3444
YOSHIROOGATAAND IWAO TABUSHI
Vol. 83
TABLE IV reactions, a+, are given by eq. 1. On the other hand, Brown and his co-workers showed the linear THE EFFECTSOF SUBSTITCENTS IN WMETHYLSTILBENES k / k o and l o g ( k / k o ) (in parentheses), in benzene at 10’; k / k a correlation 3 between the para partial rate factor IS the relative rate of substituted us. unsubstituted or-methyland the selectivity factor, Sf, defined by stilbenes for the same PBA Sr = log (pf/mr); log pi = or x Si (3) 9Sub-
Ttlt.
in 0methylstilbene
Me 0
Me0 Me C1
4 . 7 5 (0,677) 1.88 ( .276) 0 615(-0,212)
$-Substituent i n PBA
H
CI
4.84 (0.686) 2.17 ( .33tj) 0.653 (-0.185)
4 82 10.683) 1 9s ( ,297, 0,586(-0.232)
values of k/ko increase from small values for p methoxy-PBA (lowest reactivity) through the maximim for PBA, and then decrease aga& with p-chloro-PBA (highest reactivity), On the other hand, in the case of p-chloro-a-methylstilbene, k/ko decreases from p-methoxy-PBAlto PBX and increases again from PBA to p-chloro-PBA. Facts i and ii cited above may be ascribed to (a) a correlation between selectivity and reactivity, and (b) a change in the resonance contribution, which affects k/ko inversely to (a). 2 2 With p-methoxy- and p-methyl-cr-methylstilbenes, the effect of the selectivity-reactivity relationship slightly exceeds the increase of resonance effect, and thus larger values of k / k o were observed for PBA than for p-chloro-PBX. However, the decrease of resonance effect in the transition state may be more important than the selectivity relationship with p-methoxy-PBA, so that k / k o decreases from PRA to p-methoxy-PBB. With 9-chloro-a-niethylstilbene. the decrease of resonance effect is less important and its increase is important: therefore the reverse is true. Selectivity and the Nature of the Transition State.-The substituent constants in electrophilic (22) T h e change of k / k a may be interpreted as foilows: with p methyl- a n d p-methoxp-~-methylstilbenes\\here the resonance interaction in t h e transition s t a t e seems t o be larger than with or-methylstilbene, t h e effect of p-chloro substituent in PBA seems iess i m p o r t a n t , since t h e increase of t h e resonance contribution approaches a limiting value determined b y other less variable effects such a ? t h e steric effect
[COXTRIBUTION FROM THE
Hence, the following equations are derived, where a1 and C R are the inductive and resoiiLmcc contribution in u value, respectively. log pf P(U1 UR YAK) (4a)
+ + + YAUR)
log mi = p u ~ Si =
On substitution of 4 into 3, LY=
(4b) (4c)
P(UR
UI
+
CY
is given in 5.
f U R -I- ~ A C R UR YAuR
(5)
The value of a in eq. 5 should vary with Y. Hence, with electron-releasing substituents where Y AuR is comparatively important, a decreases with an increase of r , since UI, U R and ACR are all negative and Y is positive. In other words, in a series of electrophilic reactions, an intermolecular selectivity pf becomes smaller compared with an intramolecular selectivity Sf with an increase of Y, or the resonance contribution in the transition state. Equation 5 may be transformed into 6. On substituting Brown’s value of a = 1.36 into eq. 6, the parameter Y is calculated to be 0.92 which is
smaller than Y in Deno’s’? a+ (evaluated by Yukawa16 1.00) and than r in Brown’sI5 a+(1.33).I6 This change in Y agrees with the above discussion. For the epoxidation of a-methylstilbenes, r was estimated to be 0.5-0.6, and this smaller value may be due to smaller electron demand of the reaction where no carbonium ion is involved. Acknowledgment.-The authors are grateful to Prof. R. Oda for his aid in performing these experiments.
DEPARTMENTO F INDUSTRIAL CHEMISTRY, THE FACULTY OF ENGINEERISG, KYOTO UNIVERSITY,
YOSHIDA, KYOTO, JAPAN]
Kinetics of the Peroxybenzoic Acid Epoxidation of Substituted Propenes and 2-Butenes in Benzene BY
1-OSHIRO OGAT-4 AKD IWAO T A B U S H I
RECEIVED FEBRUARY 20, 1961 The rates of the epoxidation of 3-substituted 1-propenes and 1,4-disubstituted 2-butenes with peroxybenzoic acid in benzene hare been measured. I n general, electron-releasing groups accelerated the reaction, while no appreciable steric effect was observed with the substituents used. The substituent effect in propenes yields two straight lines in plots of log k us. the u* of Taft, which is explained by a less efficient transmission of polar effect of electron-releasing groups and by an asymmetric attack of peroxyacid on the double-bonded carbon atoms. The substituent effect in 2-butenes is discussed in terms of additive hyperconjugation of hydrogen atoms at the 1- and 4- positions as well as by an inductive effect. The ahnorinally high reactivity of allyl alcohol is ascribed to activation by hydrogen bonding.
It is well known that the peroxyacid epoxidation of ethylene derivatives involves an electrophilic attack by the peroxyacid” Our paper On the epoxi(1) For examples, see (a) D. Swern, Chem. R e m , 45, 1 (1949); h ) P Swt-rn. “Orranic Reactions,” Vcrl. 7 , John Wile? and S o n s , Tnc ,
dation of a-methylstilbenes’ reported that the effects of substituents showed deviations from New Y o r k , N. Y , 1953, p. 378; ( c ) Ya. K. Syrkin and I. I. hloiseef, Prog. Chem. U S S R , P9, 425 (1960). ( 2 ) Y .O r e t a nnd T . Tabushi, J . a m . C h p m Soc., 83,3440 ( 1 9 6 1 ) .
Aug. 20, 1961
PEROXYBENZOIC ACIDEPOXIDATION OF PROPENES
Hammett's equation and the results were explained by resonance interaction3 and asymmetric attack. The present paper describes substituent effects in the epoxidations of 3-substituted propenes and 1,4-disubstituted 2-butenes by peroxybenzoic acid (PBA). These compounds have methylene insulation of the resonance effect. The observed effects are discussed on the basis of the substituent constants obtained from aliphatic compounds4and hyperconjugation effect.
Anal. Calcd. for C8HsOBr: C, 26.33; H, 3.68; Br, 58.34. Found: C, 26.75; H . 3.81; Br, 58.32. Allylbenzene," l-heptenela and allyl alcohollg have been reported to give the corresponding epoxi 2es on peroxyacid oxidation.
Results and Discussions The rate data satisfy the second-order equation, v = K[ethylenic compound][PBA], and no appreciable effect of added benzoic acid on the rate was observed. The K-values for ethylenic compounds are listed in Table I.
Experimental Materials.-Allyl cyanide was prepared from allyl bromide and cuprous cyanide,l allyl phenyl ether from allyl bromide and sodium phenoxide,e allyl isopropyl ether from allyl bromide and alkaline aqueous isopropyl alcoh01,~ allylbenzene and 1-heptene from allyl bromide and phenylmagnesium bromide8 and butylmagnesium chloride: respectively. These compounds were purified by fractional distillations; their boiling points agreed with those in the literature. Commercial allyl bromide and chloride were sometimes found to be contaminated with small amounts of an impurity which was easily oxidized and caused slightly increased rate constantslo a t early stages; hence the puri' of fication by the preliminary oxidation with 3 mole % crystalline PBA was necessary between the fractional distillations. trans-l,4-Dibromo-2-butene ( I ) was prepared by the addition of bromine to butadiene (99.9% purity) in chloroform below -30"; m.p. 52.5".11 trans-1,4Diphenoxy-2-butene, m.p. 84', was prepared from I on treatment with an excess of phenol and concd. aqueous alkali.12 trans-l-Phenyl-4-bromo-2-butene, b.p. 113-1 14O, was prepared from I and phenylmagnesium bromide.l8 trans-1,4-Diphenyl-2-butene,m.p. 45.5", was prepared by the sodium amalgam reduction of 1,4-diphenylbutadiene14 obtained by the condensation of phenylacetic acid and cinnamaldehyde.16 The purifications of PBA and solvent benzene are described previously.2 The Kinetic Study.-The procedure for rate measurements is the Same as in our previous report.2 Products.-As an example of the isolation of the product, the epoxidation of allyl isopropyl ether is described. Peroxybenzoic acid (7 g.) was added to allyl isopropyl ether (5 g.). The mixture was allowed t o stand a t 10-15" for 2 days until the peracid was practically consumed; then the produced benzoic acid was removed by filtration. Distillation of the filtrate, after recovering the unreacted material (ca. 1 g.), gave a liquid boiling a t 130-137' (3.1 g.) and a small amount of liquid of higher boiling point. The main fraction was the epoxide (lit.16 b.p. 131-134'). Anal. Calcd. for C a I 2 O 2 : C, 62.04; H, 10.40. Found: C, 61.67; H , 10.20. Similar treatment of allyl bromide gave the corresponding epoxide, epibromohydrin, b.p. 135-138' (lit. 138"). (3) (a) Y . Yukawa and Y . Tsuno, Bull. Chem. SOC.J a p a n , 32, 965, 971 (1951); see also (b) H. C . Brown and Y . Okamoto, J . A m . Chem. Soc., 79, 1913 (1957), and (c) R . W. Taft, Jr., ibid., 81, 5343 (1959). (4) (a) R . W. Taft, Jr., i b i d . , 7 4 , 3 1 2 0 (1952); (b) 7 4 , 2729 (1952); (476, 4231 (1953); (d) 75, 4534 (1953). (5) J. V. Supniewski and P . L. Salzberg, "Organic Syntheses," Coll. Vol I , John Wiley and Sons, Inc., New York, N. Y., 1948, p. 46. (6) L. Henry, Ber., 5, 455 (1872). (7) R . Skrabal, 2. p h y s i k . Chem., 185, 81 (1939). (8) M. Tiffeneau, Compl. t e n d . , 139, 482 (1904). (9) The preparation is analogous t o that from butylmagnesium bromide: see G . B . Kistiakowsky, J. R . Ruhoff, E. A. Smith and W . E. Vaughan, J. A m . Chem. Soc., 58, 137 (1936). (10) Since in most of the kinetic runs a large excess of allyl compound t o PBA was used. (11) E. M. Shantz, J . A m . Chcm. Soc., 68, 2553 (1946). (12) J. Braun and G. Lemke. Ber., 66, 3548 (1922). (13) J. Braun and G. Lemke, ibid., 65, 3536 (1922). (14) F. Strus, A n n . , 342, 255 (1905). (15) B. B . Corson, "Organic Syntheses," Coll. Val. 11, John Wiley and Sons, Inc., N e w York, N. Y . , 1948, p. 229. (16) F. G. Ponomarev, Z h u r . Obshchli K h i m . , 2 2 , 929 (1952). C . A . , 47, 3794 (1LJ53).
3445
TABLE I SECOND-ORDERRATECONSTANTS OF
PBA EPOXIDATIONS OF SUBSTITUTED ETHYLENES IN BENZENE AT 25"
Ethylenic compound
(M)
PBA, A4
k z X 10:. I . mule-' sec.
-1
1.73 1.38 0.0375 5.77" 0.468 .0200 5 . 79" ,0198 .936 5.77" ,0390 .932 Allyl chloride 1.71 1.96 ,0371 50.8 Allyl alcohol 0.177 .0388 50.5 .171 .0353 51.2 .344 ,0353 48.4 .0895 ,0367 Allyl cyanide 0.442 ,0358 1.65 0.433 3.22 ,0353 Allyl phenyl ether 7.94 0,367 ,0362 16.9 Allyl isopropyl ether .0505 ,0361 16.5 ,104 ,0363 15.9 .202 .0363 .202 16.1* ,0346 Allylbenzene .ow .0358 26.6 ,0352 .167 26.4 1-Heptene .0520 49.9 ,0344 .103 51.9 .0345 1,2-Dibenzylethylene .0721 .0183 186 191 .0479 .0182 ,0483 189 .0184 1,4-Diphenoxy-2-butene 12.4 ,2080 .0348 .0220 13.0 ,3336 l-Bromo-4-phenoxy-2,2378 .0358 7.52 butene .4740 7.46 ,0365 1,4-Dibromo-2-butene 1.218 0.701 .0315 0.657 0.693 ,0357 a At 40.0'. Benzoic acid (0.082M) was added.
Allyl bromide
In contrast to stilbenes and/or a-methylstilbenes having conjugation between ethylenic bond and benzene ring, these propenes and butenes have no such resonance because of the separation of the ethylene bond from the functional group by a methylene group. The difference in steric effect among 3-substituted propenes may be negligibly small, since they have a common methylene group. The constancy of steric effect is seen in the approximate constancy of steric substituent constants, E,, by Taft4: n.-BuCH*, -0.40; PhCH2, -0.38; Ph(CH& (and probably i-PrOCH:, and PhOCH2), -0.43; n-Pr (and probably CH2CN), -0.36; CH2Cl (and probably CH2Br), -0.19. (17) J. Boeseken and J. S. P. Blumberger, Rec. Ivan. chim., 47, 839 (1928). (18) J. Levy and R . Pernot, Bull. soc. chim. France, [SI 49, 1838 (193 1). (19) N. Prileschajew, Ber., 42, 4811 (1909); J. Russ. PhyJ. Chem. Soc., 4b, 609 (1911).
YOSHIROOGATAAND IWAO TABUSHI
3446
VOl. 83
A considerable upward deviation from the line was observed with allyl alcohol and also a slight deviation with allyl phenyl ether. The Difference of p-Values between Electronwithdrawing and Electron-releasing Groups in Propene%-The observed difference between values of electron-releasing and electron-withdrawing groups seems to be ascribable partly to the low transmission coefficient of polar effect of electronreleasing groups through the methylene group. Examples of this low transmission are shown in the comparison of Taft's polar substituent constants u* for a substituent X itself and that for CHZX. The ratios UX*/UCH~X* are with X: methyl, 4.90; ethyl, 5.13; n-propyl, 4.66; acetyl, 1.93; chloromethyl, 1.46. It is apparent that the transmission ratio of the polar effect of electron-releasing groups is ca. l/6 on the average, while that of electronwithdrawing groups is 2/8-1/2.23 - 7.0 __ I I I I n the transition state of epoxidation, the double bond may have partial single bond char- 0.50 0.00 +0.50 f1.00 acter; hence a substituent RCHz on C1 may inU*. fluence CZ as a substituent RCHzCH in part with Fig. 1.-Plots of u* us. log k for the PBA epoxidation of the above-decreased efficiency. Thus with elecCHFCHCH~S by PBA. tron-releasing groups, the effects are smaller beFurthermore, rates of peroxyacetic acid epoxida- cause of lower transmission efficiency. A further probable explanation for the diftion of ethylenes show a regular increase on the introduction of one or two methyl groups as is ference in the p-values is asymmetric attack. The obvious from a comparison of ethylene, propylene possibility of an asymmetric transition state (IIb) and isobutene or 2 - b ~ t e n e . This ~ ~ means that the instead of symmetric state I I a previously has been The steric reaction constant 6 is small in Taft's equation4 discussed in terms of the resonance 1 log (klko) = P * U * 6Ea (1) R - ('tT,-CH=-=,CHI t - C H ~ - - ~ ~ H , -1 1~2 f Therefore, the variation of 6Es is negligible in 3substituted 1-propenes and, inasmuch as the inductive effect only is operating, the equation IIa IIb should be simplified to transition states for allylic compounds with log ( k / k o ) = P*U* (2) electron-releasing groups seem to be more asymJust as the U-values in aromatic electrophilic metric than those of electron-withdrawing groups substitution vary according to the supplementary because of the easier stabilization of the partial resonance contribution, the a*-values in aliphatic positive charge in the central carbon atom in electrophilic reactions such as this epoxidation propene. Hence, in the transition state 11, the may vary with supplementary hyperconjugation, inductive effect of CH2R is less effectively transalthough Taft paid little attention to this because mitted when R is an electron-releasing group, most of the reactions selected by him were nucleo- thus lowering the p-value. This assumption of philic. Hence i t is not surprising that the u*- an asymmetric transition state is also supported values vary in this reaction. by the orientations in electrophilic additions of Plots of log k a t 25' vs. Taft's a*z1gave two dif- ethylenic compounds. The distribution of charges ferent straight lines with slopes of - 0.76 and - 2.44 in the transition state may vary with the change and with a bend in the neighborhood of unsubsti- in polar effect of substituent and the activity of tutedethylene (u*=O), asshowninFig. 1. Asimilar reactants. According to Shelton and abnormal substituent effect was observed in the hyperconjugation is important when the reactants relative rates of halogen addition to double bond.22 have low activity; e:g., the addition of hydrogen For example, the rate increase with the introduc- bromide to allyl chloride which obeys the Markowtion of a methyl group (u* = -0.490) is compara- nikoff rule,24but the inductive effect becomes imtively small (A log k = +0.302), whereas the rate portant when one of the reactants has high activity decreases with that of CHzCl (u* = 0.560, A log (e.g., the addition of hypochlorous acid (HO-Cl+) k = -1.721), CHzBr (0.540, -l.SSS), C H C N to allyl chloride.2s (0.510, -2.568) or (CH2C1)2(XU*= 1.120, -3.620) I n general, peroxyacids have lower activity ; are much larger. hence the transition state for epoxidation may (20) See Swern's list of k-values for the epoddatlon; D . Swern, resemble that of normal (Markownikoff) addition.
i
f
+
/'
Chcm. Revs., 46, 49 (1949). (21) T h e Taft's value of u* estimated on the basis of methyl substituent was corrected t o use the value for hydrogen as the standard. (22) J. R. Shelton and L. Lee, 1.Org. Chem., 16, 428 (1960). See also, S. V. Anantakrishnan and C. K . Ingold, J . Chcm. SOC.,1396 (1935); P. W. Robertson. J. K . Heyes and B. E. Swedlund, ibid., 1014 (1952).
, /
(23) On the other hand, phenyl and o-phenylalkyl groups show abnormal transmission ratios; L e . , UCH,X* is larger than UX* in pheuyl and UCHS* is positive and ox* is negative in uphenylalkyls. (24) J. R. Shelton and L. Lee, J , Ow. Chem., as, 1876 (1958). (25) J. R. Shelton and L. Lee, i b i d . , 24, 1271 (1959). (26) J. K. Shelton and L. Lee, i b i d . , 15, 907 (1960).
Aug. 20, 1961
3447
PEROXYBENZOIC ACID EPOXIDATION OF PROPENES -
C1C 14&HC 1 C H20H 70%
OH -
CICI-I,CH(OH) CHzCl 30%
Furthermore, 1-propenes having electron-withdrawing groups in the 3-position, e.g., allyl chloride, may have rather symmetric transition states in comparison with propene and 1-propenes having electron-releasing groups in the 3-position, since the direction of the inductive and hyperconjugative effects of CHzX is reversed in the former. The Application of Aromatic Substituent Constants to Aliphatic Reactions.-The applications of Hammett’s equation for aromatics to the addition of olefins has been attempted by assuming 1- and %carbon atoms as m- and p-carbon atoms in the benzene ring, respectively,27a using u or o-+.?:~ However, this treatment is valid only in reactions with the most asymmetric transition state and is unsatisfactory for the relatively symmetric epoxidation which should have u = ( X ) ( a m + ) 4- ( p ) ( u p + ) , where X and p are unknown parameters. M‘e use here Taft’s polar substituent constants for the aliphatic series, u * , * , * ~the value having the same significance as the inductive substituent constant for aromatics, u1,Z8so that it includes no supplementary hyperconjugatioii effect of 111.
s
I
CH~=CH-CHIT
64-
s
H +--+
H
- 0.50
Fig. 8.-Plots OF U * zs. log k for t h e PBA epoxidation of XCIX2CH.=CHI (a) and SCHsCH=CHCH*Y (0).
The deviation, A log k, is nearly constant and corresponds to an increase of hyperconjugation by the introduction of two a-hydrogen atoms. Hence log k for ethylene (XCHz=YCHz=H) should deviate downward by 0.57 because of the lack of a-hydrogen atom, and i t is shown in Fig. 2 as H. Figure 2 shows that linear relations exist between log k and U* for a set of HI Ph and Phz,
~H,-CH-CHY
6-k
IIIa
IIIb
Trm CALCULATED
The hyperconjugative effect on orientation, as describcd in the previous paragraph, suggests that this effect also contributes to the rate. Hence, the relative rate may be developed as log (klko) =
PU*
+
pnm
4-8En
(4)
Here, p, u * , 6 and E , have the same meanings as those of Taft, UII is the hyperconjugation constant and n is the number of a-hydrogen atoms. In a series of 1-propenes, the hyperconjugation term, p % U H , and the steric term, 6E,, are both constant, and the equation log ( k / k a ) = pu* is applied. The Estimation of the Extent of Hyperconjugation.-In 3-substituted 1-propenes, CH2=CHCH2X , the hyperconjugative effect is essentially constant as shown previously. But in lI4-disubstituted 2-butenes, XCH2CH=CHCH2Y (IV), the values of kz observed in benzene a t 25’ were considerably larger than those expected from kz for the propenes (Tables I, I1 and Fig. 2 ) . Table I1 shows the deviations of log kobs from log kcalc calculated from the data of 3-substituted propenes, assuming additivity of u’. 29 (27) (a) For example, see F’. B . D. d e la Mare, J. Chen. Soc., 3823 (1900); (b) H C. Brown and Y.Okamoto, J . A m . Chem. Sor., 80, 4978 (1958); J . Org. Chem., 22, 485 (1957). (28) R. W. l a f t , Jr., and I. C. Lewis, i b i d . , 81, 5343 (1939). (29) u’ was obtained b y extrapolation of a log k-u* plot of allylic compounds having the electron-withdrawing group X. These values seem t o correspond to the total polar substituent effect except for the byperconjunction. The values of u’ are: n-Bu, -0,055; Ph, 0.OG5; i-Pro, 0.150: PhO, 0.290.
+ 1.00
+0.50
0.01) U*.
TABLE I1 OBSERVED RATE CONSTANTS FOR 2-BUTENES, XCI-IsCH=CHCHzY
AND
1,4-DISUBSTITUTED
k Substituent x Y
Ph PhO Ph Br
Pli PhO Br Br
-
A log log
koba
Xu’
log k d o
log kob.
0.130 .580 .BO5 1.080
-3.73 -4.80 -4.86 -5.99
-2.2 -3.91 -4.12
-5.16 Av.
-
log k d o
1.01 0.89 .74 .83
0.87
a set of 13, Br and Brt, and a set of H, OPh and
(OPh)r, which indicate that the contribution due to steric effect is still negligible in these disubstituted ethylenes and that the hyperconjugativc effect is additive.30 Hence eq. 4 is simplified to log ( k l k o ) =
P(U*
+ moa)
(5)
Therefore, the hyperconjugative deviation of log k is given by A log k = p A n m
(6)
where A indicates the difference of two valucs with different numbers of hydrogen atoms. The value of p was estimated to be -2.38 from the plots of u* vs. log k with propenes and the average value of A log k in Table I1 was 0.87 and n was 2; hence the value of U E was calculated to be -0.183. (SO) The effect of an asymmetric transition state on A log k is apparent in Table 11. Asymmetric phenylpropene with decreased inductive shift and symmetric diphenylbbtene give a larger A lop k , while asymmetric phenylpropene and more asymmetric phenylbromobutene with increased inductive shift of bromomethyl give a smaller A log k.
Vol. 83
YOSHIROOGATAAND IWAOTABUSHI
3448
and Mez ( 2 u * = 2 X (-0.490), n = 2 X 3 ) ; more negative G * will correspond to larger n. The values of A log k for monoalkylethylenes, 0.28 (n-Pr, U * = -0.605) and0.26 (n-Bu, u* = -0.620)) give an average value of 0.27 (u* = -0.61). The n-value corresponding to u* = -0.61 is 3 X (0.61/0.49) and the actual n is 2, giving A n = 1.74. Substituting these values of A log k , An and n / u * = 3/(-0.490) into eq. 6 and 8 give a UH of - 0.086 by means of eq. 9
0.0
- 1.0 \
&
-- -2.0 M
The values of U H calculated similarly for other ethylenes are listed in Table 111. These GH values TABLE 111 THEl rOF UK ~ CALCULATED ~ FROM ~ THE EPOXIDATION ~ ~ OF ETHYLENIC COMPOUNDS BY PEROXYACETIC ACID
-3.0
SubstitLent on
-4.0
- 1.5
- 1.0
-0.5
0.0
U*.
Fig. 3.-Plots of u* us. log k for the peroxyacetic acid epoxidations of ethylene derivatives, XCH=CH* and XCH=CHY in acetic acid; k in 1. mole-' min.-l.
The Application of the Hyperconjugative Constant for Ethylene Derivatives in the Peroxyacetic Acid Epoxidations.-As shown in Fig. 3, a straight line was obtained from plots of Taft's u**I us. log k2 for the peroxyacetic acid epoxidation of ethylene derivatives, ethylene, propylene, isobutene and 2butene. Most of the other 3-substituted l-propenes except allylbenzene show downward deviations from the line. Since the line involves both inductive and hyperconjugative effects,31the slope of the line, p.,,,,, is given by
Hence
The extent of the deviation of log k from the line depends mainly on the number of a-hydrogen atoms (Table I11 and Fig. 3 ) . From this deviation, the hyperconjugative constant UH is calculated as follows : The deviation, A log k , is given by eq. 6 with given u * , p being estimated from papp by means of eq. S. Here, An is the difference between the number of a-hydrogen atoms corresponding to the line and the number of actual a-hydrogen atoms. For example, the points on the line are H (CT* = 0, n = 0), Me (u* = -0.490, n = 3 ) ,31) A slight deviation of log k for trimethylethylene m a y be due t o steric effect, since t h e rate increase by t h e introduction of methyl group t o cyclopentene ( 4 log k = 1.068)is smaller t h a n t h a t to ethvle n s a n d propylene ( A log k = 1 345).
c1
C2
It
n-Pr n-Bu Me
H H Et n-Pr n-Pr n-Bu
2 2
Me
Et n-Pr
4log
-0
k
27 .265 .28 .47 - .60
-
3 0
4 4
Oil
086 - ,090 - ,092 - ,082 - ,097 -Av. -0.089 -0
are nearly constant, but smaller than uEiobtained from PBA epoxidation of 1-propenes and 2-butenes, suggesting G= may depend on the solvent and,'or the attacking agent. Hence eq. 4 is extended as log ( k l k o ) = pu* f f'pnUH -I- 6Es (10) where Y is a parameter corresponding to the coritribution of hyperconjugation. The Deviation Observed in Allyl Alcohol and Ethers.-Allyl alcohol shows a large deviation upward as shown in Fig. 1. It is probable that this deviation is the activa.tion of PBA due to hydrogen bonding as in IV, as suggested by the activation of PBA itself by the intramolecular hydrogen bonding.32 The possibility of a change in reactivities of allyl ethers by the formation of complexes with benzoic acid was eliminated by the observation that the rate showed no appreciable change on the preliminary addition of benzoic acid to the reaction mixture.
--___
IV
(i12) (a)A . Knhertson and W. A. Waters. J . C / w m SOC.,1674 (19481, (h) B. hI. Lynch a n d K. H . Pausacker, ibid., 162.5 (1955), (c,) for t h r evidences from infrared studies, see H. T . Davison, ibid., 24.56 (1951) and G. J. Slinkoff, Proc R o y . S O L .(Londoiz), A224, 17R (1954)
