TRICHLOROMETHYL RADICAL ADDEDTO 3-PHENYL-1-PROPENES
Jan. 20, 1964
233
ORGANIC AND BIOLOGICAL CHEMISTRY [CONTRIBUTIOS FROM
THE
DEPARTMEST OF CHEMISTRY, THE CNIVERSITY OF MICHIGAN,ANS ARBOR,MICH.]
The Addition of the Trichloromethyl Radical to Substituted 3-Phenyl-1-propenes and 4-Phenyl-1-butenes BY MICHAEL M. MARTIXAND GERALD J A Y GLEICHER' RECEIVED J U N E 14, 1963
-4 Hammett study of the addition of the trichloromethyl radical t o 3-phenyl-1-propenes and 4-phenyl-lbutenes has led t o p-values of -0.29 and -0 20, respectively These unexpectedly high values for reactions which produce alkyl radicals insulated from the substituted phenyl group by one and two methylene groups are explained as being the result of the formation of, and intramolecular addition within, a r-complex of the trichloromethyl radical with the phenyl group of the w-phenyl-1-alkene.
Introduction Although free radicals are neutral species in a formal sense, it has long been recognized that polar substituents in a molecule can have a considerable effect upon the rate and course of a free radical reaction. fiumerous studies2-' of hydrogen abstraction reactions from substituted toluenes (reaction I ) and cumenes have provided very useful information on the polar effects
operating in homolytic processes. A summary of published Hammett up correlations involving hydrogen abstraction reactions, including the results of several unpublished investigations, has recently appeared.8 The p-values for abstraction of a benzylic hydrogen are negative, and correlation is generally better using the electrophilic u +-values than the normal U-values. Benzyl radicals have also been generated by the one-step decarboxylation of a series of substituted t-butyl phenylpera~etates~ (reaction 2 ) . V
V
The data correlate better with u + than u, and lead to p-values of - 1.04 a t 100' and - 1.20 at 56'. 'Studies of free radical addition reactions to styrenic compounds t o produce benzyl radicals (reaction 3) have
+ Y.
-
+ C1,C
'
I a , n = 1; b, n = 2
Experimental
(CH3)3 CO a
X D ( R ) = C H ,
donating groups tend to accelerate and electron-withdrawing groups decelerate homolytic processes in which a benzyl radical is generated by the reaction of an aromatic substrats with an electrophilic radical. In this paper the polar effects operating in reactions in which a radical carbon atom is generated two or three carbon atoms removed from the substituted phenyl group are reported. Since the radical carbon atom so generated is insulated from the substituted phenyl group by one or two methylene groups, the influence of the substituent would be expected to be slight. The reaction chosen for study was the addition of the trichloromethyl radical, generated from bromotrichloromethane, to series of substituted 3-phenyl1-propenes (Ia) and 4-phenyl-1-butenes (Ib) (reaction 4).
x 0 c ( R ) C H 2 Y (3)
not led t o very linear Hammett correlations, but acceleration by electron-supplying groups is apparent in all cases in which the attacking radical is electrophilic. l o The precise origin of these polar effects is a matter of current interest, and two explanations have been advanced, but the established fact is that electron(1) Petroleum Research F u n d Predoctoral Fellow, 1961-1962 and 19621963 ( 2 ) E C Kooyman, R Van Helden, and A . F Bickel, K o n i k l . N e d . A k a d . W e l e n s h a p . P r o c . , B66, 75 (1953). (3) G A . Russell, J . A m . C h e m . Soc., 78, 1047 (1956). ( 4 ) C Walling and B. Miller, ibid , 1 9 , 4181 (1957). ( 5 ) E . S . Huyser. i b i d . , 82, 394 (1960). (6) C . Walling and B. Jacknow. i b i d . , 82, 6113 (1960). ( 7 ) R I:. Pearson and J C . Martin, i b i d . , 86, 354 (1963). (8) J A Howard and K . U. Ingold, Can. J . C h e m . , 41, 1744 (1963). (9) P. D Bartlett and C. Ruchardt, J A m . C h e m . Soc., 82, 1756 (1960). (10) C Wslling, D Seymour, and K B. U'olfstirn, i b i d . , T O , 1537, 2559 (1948).
Materials. Bromotrichloromethane (Matheson Coleman and Bell) was distilled through a 30-in. helices-packed column, taking only a middle cut, b.p. 103-103.5". This distilled portion was washed with water, dried with sodium sulfate, and redistilled a t reduced pressure, b.p. 43" (80 mm.). Gas phase chromatography on a 15-ft. 10% silicone gum rubber column indicated t h a t the material was 99.2% pure. The impurity had the same retention time as carbon tetrachloride and a different retention time from chloroform. Chlorobenzene (Eastman) mas shaken with three portions of sulfuric acid, once with water, three times with 5% sodium bicarbonate, again with water, dried successively over calcium chloride, calcium sulfate, and phosphorus pentoxide, and distilled; b . p . 130-130.5". The 3-phenyl-1-propenes and 4-phenyl-1-butenes were prepared by the coupling reaction of allyl bromide with the appropriate phenylmagnesium bromide" or benzylmagnesium chloride.'s 4-(4-Methoxyphenyl)-l-butene was prepared by the reverse process, that is, from allylmagnesium bromide and pmethoxybenzyl chloride. Coupling of p-methoxybenzylmagnesium chloride with allyl bromide gave only a 5 5 yield of 4-(4methoxypheny1)-I-butene, leading mainly t o 4,4'-dimethoxybibenzyl. Yields, boiling points, and analytical data on all new compounds are presented in Table I /3-phenyI-l-propenes) and Table I1 (4-phenyl-I-butenes). .ill of these compounds were purified by distillation and were adjudged pure when gas phase chromatography on a diethylene glycol succinate column resulted in only one peak. In some instances, the initially distilled 4-phenyl-1-butenes contained traces of unreacted benzyl chloride. This was removed by heating the liquid product with triethyl(11) C. Walling, "Free Radicals in Solution," John Wiler and Sons, I n c . , S e w P o r k , K.Y., 1957, Chapter 4. (12) G A. Russell, J . O V R .C h e m . . 23, 1407 (1958). (13) R . W. T a f t , J. R Fox, and I , C . Lewis, J . A m C h e m . S o c , 8 3 , 3349 (1961). (14) E B. Hershberg, H t l v C h i m A h , 11, 352 (1934) (15) P. Kozacik and E . E. Reid, J . A m . Chem Soc , 6 0 , 2436 (1938).
MICHAEL M. MARTINA N D GERALD JAY GLEICHER
234
Vol. 86
TABLE I SUBSTITUTED 3 - P H E N Y L - l - P R O P E K E S
Yield, %
X
H p-CH3 m-CH3 p-Cl m-CI P-CHaO WZ-CH~O p-CeH6
P-F
76 61 60 74 70 67 50
55
47 60 m-CFa 52 a Lit.14153-154".
p-CFa
B . p . , "C.
153-154" CsHio 180-183b CioHiz ... 180- 183c CioHi2 199-201 CgHgC1 70 83 198-202 CgHsC1 70.83 213-216d CioHizO ... 213-216 CiaHizO 81.08 132-135 (1 m m . ) Cl6Hl4 92 74 157-160 CsHsF 79.42 168- 17 1 CioHgF3 64 51 164-166 CioHsFa 64.51 Lit. 180-181°,16180-182°.17 L i t i 7 60" (11 m m . )
--
yo---
---Carhon, Calcd.
Formula
Found
Hydrogen, 7,Calcd Found
--Other, Calcd.
70 81 70.70
5.94 5.94
6 00 5 91
23 23 2 3 23
23 40 23 10
14 08 30 64 30 64
13,95 30 46 30 50
81 92 79 64 64 Lit
09 8 71 '7 39 6 36 4 39 4 215216'
11 26 70 85 85
8 7 6 4
29 42 66 93 4 97
%-Found
TABLE I1 4-PHEXYL-l-BUTESES
Yield,
X
H p-CH1 m-CHn
a
%
70 42 58 p-c1 68 m-C1 76 P-F 64 m-F 59 P-CHIO 26 m-CHIO 45 Lit.lb b.p. 181-182".
Bp--
7 -
OC
176 5-177" 83 91-92 93-93 5 91-92 85-86 76-77 110 5-111 105106
amine for 10 hr. a t 85', then washing with water and dilute hydrochloric acid, drying, and distilling the residue. T h e absence of any benzyl chloride was confirmed by v.p.c. and b y testing for active halogen with alcoholic silver nitrate solution. A11 of these olefins exhibited an infrared maximum between 1640 and 1653 cm.-', characteristic of an unconjugated double bond, and a maximum between 905 and 925 cm.-' and one between 985 and 1000 cm.-', characteristic of a terminal vinyl group. No bands attributable t o compounds in which the double bond had rearranged by migration into conjugation with the benzene ring were evident. The ultraviolet spectra also indicated the presence of a substituted phenyl group, unconjugated with the double bond of the side chain. These compounds are definitely not contaminated with a n y stprenic products. Products.-A mixture of 1.197 g . (0.0101 mole) of 3-phenyl-lpropene and 8.729 g. (0.0443 mole) of bromotrichloromethane, sealed under nitrogen in a Pyrex ampoule, was placed in a horizontal position just beneath the surface of a n oil bath maintained a t 69.5 i 0.2", and irradiated for 19 hr. by a General Electric 2it5-w. sunlamp placed 19 cm. from the surface. At the end of tluo time, v.p.c. analysis on a diethylene glycol succinate column indicated complete utilization of the 3-phenyl-1-propene. Excess bromotrichloromethane was removed by heating on a steam bath a t aspirator pressure, after which there remained 3.04 g. of transparent brown residue, consisting of the 1 :1 adduct, l , l , l trichloro-3-bromo-4-phenylbutane, and the bromination product, 3-bromo-1-phenylpropene. Making use of the ratio for the rate of addition of the trichloromethyl radical t o the double bond t o the rate of abstraction by the trichloromethyl radical of a labile hydrogen atom, determined as described below, it can be calculated t h a t this corresponds t o a yield of 2.93 g. (93.87,) of l : l adduct, and 0.11 g. (3.3Yc) of bromination product. Extensive decomposition and fuming occurred when this residue was distilled, so t h a t purification of the entire sample was not achieved. The physical properties of a sample of distillate agreed satisfac______~ (16) C D H u r d and H . T . Bollman, J . A m . Chem. Soc , 66, 447 (1934) ( 1 7 ) K. Y 1-evina. . I . (;en Chem U S S R , 7 , 684 ( 1 9 3 7 ) . Chem A b s f v ,
31, 57726 (1437) (18) I . hi Heilbron. "Dictionary of Organic Compounds," Val. 2. Oxford U n i w r s i t y Press, K e w York. N. Y , 19.53,p. 483.
---
Carbo?, 70--Calcd Found
--Hydrogen, Yo-Calcd Found
----Other, Calcd.
...
90 33 90 33 72 08 72 08 79 95 79 9 5 81 44 81 44
90 90 72 72 79 80 81 81
36 19 13 21 97
07 35 46
9 67 9 67 6 66
6 7 7 8 8
66 40 40 70 70
9 77 9 50 6 73 6 70 7 53 7 52 8 44 8 90
21 28 21 28 12 65 12 65
%-Found .
.
21 12 21 50 12 77 12 50
torily with those reported for the adduct by Kharasch and %gelg, b.p. 178' (30 mm.), %*OD 1.5640 (lit.lg b.p. 79" (0.05 m m . ) , ~ * O D 1.5650). T h e mode of addition was established as follows. A mixture of 6.91 g. (0.022 mole) of adduct and 3.57 g. (0.044 mole) of lithium aluminum hydride in 50 ml. of ether was refluxed for 2 days. The reaction mixture was hydrolyzed by the addition of 3.6 g. of water and 5 . 3 g. of 10yc sodium hydroxide solution, the mixture filtcred, layers separated, and the ether layer was dried. T h e product was analyzed by V.P.C.a t 120" on a 10% diethylene glycol succinate column. The major product peak corresponded to n-butylbenzene. N o peak corresponding t o isobutylbenzene was present. From a consideration of the minimum quantity of isobutylbenzene t h a t could have been detected, i t was concluded t h a t a t least 99% of the addition occurred at the terminal position. In a similar experiment, 6.62 g. (0.05 mole) of 4-phenyl-lbutene and 20 g. (0.10 mole) of bromotrichloromethane were irradiated for 27 hr. at 69.5 i 0.2". All of the 4-phenyl-1-butene was found to have reacted, and excess bromotrichloromethane was removed by heating the reaction mixture a t 80" and 0.2 mm. for 3 hr. T h e transparent, light brown residue weighed 16.0 g., corresponding to a 97Yc yield of 1 : 1 adduct. This liquid could be distilled with some decomposition, b.p. 122' (0.35 m m . ) . The infrared spectrum showed characteristic aromatic absorption of the type expected of a monosubstituted benzene, although the bands due to carbon-chlorine obscure this region to some extent. No bands attributable to a double bond were present. Anal. Calcd. for C1lHIZBrCIa:C, 39.96; H , 3.66; Br, 24.1'7; C1, 32.17. Found: C , 40.05; H , 3.71; Br, 24.18; C1, 32.22. Determination of Relative Rates of Addition to Abstraction.--Samples of the olefin and a stock mixture of bromotrichloromethane and chlorobenzene were accurately weighed into an amber vial. The weights used corresponded t o mole ratios of olefin : bromotrichloromethane: chlorobenzene of 8:2 : 1. The olefin was used in excess so t h a t the favored reaction of the trichloromethyl radical would be with olefin and not with bromo(19) M S. Kharasch and M. Sage, J Ovg. Chem., 14, 537 (1949)
TRICHLOROMETHYL RADICAL ADDEDTO %PHENYL- 1-PROPENES
Jan. 20, 1964
trichloromethane or a product of the reaction. Chlorobenzene served as an inert standard in the v.p.c. analyses. A sample of this solution was sealed under an atmosphere of nitrogen in a Pyres ampoule, and placed horizontally just beneath the surface of an oil bath maintained a t 69.5 f 0.2", and irradiated by a General Electric 275-w. sunlamp placed 19 cm. from the surface. The extent of reaction, t h a t is, the per cent of bromotrichloromethane consumed during a given time of irradiation, was not a very reproducible quantity, since the chain process b y which bromotrichloromethane adds to olefins is quite susceptible to trace impurities which can serve as initiators, promoters, retarders, and inhibitors. In these studies, the ampoule was irradiated such t h a t from 15 to 90% of the bromotrichloromethane was consumed. The ratio of rate constants for addition t o abstraction was independent of the extent of reaction. Reactions were run in duplicate, and in some cases triplicate and quadruplicate. The ratio of k , / k , was calculated from the expression
k, kt
-
moles of BrCC13 consumed - moles of HCC13 formed moles of HCC13 formed
where k , is the rate constant for addition (reaction 4 ) and k , is the rate constant for abstraction of a labile hydrogen. Analyses were conducted in duplicate or triplicate by gas phase chromatography ( v . P . c . ) on a 6-it. 10% diethylene glycol succinate column (L.lC-728). Relative areas under curves were determined with the aid of a disk integrator. In general it was observed t h a t in experiments in which k J k t ratios were determined for the 3-phenyl-l-propenes, more olefin was consumed than bromotrichloromethane. This is most likely a result of reactions between the alkyl radical produced by reaction of the trichloromethyl radical and the 3-phenyl-1-propene with other olefin molecules, such as polymerization (reaction 5 ) or hydrogen abstraction (reaction 6 ) . Such a reaction sequence would require the consumption of more olefin than bromotrichloromethane, and is very reasonable in a medium in which olefin is present in a fourfold excess over bromotrichloromethane. I t does not inva!idate the calculations of k J k , since it provides no alternative modes of consumption of bromotrichloromethane or formation of chloroform. With the 4-phenyl-l-butenes, no chloroform could be detected in the irradiated mixtures of reactions carried out as described above. Thus, accurate k J k t ratios could not be calculated. Minimum values of the order of 200/1 can be estimated from the limits of the experimental method. The real values are probably much higher. With the 4-phenyl-l-butenes, the number of moles of bromotrichloromethane and olefin consumed were generally within 59; of each other, indicating much longer chains than in the reaction involving 3-phenyl-1-propenes and less reaction of intermediate alkyl radicals with olefin. Determination of Rates of Disappearance of Substituted Relative to Unsubstituted w-Phenyl-1-alkenes.-Samples were prepared and irradiations conducted in the same way as described for the determination of k , / k t . Mole ratios of substituted olefin: unsubstituted olefin : bromotrichloromethane : chlorobenzene were 1 : 1 : 4 : 0.1 in the case of the 3-phenyl-1-propenes and 1 : 1:8:0.2 in the case of the 4-phenyl-1-butenes. Bromotrichloromethane was used in excess to favor chain propagation by reaction of the alkyl radical produced in the addition or abstraction step with bromotrichloromethane and to suppress polymerization or termination involving abstraction of a hydrogen from olefin, t h a t is, reactions of the type depicted in 5 and 6. Relative rates of disappearance of the two olefins b y all reaction paths was calculated from the expression
235
determined rate constant for disappearance is equal to the sum of the rate constants for addition and abstraction, t h a t is, k = k, k,, and k , / k , is known for each individual compound. Since abstraction was negligible with the 4-phenyl-l-butenes, the measured k x / k a ratio for disappearance is equal to the ratio for addition. The Hammett p-values were calculated from least squares plots of log kx/kH v s . u 0 ,according to the procedure recommended by Taft.20-22 The UO-constant reflects all polar effects exerted by a substituent, both inductive and mesomeric, which are ultimately transmitted to the reaction center by induction, a n d excludes polar effects resulting from direct resonance interactions between substituent and reaction center. The correlation coefficients were calculated according to standard statistical methods.23
+
Results Course of the Reaction.-In any kinetic study in which the disappearance of a compound is followed, it must be established with certainty t h a t the substance is disappearing by the reaction which the investigator claims to be studying. In the case of the disappearance of w-phenylalkenes upon reaction with trichloromethyl radicals, several paths other than addition to the terminal vinyl group are conceivable: (a) addition to the ring, (b) polymerization (reaction 5 ) or termination (reaction 6) involving the olefin, and (c) abstraction of a labile hydrogen by the trichloromethyl radical (reaction 7 ) . r--+ C13CCH2CHCH2CH. I I CH24r CHzAr
CIaCCH2CH.
+ CH2=CH
j
CH2Ar
I
CH2Ar L+C13CCH2CH2I
+ CHz=CH-CH
CH2.k ArCH2CH=CH2
+ CIJC. +hrCHCH=CH2 +
I
,4r HCCL ( 7 )
where k x is the rate constant for disappearance of the substituted compound, kH the rate constant for the unsubstituted compound, X , and Hi the initial numbers of moles of substituted and unsubstituted olefin, respectively, and Xf and Hi the final numbers of moles of substituted and unsubstituted olefin, respectively. Analyses were conducted by V.P.C.on a 6-ft. lo%, diethylene glycol succinate column. The fluoro compounds were not completely separated from the unsubstituted compounds on v.P.c., and in those cases competitive experiments involving the fluoro compound and a methyl or chloro compound were run. Then, kF/kkr was calculated from an expression such a s
A survey of the literature indicates t h a t ring addition (a) can be ruled out as a complicating reaction. The synthetic aspects of the addition of bromotrichloromethane to olefins has been extensively s t ~ d i e d , 'and ~ , it ~ ~has ~ been ~ ~ found to proceed much more readily than the much slower attack on an aromatic s y ~ t e r n . ~ ~ ~ ~ ~ Polymerization (b) can be ruled out on several grounds. The high yields of 1:l product (adduct plus bromination product), 9770 from both 3-phenyl1-propene and 4-phenyl- 1-butene, is a good indication that olefin is not being consumed by polymerization or termination as depicted in reactions 5 and 6. X!so, in several of the kinetic runs in which the number of moles of bromotrichloromethane consumed was carefully determined, it was found t h a t the total number of moles of olefin consumed corresponded very closely to the total number of moles of bromotrichloromethane consumed. The failure to detect any chloroform in the reactions of bromotrichloromethane with the 4-phenyl-1-butenes rules out abstraction of a labile hydrogen atom as a complication. Small quantities of chloroform were detected in the 3-phenyl-1-propene series, and a ratio of the rate constant for addition, ka, to the rate con-
kF/kH = kF/kcna X k c d k n The analyses were not conducted in such a manner which gave high precision to the measurement of the area of the bromotrichloromethane peak. This material came off the column very rapidly, giving a spikelike peak. However, in several instances in which the number of moles of bromotrichloromethane consumed was carefully determined, it was found t h a t it was very close to the total number of moles of olefin consumed, confirming t h a t the olefin is being consumed only in the formation of 1: 1 adduct and is not disappearing by polymerization. Calculations.-The rate constant ratios, k x / k A for addition a n d abstraction are readily calculated, since the experimentally
(20) R W. T a f t and I C Lewis, J A m Chem Scc , 81, 53-18 (1959) (21) R. W. T a f t . S Ehrenson, I . C Lewis, and R E Glick, ibtd , 81, 5352 (1959). (22) R . W T a f t , J . P h y s Chem , 64, 1805 (1960) (23) W . L . Gore, "Statistical Methods for Chemical Experimentation." Interscience Publishers, I n c , New York, S Y . , 1952, Chapter 6 (24) Reference 11. Chapter 6 (25) J I G. Cadogan and D H Hey, Quart. Rev (London), 8 , 30% (1954). (26) M S . Kharasch, 0 R e i n m u t h , and W H Urry, J A m C h e m Soc , 69, 1105 (1947). (27) E. C . Kooyman, Discussions F a r a d a y Soc., 10, 163 (1951) (28) E C Kooyman and E Farenhorst, T r a n s . F a r a d a y Soc , 49, 5 8 (1953)
kx/ka
=
log Xi/Xr/log Hi/Hr
M. MARTINA N D GERALD J A Y GLEICHER
2Xli
TABLE 111
TABLE Y
RELATIVE RATESOF ADDITION TO ABSTRACTION I N THE REACTION O F XC6H?CH2CH=CH? 1VITH c13c. AT 69.5" k,'kt
Substituent
'2
VOl. 86
~IICIIAEL
P-CHaO 2 3 . 8 f 1 5'" P-CHa 25 3 f 0 . 3 m-CHa 28.5 f 1.9 H 28.9 f 1 0 P-CeH, 29 1 f 1 . 7 P-F 30 3 f 1 . 0 Average deviation
Substituent
ka,lkt
WZ-CHaO p-Cl m-C1 m-CF.1 p-CFa
3 2 . 5 f 0 03 32 9 i 1 . 1 40 1 0.85 42.9 f 1 6 44 9 f 1 . 9
+
stant for hydrogen abstraction, kt, was determined for each compound. The results are summarized in Table 111. These values of ka/kt ranging from 24 to 45 for the 3-phenyl-1-propenes and in excess of 200 for the 4-phenyl-1-butenes are unexpectedly high. HuyserZ9 observed a ka/kt ratio of 43 for 1-octene a t 77.8'. Considering that 4-phenyl- 1-butene contains benzylic as well as allylic hydrogens, a value of ka/kt of 200 is unexpectedly large. Also, considering the great lability of the aliphatic hydrogens in 3-phenyl-1-propene, values of ka/kt less than 24 would have been expected. 3-Phenyl-l-propene, 4-phenyl-l-butene, and 5-phenyl- 1-pentene have all been observed to form unexpectedly high molecular weight alternating copolymers with maleic anhydride,x' also pointing to a high rate of addition relative to hydrogen abstraction. Data in the l i t e r a t ~ r e ~ ~indicate - - ? ~ ~ that ~ ~ the addition of the trichloromethyl radical to a terminal olefin is practically entirely, if not exclusively, a t the terminal position. In this work, i t was confirmed that terminal addition was favored over nonterminal addition in the case of 3-phenyl-l-propene, by a factor of 99 : 1. Thus, it is concluded that the rate of disappearance of olefin in these reactions is a measure of the rate of addition to the terminal position of the double bond (reaction 4), with a small correction for hydrogen abstraction from the Zphenyl- 1-propenes. Relative Rates of Addition.-In Table IV are presented the rates of addition of trichloromethyl radicals to substituted 3-phenyl-I-propenes, relative to the unsubstituted compound, and the relative rates of hydrogen abstraction by trichloromethyl from these compounds. In Table V are the relative rates of addition to the series of 4-phenyl-I-butenes. Negligible abstraction was observed. TABLE I\' RELATIVE RATESOF .ADDITION A N D ABSTRACTIOS I N REACTION X C ~ H I C H ~ C H = C H ~ ChC. A T 69.5"
THE
+
Substituent
6'21
( k x ' k ~ ) ,addition
( k x , / k H ) , abstraction
1 20 f 0 06" 1 46 f 0 07" 1 2 1 f 02 1 42 f 03 1 13 f 04 1 15 f 04 00 1 000 1 000 1OPfO08 102fO08 00 0 93 f 01 0 82 f 01 13 99 f 01 94 f 03 17 88 f 05 78 f 05 Pi 37 86 =I= 02 60 f 01 4Ph 80 f 03 54 f 02 52 73 i 02 1 7 f 01 Average deviation u"-values for m- and p-CF, are not available, but for strongly electrcin-wEtlidrawing groups, u = 6'. The values used were a-values 3 1
-0
16 15 07
+
+ + +
+ +
(29) 1.: S Huyser. J 0 i . p C h r m , 36, 326.1 ( 1 9 6 1 ) . ( 3 0 ) M hl hlartin and S P Jensen. i b % d , 2 7 , 1201 (1062) ( 3 1j J Hine, "Physical Organic Chemistry." AlcGraw-Hill Inc , New, York. K Y . , 19.56, p 7 1
Book Co..
RELATIVE RATESOF ADDITION ISTHE R E A C T I O ~ X C G H I C H ~ C H ~ C H = C H ~CIsC. A T 69 5'
+
Substituent
QozL
P-CHsO P-CHa m-CH3 H m-CH30 P-F p-CI m-F m-C1 Average deviation
-0 -
+
+ + + +
16 15
07 00 13 17 2i 35 37
( k X / k H ) addition
1 07 f 0 1 10 1 08 f 1 000 0 96 f 0 94 i
*
88 f 90 f 84 f
05. 02 01 04 06 02 02 01
I n Table VI the Hammett p-values are recorded. In Fig. 1 and 2 the graphical plots of log ( k x / k ~ ) for addition vs. u o are represented. TABLE VI HAMMETT P-YALUES Reaction (69.5')
P
r'
Addition of C h C , t o XC~HICH~CH=CH~ - 0 . 2 9 f 0.02" -0 98 Addition of C13C. to XCaH4CHzCHzCH=CHz - .20 f 02' - .9T Abstraction of H by CI3C. from XC~HICH~CH=CH; - .58 02 - .99 Correlation coefficient. * Calculated according t o the procedure recommended by Taft,lg-*lin which the p-value is calculated from the points corresponding t o m-CHs, H , m-CH:lO, m-CI, m-CF,, and p-CF,. The other points, corresponding t o p-CH,O, p-CHs, p-F, p:CGH3, and p-CI, are capable of direct resonance interaction with the radical center adjacent t o the phenyl group. The deviations exhibited by these points indicate t h a t such a resonance effect is significant. The calculated Rvalues are: p-Cll,O (+0.31); p-CH, ( + 0 . 1 2 ) ; p-C6Hi ( + O . O T ) ; p-F ( + 0 . 2 8 ) ; p-CI ( + 0 . 1 6 ) . The p-values, calculated using ordinary Hammett U-values are -0.28 ( Y = -0.97) for the 3-phenyl1-propenes and -0.18 ( r = -0.96) for the 4-phenyl-1-butenes. The use of u+-v:~luesis not appropriate since there is no direct resonance interaction between substituent and reaction center.
+
The p-value for hydrogen abstraction by trichloromethyl radicals from toluenes is - 1.46,j considerably higher than the -0.58 observed for the allylbenzenes. This is as it should be, since the abstraction of a hydrogen atom from a 3-phenyl-1-propene is a lower energy process than from a toluene, as a result of the greater stability of an a-vinylbenzyl free radical compared to a benzyl radical. This compression of the energy scale results in a corresponding diminution of the effects of a substituent in the ring.
Discussion In common with the other Hammett correlations discussed in the Introduction,2-10negative p-values were observed for the addition of the trichloromethyl radical to 3-phenyl-1-propenes ( p = -0.29) and 4-phenyl1-propenes ( p = -0.20). However, these p-values seem rather large for homolytic reactions of relatively low activation energy which produce radical centers two and three carbon atoms distant from the substituted phenyl group. A p-value probably no greater than -0.1 would be expected for the addition of the trichloromethyl radical to the 3-phenyl-1-propenes on the basis of the following considerations. If two reactions series involve species of comparable polarity, that reaction series associated with the lower activation energy will be characterized by the p-value of lower absolute magnitude. Thus the p-value for hydrogen abstraction by the benzylperoxy radical from toluene is -0.705 and from cumene is -0.375.3,s . The p-value for hydrogen abstraction by the trichloromethyl radical from toluene is - 1.46" and from 3-phenyl-1-propene
TRICHLOROMETHYL RADICAL ADDEDTO 3-PHENYL-1-PROPENES
Jan. 20, 1964
237
0.15, 0.10
0.15,
9
3(
I
0.05 0.00
-0.05
= -0.10
UP
co
Fig. 1.-Correlation
of log k x / k o and u 0 for the addition of t o XCBHICHZCH=CHZ.
-
CLC.
is -0.58, and for abstraction by the bromine atom is - 1.39 for toluene7 and -0.76 for 3-phenyl-l-pr0pene.~~ The p-value for abstraction of a hydrogen atom from toluene by a bromine atom is - 1.3g7and by a chlorine atom is - 0 . G G . 4 , 8 Temperature differences among the various studies cited complicates direct comparison of the p-values reported, but there can be no doubt t h a t in each case the reaction series involving the production of the more stable radical type (cumyl vs. benzyl, a-vinylbenzyl v s . benzyl) or the one involving the more reactive atom (chlorine DS. bromine), that is, the reaction series of lower activation energy, is associated with a lower p-value. Now, the high values of the ratios of rate constants of addition to abstraction (24 t o 45) observed for the 3-phenyl-1-propenes indicates that the addition of a trichloromethyl radical to the terminal double bond has a lower activation energy than abstraction of a benzylic hydrogen atom. Consequently, the p-value for addition would be expected t o be less than the p-value for the abstraction reaction. The exact magnitude of this difference is difficult t o specify, but a factor of two or three would seem reasonable. However, in addition to these features of the comparative energetics of the two processes, the addition reaction produces a radical center one carbon atom further distant from the substituted phenyl group than does the abstraction reaction, which will reduce the p-value by another factor of about 2 . 5 . 3 3 , 3 4Thus, from these considerations, a p-value roughly of the order of one-sixth t o one-ninth t h a t of hydrogen abstraction from 3-phenyl-1-propenes by the trichloromethyl radical (-0.58) would be anticipated for the addition of the trichloromethyl radical to the terminal double bond of the 3-phenyl-1-propenes. Thus, a maximum value of -0.1 would be expected for addition to the 3-phenyl-1-propenes and of -0.05 for addition t o the 4-phenyl-1-butenes, Several explanations for the origin and mode of transmission of the observed effects must be considered. One possibility is that the formation of bridged radicals, analogous to phenoninum ions,36could result in resonance stabilization of the P-phenylethyl and y-phenylpropyl radicals formed by the addition of trichloromethyl radicals to 3-phenyl-1-propenes and 4-phenyl-1-butenes. However, other s t ~ d i e P -have ~~ shown that such nonclassical radicals are not formed. Also, the excellent correlation with u 0 without deviations by strongly electron-donating groups argues against a direct resonance effect. Neither does a pure inductive effect, operating through the intervening methylene groups, seem to be a satisfactory explanation for the observed polar effect. As has already been ( 3 2 ) M . ?A. Martin and G. J Gleicher, J . Org. C h e m . , 28, 3266 (1963). ( 3 3 ) H Jaff6. Chem Rev. 63, 216 (1953). (34) M h l Martin and
