April 5, 1957
PHOTOOXIDATIONOF
technique t h a t the C3 fraction contained propylene as well as propane. KO attempt was made t o isolate hexane.
Acknowledgment.--In conclusion we wish to express our gratitude to the Office of Ordnance Re[CONTRIBUTION FROM THE
DIETHYL KETONE
1337
search, U. S.Army for their generous support of this investigation. SYRACUSE 10, S . Y.
DEPARTMENT OF CHEMISTRY,
U N I V E R S I T Y O F ROCHESTER]
The Photooxidation of Diethyl Ketone1 RY J. ERICJ O L L E Y ~ RECEIVED OCTOBER 4, 1956 The photochemical reaction between diethyl ketone and oxygen has been studied a t temperatures from 35 t o 150" at several oxygen pressures and over a wide range of intensities. At all temperatures t h e carbon monoxide yield becomes small a t high oxygen pressures, thus indicating t h a t the propionyl radical reacts with oxygen without producing carbon monoxide. All yields pass through maxima (except possibly carbon monoxide at 35"), and oxygen acts as a n inhibitor a t high oxygen pressures. At room temperature quantum yields are independent of intensity, but at 150' they are a n inverse linear function of the square root of t h e intensity. Oxygen at very low pressures strongly suppresses the formation of Cp hydrocarbons so t h a t t h e reaction between ethyl radicals and oxygen must be very rapid. A fairly complete mechanism for t h e reaction can be suggested and the relative values of several rate constants estimated. Chains must be propagated mainly by ethoxy and pentanonyl radicals. Since rates of formation of acetaldehyde and carbon dioxide are closely parallel t o each other, possibly pentanonyl radicals react with oxygen to give acetaldehyde, carbon dioxide and ethyl radicals.
Introduction The photolysis of diethyl ketone has been studied, and a mechanism which satisfactorily accounts for the experimental observations has been put for~ a r d . ~Information - ~ about some of the reactions of radicals with oxygen was obtained7 in an investigation of the photolysis of mixtures of diethyl ketone and oxygen. The reaction was shown to be a complex radical chain of short length, and various possible radical reactions were discussed. The primary process is almost certainly reaction 1, and it is known that the propionyl radical dissociates almost completely even a t room temperature (reaction 2) in the absence of foreign substances. In the presence of oxygen the low yield of carbon monoxide indicates that reaction 2 is largely suppressed. Reaction 3 has been suggested for the reaction of propionyl radicals with o ~ y g e n . ~ ~ ~ The present paper gives further data on the diethyl ketone-oxygen reaction. In general, oxygen suppresses formation of Cz hydrocarbons more completely than hitherto indicated. Due to the chain character of the reaction, great care must be exercised to avoid local depletion of oxygen when the oxygen pressure is low. The higher values for Cz formation previously reported7 are thought to be due to such local depletion of oxygen. Experimental The reaction was carried out in a high vacuum system consisting essentially of a reaction cell, fractionation spiral, (1) This work was supported in part by contract with t h e Office of Naval Research, United States Navy and t h e Chemistry Division,
Air Force Office of Scientific Research, United States Air Force. T h e material in this article may he reproduced or used in any way by t h e United States Government. (2) Postdoctoral Fellow 1953-1955 under a grant from t h e Camille and Henry Dreyfus Foundation, Inc. (3) W. Davis, Jr., THISJOURNAL, 70, 1808 (1948). ( 4 ) L. M . Dorfman and Z. D. Sheldon, J. Chem. P h y s . , 17, 5 1 1 (1949).
( 5 ) K. 0. Kutschke, 11.E.J . Wijrien and E. W. R . Steacie, Tms 74, 714 (1952). (6) M . 11. J , Wijnen and E. W. R. Steacie, Can. J . C h e m . , 29, 1092
JOURNAL,
(1951).
(7) A. Finkelstein and W. A. Noyes, Jr., Disc. Favaday Soc., KO.14, 76, 8 1 (1953). ( 8 ) R . B. Martin, Ph.D. Thesis, University of Rochester, 1953.
MrLeod gage and oxidizing furnace. The parallel light beam entirely filled the cell, and homogeneous conditions were maintained by means of a magnetically driven stirrer. T h e volume of the part of the line which included the blcLeod gage was 663.6 ml., the volume of the reaction system (including stirrer) was 198.8 ml., and the volume of the reaction cell was 76.0 ml. Two light sources were used (1) a Hanovia Alpine Burner and (2) a water-cooled General Electric Co. A-HB arc. The beam was collimated by a quartz lens, and radiation below 2300 A. was removed by a Corning 9863 red-purple Corex filter. Since thqlong wave absorption limit in diethyl ketone is about 3200 A.,-the chief absorbed wave lengths mere 3130,2650 and 2537 A . Some experiments used a chlorine g?s filter so t h a t the main transmitted radiation was 2650 A . Another series was carried out with a 2-mm. Pyrex plate to give mainly 3130 Intensity effects were studied by neutral density filters of two types-silver coated quartz plates and darkened copper screens. Transmissions were determined by a photo cell and also by a Beckman photometer. T h e measurement of the radioactivity of the C'"1abeled products was made by freezing them into a proportional counter tube of the type described by Bernstein and Ballantineg and counting on a scaler. It was found that the activities of portions of a given sample were directly propoltional to their pressures. Procedure.-A small sample of diethyl ketone, prepared as previously described3 was removed from the black storage bulb and degassed before being admitted to the cell a t 35 mm. pressure, a t which pressure all runs were carried out. The desired amount of oxygen (prepared by heating potassium permanganate) was added and the well-stirred mixture illuminated. The gases not condensed by liquid nitrogen (oxygen and carbon monoxide) were removed, measured and transferred to a bulb containing copper-copper oxide which could be heated in a furnace. The resultant carbon dioxide enabled oxygen and carbon monoxide to be calculated. A t -160" a second sample, carbon dioxide, was removed and measured, and a third fraction was obtained a t - 1 1 5 O . The third fraction was acetaldehyde and v a s readily characterized by freezing it from the system and causing it to react with p-hydroxydiphenyl with which it gives a colored compound with a characteristic absorption peak at 5700 A . 1 0 I n experiments at low oxygen pressure, oxygen was added by a "doser" to keep the oxygen pressure approsimately constant.1l Small amounts of ethane and ethylene are produced under these conditions. I n these experiments, the non-condensable gases were passed through a fractiorlatiou spiral maintained at -215" with supercooled liquid nitrogen. This prevents removal of ethane and ethylene
a.
(9) W. Bernstein and R.Ballantine. Res. S r i . ITtstr., 21, 158 (1950). (10) E. Stotz, J . Eiul. C h i n . , 148, 585 (1943). (11) F. B . Marcotte and W. A. N o y e s . J r , Disc. F a l a d a y SOL.,S o . 10, 236 (1951).
J. ERICJOLLEV
153s
with the oxygen and carbon monoxide. The C2 hydrocarbons were then removed through a liquid oxygen trap arid measured.
Results The following quantum yields are reported : oxygen disappearance, Cq hydrocarbons, carbon monoxide, carbon dioxide, acetaldehyde. They are obtained in each case by dividing the amount of product by the carbon nionoxide formed in a photolysis without oxygen, carried out either immediately before or after the oxygen run. This assumes that the quantum yield of carbon monoxide production in the absence of oxygen is unity, and the intensities quoted are also based on this assumption. This is true a t temperatures of 100 t o 200°, but may be a slight overestimate a t 35". Runs made with pure diethyl ketone in the absence of oxygen failed t o show a definite trend in the ratio of carbon inonoxide formed to incident intensity a t 35". There may be a slight decrease in quantum yield of carbon monoxide formation a t high intensities due to reaction between ethyl and propionq-1 radicals. Kutschke, JYijnen and Steaciej using higher intensity radiation than used in this work found values of 9co = 0.6 a t 23" and 0.S a t 50". Thus while no evidence indicating radical recombination was found in this work, it is possible that the quantum yields quoted a t 33" are all 10 to 20yc high. This does not affect the conclusions reached. :It the higher temperatures, the quantum yields should not be subject t o any systematic error caused by low carbon monoxide yields in the absence of osygen.4 Table I shows the c~uantuniyields as functions of oxygen pressure at four temperatures. They are accurate to within about 0.05 except for occasional inexplicable wrong values, and also excepting the quantuin yield of oxygen disappearance which may be in error by as much as 0.5 in some runs. In all runs the amount of diethyl ketone reaction was kept well below 1!2% so that secondaq- reactions of products were negligible. T h e most interesting feature is t h a t the oxygen consumption passes through a maximum a t 33" arid that the maxiinuni becomes less pronounced with increasing ternperature. The quantum yields of carbon dioxide and acetaldehyde follow a similar course, the former being the larger a t low temperatures, w!de the acetaldehyde predominates a t the high temperatures. The carbon monoxide quantum yield falls rapidly from unity a t 33', hut a t 1 50" reaches values as high as 1,s.Soinc experiments were carried out a t higher oxygen pressures (about 30 nim.) to tleterniine whether oxygen completely prevents reactio11 or not. These results are also shown in the tnblcs. Ethane ntid ethylene are pruduced a t T-ery low oxygen pressures s n d then only in very small amounts which increase with increasing temperature. 'rllt.se quantuni yields are given i n Table 11. .-ifcn- csperiiiit~iitstvc'r(Jcarrictl o u t with 35 i i i m . o f diethyl ketoiir :iiitl !iigli presmrc's o f nitric oxide tc~) act :is a "scu\-t.i!ger." 'The rcwilts arc shown i n TaI)lc 11 I , Sitrcigcii W I S one r i f thc nmin protlucts \vlicvi iiitric oxidc is preseiit. S ( Ia t t c n i p t is niatlc. to explain this iact i n l-iew of the small number of runs carried out with nitric oxide present.
1-01. 70
The quantum yield of oxygen disappearance is larger than the sum of CO, and CH3CH0, thus indicating that there is a t least one major oxygenTABI.~: I Q V A P ~ T V MYIELDS IU
DIETHYL KETONE-OXTGCU Ketone pressure = 35 mm.
~ ~ I X T ~ R E S
I*,
0 2
pressure,
mm.
Length of run, min.
einsteins,' ml./sec.
x
1012
*CO
Qo?
*co
*co? *cnacno calcd.
I' = 35" 10.50 10.20 ,195 10.40 10.50 ,400 11.00 ,745 1.56 20.00 21.2d 2.89 27.5 106.80 0.033 ,075
7.75 7.75 7.75 7.75 7.75 7.75 7.75 7.73
4.60 4.82
4.62 4.65 -1.16 3.36 2.90 .
,
0.26 .I6
2.26 2.49 .13 2.45 .15 2.42 .13 2.10 .ll 1.77 .09 1 . 4 1 .. 0.93
1.47 0.29 1.65 .19 1.18? .13 1.66 .11 1.03 .10 1.22 . 09 1.26 . OR 1.05 , 09
7' = 70"
0.033 . 06 .16 .44
. 70 1.14 2.03 34.0
10.50 8 . 8 6 4 . 8 7 1.11 11.00 8.86 5.25 1.00 12.00 8.86 a.87 0.56 7.00 8.86 6.0.5 .34 8.00 X.86 6.50 .34 10.00 8.86 5.60 .24 10.00 8.86 5.0,7 .26 80.00 8.86 .. . .
1.85 1.80 2.85 3.37 3.25 2.72 2.43 1.03
1.70 1.63 2.02 2.29 2.35 1.78 1.44 1.10
li' = loo0 0.025
. 0.3 .10 .50 1.10 1.88 4.0
3.5. (1
7.95 3.39 1.73 0 . , X 1.12 7 . 9 5 4 , 3 1 1.77 .96 1 . 4 1 7.96 5.28 1.68 1.58 1.86 7,iU 6.84 0.80 2.88 2.66 9.65 6.70 .ti3 3.11 2.50 7.70 8.51 .81 2.64 2.42 9.72 , . . A 1 2.62 2.07 85.00 8.00 . . .. 1.19 1 . 1 4 10.45 10.50 11.20 9.00 10.00 10.00 13.00
0,022
I,j,fiO
. 1*?J
12.00
Xi .91 1.OR 2.10 2.21 40. (1
7.00
,
l t l , 00
10.00 11.00
11.00 60.00
7- = 150° 7.K2 2 , X L(i9 0 . 1 0 0.97 8.3(1 i.02 7 ,62 7.ti2 i.62 7.52 7.62
.$.l(J 1.86
.i.71 1 . 6 6 fi. 88 1 , Xi 0,5,5 1.39 S . 3 9 1.48 3.41 1.40 .. 0.93 ?'auLE
.71 1.19
1.4:3 1.XO
1 ..'I2
2.24 1.9ii 2 , 7 ( l 2 . 2 5 2.4,7? 2 . 2 4 *';.Cl1 1.x3 2 . 0 4
1I
1 , 53 1.00 11. 63 1 , ill 1 . 01 0.6, $ 1
PHOTOOXIDATION OF DIETHYL KETONE
April 3 , 1937 TABLE I11
QUANTUMYIELDS IS DIETHYL KETONE-SITRICOXIDE MIXTURES~ Ketone pressure = 35 mm., alpine burner Temp.,
NO pressure, mm.
Qco *NZ 35 10 0.07 0.81 .09 6.6 35 80 150 73 * 34 2.0 a The formation of nitrogen in these experiments is interesting but was not pursued further. A similar effect was found in acetone-nitric oxide mixtures by Andei son and Kollefson . 2 0 "C.
ated product for which analysis has not been made. Products such as acetic and propionic acids, methyl and ethyl alcohols, water, diketones, etc., all have vapor pressures which would make their fractionation from a large amount of diethyl ketone impossible. 4 n attempt was made to identify some of these products by photographing their ultraviolet spectra after a 36 hour run, but the absorption curve was almost identical with t h a t of the pure ketone. The presence of acid groups was shown, however, by freezing the products of another run into 20 ml. of air-free distilled water and measuring the pH of the resulting solution. h value of 4.5 was obtained, and this when compared to 6.0 for the water corresponds to an acid quantum yield of about 3 if the value of K-4for acetic acid is assumed. Variation in light intensity has very little effect a t 35" but, a t higher temperatures, the quantum yields all increase with decreasing intensity (Table IV). At the highest intensity, with the ATABLEI V VARIATION OF QUANTUMYIELDS ~ I T H INTENSITYIN DIETHYL KETONE-OXYGEN MIXTURES Alpine burner; ketone pressure = 35 mm., 0 2 pressure = 1mm. I, 8
Temp., OC.
Length of run, min.
einsteins/ ml./sec.
x
1012
002
*co
0.27 3 . 6 0.15 35 340 1.20 3 . 3 35 70 .15 3.7 .14 35 10 7.67 0.21 8.8 1.03 100 480 40 1.57 7.7 0.73 100 10 5.70 7.8 .74 100 10 9.63 6.7 .63 100 0 . 2 1 15.4 4 . 0 150 360 1.39 1 2 . 7 2 . 8 150 50 5.58 8.9 2.1 150 19 9,52 8 . 5 1.9 150 9 5.9 1.5 150 0 . 2 1 750" A-H6 arc.
*cox
1.5 2.3 2.0 3.2 3.3 3.2 3.1 3.0 2.6 2.3 1.9 1.5.
QCH~CHO
1.2 1.4 1.3 2.9 3.0 2.8 2.5 4.8 3.0 2.8 3.0 2.0
H6 arc, the quantum yields seem to have approached a lower limit, and the variations can be represented by equations of the form @ = k'
+ k"/Ia'/z
where CP is the quantum yield in question, k' and k" are constants dependent on oxygen pressure, and I, is the number of quanta absorbed per second per inl. Diethyl ketone labeled with CI4 in the carbonyl position, prepared by Finkelstein,' was photolyzed
1539
both in the presence and in the absence of oxygen a t 150'. The amount of carbon monoxide produced in the absence of oxygen agreed well with t h a t produced by the unlabeled ketone (8.6 X mole moles mL-l set.-' compared to 8.4 X ml.-l set.-'), but the amounts of products during photooxidation were somewhat low. The carbon monoxide, carbon dioxide and acetaldehyde were found to be 93, 77 and 4% radioactive, respectively. Table V shows the effect of varying the wave length upon the quantum yields of products. TABLEV VARIATIOS OF QUANTUM YIELDSWITH WAVELENGTHIN DIETHYLKETONE-OXYGEN MIXTURES Ketone pressure = 35 mm.
35 35 35
2650 2650
0 . 5 5 0.82 1.08 0.82
4.3 3.8
1.01 1.20
1.01 1.22 0 . 5 2 0.056 0.54 .57 .57 1.10 1.85 .57
{;;;;) 3130 2650 2650 2650 2650
35 150 150 150 150
0.25 2 . 1 .23 1 . 8
1.7 1.7
3.3
,13 2.3
1.4
3.0 9.75 6.8 8.7 10.1
.10 2 . 6 3.0 1.9 2.3 1.2 2.6 1.7 2.6 1.9
1.3 3.1 2.5 3.2 3.2
Discussion To facilitate the discussion it seems advisable to list a t one point all of the steps in the mechanism which will be discussed
+
+
CzHsCOCzHs hv = CzHjCO CzH, (1) CzIIbCo = CO f CzHs; Rz = kz(CzHaC0) (2) CzH&CO*= CO CZH5; @ = cu+ (2a)
+
+
+ + +
C2HjCO 0 2 = COz CzHjO (3) R CzHjCOCzHs = RH C2HdCOCzHb (4) CzHjO CzHsCOCzHs = CzHjOH C~H~COC~H S (5) 0 2 = CHICHO COz CzHj C2HdCOCzHj (6) CzHsO 0 2 = CHaCHO HOz ( 7) CzHj0 = CHaCHO CzHjOH CzHjO (8)
+
+
+
+
+
+
+
+ + +
0 2 = C2Ha02; Rs = ko(CzH,)(Oz) CzHs CzHs = CzHe CZH4 CzHj CIHI, CzHj CiHsCOCzHi C2Hj = CzHs CzHaCOCzHj
(9) (10) (11) (12)
2CzH502 = 2CzHbO
(13)
CzHj
+
+
+
+
+
+
0 2
The photochemical reaction of diethyl ketone with oxygen must have a complex mechanism. The following facts must be borne in mind. (1) Quantum yields are essentially independent of intensity from 35" to about loo", but a t 150" they decrease as the intensity increases. The chain ending step must vary with temperature and is presumably a radical-radical reaction a t 150". (2) All yields pass through maxima (except possibly carbon monoxide a t 35") as oxygen pressure increases. This indicates that whereas oxygen is essential to chain propagation, it must also act as an inhibitor. At high oxygen pressures a t all temperatures the yields approach limiting and not very high values.
A coiiiplete iiiaterial balance has not been achieved. Products other than those quantitatively determined are formed. This statement may be easiIy verified by attempting to base a material balance on the determined products: CO, CO,, CH3CH0, CShydrocarbons. Semi-quantitative analyses indicate that the missing products are mainly acids' and ethanol. Certain conclusions may be drawn about the mechanism, and fairly good estimates made of the rates of a few of the steps. (a) The Decomposition of the Propionyl Radical.-No reliable estimate seems to have been made of the rate of this rcaction although it is generally considered to be higher than that of the decomposition of acetyl radicals. l 2 The primary process may be assumed t o bel3 reaction 1 where #J,the primary quantum yield, is probably about one in the absence of oxygen, although a weak fluorescence has been observed.'l The same is almost certainly true a t low oxygen pressures a t all temperatures, but the effect of oxygen a t pressures of several millimeters, particularly a t low temperatures, needs further investiga= 1 tion." The error introduced by assuming is probably negligible. At 35' neither oxygen nor nitric oxide a t pressures of several millimeters suppresses carbon monoxide formation completely (Tables I and 111). If reaction 1 is the sole primary process, a small fraction of propionyl radicals must retain enough energy to dissociate immediately. This conclusion resembles one which also appears to be necessary i n the case of Let us designate the fraction of propionyl radicals which inevitably dissociate as CY (all a's designate quantum yields). Thus propionyl radicals will decompose by reactions 2 and 2a where CzH5CO* indicates propionyl radicals which have retained energy from the primary process, and reaction 2 indicates that some of the propionyl radicals have reached thermal "equilibrium" and react either by ( 3 ) or with molecules in the system such as oxygen. The quantity cy will not be discussed in detail. i2t 35" with the radiation used in most of the experiirients a: = 0.09 since this is the approximate limiting value for @CO at both high oxygen and a t high nitric oxide pressures. The data in Table V show that a increases with decrease in wave length. The dependence of a: on temperature is uncertain since even at 'io" any source of CO other than (2a) may not be suppressed completely even when 2 mm. of oxygen is present. Estimates from Tables I and 111
+
(12) E . W. R . Steacie, "Atomic and Free Radical Reactions," lleinhold P u b l . C o q . , S e w Tork S Y . , 1054, p . 630. (13) W. A . S o y e s , Jr , C . B. Porter and J . E. Jolley, Cliev7. R F V S . , 56, 40 (1056). (14) > f . S. >latheson and J . TV. Zabor, J . Chem. P h y s . , 7, 536 (1939). (1.5) See W. A. h-ayes. J r . , Helw. C h i m . A d a , in press. (IO) D. S. Herr and W. A . Noyes, Jr., THISJ O U R N A L , 62, 2052 (1940). (17) TV A . S o y e s , Jr., and L. &I. Dorfman, J . Chem. P h y s . , 16, 788 (1048,. (18) J . S . A . b'nrsyth, Traits. Favnduy S o r . , 37, 312 (10.121. ( 1 9 ) R . UT.Durham and E. W. R . Steacia, J . Chem. Phqs., 20, 582 (10.5'2). (20) 11. W , Anderson a n d c r . li. Kullefson, 1'HIS J O C ' X N A L , 63, 816 (
1 o . k 1J
-
-
place a 0.15 a t 70" and -0.3 at 150". Hence a 0.2 at 100" by interpolation. Equilibrated propionyl radicals react with oxygen without giving carbon monoxide. Since Qro, = 1 a t high oxygen pressures at 35" and is but little greater than unity a t io" and at loo", one may with considerable assurance write reaction 3 with a rate R3 = k3(CSH6CO)(02). CzHjO is intruduced t o balance the equation, and the only evidence for this is the presence of ethanol in the products. Reaction 3 might be broken down into a series of steps, but there is no real basis for so doing. From eq. 1-3 one may write, if CzH6C0radicals are produced only in the primary process
+
mcl *co
+
+GO,
= 1
(lci)
Equation 14 reproduces the data at 33", from which the value of k3 'kz a t this temperature may be estimated as 20 X lo5 l./'mole. Values of @X calculated by use of this ratio and a: = 0.09 are shown in Table I. Above 35" the CO yield does not obey eq. 14, and eq. 15 is not valid a t any temperature. Equation 16 is valid a t high oxygen pressure at 35", approximately valid at 70" and niay even be valid at 100 and 150" at higher pressures than were used in this work. The following conclusions seem valid: ( a ) there is a CO or CSH5C0producing step other than reactions 1, 2 and 2a, which is teniperature dependent and becoines appreciable above 35"; (b) there is a source of COS other than reaction 3 at all temperatures; and (c) a t high pressures of oxygen these additional sources of CO and COSare suppressed. Estimates lead t o a value of about 2.2 X 10' l./mole for k3,'k2 a t 70". This gives an activation energy difference for E z - E3 of about 13 kcal., but this is uncertain to a t least 3 or 4 kcal. The ratio of pre-exponential factors a2 'a3 is about IO3 and hence to give a reasonable collision dianieter to reaction 3 , the pre-exponential factor for reaction 2 must not be greater than about l O I 3 set.-'. (b) The Chain Propagating Steps.--Quantulll yields of oxygen disappearance are such that under most conditions more than one molecule of ketone must disappear per quantum absorbed. There must be a radical attack on ketone molecules, arid this is probably an abstraction of the form shown iii reaction 4. A t very low pressures of oxygen (see Table 11) the formation of C2HS is good evidence that CzH5 Partakes in reaction 4, but it seems alniost certain that C2H& is the main hydrogen abstractor as shown in reaction 5. Quantum yields for carbon monoxide formation in pure diethyl ketone are b u t little more than unity does even a t 1 5 O O . 2 1 Hence the radical CZH~COCZH~ not dissociate appreciably under conditions employed in this work. I t may be expected t o react with oxygen prior to dissociation in thesc experiments. Reactions of pentanonyl radicals with oxygen are proposed as chain propagating steps. The ( ? I ) A p; S t r a c h a n , 1'11 r). 'l'hciis, Unirersitp of Rochester, l!l,>l.
April 5, 1957
PIIOTO~XJDATION OF DIETHYL KETONE
1.541
yields attain their maximum values. These facts are in accord with the mechanism put forward so far, as can be seen by comparing the total effect of step 2 with that of the sequence 3, 5 and 6. In each case one ethyl radical is formed, but the second sequence gives four product molecules and uses up two oxygen molecules compared to the single carbon monoxide molecule produced by step 2. Thus high quantum yields will be obtained when propionyl radicals react with oxygen before they dissociate Rco Rco, = aG kn(CzHaC0) ka(On)(C&CO) and when the oxygen pressure is low enough to make reaction 7 of minor imporfance. These conditions M P )( 0 2 ) (17) are found a t relatively low oxygen pressures a t 35" where (P) represents the concentration of penta- and a t relatively high oxygen pressures a t 150" (where radical-radical reactions largely replace ( 7 ) nonyl radicals. From the steady state for C2HsCO as chain ending steps). The carbon monoxide quantum yields increase r, = ar. M C ~ H ~ C O ) k 3 ( 0 2 ) ( ~ 2 ~ (1s) s ~ ~ ) with temperature a t any given oxygen pressure. Values over unity are observed a t 70, 100 and 150°, Hence indicating another CO or C ~ H S C Oproducing step $6 = Qco $co, - 1 with a suitable activation energy. Since @CO al(19) ways decreases with increasing oxygen pressure where @e is the number of pentanonyl radicals (except for an initial rise a t 70, 100 and 150'), the which react per quantum by reaction 6. Compari- direct formation of carbon monoxide is not possible, son of the right-hand side of eq. 19 with @CH,CHO and almost certainly propionyl radicals are pro(Table I) shows good agreement up to pressures of duced which may dissociate to give carbon monoxabout 0.4 mm. oxygen. At higher pressures ide. It is possible to suggest a second mode of re@CH*CHO exceeds @CO @CO, - 1, the excess in- action of pentanonyl radicals with oxygen (in addicreasing with oxygen pressure until a t 27.5 mm. i t tion to (6)) which would yield acetic acid and proequals unity. The appearance of another acetalde- pionyl radicals. Such a step has the formal qualifihyde producing step, increasing with oxygen pres- cations to satisfy the facts, but the kinetic exsure, coincides with the decrease of quantum yields pressions may not be used because the data are not from their maxima and suggests a chain stopping accurate enough. step of the form of reaction 7. The reaction undoubtedly increases in complexIf all pentanonyl radicals react with oxygen by 6 ity as the temperature increases. Previous worki and pentanonyl radicals are formed solely by 5, one indicates that carbon monoxide and carbon dioxide obtains are formed almost entirely from the carbonyl group a t 30' and that acetaldehyde comes from the ethyl groups. This is in agreement with the mechanism so far proposed. At 100' part of the carbon where (K) and (02) are concentrations of ketone dioxide comes from the a-carbon atoms and in the and of oxygen, respectively. Since (@CH,CEO - present work a t 150' possibly one quarter of the @Go - @co, 1) is very close to zero, small errors carbon dioxide does not come from the carbonyl in the individual quantum yields cause relatively group. These facts indicate strongly that compounds large errors in k7/k5, real values of from 5 to 16 being obtained. Substituting a value of k7/k5 = 10 in containing one carbon atom should be formed from eq. 20 and calculating (PCH~CIIOgives values which the end carbon atoms. No formaldehyde was found are in reasonable agreement with the experimental in the present work, and forinic acid could not have data. If the mechanism is correct, then ki/kb E been detected. The chain terminating step a t 35" and a t 70" ap10 f 5 . The greater efficiency of reaction 7 over 5 is qualitatively evident from the fact that 2 mm. of parently involves oxygen and reaction 7 is a t least oxygen in 35 mm. of ketone causes a considerable formally satisfactory. The HO2 radicals (formed reduction in the quantum yields. ,4t higher tem- from 7) are often considered to be inert and not to peratures the decreasing effectiveness of oxygen as propagate chains a t low temperatures, but satisa chain stopper indicates that kiilk6 decreases with factory data about these radicals are almost completely lacking.22 increasing temperature. At high temperatures (150") the effect of in(d) The Reaction 70-15O0.-The oxygen pressure a t which the maxima in the quantum yields tensity indicates strongly a radical-radical chain of 0 2 , CO, and CH,CHO occur increases with tem- stopping step. This raises the question as to what perature. At 35" the maxima occur a t very low happens to the C2H502 radicals. If these radicals oxygen pressures. At 150" the quantum yields are relatively inert, they may last long enough to are small a t low oxygen pressures and rise slowly yield ethoxy radicals by reacting with each other.23 so that they have not begun to decrease even a t Probably most C2H602 radicals yield ethoxy radi2 mm. pressure. At each temperature the CO quan(22) See ref 12 p 607 e l ff tum yield decreases nearly to its lowest value in the (27) T H Rnlev 1, R I Porter, F T Rust and W R X'dughan THIS same oxygen pressure range as the other quantum TOURNAT 73, 1 5 (1951)
mechanism seems less complicated a t 35' than a t higher temperatures. (c) The Reaction at 35".-A source of carbon dioxide other than reaction 3, which depends on the chain length and which is important a t all oxygen pressures except the very highest is required by the considerations of section (a). Reaction 6 fulfills these conditions. Evidence for reaction G is provided by the data a t 35" since
+
+
+
+
+
+
+
+
+
cals by some process. X quantitative method of determining ethyl alcohol was not available when these experiments were performed. In any case the most logical chain stopping step is reaction S. For ( 8 ) to be important a t 1.50' and not a t 35 and 70°, the concentration of ethoxy radicals must be far greater even a t the lowest intensity used. This is so because (5'1 and (7) must have appreciable activation energies whereas (S) would probably have a low activation energy. This may mean that the intermediate steps preceding ethoxy formation (such as ( 6 ) and ( 3 ) ) proceed far more rapidly a t 150' than a t lower temperatures. (e) The Reaction of Ethyl Radicals with Oxygen.-The rapid reduction of the quantum yield of C2 hydrocarbons with increase in oxygen pressure is a n indication of the speed with which ethyl radicals react with oxygen and provides a means whereby the method of competing reactions may be used t o measure the rate of reaction 9 The C2 hydrocarbons are formed along with butane by the well characterized reactions 10, l l and 12,4-6a t a rate given by
+
Rc, = 2klo(C1Hd7 k d K ) ( C * H d
(21)
The total number of ethyl radicals introduced into the system may be assumed t o be the sum of those introduced by reactions 1, 2, 2a and 6 Thus from the yields of CO and COSand the number of quanta absorbed, one can estimate the number of ethyl radicals formed. They may now be assumed to disappear only by reactions 9, 10, 11 and 12. Since those which disappear by reactions 10, 11 and 13 can be calculated from &,, the number which disappear by reaction 0 can be calculated. If one uses klo, k , and klz 'klI1/?from previous ~ o r k , ~the , ?in~ stantaneous value of (C2H5) can be calculated from eq. 21. Hence all of the information necessary for calculation of kg is available The results are shown in Table 11. Considerable scatter is to be expected due to uncertainties in &?, but the activation energy of (9) must be very low, and the steric factor is of the same order as t h a t of (12), i.e., about The steric factor for the correspond(2-1) R J I I i n a n d E 8 2 0 8 , 2 5 (1951)
W R Steacie
PliJC
R o j Sor
[COSTRIBI'TIOY FROM THE
ing methyl radical reaction has ;I similar order o f iiiagnitude. The initial product of (!I) must be CS1&O2radicals but no definite conclusions about their subsequent reactions can be drawn froni the present work a though some suggestions can be made. AitX3,for example, since the formations of CO, CO? and CH3CHO have been accounted for by the mechanism, these products may not result in appreciable quailtities from reactions of C2H502 radicals. Decoiiiposition, either directly or following hydrogen abstraction to form a hydroperoxide, would almost certainly produce acetaldehyde in much larger amounts than observed a t all temperatures. It seems probable that reaction 13 takes place'3 although reaction with oxygen to form C2Hb0 and ozone cannot be eliminated. From the mechanism the expression 2 (%o @ c o p ) - @o, should be zero if none of the oxygen consumed by the radicals is returned t o the system. The values for this expression show considerable scatter but are about 0.4 ( 3 5 O j , O . i (7OC'1, 1.0 (looo),and at 150' there are about as many negative as positive values. Support is obtained therefore for a reaction which returns some oxygen to the system. Further speculation about reaction mechanism does not seem to be warranted. Thus the photochemical diethyl ketone-oxygen reaction may be described in much the same terins as the acetone-oxygen reaction.11,25f2in28 -Ifany important questions remain to be answered, particularly as regards the detailed behavior of the IiO, intermediates. I t seems fairly evident that their main fates are not the formation of hydroperoxides. Acknowledgment.-The author wishes t o wknowledge use of part of the doctoral thesis of R. R . Martin for some of the preliminary results of this investigation and to thank Professor U'. -4. Soyes, Jr., and his group for many helpful discussions ant1 suggestions. 1532592fi
+
ROCHESTER, S E W YORK
(2.5) F. B . l l a r c o t t e and IT. A . S o y e s , J r . , T H I SJ o r ~ s . ? ~7 4. , 7 x 3 (1932). (2G) G I
