Reactions of Benzoyl Peroxide and Olefins Catalyzed by Copper Salts

Benzoyl peroxide reacts with olefins in the presence of copper salts to form benzoate esters and benzoic acid as described by Kharasch and Fono...
0 downloads 0 Views 985KB Size
15'72

JAY

K. KOCHI

Vol. 84

[CONTRIBUTION FROM SHELLDEVELOPMENT Co., EMERYVILLE RESEARCH CENTER,EMERYVILLE, CALIF.]

Reactions of Benzoyl Peroxide and Olefins Catalyzed by Copper Salts' BY JAY K. KOCHI'~ RECEIVEDOCTOBER 19, 1961 Benzoyl peroxide reacts with olefins in the presence of copper salts to form benzoate esters and benzoic acid as described by Kharasch and Fono. This reaction is compared to the andogous reaction employing t-butyl peresters. The difference in composition of products obtained with each peroxide is attributed to the formation of benzoxy radicals from benzoyl peroxide and t-butoxy radicals from l-butyl peresters with copper salts. Benzoxy radicals react with olefins primarily by addition, whereas t-butoxy radicals react by hydrogen abstraction. Products are formed as a consequence of the facile cupric salt oxidation of the ensuant radicals, benzoxyalkyl adduct radicals and alkenyl radicals, respectively. The relative rates of addition t o and hydrogen abstraction from butenes by benzoxy radicals is discussed Evidence is presented for the formation of benzoxybutyl radical intermediates.

Introduction The copper salt-catalyzed reaction of t-butyl peresters and olefins was described in a previous paper.2 The reaction was postulated to proceed vza the steps

+ +

+

Bu' 03CR CUI -+ Bu' 0 . CU" OaCR (1) Bu' 0. R'H +But OH R'. (2) R'. CuI1 OzCr -+ R'02CR CUI etc. (3)

+

+ +

The isomeric butenes (butene-1, cis- and transbutene-2) and t-butyl peresters (acetate and benzoate) with copper salts as catalysts (0.5%) produced the same mixture of a-methallyl (-90%) and The reactions conducted in crotyl esters (-10%). n-hexane and benzene solvents were not homogeneous due to the low solubility of the copper salts. The relative yields of a-methallyl and crotyl esters formed under these conditions did not differ significantly from those obtained in the completely homogeneous media (acetonitrile, methanol and acetic acid). However, in the presence of excess cupric salts (-27,) the homogeneous systems afforded a mixture consisting mainly of the a-methallyl ester (95-10070) with only small amounts of crotyl isomer (< 5yc). In similar reactions, Kharasch and co-workers first described the copper salt-catalyzed reaction of octene-1 with t-butyl perester9 and with benzoyl p e r ~ x i d e . ~Although the former was reported to yield exclusively the unrearranged 3-acyloxyoctene1, the latter yielded a mixture of allylic isomers, 3benzoxyoctene-1 and 1-benzoxyoctene-2 in approximately equal amounts, in addition to benzoxyoctanes and phenylated materials. The resolution of this apparent anomaly between the behavior of t-butyl peresters and benzoyl peroxide toward olefins and an elaboration of the mechanism of these interesting reactions is the subject of this paper. Results The isomeric butenes were chosen as representative olefinic substrates for study because of the accessibility and unequivocal structure assignments ( 1 ) Presented in part before t h e Petroleum Section of t h e Gordon Research Conference, Colby Junior College, New London, N. H., June, 1961. P a r t 11, Peroxide Reactions Catalyzed by Copper Salts. For previous papers see J. Am. C h o n . Soc., 85, 3162 (1961), and publication in press. ( l a ) Department of Chemistry, Case Institute of Technology, Cleveland 6 , Ohio. (2) J. Kochi, J . A m . Chem. Soc., 84, 1193 (1962). (3) It.Kharasch, G. Sosnovsky and S. S a n g , ibid., 81, 5819 (1959). (4) M. Kharasch and A. Fono, J . Org. Chem., 24, 606 (1959).

of their derivatives. The copper salt-catalyzed reaction of benzoyl peroxide and butenes taken in analogy with the reactions of t-butyl peresters ideally should be describable by the equation (C6H5COz)z

+ CdHs c u

+

----f

+

C ~ H ~ O Z C C B HC6H,COOH ~

(4)

The subsequent reactions will be discussed in terms of this stoichiometry. Butene-2 and Benzoyl Peroxide.-A solution of benzoyl peroxide and cis-butene-2 in benzene reacted when heated a t iO-SOo for 24 hours to yield benzoic acid (20'%) and a mixture of esters (1S070)5 (see run 4, Tables I and 11). The ester products for this uncatalyzed reaction consisted of a 94% yield of sec-butyl benzoate, 3yc of a-methallyl benzoate and 28Yc of octyl benzoate (Z-benzoxy-3,4-dimethylhexane) . Cnidentified higher molecular weight esters were formed in 59% yield. The recovered unreacted butene was unrearranged. X reaction carried out under similar conditions in the presence of catalytic amounts of copper salts (cuprous chloride, cuprous bromide, cupric acetate) yielded different results. There was formed in 75-S57G yield a C4-benzoate mixture (mainly a-methallyl and sec-butyl benzoates) together with SO-857, benzoic acid, 30y0 unidentified higher molecular weight benzoates and a small amount (< 17,) of biphenyl and also some C4-benzenes. Less than 5Yc Cs-benzoates were formed in these catalyzed reactions. In Table I the yields of benzoic acid, C4-benzoates and higher molecular weight esters formed under representative conditions are listed. In most cases there were good material balances amono, these compounds and the benzoyl peroxide charged. The yields of carbon dioxide were not measured, but in all the catalyzed reactions little gaseous products were apparent The unreacted butene-2 was isolated in each case and examined for rearrangement and isomerization. Butene-1 was not formed. Only when cuprous bromide was employed as a catalyst was czs-trans isomerization observed. Neither cuprous chloride nor cupric acetate induced the isomerization; the results were otherwise the same as those obtained from the cuprous bromide catalyzed reactions. Competition experiments were conducted using a fivefold excess of cas- and trans-butene-2 mixtures. The unreacted butene-2 was reisolated and re( 5 ) Yield calculations are based on a mole of benzoyl peroxide producing a mole of benzoic acid and a mole of benzoate ester.

May 5, 1062

1573

REACTIONS OF BENZOYL PEROXIDE A N D OLEFIXS TABLEI COPPER SALT

Run

.A €3

Butene-2, y.

r i s (52) (.is (-53)

c

lruris (50)

1)

cis ( c55)

E G

cis ( 5 5 ) cis (56) cis (52)

H

cis (53)

F

cis ( 5 5 )

CATALYZED REACTION O F BUTENE-:!

Solvent, ml

Benzene (75 1 Benzene (75) Benzene (75) Benzene (7S Benzene ( 7 5 ) n-Heptane (100) %-Heptane (1001 Benzene ( 7 5 ) .Icetic acid (25) .ketonitrile (100)

1

WITH

BESZOYLP E R O X I D E

Benzoyl peroxide, g .

Copper salt, g.

15 3 23 7 23 0 20.8 20 2 23 7 22 3 16 9

Sone CuBr ( 0 161) CuBr ( 0 163) C u C l ( 0 205'1 CU(OAC)?(0 305) CuBr ( 0 155) CuBr 10 214) CuRr (0 163)

Benzoic acid, " % ( g . )

Ca-Benzoateb

97 74 78

21 84 (8 85 (8 74 ( 7 82 ( 7 61 ( 7 (7

1) 4) 61 6) 1) 01 (10 1 )

XX

87 85 89 91

High mol. wt. c ester, %

Z,d

%

89 95 88 92 89 95

598 33 :1,3 "1 29 44

41 11

12 CuBr (0.165) 93 ( 9 . 8 ) 89 97 14 99 CuBr (0.2053 loo(11.7) 85 K cis (56) dcetonitrile (100) 24.3 Cu(OBz)? (1.241) ( CuCl (0.302) SG(13.2) 89 98 10 L cis-trans (35.0)' .icetonitrile (100) 2 4 ' 8 ICu(OBz)*(1.996) 51 cis-trans ( 3 5 . f 1 ) ~ Benzene (75) 24.8 C u C I ( 0 331) 84 (8 3) 77 27 94 Determined by potentiometric titration.; weight of benzoic acid isolated in parentheses (not quantitative). * DeterMaterial balance; mined by G.L.C. marker. < Difference between total saponification value and C4-ester by marker. percentage of peroxide charged accounted for as benzoic acid and benzoate esters; remainder is phenylated butenes, benzene, etc. e Includes %yo2-benzoxy-3,4-dimethylhexane. f Reactant butene is 38.17, trans and 61.67, cis : recovered butene-2 is 40.0Vo trans and 60.05%cis. 0 Reactant butene is 64.67, trans and 35,4!$ c i s ; recorered butene-2 is 67.6(j, trans and 32.4% cis.

J

'34 1

analyzed. The cis isomer was found to be consumed 2.1 times faster than trans-butene-2 in acetonitrile and 1.8 times faster in benzene. The reactions in benzene and n-heptane were not homogeneous in copper salt catalyst. After completion of the reaction most of the copper salt charged was present as insoluble cupric benzoate. The heptane solutions were pale blue and the benzene solutions slightly bluer, qualitatively indicating a higher cupric salt concentration in the latter system. The reactions in acetonitrile and acetic acid-benzene solvents were completely homogeneous. The reaction mixtures varied from dark green to blue depending on the solvent and cupric salt concentration. The results presented in Table I indicate that solutions (acetonitrile and acetic acid-benzene) containing higher cupric ion concentrations approach the ideal stoichiometry, viz., one equivalent of benzoic acid and one equivalent of C4-benzoate for each mole of benzoyl peroxide charged, as given in eq. 1. The accompanying change in the composition of the &benzoate mixture under these conditions is shown in Table 11. The formation of butenyl benzoates also parallels the approach to ideal stoichiometry. The side reactions represented by the cis-trans isomerization of the reactant 2-butenes (with cuprous bromide) and the formations of secbutyl benzoate and higher molecular weight esters decrease concomitantly. In these reactions containing relatively high cupric salt concentrations (acetonitrile and acetic acid-benzene solvents), the butenyl benzoate mixture formed in high yield consisted predominantly of the a-methallyl isomer. Thus, in the presence of added cupric benzoate, amethallyl ester is formed uncontaminated with crotyl benzoate ( butene-2) is also aided slightly b y t h e more favorable chain transfer constant (butene-2 > butene-1; J. Boland, Quart. Revs., 3 , 1 (1949); A. Trotman-Dickinson and E. Steacie, J . Chem. Phyr., 19, 169 (1951)). ( 8 ) (a) M. S. Kharasch, 0. Reinmuth a n d W . Urry, J . A m . Chcm. Soc., 69, 1105 (1947); (b) E. S. Huyser, J. Org. Chcm., 26, 3261

(1961).

(9) G. Hammond and L. Soffer, J . A m . Chcm. Soc.. 72, 4711 (1950).

1575

In the presence of cupric ion the benzoxybutyl radicals are oxidized by cupric ion to form butenyl esters" by an oxidation-reduction reaction described earlier12 (reaction 16 or 17). CsHiCOs.

+ CHaCH=CHCHa

---f

C~H~CO&HCHCHI (15)

I

CHa \.'

+

C~H~COZCHCHCHI Cuxl --+

I

CHI

+

CUI H f (16) I CHt CHFCHCIIZCH~ + C ~ H ~ C O ~ C H ~ H C H(io) ~CH~ C6HaCOzCHCH=CH;

CsHsCOi.

+

I1

+

CsH&O2CH2dHCH~CHs Cuxl -+ CsHsCOzCHiCH=CHCHa H+

+

+ CU+

(17)

The interception of the benzoxybutyl radicals 11 and V must compete with the chain transfer reaction 7 and addition reaction 8. The delicate balance among the rates of all these reactions is further demonstrated by examination of the Cg-benzoate mixture. Although little (< 3%) Cs-benzoates were formed from butene-2 and benzoyl peroxide and copper salt, the more easily polymerizable butene-1 yielded among other higher molecular weight products a mixture containing these Csbenzoates. In addition to the saturated l-benzoxy2-ethylhexane there were formed two isomers which were presumed to be 1-benzoxy-2-ethylhexene-3(VI) and l-benzoxy-2-ethylhexene-4(VII). (10) T h e rate of t h e cuprous catalyzed decomposition of !-butyl peresters is qualitatively much faster than t h e rate of thermal dissociation a t 80'. With di-t-amyl peroxide t h e catalyzed rate is only slightly faster t h a n the uncatalyzec! rate a t 123". Qualitatively, t h e copper salt catalyzed decomposition of benzoyl peroxide is similar t o rli-f-amyl peroxide in t h a t a t 75' it is oilly slightly faster t h a n the uucatalyzed decomposition. These observations will be elaborated further by quantitative studies. If t h e cuprous catalyzed reduction of certain peroxides comperes with t h e purely thermal dissociative process t h a n there will be a uet deficit of cupric salts (especially in t h e presence of reactive hydrogen donors with which reactions 2 and 3 are rapid). I n some reactions it has been observed t h a t little cupric salt is present in t h e final reaction mixture.18 Under these circumstances in which relatively high steady state concentrations of both oxy radicals and cuprous salts persist, t h e reduction of oxy radicals b y cuprous ion may be a relatively important reaction ( I wish t o thank I>r. Cbeves Walling fur this suggestion). Such electron transfer reactions

IiO. $- CU' +RO-

+ CU"

between oxy radicals and ferrous salts have been pustulated previuusly. (11) There are two modes by which each uf t h e benzoxybutyl radicals V and I1 can lose a 8-proton. Loss of a @-proton frum t h e benzoxy-substituted carbon atom would result in e n d e%er products, c.g.

1' I1

-

+ C U *--+ ~ + CU"

OBz

+

+

CHsCH=(!CHr H + CUI BzOCH=CHCH~CH~ II+ CU'

+

+

We could find little evidence for t h e formation of eithe; methyl ethyl ketone or n-butyraldehyde from t h e base-catalyzed hydrolysis of t h e reaction products. Kharasch and Fono' reported spectral evidence for t h e formation of some vinyl esters from octene and benzoyl peroxide. They also isolated t h e dinitrophenylbydrazone of octanal from t h e hydrolysate of t h e octenyl benzoate mixture. T h e selectivity observed in t h e oxidation of these adduct benzoxyalkyl radicals b y cupric salts t o unsaturated esters will be discussed at a later time in another connection. (12) H. De La Mare, J. Kochi and F. Rust, J . A m . Chem. SOC.,8 3 , 2013 (1961).

1576

JAY

K. KOCHI

voi. a4

I n the presence of excess cupric salt

reactions the benzoate does not arise via the butenyl radical intermediate. I If i t is assumed that all butenyl radicals formed in 111 + CU+*+CsH6COzCHzCHCH=CHCHzCHa + acetic acid yield acetates: i t is possible to determine VI the ratio of the rates of benzoxy radical addition and CHzCHa hydrogen abstraction from butene. On this basis I the relative rates of addition ( 5 ) to chain transfer C6HsCOzCHzCHCHzCH=CHCHg ( 18) (19) with butene-1 is calculated to be 3.6 to 1. VI1 With butene-2 i t is 3.8 to 1. Similarly, by considerminimal amounts of these C8-benzoates are formed. ing the difference in yield of a-methallyl benzoate Optimum concentrations of cupric salt, therefore, formed in acetonitrile and in acetic acid-benzene the must be maintained in order to avoid these (7 and relative rates of addition to abstraction with butene8) chain propagation reactions. However, in the 2 is 4.6 to 1. case of benzoyl peroxide excessive copper concen\17ith butene-1 i t is also possible to get another trations will enable the catalytic decomposition of indication of these relative rates from the yields of benzoyl peroxide, which is less efficient in the pro- a-methallyl benzoate obtained in acetonitrile in duction of benzoxy radicals, to compete with the the presence of excess cupric benzoate. Since addithermal process. The latter objection is not ap- tion of benzoxy radical to the butene-1 a t the interplicable to t-butyl perester reactions. nal position is unfavored, it is assumed that most of Further indication that benzoxy radicals react the a-methallyl benzoate arises via the butenyl with butenes primarily by an addition reaction 5 is radical by processes 19 and 20.14 The relative given by the relative reactivity studies on butene-2. rates of addition and abstraction with benzoxy In acetonitrile and benzene solutions cis-butene-2 radicals and butene-1 based on this assumption is was found to be 1.8 to 2.1 times more reactive than 5.3 to 1. Taken all together the relative rates of the trans isomer. These relative reactivity values addition and hydrogen abstraction of butene with are similar to those found by Skell, et aZ.,13 for the benzoxy radical is approximately 4-5 to 1. These addition of trichloromethyl radicals to the butene-2 crude results cannot be used to distinguish between isomers (kcis/ktrans= 2.4). the reactivities of butene-1 and cis- or trans-buteneThe butenyl ester products are not wholly 2. formed via the cupric oxidation of the adduct benThese ratios of rates of addition to hydrogen zoxybutyl radicals I. There is an alternative proc- abstraction in the range of 4-5 to 1, tentatively asess involving butenyl radicals formed from butenes signed for the reaction between benzoxy radical by chain transfer reactions such as 9 and 19. The and the isomeric butenes, are not unusual. Alformation of butenyl radicals is though more sophisticated experiments are required to establish these values with greater cerCsHjCOz. CaHs ---+CsHjCOzH f CaH7. (19) followed by the oxidation-reduction reaction with tainty, there are indications that these ratios of cupric benzoate described earlier in the t-butyl rates vary considerably with structure of the free radical. Thus, Huysersh found the ratios to be 26, perester oxidation.2 34 and 43 for the reactions between the trichloroc4.H~. + CUI1 OzCCsHn +C4HiOuCCsHs f C U I methyl radical and cis-butene-2, trans-butene-2 (20) and octene-1, respectively. These values are in The importance of butenyl ester formation by this marked contrast to the low numbers obtained by path is shown by the reaction of benzoyl peroxide M‘alling and Thaler’j” for the reaction between tand butene in the presence of acetic acid and ace- butoxy radicals and cis-butene-2 (0.20), transtonitrile under similar conditions. The butenyl butene-2 (0.036) and butene-1 (0.034) .15b The efficacy of cupric ion in oxidizing the benzacetates isolated in the former solvent had the same distribution of a-methallyl (87-947,) and oxybutyl radical to butenyl benzoate products can crotyl (G-13y0) isomers regardless of the reactant also be qualitatively interpreted as reducing the butene. This is in strong contradistinction to the effective steady state concentration of these alkyl butenyl benzoate mixture whose isomer distribu- radicals.I6 Another demonstration of these oxidation is highly dependent on the reactant olefin. tion-reduction reactions between cupric ion and This ratio of a-methallyl acetate to crotyl acetate alkyl radicals is given by the cis-trans isomerization is also the same as that obtained earlier from the t- of the reactant butene. In those reactions with butyl perester reactions. Moreover, it was shown copper bromide catalyst, the isomerization of that t-butyl perbenzoate reacted with butenes in butene-2 is due to bromine atoms, since neither the acetic acid to yield predominantly butenyl acetates uncatalyzed, cuprous chloride or cupric acetate with minor amounts of benzoates.2 This was inter- catalyzed reactions induced the isomerization. 111 preted in terms of the rapid metathesis of cupric ben- heptane and benzene solutions extensive cis-trans zoate in acetic acid solvent. The high yields of bu- isomerization of butene-2 occurred. These were tenyl benzoates obtained from benzoyl peroxide and also conditions under which the least amount of butenyl esters and the greatest amount of saturated XCU” OzCCsHs 4- H02CCH3 CHzCHj

+

XCU” OzCCH3

+ C6HjCOOH

(21)

butene in acetic acid solvent indicates that in these (13) P. Skell and I< W o o d w u r t h , J . A m . Che?n .SOL.. 77, 4638 (1955); see, h