3946
JAY
K. KOCHIAND FREDERICK F. RUST
Anal. Calcd. for C I ~ H , ~ OC,~ :71.43; H , 9.57. Found: C, 71.50; H, 9.55. Reaction of I and (11-Styrene Oxide.--A mixture of 12.0 g. (0.0345 mole) of I and 24.0 g. (0.20 mole) of styrene oxide was heated a t 170' (bath) for 24 hours. The cooled reaction mixture was triturated with hexane. The solid residue was washed with hexane and dried in vacuo to give S.5 g. (S9%) of triphenylphosphine oxide, m.p. 155-157' (lit.'* 153'). The hexane solution was concentrated and then fractionated to give 20.3 g. of recovered styrene oxide, b.p. 73" (12 mm.), atid 2.0 g. (30%) of ethyl trans-2phenl-lcyclopropanecarboxylate, b.p. 102' (0.15 mm.), (lit.'3 105-110", 2 m m . ) . In another experiment, 12.0 g. (0.03 mole) of I and 12.2 g . (0.10 mole) of styrene oxide were heated a t 190-200 (bath) for 6 hours. Following the procedure above, 90% of triphenylphosphine oxide, m.p. 156-158', was obtained. The infrared spectrum iyas identical to that of an authentic sample. Distillation afforded, after removal of hexane and styrene oxide, 1.2 g. (21y0) of ethyl trans-2-phenylcyclopropanecarboxylate, b.p. 103-105" (0.5 mm.). The ester from another run was converted to the acid by heating 1.90 g. (0.01 mole) with 15 ml. of 6 Npotassium (12) G. M. Kosolapoff, "Organophosphorus Compounds," J. U'iley and Sons, Inc., Xew York, X, Y.,1960. p. 114. ( 1 3 ) A . Burger a n d 1%'. I.. Yost, J . A m . Chein. .Soc., 70, 2198 11948).
Vol. 84
hydroxide and 6 ml. of methanol for 6 hours. Acidification gave the acid which was recrystallized twice from aqueous ethanol containing a little hydrochloric acid; 111.p.90 4 1 O (lit.l3 92-93'). The amide was prepared by boiling 1.0 g. of the acid in 5 ml. of thionyl chloride for 20 minutes. The mixture was poured into 15 ml. of ice-cold concentrated ammonium hydroxide. The amide was recrystallized from aqueous methanol; m.p. 190.5-192' (lit.'3 190-191'). Reaction of I and l-Styrene Oxide.-A mixture of 10.0 g. (0.029 mole) of I and 10.0 g. (0,084 mole) of I-styrene oxide,14 [ c z ] ~ ~ D-12.45', c 20 in chloroform, was heated a t 200" (bath) for 7 hours. The cooled mixture was dissolved in 30 ml. of methanol and 15 ml. of 6 .V potassium hydroxide was added. T h e mixture was boiled for 12 hours. E x traction with chloroform followed by acidification gave 10 g. of an amber semi-solid which was taken up in ether and esterified by adding a solution of diazomethane in ether. The resulting ester was molecularly distilled, b.p. 90" (block) (25 mm.), to give 1.5 g. (2SY0) of methyl trans-2phenylcyclopropanecarboxylate, [CY]*OD -31.3", c 13 in chloroform. The ester was saponified and the acid was converted to the amide, m.p. 192-193" (lit.'3 190-191" for the &amide), [cyIz3D +S.l5", c 2 in 95% ethanol. (14) Prepared according t o t h e procedure of 13. I.. TSliel am1 1 ) . XV. Delmonte, .I. 0 i . f . Chrm., 21, 596 (1956).
[CONTRIBUTIOS FROM EMERYVILLE RESEARCH CESTER,SHELL DEVELOPMENT Co., EMERYVILLE, CALIF.]
Oxidation of Free Radicals from Unsaturated Compounds by Cupric Salts BY JAY K. KOCHI'AND FREDERICK F. RUST RECEIVED APRIL23, 1962 Cupric salts react with alkyl radicals by two oxidation-reduction processes which have been described as electron transfer and ligand transfer. We have studied the oxidation of substituted alkyl radicals by examining the competitive rates of oxidation of 5-(methoxpcarbony1)-pentyl and the radicals resulting from its addition t o a variety of monomers. By regulating the concentration of monomer and cupric salt, it is possible t o oxidize selectively allylic radicals in the presence of the primary hexanoate ester radicals. 5-( Methoxycarbony1)-pentyl reacts with butadiene in methanol under these conditions t o produce methyl methoxydecenoates. Isoprene, chloroprene, styrene and acrylonitrile are also effective radical acceptors. T h e mechanism of the oxidation of substituted alkyl radicals by cupric salts is discussed,
Introduction The oxidation of relatively simple free alkyl radicals by cupric salts was previously described2 as involving either a ligand transfer process or an electron transfer process depending on the particular cupric salt employed. Part of the problem encountered in describing these reactions in a more quantitative manner is the difficulty in obtaining meaningful rate data. Competitive experiments based on product isolation a t present forms the basis of obtaining some semi-quantitative kinetic information. In this paper we wish to present studies on the competitive oxidation of substituted carbon free radicals by examining the products from the oxidation by cupric salts of 5-(methoxycarbonyl)-pentyl (I) in the presence of unsaturated compounds. The competitive reactions are (a) the oxidation of the 5-(methoxycarbony1)-pentyl, (b) the addition to an unsaturated compound and (c) the oxidation of the resulting adduct radical. 5-(Methoxycarbonyl) -pentyl was chosen because of the relative ease of isolation of its oxidation products. (1) Department of Chemistry, Cese I n s t i t u t e of Technology, Cleveland 6, Ohio. (2) H. E . De La Mare, J. K. Kochi and F. F. Rust, J . A m C h e m . SOC.,83,2013 (1961); t o be published; J. K. Kochi, i b i d . , 84,774.1572 1192, 2121, 2785 (1962); Tetrnhedum, 18, 483 (19G2).
CH30nC(CH?);.
+ CU'~X,
I
oxidation products
+ CurXV),
b
+
C H ~ O ~ C ( C H ~ )CH~=CHX ~. +C H ~ O ? C ( C H ~ ) & H X
+
C H ~ O ~ C ( C H ? ) ~ C H XC ~ I I X C , _ oxidation products
+ Cu'X,
Results Cyclohexanone and hydrogen peroxide react to form a complex mixture of adducts3 In methanol this mixture reacts with ferrous sulfate to produce the radical 5-(methoxycarbonyl) -pentyl.
oo
+ H202-
CH30H
CH,02C(CH,),.
cH3?52H Fe i-*
+
Fe(OH)+'
This radical will react with a variety of uns a t u r a t e ~ . In ~ the case of butadiene] for example, 5-(methoxycarbonyl)-pentyl adds to produce an (3) A I . S. Kharasch and G. Sosnovsky, J . Org. Choii., 23, 1322 (1558). (4) hf. S. Kharasch and W. Nudenberg, i b i d . , 19, 1921 (1954). D.D. Coffrnan and H. 9.Crlpps, l?.S. P a t e n t 2,811,5>1 (til dii P m t ) , Oct. 29, 1957.
CUPRICSALTOXIDATIONOF FREERADICALS
Oct. 20, 1962
3947
allylic radical, 11,which dimerizes to form a mixture products are the Cno-dibasicacid esters 111. At of eicosadienedioic acid esters, 111, in 6E1-757~ low cupric ion and high butadiene concentrations yield.4 Our earlier studies? showed that, i n the the product is still predominantly the C2"-diester. At high cupric ion and low butadiene concentraCH302C( CH?)5. + CIHs + tions the important products are the CB-estersVI I CH$OzC(CH )GCH-CH-CHZ and VII. The yields of Clo-esters IV and V can be I1 optimized a t critical cupric ion Concentrations. 2CH302C(CH2)6CH-CH-CH + The results are summarized in Table I . In these 8 isomeric dimethyl eicos~dienedio:ite.; reactions a mixture of cyclohexanone and 3070 I11 hydrogen peroxide is stirred for ten minutes and presence of C U + ~saturated , primary radicals are then allowed to react with a chilled methanolic solution of sulfuric acid. The cupric salt (sulfate) RCH:?CHz. + C U +--f ~ RCH=CH? + CU' f H i (1) and butadiene are added quickly with subsequent oxidized to terminal olefins. Likewise in the presence of C U + ~allylic , radicals derived from t-butoxy dropwise addition of ferrous sulfate. and butadiene (t-C4HeOCH2CH-CH-CH?) were TABLE I oxidized to allylic carbonium ion intermediates (see YIELDOF CIO-ESTERS A N D CZ~-DIESTERS AS A FUNCTION OF ref. 13), which in the presence of a protic solvent C U + AND ~ BUTADIENE CONCENTRATION' led to the alkoxylated p r o d ~ c t . * ~ ~ ~ CuSOa, Butadiene, Cm-Diester;
+ +
ROCHpCH-CH-CHz CU+' + [ROCHzCH-CH-CH2] + CU' [ R O C H P C H C H C H ~ ] ROH ---+ ROCHPCH=CHCH~OR ROCH2CHCH=CHz +
+
+
1
g.
(2)
+ H+
OR
Butadiene.--We have now intercepted with cupric ion the allylic radical I1 produced from the addition of 5-(methoxycarbony1)-pentylto butadiene-and other unsaturates-and related the findings to a number of reaction variables in order to understand better the mechanisms involved.
+ +
CU-' ---f [ CH~OYC CHZ)~CHZCH-CHI] ( [CH~O~C(CHZ)GCH-CH-CH~] + Jr Cui (3) [ CHBOZC( CH~)GCH-CH-CH*] + CH30H + CH302C( C H Z ) ~ C H = C H C H P O C H ~ (4) 11+CHsO?C(CHe)sCHCH=CHz ( 5 ) 1
OCHs
I, In addition, there are obtained products from the reaction of C u i y with the 5-(methoxycarbony1)pentyl.
+ CU-' +
CHsO$C(CH?)j.
CH~OZC(CHZ)~CH=CH (6) ~ VI CHpOH
+CHaOzC(CH2)4CHz -+ +
CH30aC( CHz)jOCHs (7) 1711
The distribution of products 111, IV, V, VI and VI1 is highly dependent on the cupric ion concentration and to a lesser degree on the butadiene concentration. If the hexanoic ester radical I is generated in the absence of butadiene, termination occurs a t this stage. Thus, i t reacts with cupric sulfate in methanol solution to produce methyl 5hexenoate and methyl w-methoxyhexanoate in 47% and 27y0yields, respectively. With cupric chloride under the same conditions methyl w-chlorohexanoate is formed in 69% yield and neither methyl 5-hexenoate nor methyl w-methoxyhexanoate is formed. I n those reactions containing butadiene, extreme conditions of cupric ion and butadiene concentrations can be considered. I n the absence of cupric ion, the predominant
moles
%
70
10 65 2.0 23 43 1.5 0.4 51 11 61 6 1.5 1.0 1.5 2.0 41 20 1.5 3.0 18 29 3.0 1.0 47 15 6.0 0.4 32 5 6.0 1.0 37 16 1.0 31 14 12.5 a In reactions containing 0.20 mole of cyclohexanone, 11.5 g. (0.10 mole) of (30y0)hydrogen peroxide, 20 ml. of concd. sulfuric acid and 41.7 g. of ferrous sulfate in 750 ml. of methanol. * Mixture of allylic isomers II',I', Mixture of Cno-diesters 111. ?;one 0.5
1. 0
The best yield of Clo-esters obtained in our runs was Gl7( (based on hydrogen peroxide charged). The results in Table I show that a t optimum cupric ion concentration it is possible to intercept the (&ester radical '11 quite selectively. The formations of the Clo-adducts IV and V depend on the relative concentrations of the C6-radical 1, the CIoallylic radical I1 and cupric ion. Saturated alkyl radicals such as I, in general, show enhanced reactivity relative t o allylic radicals such as I1 in free radical reactions involving hydrogen transfer, addition, disproportionation, etc. It could, then, be expected that the steady state5concentration of radical I could be lower than radical 11, especially in systems containing an excess of butadiene. By keeping the cupric ion concentration sufficiently low to minimize reactions such as G and 7 with radical I, yet high enough to react with allylic radicals, it is possible to effect oxidation primarily a t the allylic radical I1 stage. The selectivity arises, we believe, not primarily due to the difference in the reactivity of the two types of radicals toward cupric ion but in a less direct manner by the difference in their "steady state" concentrations. If one assumes that the electron transfer reactions represented by G and 7 and 3 are equally fast, then i t is possible to account for the selectivity that has been observed in the cupric ion oxidations on the basis of the relative rates of the reactions 6, 7 and 3 being dependent mainly on the concentrations of I and 11, respectively. ( 5 ) T h e term "steady state" is employed loosely here.
3948
JAY
K. KOCHIAXD FREDERICK F. RUST
VOl. 84
Earlier studied on the reactions of free radicals with cupric ion indicated that in the presence of chloride ion an additional oxidation-reduction reaction involving ligand transfer occurred; e.g.
allylic product could be isolated. Due to limited solubility in acetic acid the corresponding metal sulfates could not be used. Approximately 6070% of the cyclohexanone is recovered. The reaction in t-butyl alcohol is similarly not encouragKCHzCHz. CUCI:!---+ KCH2CHzCI CUCl ( 8 ) ing. Approximately 30-4Oy0 yields of carboxylic I n order to prepare the chloro analogs of the Clo- acids are produced. These acids decompose readily esters I V and V such as LrIII, 5-(methoxycarbonyl)- on attempted distillation and presumably coritaiti pentyl was added to butadiene in the presence of unreacted peroxides. They were not characterized ferrous and cupric chloride salts. The product is further. not the expected allylic chloride VI11 but consists Isoprene.--3-(Methoxycarbonyl)-pentyl (I) adds [ CH,O2C-( CHz)sCH-CH-CH2] $- CUC12 + to isoprene in a manner similar to butadiene. The CR302C(CH,)BCH=CH-CH?Cl (9) mode of addition appears to be predominantly VI11 1,2 to produce the allylic ether I X a (33%) and of the same isomeric mixture of allylic ethers IV minor amounts of the isomer I X b (7yo). In addiand V found in the cupric sulfate oxidations; they c u +2 + are also formed in the same yields. When the re- CHaOzC( CH9)b. + CbHa CHBOH action is repeated using concentrated hydrochloric CH3 acid instead of sulfuric acid in order to maximize I the chloride ion concentration, the yields of Clo- cIC(CH2)bCHCHXH.
+
1
c H 3 0 2 c (CH2)&CH,cHOCI11 XXI
tration is increased and the styrene concentration decreased, the yields of X X and X X I are decreased. Haloolefins-Simple vinylic haloolefins are known to undergo free radical addition reaction^.^ CHaOpC( CH2)6. CH-CH-CN + Vinyl bromide reacts with 6-(methoxycarbony1)cs Cn. pentyl readily but similar to the adduct radical XV I cuc12 I CH~OIC(CH~)~CH. CH~O~C(CH~)G-CHCIfrom acrylonitrile ; extensive polymerization of the vinyl bromide occurs when only cupric sulfate xv XVI chloride (0.06 mole), ferrous chloride (0.23 mole) is present. The indications are that the bromoalkyl Br and ferrous sulfate (0.25 mole) a 46y0 yield of I adduct is formed. The adduct XVII contains CHa02C(CH2)s. + CH2=CHBr --+ CH102C(CHz)&H. predominantly two acrylonitrile moieties ( n = 2) XSII associated with each hexanoate radical. radical X X I I is not easily oxidized to the corcs respondiiig carbonium ion by cupric ion. I When vinylidene chloride is allowed to react iii CI-Is02C(CH~)6CII.+ =L!IlCN -4 the presence of the Ce-ester radical and cupric chloride it shows no indication of undergoing an C11a02C(C H i ) s CHzCH nC1 addition reaction. The sole isolable product is XI71 methyl a-chlorohexanoate (25% yield). The adduct radical XV thus exhibits marked Methallyl chloride reacts readily with the 5differences in selectivity to oxidation by cupric (methoxycarbony1)-pentyl and cupric sulfate to species. Oxidation by cupric ion to the carbonium form adducts in approximately 55% yield. The ion intermediate XVIII appears to be an un- products boiling over a range were not completely favorable process since extensive homopolymeri- identified, but elemental analysis and their inzation of the monomer occurs. Moreover, the frared spectra suggest that the products do not include the ether product XXIV obtained by the CN oxidation of the chloroallyl radical X X I I I to the I CHsOzC(CH2)6CH. f CU'? + corresponding carbonium ion. The product most CK compatible with the analysis is the unsaturated
+
( ';">
CII,o,C(C'II*),JIIa XVIII
+ Cu
+
( 7 ) C . Wdling, "ITrcc Ilatlicals Inc., Kcw York, N. Y.. 1957.
iii
Sctliition," J r h n Wilry and Sons,
3950
JAY
K. KOCHIASD FREDEKICK F. RUST
Vol. 84
(1) copper ion concentration and ( 2 ) olefin concentration. A t high cupric ion concentrations the reaction is terminated at the initial free radical stage [CH302C(CHJ6. 1. The products are methyl 5-hexenoate and methyl w-methoxyhexanoate when cupric sulfate is used and methyl w-chlorohexanoate when cupric chloride is present. At high olefin concentrations the telomers of the general structure
CH3
CH,OpC( CH2)6CC&Cl );XI\-
I
OCH3
chloroester XXV. The vinyl chloride structure XXV is preferred to the alternative isomers XXVI CH3
XXT'
in which the double bond is allylic to the chlorine CHz 'I CHJO~C( CH2)6CCHzCl XX\%
CI13
I
CHaOC>(CHr)aCH=CCH C1 XXVIb
because conditions under which the ester is saponified do not liberate chloride ions. There is no indication that the unsaturated ester XXVII is formed which would result from the expulsion of the chlorine atom from the chloroallyl radical X X I I I by a process similar to the free radical dimerization of methallyl chloride.8
CH 3
CH:iOZC(CH,),-C=CH, XXVII
+ C,'1.
Discussion The addition of 3-(iiiethoxycarboiipl)-j)entyl 1.0 olefins of various structural types has been examined. The general reactivity sequence of olefins to addition by this radical is in the decreasing order CH i c l ~ 2 - C l l C ~ I - c l l ~ ,cII,-CCII--
c112
>
C1
I C1I2--ccI1-CII,
,
i c,;ll!,cll-cll~ i e112 ~ - C l l c T i CI 1.3
C I ~ ~ - - . ~ C Ii I ~cirl--C I C I I >> ~ c i i 2 -CCI, -
'l'liis order is based
011 thc total yields of adduct 1)rotlucts obtained by adding the 3-(inethoxycarbony-1)-pentyl to the respective olefin under standard conditions of cupric ion and olefin concentration. This method is applicable since there is an independent alternative route by which the 5(methoxycarbony1)-pentyl can also react, viz., to form the variety of Cs-ester products mentioned earlier. There are two variables which must be considered in determining the products of the reaction;
(SI Relerciice 7 ,
11.
269.
XXVIII
XXIX
XXVIII are formed from olefins X X I X in addition to diesters of type 111. It is apparent from the pattern of reactions presented earlier that there are a t least two main considerations to be made with reference to the receptivity of the olefin to 5- (methoxycarbony1)pentyl addition: (a) the capacity for free radical addition and (b) the oxidation by cupric ion of the resulting adduct radical. With respect to the former, olefins show the same behavior to the 5-(methoxycarbony1)-pentyl as they do to other free radicals. Thus, conjugated dienes are the most reactive followed by olefins conjugated to unsaturated groups such as phenyl, carbonyl and cyano, and alkyl substituted olefins and followed lastly by halogenated vinyl compounds. In this respect it should be mentioned that 5-(methoxycarbonyl)-pentyl, like alkyl radicals in general, exhibits electrophilic propertiesg in that it adds relatively slowly to olefins possessing electronwithdrawing groups. The ease of oxidation of the adduct radical is also a function of the substituents on the olefins. In olefin XXIX if I ' is an electron-withdrawing group such as cyano, carbonyl or halogen, oxidation by cupric sulfate involving an electron transfer reaction does not occur. If the olefin is prone to homopolymerization, extensive polymerization will occur. This seems reasonable in view of the general observation that positions alpha to these substituents do not tolerate a positive charge"' (e.g., slow solvolysis of a-haloketones, nitriles, etc.). However, if the adduct radical can undergo an oxidation without generation of a positive charge as, for example, in the chloride ligand transfer reaction'l (ey. 8) alluded to earlier,': the reaction proceeds smoothly to produce the corresponding chloride, without polyineriziiig the olefin. These 1)Iieiioiiienahave been observed with acryloiiitrile, xrolein*3and vinyl broinitle. I11 the cases where Y is ai1 d k y l , vinyl, aryl or a siiiiilar moiety cup;tble oi stabilizing ail adjacent positive charge, oxidatioll by cupric suliateI3as wcll as cupric chloride, occurs quite readily. I n c , S e w Ymk, S .T. (11) Even in ligand transfer reactiuns t h e r e is a mudicum of positive charge generated in t h e free radical moiety l h i s is horne o u t hy t h e f a c t t h a t uxy a n d thi3-I radicals a r e n u t oxidized by cupric chluride (sre refs. l , l : 3 ) . [ 121 J K . Kochi. iinpublished observatic,ns. (1:i) I n t h e 1)reviou': +ectii,ni t h e oxidatii,n , inviil\ i i i x t h c f o r n ~ j ~ t i c of , i i carhimiurn i o n i interme~lixtcs. l'hc ~ircienlatiuii w:i> niaCftJre u c ; Acrylonitrile, E:tstiiian Krlcktk reagent grade, tlistilletl k f o r e use; Vinyl bromide, The h1:ttlteson Co., 99 tlistillctl from cylinder; Styrene, Eastman K o redistilled before use; Vinylidene chloride, Uow Chemical e o . , redistilled before use; Methallyl chloride, Shell Chemical Co., redistilled before use. 5-(Methoxycarbonyl)-pentyl with Cupric Sulfate.-Cyelohexanone (19.6 9.) and 30'2 hydrogen peroxide (11.5 implication. For a discussion of substitution reactions which accomp a n y electron transfer processes, see ref. 1. (14) Although t h e difference in reactivity is a t t r i h u t e d mostly t o t h e oxidation potential
