla., April, 1957. p. 45-0.
(1F) The extrapolation of t h e Ivanov stereochemistry t o the Perkin reaction is imperfect due t o a solvent difference and t h e fact t h a t the Ivanov reaction involves a dienolate while only a monoenolate occurs in the Perkin reaction. (17) H. E. Zimmerman and T. W. Cutshall, T i m J O U R N A L , 80, 2803 (1958). and earlier papers cited therein.
2087
REACTION
STEREOCHEMISTRY OF T H E PERKIN
/
c=c
COOH
/
H
\
C"s,
/
+
4
C&
'C=C'
H
/
CsHc
\
(3) COOH
IIb
113
The greater stability of the isomer (IIb) having &-phenyl groups is unusual and requires comment.19*20 I t is clear from inspection of models that in each stereoisomer only the trans related groups may approach coplanarity with the double bond with consequent resonance interaction. Thus the carboxyl group in IIa and the a-phenyl group in IIb are forced out of the molecular plane.% Being electronically insulated from the rest of the molecule, these groups cannot influence resonance stabilization. Furthermore, since each of these groups is perpendicular to the molecular plane, and since the half-width thickness of a carboxyl group should not differ appreciably from that of a phenyl group, it may be concluded that the van der Waals repulsive forces due to the perpendicular carboxyl in IIa do not differ from those engendered by the perpendicular a-phenyl group in IIb. Thus from both an electronic and a steric viewpoint the systems (perpend.)
COOH
CaH5 \
(perpend. ) cas
/
B as occurring in I I a and I I b are essentially the same, and therefore the observed energy difference must be attributed to the remaining portion of the system (Le., the trans-a-phenyl group in I I a and the trans-carboxyl group in IIb). A
(18) D. Y.Curtin, Abstracts of t h e Thirteenth Organic Symposium of the A.C.S., Ann Arbor, Mich., 1953, p. 40. (19) It was erroneously assumed by Crawford and Moore (ref. 4) t h a t m-phenyl-cis-cinnamic acid is the stable isomer. (20) I t should be noted t h a t in the present case the isomerization involved is actually t h a t between the mixed acetic a-phenylcinnamic acid anhydride stereoisomers, the equilibrium mixture being converted t o the corresponding mixture of I I a and I I b during work-up. (20a) The non-coplanarity of the carboxyl in I I a is evidenced by its shorter infrared carbonyl absorption wave length ( 5 . 8 7 ~ in C& us. 5.93 for Ilb).
2088
HOWARD E.
ZIMhlERMAN P ND
Although the literature suggestsz1 greater resonmcc stabilization should be available from introduction of a coplanar phenyl group to system A than addition of a coplanar carboxyl group to system 13, the carboxyl group, because of its smaller steric requirements and lesser van der Waals interaction with the perpendicular group, finds i t easier to assume the planar conformation required for maximum resonance stabilization. Thus, the observed stability order can be attributed to ineffective resonance interaction by the a-phenyl group in IIa due to difficulty in attaining a planar conformation. T h e next step in the present research was investigation of the behavior of each of the diastereotiieric 3-hydroxy-2,3-diphenylpropionicacids under Perkin conditions. Each of the diastereomers (Ia and Ib) was treated with refluxing 1 : 1 acetic anhydride-triethylamine for 35 minutes, under which conditions the interconversion of stereoisomers had been demonstrated to he only a minor Infrared analysis of the reaction mixture (note Table I) T.4BLE
1 a-
Phenyl-
tuanscin-
Run
Ileac tan t s
1 threo-3-Hy~lros);-2,:li-dii,llc11ylpropionic acid 2 c,iythro-~-Hydrox~-~,~diphenylpropionic acid 3 Plirnylacctic acid ant1 beuzaldehyde 4 ~-Pliciiq.l-cis-cirinaiiiicacid 3 rj
u-Plien)l-tis-ciiiii~iiiicx i ( ! ru-I’I:cii?-l-si’s-ciiriinniicacitl
Cotiditions
namic acid. ‘X
1 1 Ac?O EtjN 33 min. reflux 1: 1 X C ~ OEt3S
99
35 inin. reflux 1f)o I : 1 .\c20 EtrX :i,7 riiin. reflus 96 1 : 1 .ic20 E t 3 S 3Ljinin. reflux 21 2 : 1 .ic20 Et3N - [ I . .) hr. reflux 75 1 : 1 .Ic?O E t 3 N 22 lir. rcflus 81
indicated this to coiisist in each case of 99 + 2% a i)lietiyl-truns-ciiiiiaiiiic acid (IIa). This result is striking for a t least two reasons. First of all, it iiegatcs the nicclianisrii 1)roposed by Crawford and IIoore,4 for the eliiiiination is not stereospecific ;LS was presumed by thesc authors. Secondly, dcspite the lack of stereoslic‘cificity, the eliniination is highly stereosclcctive2:; thus, much more of the stable isomer I l b results than accountable 011 a tlierniociyIiaIriic basis. The observation having been made t h a t under mild conditions 99yo of a-pheIiyl-trans-cinnamic (21) T h e d a t a of I
