Nov. 5, 1950
5510
REllCTION O F PERESTERS WITH OLEFINS
and cyclopentanecarboxylic acids from ethylene and tetramethylene bromides, respectively. The hydrolysis of the ester resulting from ethylene bromide was effected by means of an aqueous solution of potassium hydroxide in order to avoid the cleavage of the cyclopropane ring which might have been formed, while a mixture of dioxane and concentrated hydrochloric acid was used for hydrolyzing the ester from tetramethylene bromide. Although suberic acid was obtained with an excellent yield in the latter reaction, the ethylene bromide failed to afford any acidic product other than acetic acid by means of this amide. It seems safe to conclude from these results that the direct alkylation of t-butyl acetate is of value as an alternate of malonic ester synthesis for preparing mono- and dibasic carboxylic acids. Experimental6 Alkylation.-To a suspension of 0.1 mole of lithium amide in 200 ml. of liquid ammonia 0.1 mole of t-butyl ester was Since the added rapidly with continuous stirring at -40". use of anhydrous ether as a solvent of the ester lowered the yields of alkylated products, no solvent other than ammonia was used at this stage of the raction. After 15-40 minutes stirring at -40" 0.05 mole of organic halide (neat) or its solution in 20 ml. of anhydrous ether was added during 3-30 minutes. T h e stirring was continued for a n additional 40(8) All temperatures are uncorrected. Ogawa.
Analyses by Miss Kenko
180 minutes and then ammonium chloride \+asadded. Ammonia was evaporated by external heating, while 50 ml. of wet ether was added slomlp. The reaction mixture was treated with 50 ml. of water, extracted with ether, the combined ethereal solutions were washed with water and dried over anhydrous sodium sulfate. Upon removing the solvent, followed by fractional distillation of the residue in the presence of magnesium oxide, the alkylated product could be isolated. The details of each experiment were summarized in Tables I and 11. li'hen the isolation of the t-butyl ester was accompanied with difficulties, the evaporation residue of the ethereal extracts was directly subjected to hydrolysis. Hydrolysis of the &Butyl Ester.-A mixture of 0.02 mole of the alkylation products, 6 ml. of dioxane and 10 ml. of concentrated hydrochloric acid was refluxed for 2-5 hours. ilfter cooling t o room temperature the mixture was extracted several times with ether, the combined ethereal solutions were washed with a small amount of water and then thoroughly extracted with 10% aqueous sodium hydroxide solution or saturated sodium bicarbonate solution. T h e combined alkaline solutions were washed with ether and acidified with concentrated hydrochloric acid. The carboxylic acid which separated was taken up in ether or collected through filtration and purified through distillation or by recrystallizations. Yields were almost quantitative in each case. The condensation product obtained from ethylene bromide in t h e presence of sodium amide was hydrolyzed by heating with 507, excess of 50y0aqueous potassium hydroxide solution. The reaction mixture was washed with ether and acidified with concentrated hydrochloric acid. T h e solution was extracted with ether, b u t failed to afford any acidic product. Though the non-saponifiable portion x a s examined t o recover the condensation product, if any, no appreciable amount of neutral ester could be isolated. KY~TO JAPAN ,
[CONTRIBUTIOS FROM THE INSTITUTE O F ORGANIC CHEMISTRY,
Reactions of &Butyl Peresters. BY &I.
THEUNIVERSITY O F
CHICAGO]
I. The Reaction of Peresters with Olefins]
s. KHARASCH,~ GEORGE sOSNOVSKY3 AND N. c. YANG3 RECEIVED MAY23, 1959
&Butyl peresters react smoothly with olefins in t h e presence of copper or cobalt salt catalysts to form allylic esters in moderate t o good yields. With terminal olefins, such as 1-octene, 1-hexene or allylbenzene, t h e only products isolated are the allylic esters with a terminal double bond. When the reaction of cyclohexene, t-butyl perbenzoate and cuprous bromide is carried out in an aliphatic acid as the solvent, the aliphatic ester of cyclohexenol instead of t h e benzoate is obtained. T h e scope and limitation of this reaction are discussed.
Tertiary alkyl peresters were first prepared by Milas and surge no^.^ However, the chemistry of these substances, except for several kinetic studies, has received little attention. Blomquist5 and his associates studied the decomposition of t-butyl perbenzoate in great detail. They found that t-butyl perbenzoate decomposes about as rapidly a t l l j 0 as benzoyl peroxide does a t 80". In most aromatic solvents the rate of decomposition is of first order; in aliphatic solvents the situation is more complex, and i t appears that an induced chain reaction (involving free radicals derived from solvent) takes place in addition to the unimolecular cleavage of the perester. More recently, Bartlett and Hiatt6 examined the decomposition of various (1) This investigation was supported by a grant from t h e Office of Naval Research, Contract No. NGori-02040. It was presented in p a r t a t t h e 134th Meeting of t h e American Chemical Society in Chicago, Ill., September, 1958. (2) Deceased. (3) T o whom inquiries of this publication may be sent. (4) N. A. Milas and D. M. Surgenor, THISJ O U R N A L , 68, 642 (1948). (5) A. T. Blomquist, A. F. Ferris and I. A. Berstein, ibid.,73,3408, 3421, 5546 (1951). ( 0 ) P. D. Rartlett and R.R . H i a t t , i b i d . , 80, 1398 (1958).
t-butyl peresters in chlorobenzene. They found that most of the peresters examined give a mixture of products with carbon dioxide as the major constituent. In view of these findings, i t is not surprising that in a variety of substrates the decomposition of t-butyl perbenzoate gives a complex mixture of products; and that few useful preparative results have been obtained from such reactions. In a previous communication7 it was shown that the addition of catalytic amounts of copper or cobalt salts to the reactions of perester and olefins has a decisive effect on the composition of the endproducts. Thus, t-butyl perbenzoate reacts with ethylenic compounds to give esters of benzoic acid. These reactions are stereospecific, namely, the benzoyloxy group enters the position alpha to the double bond, and there is no rearrangement of the ethylenic system. The present paper contains experimental details which amplify the original observations. The possibility of utilizing t-butyl peresters as synthetic reagents was first noted when t-butyl (7) bl. S. Kharasch and G. Sosnovsky, ibid., 80, 750 (1958).
5S90
11
S.KHARASCH, GEORGESOSXOVSKY AND N. C. I-AXG
perbenzoate decomposed in the presence of cuprous bromide to give methyl benzoate. When this decomposition is carried out in solvents containing activated hydrogen atoms, benzoic esters are obtained. Of particular theoretical and practical interest are the reactions of t-butyl perbenzoate with allylic systems. 9lthough radical reactions with allylic compounds usually result in a mixture of isomeric products,8 the reaction of t-butyl perbenzoate and allylic compounds in the presence of cuprous bromide forms mainly one product. Thus, 1-hexene, 1-octene, allylbenzene, "octene, &vinylcyclohexene, 2,4,4-trimethyl-l-pentene, cyclohexene, cyclohexenyl acetate and a-pinene react smoothly to give monobenzoates. In the case of acrylonitrile, a very reactive olefin, high molecular weight substances were the main product. The position of the benzoyloxy group has been established in the esters derived from 1hexene (Ia), 1-octene (Ib), cyclohexene (IIa) and allylbenzene (IC),but there is still some uncertainty about the position of the benzoyloxy group in the esters derived from 2-octene (111), 2,4,4-trimethyl1-pentene (IV), cyclohexen-1-yl acetate (V), 4vinylcyclohexene (VIa) and a-pinene (VIIa). R, H C H= C H, OCOK,
CH=CH,
OCOCH.
A3 IV
v
T.01. 51
The main limitation of this reaction is the teinperature a t which it may be performed. The most suitable temperatures lie between SOo and 123'. The reactions proceed slowly a t 60-io"; highly volatile substances such as 1-pentene (b.p. 40") do not react. However, the reaction with such volatile substances may be feasible a t a higher temperature under elevated pressure. Substances of very low volatility (e.g., benzyl benzoate, b.p. 353") yield very high boiling esters which are difficult to separate from the polymeric by-products. In order to obtain the best results, the reaction between a perester and a compound containing a labile hydrogen atom should be carried out in the presence of an excess of the latter. \\'hen equimolar quantities of the reagents are employed, lower yields are obtained. Xnother limitation of the reaction is the availability of the type of t-butyl peresters: only t-butyl perbenzoate and t-butyl peracetate are conimercially available. This limitation may be partly overcome by using the desired acid as the solvent for the reaction. n'hen cyclohexene dissolved in glacial acetic acid is treated with t-butyl perbenzoate in the presence of cuprous bromide, cyclohexenyl acetate (IIb) is formed instead of the corresponding benzoate (IIa). This niodification has been applied successfully for the preparation of a number of cyclohexenyl esters (11) such as cyclohexenyl formate, acetate, propionate and n-butyrate. The yield of cyclohexenyl formate (IIe) is comparatively lower (30%) than the other esters, and there appears to be an extensive reaction between IIc or formic acid and the perester under the reaction condition.
VIa. R = CeHs b , R = CH,
IIb, R = CH3; d, R = C2H5 c , R = H; e, R = n-C3H7
VIIa, R = C5Hj b. R = CH?
The nature of the reaction described is independent of the solubility or valence of the copper salt Likewise, t-butyl peracetate reacts with 1-hexene, catalyst introduced into the reaction mixture. 1-octene, cyclohexene, pinene and vinylcyclohexene Cuprous bromide and cupric 2-ethylhexoate are to give the esters of acetic acid, Id, Ie, IIb, VIIb and equally effective. When cuprous bromide is used, i t enters the solution slowly, and the catalysis is VIb, respectively. The general applicability of this synthetic method homogeneous. Cobaltous 2-ethylhexoate is someis demonstrated by the reactions of t-butyl per- what less effective giving a lower yield of the allylic benzoate with compounds containing benzylic ester, but zinc chloride and magnesium bromide are hydrogen atoms. With cumene, i t forms 2- ineffective catalysts. phenyl-2-benzoyloxypropane (VIII) , and with 2Discussion phenylbutane, 2 - phenyl - 2 - benzoyloxybutane (IX).lo Triphenylmethane, however, failed to A. Specificity.-Terminal olefins react with react under similar conditions. t-butyl peresters in the presence of a copper salt catalyst yielding only the unrearranged secondary allylic esters. The specificity of this reaction was first demonstrated by rigorous purification of the CH3(iCH*CHa CH3YCH3 reaction product. SVhen hex-1-en-3-yl benzoate OCOCeHs OCOCGHs (Ia), isolated by simple distillation of the reaction RCO,
3
~7111
6
IX
u + CsHhC000-t-Bu -c + RCHCHECHL +
(8) (a) A. L. Bateman and J. I. Cuneen, J. Chem. Soc., 941 (19501 (b) hI. S. Kharasch, P. Pauson and W. Kudenberg, J . Org. Clzem., 18, 322 (1953); (c) M. S. Kharasch. R. hlalec and K. C . Yang, ibid., 22, 1443 (193.7); (d) I f . S. K h a r a s c h a n d A. Fono, ibid.. 24, 7 2 (1959). (9) hl. S.Kharasch and A. Fono, i b i d . , 23, 321 (1858). (10) P. A . Hallgarten, unpublished result. ~
RCHICHZCH~
I
OCOCsHs
mixture, was carefully fractionated through a Piros-Glover spinning band column rated a t SO
Nov. 5, 1959
REACTION OF
theoretical plates with a 40:l reflux ratio,ll a single fraction (87y0of the total charge) was obtained which has an infrared spectrum identical with that of the original distilland. When oct-1-en-3-yl acetate (Ie), isolated in a similar manner, was analyzed by gas chromatography, there was detected only one component. However, small amounts of the rearranged ester may be formed in the reaction, but so far they have escaped detection. The specificity was further investigated with an olefinic system which is readily susceptible to rearrangement. Allylbenzene ordinarily undergoes free radical reactions accompanied by rearrangement; for example, the bromination of allylbenzene with N-bromosuccinimide gives only the rearranged cinnamyl bromide (X).12 If cinnamyl benzoate (Xb) is formed in the reaction of allylbenzene and t-butyl perbenzoate in the presence of cuprous bromide, it may be easily detected by its characteristic ultraviolet absorption spectrum. Consequently, the ultraviolet absorption spectrum of the product was compared with the spectrum of authentic cinnamyl benzoatel3 and that of l-phenylprop3-en-1-yl benzoate (IC).13 The spectrum was found to be virtually identical with that of IC. The absence of a strong absorption in 250 mp region indicates that the product is contaminated with less than 1% of the rearranged cinnamyl benzoate. The infrared spectra of the product and IC are superimposable. CeHsCH=CHCHzX Xa, X = Br 0
II
b, X = O C C ~ H S
B. Interpretation of Infrared Spectrum.-Compounds with trans-vinylene double bonds, HC=CH, absorb in the 960-970 cm.-l (strong) region, and compounds with vinyl double bonds, -CH=CHa, absorb a t 985-995 (medium) and 905915 cm.-' (very strong).I4 The present work shows that when an acetoxy or a benzoyloxy group is introduced into a position alpha to the vinyl group the 905-915 cm.-l band is shifted to 925935 cm.-l region,15 and compounds such as hex1-en-3-yl benzoate, hex-1-en-3-yl acetate, oct-len-3-yl acetate and 1-phenylprop-3-en-1-yl benzoate exhibit additional absorption bands in 955980 cm.-l region. At first, this additional band was attributed to the presence of the vinylene isomer; but experimental investigation failed to detect any vinylene isomer. Therefore, this anomalous band is probably an intrinsic property of the esters and the interpretation of infrared spectra of allylic esters is to be done with care. C. Possible Reaction Mechanisms.-In the absence of catalysts, the decomposition of t-butyl peresters in a variety of solvents proceeds by multiple cleavages to give a complex mixture of prod(11) Using t h e same column and similar operating conditions, variallylic isomers have been separated in this Laboratory; see ref. 8c and 8d. (12) E. A. Braude and E. S. Waight, J. Ckem. SOC.,1116 (1952). (13) E. A. Braude, D. W. Turner a n d E. S. Waight, ibid., 2396 (1958). (14) L. J. Bellamy "Infrared Spectra of Complex Molecules," 2nd edition, John Wiley and Sons, Tnc., New York. N. Y., 19.58. p. 34. (1.5) L. Bateman. ef a i . , J. Chem. SOC..915 (1050). ous
5821
PERESTERS WITH OLEFINS
ucts6 I n the presence of copper salt catalysts the type of the reaction is different. t-Butyl perbenzoate gives methyl benzoate and a little carbon dioxide beside much benzoic acid. The formation of these products may be accounted for by the reaction sequence 0
It
CcHbCOO-t-BU
+ CU'
----f
It
+
C ~ H ~ C O - C U ' *OC(CH3)3 f 1) (CH3)aCO. --f (CH3)2CO
3-
CHI.
(2)
0
CH3.
I! + C~HECOO-L-BU + C6HbCOOCHs $- .OC(CHv)a
(3)
Methyl benzoate is probably formed by a chain reaction of which equations 2 and 3 are the propagating steps. The metal salt-catalyzed decomposition of peresters in the presence of olefins containing allylic hydrogen atoms-may be summarized in the steps 0
I/
CeHbCOO-t-Bu
+ CU' + 0 /I C&CO-CU+
(CH3)aCO.
+ .OC(CH3)3 + R H +R.+ (CH3)aCOH
(1)
(4)
0
+
II
R. C ~ H S C O - C ~+ + R O C O C B H -t ~ Cuf (5) The absence of the primary allylic benzoates from the reactions of terminal olefins provides evidence against conventional free radical or carbonium ion intermediates which generally result in isomerization of the allylic system. This observation suggests that the displacement of the allylic hydrogen by the benzoyloxy radical (equations 4 and 5) probably occur in a concerted manner. The detailed mechanism of this reaction is now being investigated by kinetic and tracer16 techniques. This reaction requires as catalyst a metal salt which is capable to engage in rapid oxidationreduction reactions. Transition metal salts, such as those of copper and cobalt are effective: ordinary metal salts, such as zinc chloride or magnesium bromide, are ineffective. The failure of triphenylmethane as a substrate may be attributed to the steric hindrance about the activated hydrogen atom. When the reaction of cyclohexene, t-butyl perbenzoate and cuprous bromide is carried out in an excess of an aliphatic acid, the allylic ester of the aliphatic acid is obtained instead of the benzoate. Cyclohexenyl benzoate is converted t o the acetate when it is heated in glacial acetic acid in the absence of a catalyst. This finding indicates that ester interchange occurs during these reactions. Acknowledgment.-The authors are indebted t o Professor T. P. Rudy for his assistance in analytical distillation and gas chromatography, to Mr. \Til(IO) Private communication. Professor Donald B. Denney, Rutgers University, New Brunswick, A'. J.
11.S. KHARASCH, GEORGE SOSNOVSKY AND &. C.YANG TABLE I REACTIONS O F I-BCTYLP E R B E S Z O A T E Mole olefin
Olefin
ti
Alole perbenzoate
Mmole catalyst
1-Octene 2-Octenc 1-Hexene Cycloheseue Cycloheseiie C>-clohexene Cj-clohexene 2,2,~-TrimctIij-lpeiit-l-ene Cyclohexanone enol acetate 4-I~in~lcycloliexene n-Pinene .3 . -> A411~-lbenzer~e .4 . -> Cupric 2-ethylhesoate. Cobaltous 2-cthyll~exoate
.33 ..%,j
liam Saschek for microanalyses, and to N r . Ihor 1Iasnyk for molecular weight determinations.
BSD
Teomp.,
C.
Reflux Reflux Reflux Reflux Reflux Reflux Reflux Reflux 105-1 15 105-113 106-115 105-115
Vol. s1
OLEFISS
_-____ Benzoic acid, g . ( % )
8.5(57) 9 . 4 (39) 9 . 0 (37) 2 . 2 (9) 7.@(29) 2 . 2 (9) 9 . 0 (37) 6 . 5 (27) 2 . 8 (121 9 . 0 (3;) (i.5 (27)
--([Z) Ib, 10 0 (35) 111, 13 0 (53) Ia, 16 0 (40) IIa, 29 5 (73) IIa, 31 0 (77) IId, 22 5 (56) I I a , 2 0 0 (50) IY,17 0 ( 3 7 ) I-, 15 0 (29) T-Ia, 28 0 (61 IT&, 1 7 . 0 (33j IC, 2 6 . 0 ( 5 5 )
- -Products-
Ester, g
Residue. 6.
5
7 3 4.5 5 10 6.5
G 46 5.5
ated in a potassiuin iodide trap connected after the ozonization vessel.z1 Xt the beginning of each series of deterniinations, the concentration of ozone in ozonized oxygen w a s determined iodimetrically using a ferrous-ferric system and tlie Experimental :ea result was cheeked by ozonization of a standard sample (pure 1-octene). The accuracy of this method was found to be Starting Materials.-1-Octene, 1-liexene, 2,4,4-triinethj-l+ 2 . 5 7 0 . T h e sueeess of this determination depends on the I-pentene and 4-vinylcj-elohexene were pure grade (99 mole nature of olefins. Certain negatively substituted olefins per cent. minimum) Phillips hydrocarbons lvhich were used without purificatioi~. 2-Oetene used was Practical Grade (95 such as enol acetates require a higher temperature (O'), a n d the method is not applicable for o,p-unsaturated carbonyl mole per cent. niinimurn) Phillips hydrocarbon. a-Pinene, compounds, aromatic hydrocarbons and polyhalogenated supplied b y Matheson, Coleman and Bell Co., \vas freshly olefius. distilled before use. Cycloheuene and acrylonitrile, purDecomposition of &Butyl Perbenzoate in the Presence of chased from Eastrnan Kodak Co., were freshly distiled beCuprous Bromide. Formation of Methyl Benzoate .-:\ mixfore use. C~-eloIiex-l-enylacetate was prepared according to ture of t-butyl perbenzoate (0.1 mole) and cuprous broniide the method of Hagemeyer and Hull," 1 ~ 1.4573. ~ 0 ~ Allylbenzene T W S prepared according to thc method C J f kIers1i- (0.1 g., 0 3 5 millimole) in beuzene (100 ml.) was refluxed f o r 38 hours. During this period carbon dioxide (270 nil., lo(,:&) berg.]* evolved. The reaction mixture was extracted with :t 2 ,\i t-Butyl perbenzoate and l-butJ.1 peracetate as a 75TCbensodium carbonate solution to remove the copper salt and tlie zene solution (Lupersol $7) were purchased from Lucidol benzoic acid, 8 g. (5:355). T h e benzene solution wis conDivision, \&a' lace Tiernan, Inc. Unless otherrvise specified centrated a t reduced pressure. Distillation of the residue the peresters were used nithout purification. gave methyl benzoate (3.7 g., 3457, b.p. 197-198", X ~ " D .inhydrous reagent grade cuprous bromide was dried and 1.5175), the infrared spectrum of which was identic:tl to t h a t kept in a desiccator before use. Cupric or cobaltous 2of a n authentic sample of methyl benzoate. ethylhexoate was prepared from equivalents of potassium 2]Then t-butyl perbenzoate was refluxed in ?z-octane for 5 ethylhexoate and the corresponding sulfate. The precipitate hours in the absence of copper catalyst, isolated were carhvii thus obtained was mashed, dried and dissolved in benzene to dioxide ( 3 0 % ) , benzoic acid (2@%), acetone (l.8%, 2:4make a l@yc solution. dinitroplienylhydrazone, xn.p. 1 2 5 O ) , and a n azeotroplc mix-1uthentic cinnamyl benzoate and 1-phenylprop-a-en-1-yl ture of t-butyl alcohol and benzene, 8 g., b.p. 73." Only benzoate were prepared by the method of Braude, Turner traces of methyl benzoate could be detected bl- infrared and Waight.ls Analytical Methods.-&Butyl peresters were analyzed b>- analysis. The Reaction of t-Butyl Perbenzoate and Olefins in the tlie addition of a weighed sample of the perester to a reagent Presence of Cuprous Bromide.-The reactions were carried consisting of glacial acetic acid. sodium iodide and traces of ferric chloride. T h e liberated iodine was titratctl with 0.1 o u t under an atmosphere of Linde liigli purity nitrogen. t-Butyl perbenzoate was added over a period of one hour into -Vsodium tliiosulfatc , l o The extent of unsaturation of the allylic esters formed ill a stirred mixture of the olefin and a small amount o f the catalyst maintained a t the reaction temperature. T h e progtlie perester reaction cannot be determined satisfactorily by ress of the reaction was followed by periodic exaniination of conventional methods. Both hydrogenation and bromiclethe infrared spectrum of the mixture. Heating was conbromate titration gave values higher thau those of theoretical, tinued for 0.3 hour after the disappearance of the perester presumably due to h>-drogenolysis arid allylic bromination, band froin its infraied spectrum. The total reaction time ectivcly. T h c procedure used is a modification of t h e IKLS :tpproxirnately 1-2 hours a t llOo m d 2-4 hours i t t 90' ting methods20 developed in this Laborator>- by Dr. after the addition o f the perester. There was little or no gas l'i-alter Sitdenberg. -4 constant stream of ozonized oxygen evolution. Tlie mixture was next extracted with a 2 N soluo f k i i o i n i ozone emiceritration ( 0 . 3 5 ~ 0 . 4 0niini~Ie/:300 1111. of iixJ-gcn/itiin.) w:is paswti into it solutioti of the sainple ( t ( b tion of sodium carbonate to reinnve the catalyst and tienzoic. acid. I I I ccrt:tin cases, tlie mixture \vas dibtilled prior t o t h e 30 uinioles) i i i ethyl acetate (25-50 rnl.) niaintaiiietl at -40' estractioii in order to isolate t-butyl alcohi>l. Tlie yields rif c ~ ~ t l - p ~ )was i n t t:tkcn \vlien iodinc was lihcr/-butyl :ilcoIioI were 76 and R8qG iii thc reactioiis o f I - i ) c t c t ~ e oints and melting lx,ints are uncorrected. U d e c u l a r and "-octeiic, respectively. Benzoic :tcid TVRS recovered f r o i i i weights were determined b y the cryoscopic method in benzene. I n f r a tlie carhonate extracts 11). neutralizaticin. T h e c~rganicpliiise red spectra were determined on a Perkin-Elmer model 2 1 spectrophowas w:tshetl lritli water, dried over sodiuni sulfate, a i i t l tiistometer and ultraviolet spectra were determined on a Warren Spectrotilled under reduced pressure. I n all the runs, appreciablr cord. quantities of starting olefins were recovered and the yields (17) H. J. Hagemeyer, Jr.. and D. C. Hull, I f i d . Eizg. Chein., 41, 2'3'20 allylic benzoates recorded were based on the amount r i f per (1949). benzoate used. Tlie results \{o-o\c;CH3
I
CzHj
?
CaHs
0
l
0 0 \ / C /'\
~-~
CHI CZHS I, Rr 0.960
~
( 1 ) 1,uridol Research Assistant. ( 2 ) S . A . l l i l a s a n d A . C,olubr,vii., 'I'EIIS 1 O L ' R N A I . , 8 1 , :3:jli1 (l!)z!l). ( 3 ) S . A. Milas aiid I. Belit, zbrd., 81, 3858 (1H50).
Fig. 1.-Peroxides
C2& Y I I , Ri 0.012 derived from methyl ethyl ketone a11d hydrogen peroxide.
The percentage of each peroxide present in the mixture was estimated by paper chromatography to be approximately: I, 2 3 ; 11, 1; 111, 2; IV, 5 : V, 1 2 ; \T, 45; and 1711, lOyc,respectively. Peroxides I, VI and VI1 were separated from the mixture without resorting to cellulose powder chromatography, while the other four peroxides were
