PHOTOCHEMICAL CXIDATION OF ACETONE
Oct. 20, 1964 [CONTRIBUTION FROM
THE
4249
DEPARTMENT O F CHEMISTRY, UNIVERSITY O F CALIFORNIA, BERKELEY 4, CALIFORNIA]
Photochemical Oxidations. 111. Acetone BY HAROLD S. JOHNSTON
AND JULIAN
HEICKLEN
RECEIVEDFEBRUARY 28, 1964 T h e room-temperature photooxidation ;f acetone (0.25 t o 17 mm.) in oxygen (0.09 to 9.7 m m . ) with continuous ultraviolet radiation above 2200 .4.has been studied by the method outlined in p a r t I of this series; observations were made by leaking t h e reaction mixture directly into the electron beam of the mass spectrometer during photolysis. T h e principal products of the reaction were H X O , H20, CHaOH, and CHsOOH; minor products were CHaCOOH, HCOOH, CH300CH3, and higher molecular weight products which were probably CHBCOCHOand CH3COCH20H. Because of the cracking pattern of the reactants, it was impossible t o establish t h e presence or absence of CHa, CO, CHsCO, and CO,. From the identified products a t least 14 free radicals are inferred t o be intermediates in this system. An exhaustive and impartial examination of all possible radical-radical disproportionation and recombination reactions indicates that the data exclude a large number of possible reactions but t h a t 140 reactions could still be occurring. Formaldehyde, so far as these studies go, could be formed by 39 diEerent reactions. Thus, this experimental method cannot give a complete mechanism, nor can any method t h a t simply analyzes all molecular products.
Introduction The photooxidation of acetone above 120' has been studied by several investigators. lP5 The radicals formed under these conditions abstract readily and undergo unimolecular decompositions with the result that the nature and amounts of products may be quite different from those a t room temperature. Srinivasan and Noyes6 and Osborne, Pitts, and Fowler7 studied the photooxidation at, room temperature with incident radiation a t 3130 A. Under these conditions the principal species that is oxidized is the triplet state of acetone. Kirk and Porter8 have also investigated the photooxidation a t room temperature with 3130 and 2800 A. incident radiations. Pearsong has investigated the reaction between 3G and 100' with both 3130 and 2537 fi. radiation. In the work reported here the room-temperature photooxidation of acetone vapor was studied a t low pressure and with incident radiation extending to 2200 A. Under these conditions CH3 and C H 3 C 0 radicals were produced and their subsequent oxidations were studied. Experimental T h e apparatus and experimental procedures have been fully described in the preceding articles of this series.'O Twelve series of runs were made with variation of the oxygen pressure and acetone pressure (and thus the absorbed intensity). The incident intensity was similar in all cases except series 5 . The incident radiation passed through a Corning 9-54 glass befofe entering the reaction cell to remove all radiation below 2200 A . Replicates of all series were run and agreement was to within 10 or 157, except in a few scattered cases. Matheson tank oxygen was used and impurities were 0.37, argon and 0.77, nitrogen. Eastman Spectrograde acetone was used. I t contained no impurities in amounts greater than 0.1%. Because of the background cracking peaks, analyses could not be made for COa (441, CH2O (421, C 0 . ( 2 8 ) , and CHa (16). (1) F . B. Marcotte and W. A. Noyes, Jr., Discussions F a r a d a y Soc., 10, 236 (1951). (2) F. B. Marcotte and W. A . S o y e s , J r . , J . A m . Chem. Soc., 74, 783 (1952) (3) 11. E. Hoare, T r a n s . F a v a d a y S o c . , 49, 1292 (1953); J. Caldwell and D. E . Hoare, J . A m . Chem Soc., 84, 3990 (1962). (4) M . I. Christie, i b i d . , 76, 1979 (1954). ( 5 ) J. R. Dunn and K . 0. Kutschke, C a n . J . Chem., 86, 42 (1958). (6) R. Srinivasan and W. A. Noyes, Jr., J . A m . Chem. Soc., 82, 5591 (1960). (7) A. D. Osborne, J. S . Pitts. Jr , and S. L. Fowler, J . P h y s . Chem., 66, 1622 (1961). (8) A. D . Kirk and G. B. Porter, ibid., 66, 556 (1962) (9) G. S. Pearson, ibid., 67, 1686 (1963). (10) J. Heicklen and H. S. Johnston, J . A m . Chem. Soc., 84, 4030, 4394 (1962).
During the course of the experiments the gold foil in the reaction cell was inadvertently damaged, and had t o be replaced. T h e gold foil associated with each series is included with the data, Tables 1-111. From the data in Table 111, it is apparent t h a t both gold foils have similar sized pinholes as the leak rate is the same for both.
Results During irradiation product peaks were observed a t m/e = 17, lS, 30, 31, 46, 48, 60, 62, 72, 73, and 74. The first eight peaks are readily identified with HzO, CH20, CH30H, HCOOH, C H 3 0 0 H , CHsCOOH, and CH300CH3. The last three peaks are not so easily recognizable, but in all likelihood they are from CH3C ( 0 ) C H O and CHaC(O)CH20H. The absolute pressures of the products were estimated from calibrations, comparison with the literature, and interpolations among related compounds. However, it is re-emphasized t h a t some of the calibration values may be in error by as much as 30 or 40%. The experimental results are listed in Tables 1-111. For all of the products there are listed the initial rates, Ri ; the steady-state partial pressure, P s s ; the half-time of build-up to the steady-state pressure, T L ; and the half-time of decay from the steady-state value after the light was turned off, T D . The results can be summarized in phenomenological terms, using the classification of part I of this series, without reference to mechanism. (1) In decreasing order of abundance the initial measured products are H2C0, HzO, CH30H, CH300H, CHaCOOH, HCOOH, CH3COCH0, CH&OCH*OH, and CH300CHs. (2) Relative quantum yields for important products are plotted as solid circles in Fig. 1 and the corresponding quantities found for the photooxidation of methyl iodidelo are given as open circles. These relative yields are plotted on a log-log scale against the ratio of reactant to oxygen. The relative quantum yield of CHsOH is about the same in the two systems. The yield of HzCO is almost twice as great in acetone as in methyl iodide, but CH3OOCH3 is much more abundant in methyl iodide than in acetone. The biggest difference between the two systems is in the rate of formation of H20, the yield being almost ten times as great in acetone as in methyl iodide. No CHBCO(OOH) was observed, and both CH3COOH and HCOOH were very minor products. The relative quantum yields are not sensitive functions of the reactant-oxygen ratio; however, the yield of CH300H and HzO
4250
HAROLD S. JOHNSTOS
XNL) JULIAN
HEICKLEN
VOl. XG
TABLE I 1 N I . r I A L R A T E S O F I;OKMATIOh' O F P K O D 1 : c r S ~~...~ ---.Sei-ies~~ .....
.~
,
~~~
1
9 7 0 25 1
[Ozl" [CHyCOCH,]a Gold foil (Ro X 1061)/[CH3COCH~1' HD CHzO CH30H
1.0 0 23 1
a =
9 3 0 7 2 27 49 13 8
47 34 16
HCOOH CHiOOH CH3COOH CH300CHs CH3COCHO CHBCOCH~OH ( R > ( Z H )X lofi)/ 2[CH3COCH31h
4
3
2
2 5 0 9
->
18.9 39 11.7
16 7 1 2
... 12.1 1.5 ...
0.73
5
I 0 0 9 1
0 8 0 85
-1
21.5 42 15.3
241 260 75 4.0
3 2 11.3
10
9 2 3.6 1
2.6 3 5 2
0 9 3.9 1
15 6 39 13 9
24 43 12.2
16 0 39 18.5
24 44 %8
...
...
...
14 5
11.6 1.1 0 83 0 83
7.7 2 1 14 0.5
. . ,
1,12
9
S
...
11.7 1 2
44 ...
~-~ -
7
A
0 65
..
...
0.09 3.6 1
6
11
12
0 9 16 2
1.0 17 1
18 2 41 26 16 7
17 6 37 1% 2
8 4
8 8
0 95 0 56
0 83 0 46
...
1.7 , , .
129
133
109
782
123
116
122
118
144
116
103
59
79
68
425
76
72
74
76
87
70
63
70 40
51 39
41 37
357 253
47 40
44 38
48
40
42 38
57 46
46 37
40 33
39
38
36
249
39
37
39
37
45
36
32
{ l/$l(ZH) - R,(CHBCOCHO)- R,(CHaCOCHzOH) /[CH3COCHa]
Millimeters
5
Set.-'
c
Calculated from difference in mass balance of hydrogen and carbon
tended to decrease and CH30H tended to increase with increasing reactant-oxygen ratio. (3) The curves of growth of many of the products fall into case 1 of article I (i.e., the initial product is unaffected as ir-
composes or is attacked). (4) In the dark, CH300H decomposes but all the other products are stable. Discussion The Primary Process.-The primary process in acetone in the absence of oxygen is complicated. A molecule excited by radiation can either decay into free radicals soon after excitation or it may pass through low vibrational levels of the relatively longlived triplet state before decomposing.l 1 This latter process is inhibited if biacetyl is present" or altered if oxygen is present.16-9 However, it has been shown that a t low pressures (but still greater t h a c those used here) and with incident radiation of 2800 A. (and presumably below this) the long-lived triplet state is not important.I2 From the spectral distribution of the lamp, the transmission characteristics of the Corning glass, and the absorption coefficients of acetone, l 3 i t was estimated that about i5yo of the incident radiation was at wave lengths below 2800 in this work. T o confirm this estirnate a photooxidation was performed using a Corning 9-53 glass which only passes The decomposition in this radiation above 2600 case was about one-third t h a t obtained when the 9-54 glass (which passes radiation to 2200 was used. Consequently i t is probable that the triplet state oxidation is not dominant, but it may occur to some extent in this work. The other decomposition might give either methyl and acetyl radicals or carbon monoxide and two methyl radicals. Soyes, Porter, and Jolleyl" have estimated from the data of Herr and NoyesIb that about SO?:
a.
a.
4-
CH,@@H
2-
001
I
!
"I
I
I
I I 1
I
~
I I 1
I
I
I
I
1
radiation proceeds) However, H 2 0 is an example of case 4 (initial product is also produced from other product) and H 2 C 0 , CHJOOH, CHICOCHO, and CHJCOCH,OH are examples of case 2 (initial product de-
( l l j J Hcicklen anti u' A N o y e s , J r . , J . A m . C i i r i n Soc . 8 1 , 38hS (1959).
Hcicklrn,
(12)
i b M
81, 3863 (19G9).
IT,Porter a n d C I d d i n g s . ibigi.. 48, -10 (19261 IV, A Xoyrs. J r , Cr H Porter. and J E . Jolley, Chriii. Rev,, 6 6 , 49
113) C .
(1%
(1RBB). ( 1 5 ) I ) . 3. Herr and
IV. A Noyes, J r . .I. A V . C h , 'VI. Sor , 62, 2 0 5 2 i I O l 0 ) .
PHOTOCHEMICAL OXIDATION OF ACETONE
Oct. 20, 1964
425 1 m
L?
m
a0IC - 3
M 0
$1
*
0 N
co 3 $I
rC m
m N
0 N
2c
t-e i o
0
3 0
N
ax.
m tc
d
OD*
a m
N N N O
5:
m -+
c
oooc
IC t-
3 0
*
m
t-
t-t-
i
t-
N O D
OD
OD-
X
coo0
m N N m
'5:
Mu:CO*
i
$1 2
% C V % &
-
N 0
N y:
x m
N
at-
0
m
$1
z
H o e d
N 2
+
m N i 2
0 2 2 2
N 3
a c m -0-c oca0
N
1
V
m a r * x m c O O O O
=
v
m x m
t.1xt-m
% c? X
i L? i
i c o c
V
d
m
t-rr)
i
o
c 2 3 m x.0 m - d N 2 0 3 0
v v v
t-
a i
c C
i
H
d C
d C
0 a N
m 0
$1
of the time a methyl and an acetyl radical are formed with 2537 A. incident radiation. *It longer wave lengths this percentage should be larger. However, since then it has been found by Roebber. Rollefson, and Pimentel, l 6 and recently confirmed by Shilman and Marcus,Li that in a flash photolysis with 2800 k . radiation acetone decomposes to two methyl radicals and a carbon monoxide molecule. In order to see what the primary split was in our system, acetone a t pressures varying from 1 to 15 mtn. was photolyzed in the near absence of oxygen (unfortunately, a small leak prohibited the complete removal of oxygen). Large amounts of biacetyl were produced as could be deduced from the mass spectral product peak a t m,'e = X(j. The results indicated t h a t a t least -i a G of the primary split was to methyl and acetyl radicals. Furthermore, the mass spectral peaks corresponding to (CH3COCH2)?,C H R C O C H ~ C O C Hand ~, CH:{COCHrCH:jwere completely absent. thus eliminating the possibility of a primary process leading to hydrogen atoms and acetonyl radicals. The previous discussion indicates that the most important primary process iri this work was CHiCOCHa
-C
hv
+ CH3
+CHsCO
+ R,(CO?'I = 0 . 5 R , ( Z € I )
-
K,(ZC)
(2)
where Ri(ZH) and & ( BC) are, respectively, the initial rates of formation of all the hydrogen atoms and carbon atonis in the products. The appropriate quantities are listed in Table I and it is apparent that CO and 'or CO, are indeed major products. If the assumption is made that all the products except CH:+C(O)CHO and CHaC(0)CHnOH result from the primary decomposition, then the rate of acetone decomposition, a[CH:+COCHa], can be estimated from a
=
[ ' ',IZ,(H)
-
R,(CHxCOCHO! I? (CHICOCH?OH)],'[CH~COCH~] (3)
The values of CY. which are proportional to the incident intensity, are listed in Table I . Except for series 3, the intensity is similar in all runs; in series 5 it is some eight times larger. The rate of CO and CO2 produced from the carbonyl of the acetone molecule is then given by n[CH:jCOCH.,] - Xi(CH3COOH). These values divided by the acetone pressure are also listed in Table I . .It high pressures niost. but not all, of the C O arid C O , conies from the carbonyl carbon. However, a t low pressures, a large fraction of the CO and CO:! comes froni the end carbons of the acetone molecule. The relative quantum yield is the rate of formation or^ a given product divided by a[CH&OCH3]. The large yield of HZO in this system indicates that HO and i l f i ) J . I , K o e h h e r , C, 50,- . 8 0 , 2.5.5 ( I 9.78'
K
R < ~ l I e f w nand G C . P i r n e n t e l . .I
, 1 7 1 A Shilrnan a n d R . A M a r c u s , .I. C h r m P h y s
surface
A m (~'kunz
39, O!)H ( l 9 0 3 )
-+ HICO
CHaOOH
surface
Hi01 W HJO
+ H10
+ 0 501
(4) (5)
In the photooxidation of methyl iodide in this apparatus the main source of H2O was thought to be from HO radicals produced directly by oxygen and methyl radicals CH,
-
+ 01
H1CO
+ HO
(6)
In the photooxidation of acetone the very large increase in H20 suggests that either oxygen reacts directly with photoexcited acetone
+0 2
(CH*COCHa)*
---f
HOr
+ CHICOCH,
('7)
+ H?CO + COCHI (8) --+ H O + H?CO + C O + CH3 ( 9 ) +HO
or that oxygen attacks the acetyl radical in a reaction analogous to eq 6 CHjCO
(1)
Initial Oxidation.--In this system CH4,CO, CH2C0, and CO,could not be measured. From the literature it is known that small amounts of oxygen completely suppress methane formation in systems with methyl radicals. On the other hand CO and COa are undoubtedly present. Assuming that C H 2 C 0 and CHr are absent, the initial rate of production of CO and COZ can be estimated from the carbon--hydrogen mass balance. The relationship is R,(CO'I
or HOz radicals are formed as intermediates, but it is known also that H?O is rapidly produced by the heterogeneous decomposition of CH:jOOH or H 2 0 2 ,
+ On
.--)
H,CO
+ H O 4 CO
(10)
A reaction which has often been postulated for CH,CO radicals is CIlrCO
+ 02 +CHjO + CO,
(11)
However if this reaction were dominant in our system, the results would be similar to those in CHjI and it would be difficult to explain the large excess of water and formaldehyde In other kinetic systems acetyl peroxides and acetyl peroxy nitrites are forinetl, indicating that acetyl radicals do indeed add oxygen molecules
+ 0,
CHiCO
---f
CHiCO(O0)
(12)
No acetyl hydroperoxide and very little acetic acid were found here Thus the acetyl peroxy free radical must either be very stable or very vulnerable to attack in this system If it is highly stable, its fate is probably heterogeneous destruction on the walls of the cell surface
CH;iCO(,OO)
__f
H?CO, HrO, CO, . . .
(13)
If i t is very vulnerable to attack in this system, the large yield of methyl hydroperoxide suggests CII3OO
+ CHjCO(OO) +CH:iOOH +
(14)
and similar attack by other radicals. The large yield of CH300H and the small yield of CHPOOCH:{, relative to methyl iodide oxidation, strongly calls for some very easily abstracted hydrogen atom CHI00
+ HX
----f
CHsOOH
+X
(13)
to compete strongly with the formation of inethoxy
radicals from peroxy methyl radicals 2CH300
+2CHaO + Or
(lfi)
At this low temperature the substance H X with an easily abstracted hydrogen must surely be a free radical and ey. 15 must be a radical -radical disproportionation reaction. A very likely candidate for H X would be H 0 2 which could be formed from the reaction KH.
+ 01---f I