The Vapor Phase Photolysis of Trifluoroacetone at Wave Length 3130

McPherson Chemical Laboratory, The Ohio State University]. The Vapor Phase Photolysis of Trifluoroacetone at Wave Length 3130 A.1. By Robert A. Sieger...
0 downloads 0 Views 559KB Size
VAPORPHASEPHOTOLYSIS OF TRIFLUOROACETONE

Oct. 20, 1954

carried out during the course of this study and to John Keil, who determined some of the physical

[CONTRIBUTION FROM

THE

5197

properties reported in this paper. SCHENECTADY, NEW YORK

MCPHERSON CHEMICAL LABORATORY, THEOHIO STATE UNIVERSITY]

The Vapor Phase Photolysis of Trifluoroacetone at Wave Length 3130 A.l BY ROBERT A. SIEGERAND JACK G. CALVERT' RECEIVED JUNE 5, 1954 A kinetic study is made of the photolysis of CFICOCHJ vapor a t various temperatures, concentrations, and light intensities to determine the relative reactivity of CFJ and CHI toward H-abstraction reactions. The gaseous products are: CO, CHI, and high I. the methanes and ethanes are formed only CFJH, CZH6, CFsCHa and C Z F ~ .In experiments a t low [CF~COCHB] CFICOCHI -+ CH4 CFpCOCHz (1) ; 2CHs + CzHe ( 2 ); CFI fCF3COCH3 +CFaH CFzCOCH, in the reactions: CHI (3); 2CFa -P CzFe (4). The activation energy differences are: E1 E2/2 = 8.9; E3 - &/2 = 6.6 kcal./mole; the ArX lod4,PJ/P4'/a= 1 X lo-'. The quantum yields of CO, CHI and CZH6 are greater rhenius P factors are: P I / P ~=~ 5/ ~ than 1.0 at high temperatures and increase markedly with temperature increase. The quantum yields of CFaH, CzFs and CFaCHj are less than 1.0 a t all temperatures and decrease a t high temperatures. In explanation of these unusual results a chain mechanism is proposed which involves C2He formation in a reaction other than ( 2 ) a t high temperatures. This may be a rare example of a methyl-abstraction reaction. The results provide evidence for a rarely postulated, gas phase, radical addition to a carbonyl double bond. It is suggested that CF3 radicals generate CH3 radicals a t high temperatures by the reCF~COCHJFi (CF3)2COCH3, and (CFJzCOCHa + CH, CF&OCFt. actions: CFa

+

+

+

+

-

+

In view of the results of the many detailed studies of acetone photolysis3 it seemed probable that trifluoroacetone would undergo photodecomposition forming in part methyl and trifluoromethyl radicals. The present study was initiated in an attempt to gain information concerning the relative reactivity of the CH3 and CF3 radicals and to determine the effect of fluorine atom substitution in the acetone molecule on the mechanism of acetone photodecomposition.

Experimental Apparatus.-The all-glass photolysis system consisted of a quartz photolysis cell (50 mm. long, 30 mm. diam.), a glass circulating pump, trap and mercury manometer. This system was isolated from stopcocks by mercury valves. The cell was suspended in an aluminum block furnace (regulated to & l o ) . Radiation from a Hanovia type A(S-500) burn.er (run on a regulated 3.00 amp. a.c. current) was filtered to isolate wave length 3130 The light beam was well collimated by a series of lenses and stops so that a fairly homogeneous beam of radiation filled the cell volume (35.3 cc.) almost completely. The geometrical arrangement of all components in the light train remained fixed during the entire study. A photomultiplier-amplifier system was used to measure the fractions of light absorbed. Absolute intensities were estimated by chemical actinometry a t five regularly spaced intervals during the course of the photochemical runs. Estimates of the total quanta entering the were: 1.14, 0.93 (uranyl oxalate accell per sec. X actinometry6). tinometry6), 1.06, 1.19, 1.14 (KsFe(Cz04)~ No aging of the light source was apparent, so an average value of 1.09 X 1015 q./sec. was used as the incident light intensity in quantum yield calculations for all runs. Calihrated uniform density filters were placed in the light path to obtain lower incident intensities in some experiments. Materials.-Carefully fractionated CFaCOCH3 was purchased from the Caribou Chemical Co. (b.p. 21-22' (750 mm.)) and further purified by fractionation a t reduced pres(1) Taken from the thesis ofR. A. Sieger submitted for the Master of Science degree, The Ohio State University, 1954. Presented in part before the Division of Physical and Inorganic Chemistry of the American Chemical Society, New York, September, 1954. (2) Author to whom communications should he addressed. (3) For reviews of work see: (a) W. Davis, Jr., Chem. Revs.,40,201 (1947): (b) W. A. Noyes, Jr., and L. M. Dorfman, 1.Chem. Phys., 16, 788 (1948). (4) R. E. Hunt and W. Davis, Jr., THISJOURNAL, 69, 1415 (1947). ( 5 ) W.G.Leighton and G. S. Forbes, ibid., 62, 3139 (1930): G.S. Forbes and L. J. Heidt, ibid., 66, 2363 (1934). (6) C. A. Parker, Pioc. Roy. Soc. (London), A220, 104 (1953).

sure in a high vacuum system. The mass spectrum of the purified sample was consistent with that expected for pure CF3COCHs; approx. mol. wt. (Dumas method), 114 g./ mole; approx. v.p. in mm., 6 (-65'), 18 (-51'), 98 (-23"), 200 ( - g o ) , 385 (5"), 537 (12'). High purity reference samples of the gases CFaH, CzFe, CHaCF3 and CF, were provided by the Organic Chemicals Department of E. I. du Pont de Nemours & Co. Additional samples of these gases were supplied by the Minnesota Mining and Manufacturing Co. and Professor J. D. Park, University of Colorado. C Z H and ~ CHI reference gases were Phillips Research Grade. CO was prepared by the action of H&O, on NaOzCH and purified in the conventional manner. Product Analysis.-A maximum of 2% photodecomposition of the original ketone was allowed in all runs. CO and CH4 products were removed following photolysis using a Toepler pump with the other products and excess ketone condensed a t liquid N2 temperature. A second fraction of the products was removed from the ketone excess using a The modified Ward still' maintained a t -110 to -100'. analysis scheme was checked with known mixtures; almost complete recovery of CHI, CO, CzFe, CZH6, CFlH and about 90% recovery of CFsCHa could be effected from the excess ketone. (CFsCHs is the product with the lowest vapor pressure.) CO was analyzed by chemical means8 using a Blacet-Leighton gas analysis system. The ethanes and methanes were determined using a General Electric analytical mass spectrometer.

Results Products.-CO, CH4, CzHs, CH3CF3and CzF6 were identified as the major gaseous products. Proved absent were the possible products CH3F, CF4, HF, SiF4 and Fz. An unsuccessful attempt was made to identify other products froni the mass spectrum of the condensable fraction containing excess ketone. Molar Extinction Coefficients.-& a fixed temperature the absorption of 3130 A. radiation by CF3COCH3 vapor followed Beer's law within the experimental error of the determinations over the range of concentrations, 17.3 to 0.84 X lo1*molec./ cc. The coefficient showed an almost linear increase with increase in temperature. Representative values of E for CF3COCH3 vapor a t 3130 A. are: 3.6 (25') and 6.4 (350°), where logl@o/l) = (7) E. C. Ward, Ind. Eng. Chem., A n d . E d . , 10, 169 (1938). (8) F. E. Blacet, G. D . MacDonald and P. A. Leighton, ibid., 5, 272 (1933).

5198

ROBERTA.

SIEGERAND

QUANTUM YIELDS OF PRODUCTS A T l a , q./ [CFrCOCHs], cc.-sec. molec./cc. 7

Run

1 2 3 4 5

0

7 8 9

x

lo-"

1.54 0.864 ,592

.251

x

10-1s

8.20 8.24 8:20 8.21 17.3 8.20 3.12 0.841 3.121

2.37 1.53 0.711 .210 .711 [COZ] = 10.0

1

JACK

G. CALVERT

Vol. 7G

TABLE I 117' AS A FUNCTION OF 1%AND [CF,COCH,]

co

CHI

0.52 .54 .56 .58 .38 .51 .65 .77

0.138 ,172 .174 .208

.63

Quantum yields CFsH CaHs

.116 .153 .099 .069 ,078

€61; c is concentration in moles/l., 1 is the path length in cm. Effect of Variables on the Quantum Yields of Products.-The data at different temperatures are summarized in Fig. 1 ([CF3COCH3] is approxirnately constant at 8.2 X lo1* molec./cc.; 10 is constant at 1.09 X 1015 q./sec.). @co values of Herr and NoyesQafrom CH3COCH3 photolysis (at 110 mm.) are shown also for comparison. Other experimentsQbhave shown that CDCO remains near unity up to 400' (shown by the dotted line in Fig. 1). A slow thermal reaction in the run at 350" necessitated a small correction to the total rate of product formation. (This correction amounted to

CaFs

--.

CFtCHi

0.258 0.072 0.0075 0.027 ,312 .097 ,0052 ,025 ,323 ,116 .0062 ,019 ,294 ,100 ,0051 ,010 ,189 ,062 ,0039 ,0093 ,250 I086 ,0061 ,028 ,291 ,132 .020 ,068 I305 ,180 ,055 ,099 ((20% excess prevented analysis for these products)

the following 70 of the total rate of each product: C ~ H G95%; , CO, 4';;b; CFsH, 3%; CH4, 0.4%; C2H4 was a major product of the thermal reaction; none was formed photochemically.) There was negligible thermal reaction in the runs a t all the lower temperaturc-s. The absolute and the relative 1-aluesof the quatitum yields are probably accurate to about *1.5(,;-&and +5'&, respectively, for all products other than C2F6and CF,+CHa. The very small yield of CaF6liinited greatly the accuracy of its determination. CZF6yields are too small to be represented with the data it1 Fig. 1. Representative values of @clFBa t LTarious temperatures are: 0.0003 (26"), 0.0010 (71Io), 0.0040 (113"), 0.0041 (136'), 0.0050(153'), 0.0025 (205"), 0.0018 (254'), 0.0003 (?OSo), 0.0006 (351"). @CF+H~ values are probably consistently low by 1Oyo (incomplete removal. @'s of products as a function of [CFtICOCHa] and I, a t 117" are given in Table I. Run 0 was identical to run 7 in every respect except that COP was added to a concentration of 10.0 X 10l8 molec. /cc. Discussion The Relative Reactivity of CH, and CF3 in HAbstraction Reactions.-By analogy with CH4 and CZH6 forming reactions demonstrated in CH3COCHz photolysis3 one might suggest the following reactions for the formation of the methanes and the ethanes in CF3COCH3 photolysis

+ CFaCOCH, + CH, + CF3COCH2 2CH3 +CiHo CF3 + CFsCOCHs +CFaH + CF3COCHt 2CF3 +CzFo CFa + CHo +CFlCHs

CHI

(1) (2) (3) ( 4)

(5)

A simple test of the mechanism is applied to the data by the method of Dorfman and Soyes."' Assuming the mechanism 1-4 the relations (G and h can be derived. In Fig. 2 the functions &x4/ @ c ~ H ~ ' and /~ @ ~ F ~ w / @ c ~are F ~ ~plotted /~ as. [CFsQCH4/'3C*Ual/'

T E M P E R A T U R E , OC.

Fig. 1.-The quantum yields of products of CFsCOCHa photolysis as a function of temperature: [CFsCOCH31, about 8.2 X 10'8 molec./cc.; lo,about 1.09 X 10l6 q./sec. The co yields from CHaCOCH3are from Herr and ~ o y e s , g 8 (9) (a) D. S. Herr and W. A. Noyes, Jr., THISJOURNAL, 62, 2052 (1940); (b) E. I. Akeroyd and R. G. W. Norrish, J . Chcm. S O L . , 890 (1936); H.S. Taylor and C. Rosenblum, J. Chem. Phys., 6 , 119 (1932); J. A. Leermakers, THISJOURNAL, 66, 1899 (1934); C. A. Winkler, Trans. Faraday SOL.,81, 761 (1935).

= kl[CFoCOCI-Ia]/k?'/ZLI/%

@ c F ~ B / @ c ~ F ~ '= /'

kg[CFaCOCHa]/kr1/21.'/1

(a) (b)

COCH3]/1,'/z using the data from runs a t 117" (Table I). At values of [CF3COCH3]/Ia1/' < 3 X 10" molec. (sec./cc.-q.)'/l the data show the linear relationship between the variables expected from the relations a and b. At high COnCelltratiOllS and/or low light intensities, deviations from the (10) L. M. Dorfman and W. A. Noyes, Jr., J . Chcm. Phys., 16, 557 (1948).

5199

VAPORPHASE PHOTOLYSIS OF TRIFLUOROACETONE

Oct. 20, 1954

9-0.8

2.0-

0.

- 0.6 - 0.4 +

- 0.2

1.5!i

-+

n

? a

1.0-

i! I

2

[CH~COCF$

(

3 4 l 5 x IO-'' mo iec .(set. /cc.- q.P.

Iig. 2.-Quantum yield functions vs. [CFsCOCHa]/Ia"'; temperature, 117". The open and half-darkened circles refer to the CH4-C2Heand CFaH-C2F8 functions, respectively.

P

z z

2 0.5I-

V

z 3

LL &

simple linear relation are apparent. Acetone re0 0.0sults show no such marked deviation over a wide c3 range of variables.lO.ll Following the system used in the acetone studies activation energy differences were evaluated; the experimental conditions were -0.5chosen so that kinetics consistent with reactions 1-4 were obtained. In-Fig. 3 a plot of the logarithm of the rate functions RCH,/RC~H$'/~[CF~COCH31 and R c F ~ H / R c ~ F ~ ' / ~ [ C F ~ Cvs. O C1/T H ~ is ] given. A typical linear plot of the C F ~ H - C Z F ~ -1.0function is obtained for the entire temperature I range, but a sharp break-off is seen in the plot of the C H ~ - C ~ Hfunction. C At high temperatures (as well as high [CF3COCH3]and low 1,)a reaction in addition to 2 which forms becomes important (discussed with the high temperature reactions). A correction to the CzH6 rate data for the four low temperature runs was made to eliminate the high temperature C& forming reaction. (This correction to the rate function in the runs a t different temperatures amounted to 2075 (113"), 3% ((iso), 1% (46") and 0.6% (25'). A least squares treatment of the corrected rate data gave E1 - Ez/2 = 8.9 and E3 - E4/2 = 6.6 kcal./mole. The Arrhenius P factors were calculated to be PI/Pz'/~ = 5 X and P3/P4'/2= 1 X (assuming U C H ~= U C F ~= 3.5 A. and U C F ~ C O C H= ~ 5.5 ik.). Compare these estimates with results from acetone studies. E,' - E 2 / 2 = 9.712 (9.6") kcal./niole and P 1 ' / P 2 ' / 9 = 10 X (where El'and Pl'refer to the reaction: CH3 CH3COCH3 CH4 CHzCOCH3 (1')). I t is interesting but probably fortuitous that PI/ P Z ' / is ~ one-half PI'/P~'/~, a result consistent with the simple collision picture of the two reactions 1 and 1'. Certainly the similarity of the two results from the different systems adds credence to the kinetic interpretation given. It is probable that E2 and E4 are both near zero, and thus E1 - E3 is about 2.3 kcal./mole. The high reactivity of CF3 for H-abstraction compared to CH3 may reflect a higher strength of CF3-H bond compared to the CH3-H bond. There appear to be no bond strength data available to test this suggestion.

s

+

-

+

J. C. Nicholson, THISJOURNAL, 73, 3081 (1951). (12) A. F. Trotman-Dickenson and I3 W. R. Steacie, J . Chem. Phys., 18, 1097 (1950). ( 1 1 ) A.

t I

I

,

I,

ROBERTA. SIEGERAND JACK G. CALVERT

5200

Vol. 7G

unity. The simplest condition for the E and P 2) which becomes important at high temperatures, functions to be 0 and 1, respectively, is for E8 = high [CF3COCH3]and low I,. T o determine the E4 = E Land for P6 = P4 = Ps, but obviously other kinetics of this ethane forming reaction three exvalues would fit the data as well. All of the results periments a t 265" were made a t varied [CFaseem most consistent with the interpretation that COCH,]. The values of R C H ~ / R C ? (corrected H~ CF3 has a high H-abstraction reactivity compared for minor contribution to R c ~ H from ~ 2 ) at various to CH,. pressures were: 1.64 (44.1mm.), 1.45 (202 mm.), It is apparent from the lack of CFq and CH3F in 1.83 (100 mm.). The relative constancy of the the products that F-abstraction reactions of CF3 ratio with variation in [CFaCOCH3]shows that C2and CHJ are unimportant up to 350". & and CH4 rates are proportional to the same power High-Temperature Reactions.-There are sev- of [CF3COCH3]. If one assumes that CH4is formed eral striking differences between the high tempera- only by reaction 1 then CzH6 formation by li or 7 ture photolyses of CF3COCH3 and CH&OCHd. and 8 is consistent with these results. (1) CO is formed in a chain process in the high CHa + CFaCOCHa +CaH6 + CFaCO (6) temperature photolysis of CF3COCHd; acetone CHI CFzCOCHa e CF3CO(CH3)? (7) shows no CO chain up to 400". (Compare @CO CF3CO(CH3)2 ---+C2Hs + CF3CO (8) from the acetones in Fig. 1.) ( 2 ) -2chain forming CH3-containing molecules is present a t high tem- Evidence for reactions of the type of 6 is rare inperatures in the CFjCOCH? system as seen from deed, and that which exists is not compelling. the quantum yield summations given i n Fig. 4. Trotman-Dickenson'? cites only two known examples of a reaction in which a CH, radical abstracts (Z@CHa = @CHa ~ @ C V H B @CFCHa; S @ C r a = @CF>H 2@C?Fs @ c ~ ~ c H ~-icetone .) shows no a radical from a molecule. Recent data from (CH3)Zsuch marked chain for CH3-containing species. Hg photolyses indicate that one of the stated examples, CH3 (CH3)SHg+ C2Hs CHaHg, is un( 3 ) Fig. 3 shows a sharp break in the CH&& rate function for CFdCOCHjphotolyses a t temper- important up to 20O".l5 The other example is that atures above 130". Acetone data show no break proposed by Rlacet and Be1116 in explanation of up to 400". (4) The data in Fig. 2 show marked the results from biacetyl photolysis a t high tempera(CHRCO)? + CH~COCH:I CH3 deviations from the linear dependence of the varia- tures: CH3 alternative explanation of the biacetyl bles expected from the simple reaction sequence 1-4. CO. Some of these uniisual results can be explained results may be given in terms of the reaction se(CH3CO)2 a (CH3)2COCOCHa; in terms of a reaction forming CLH6(in addition to quence: CHa (CH3)2COCOCHa ---f CHaCOCH3 CH3C0. (Similar reactions of radical addition to a carbonyl double bond are consistent with the results of the present study.) It appears that there are few, if any, known reactions similar to 6. One might expect the analogous reaction, CF3 CF3COCH3 -+ CF3CH3 COCF,, in this system if 6 were important. In Fig. 1 it is seen that C J C F ~ C Hdecreases ~ a t high temperatures in contrast to the increase for @ C % H ~ . Reactions T and S provide an alternative scheme for C2H6formation a t high temperatures. Reaction T is analogous to that suggested by Rust, Seubold and Vaughani7in explanation of 4-heptano1 formation in the peroxide initiated thermal decomposition of gaseous n-butyraldehyde. However reaction 8 is unattractive because it involves a hitherto unobserved and unexpected radical rearrangement. Definite conclusions regarding these reaction paths must await further studies of this and related systems. The temperature dependence of R C H J R C ~ from H ~ runs a t temperatures above 150" leads to E l --- E6 or El - A1Ii - E q (depending on the mechanism choice) = --.? * 1 kcal./mole. (In obtaining this estimate minor corrections to measured data were made to eliminate C P H for~ mation from 2 . ) If 1, 2 and 0 (or T and 8 ) describe CH4 and C2H6 formation completely then the quantum yield data from all the runs should be described by the relaI I 1 Jtion c.

+

+

++

+

+

+

+

+

+

+

+

+

15

90

I80

270

360

+

(14) A. P. 'Trotniatl-l)ickrns(,n, Q u w i . Rev. ( I A J I L ~ O U ) 7, , 198 (1953). (1.5) K . A. Hulroyd aud \V, A . Nuyrs, J r . , 'I'HIS JUTJKNAI., 7 6 , 1583 ( 1 9 5 4 ) ; I< 1%: K r l i t i r i l . i i i i l t' \\' I < . S I r d v l c . L ' r l t i . .I ( ' h r m , 31, ( j 3 1 ~l!4.5:31. 1 I W I: I: BI.M 1 1 1 ( ~ LV.E . Hell, U ~ K 1~, , (17) 1:. 1;. R l l i l , 1'. I f . S r u l l u l < l :,I111 li \ ~ . n , c l l l I l ,' I ' t I l S I < l l J K \ i l 7 0 . I::< fI'l4Xl

\v

~

VAPORPHASE PHOTOLYSIS OF TRIFLUOROACETONE

Oct. 20, 1954 % * H ~ / @ ~ C H= ~

+

(kJ~/ki2[CF~COCH31 ')

~ s / ~ I @ c H (~c )

At low temperatures, high I,, or low [CF3COCH3] the first term of c predominates, and c reduces to a. At 117" the second term becomes important for values of [CF3COCH3]/I,'/2 > 3 X 1 O I 2 molec. (sec./cc.-q.)'/2(see Fig. 2). The two terms are of about equal size a t 150" for the experimental conditions used in obtaining the data plotted in Fig. 3; a transition in the kinetics is evident from the break in the CH4-C2He rate function plot. At 265" the second term predominates. Although relation c gives a satisfactory description of most of the data, it fails to explain the results from experiments a t 117" and a t very high [CF3COCH3] or low I,. Under these conditions i t appears that heterogeneous reactions forming CzHs and C2F6 become important. An additional sequence of reactions must be suggested to account for the chain forming CH3containing molecules (4CHs/quantum a t 350" ; see Fig. 4) and the marked deficiency in CF3-containreaches ing species in the gaseous products; [email protected]~ a maximum of 0.6 a t about 280" and decreases with further temperature increase. The reactions 6 or 7 and 8 together with 9-12 are consistent with these results. Reactions 9 and 10 generate CH3 radicals from CF3.

5201

additional information concerning the possible reactions 7, 8, 9 and 10 through the thermal and photochemical generation of CH3 and CF3 radicals in the presence of CF3COCH3 and CH3COCH3. Primary Processes.-The light-activated state of CF3COCH3formed a t 3130 A. appears to have a greater stability than that formed from CH3COCH3 a t this wave length. This is evident in the very low quantum yields of products a t low temperature (see Fig. l), the increase in quantum yields a t low concentrations (compare @eoin runs 5-8 in Table I) and the decrease in @CO and @ . c H ~when COZ gas is added (compare runs 7 and 9 in Table I). Collisional deactivation (reaction 13) is probably important following activation a t low temperatures and high concentrations.

+

CF3COCH3 hu ----f CFsCOCHs* M' CFaCOCH3* M --+ CF3COCHs

+

+

( 1)

(13)

Three possible primary decomposition modes which might be expected in this molecule are CF,COCH,*

--+ CF3CO

+ CHa

+CFa + CHaCO +C F I C H ~+ CO

(11) (111) ( IV)

Reaction IV must be unimportant since the magnitudes of the different ethane yields are in reasonable agreement with that expected from purely random recombination of CH3 and CF3 radicals. CFD CFaCOCHo 5 (CF3)zCOCHa (9) Process11 might be expected to predominate over I11 (CFa)aCOCH3---+ CHD CF3COCF3 (10) on the basis of probable bond strength differences. CFBCO+CF3 + CO (11) Analysis was not accomplished for (CF3C0)2, CHaCOCFs --+ CHaCO CF3 (12) (CH&0)2 and other such products which might be Reactions 11 and 12 are proposed to complete the helpful in deciding this question. However a t every chain sequence. The reverse reactions in 7 and 9 temperature the quantum yield summation, [email protected]*, and the reaction 10 are analogous to the t-butoxy is a t least twice the magnitude of Z@cp8 (see Fig. radical decomposition. Reaction 10 might be 4)) a result which is consistent with the predomifavored over the reverse of 9 (because of bond nant occurrence of 11. Acknowledgment.-The authors wish to thank strength differences). To date analytical difficulties have prevented analysis for CF3COCF3 and CH2C0 The Ohio State University Development Fund which are expected major products of the high which provided one of us (R.A.S.) with a research temperature photolysis and the suggested reaction fellowship. Grateful acknowledgment is made to scheme. However the results presented give strong Professor Albert L. Henne and his co-workers for indirect evidence for the rarely postulated gas phase their cooperation in the pursuance of this study, the reaction of free radical addition to a carbonyl bond. Organic Chemicals Department of E. I. du Pont de The reaction sequence 1-12 describes well the re- Nemours & Co., the Minnesota Mining and Manusults of CF3COCH3 photolysis, but definite conclu- facturing Co., and Professor J. D. Park for samsions regarding many points in the mechanism must ples of fluorocarbons which they provided. await further experimentation. We are seeking COLUMBUS, OHIO

+

+

+