4030
JULIAN
[CONTRIBUTIONFROM
THE
HEICKLEN AND HAROLD S. JOHKSTON
Yol. s4
DEPARTMENT O F CHEMISTRY, UNIVERSITY O F CALIFORNIA, BERKELEY 4, CALIF.]
Photochemical Oxidations.
11. Methyl Iodide
BY JULIAN HEICKLEN AND HAROLD S. JOHNSTON RECEIVED MAY31, 1962 The room-temperature photo-oxidation of methyl iodide (0.2 to 3.0 mm.) in oxygen (0.030 to 10 mm.) with continuous ultraviolet radiation above 2200 A. has been studied by the method outlined in article I of this series: observations were made by leaking the reaction mixture directly into the electron beam of the mass spectrometer during photolysis. The principal products of reaction were Iz, HzCO, CHaOH and under some conditions CHaOOH; minor products were HzO, COZ,HCOOH, CH300CHa and CHaOI. Because of the cracking pattern of the reactants and major products and the background air peaks, i t was impossible to establish the presence or absence of CH4, CO and HI. There are conflicting claims in 0 2 ( + M ) -* CH~OZ( M ) (followed the literature as t o whether methyl radicals react with oxygen according to CHI HO (followed by HO attack on loosely bound hydrogen atoms). This by the Vaughn mechanism) or CH3 02 + H,CO study indicates both processes do occur, with the first being more important under conditions used here. Furthermore, it seems probable that oxygen molecules abstract hydrogen atoms from CH3O radicals to produce H C O and HOZ. A fairly complete and internally consistent mechanism is developed for the initial reaction, typically the reaction of about I2 -* CHJ I becomes very proor 10-3 of the methyl iodide. As the reaction progresses, inhibition caused by CHI nounced, radicals abstract from HzCO, a large number of other secondary reactions seem to occur, and the mechanism proposed is regarded as exemplary rather than established. In terms of the relatively simple initial reaction, many ratios of rate constants are evaluated.
+
+
+
+
+
Introduction For convenience in reference, all mechanistic steps considered in this article are presented here
+ + + + + +
+
CHoI h v +CH3 I 02 M +CH302 M I I M +1 2 -t M 1 2 +CHoI I CHr 0 2 2CHsOz +2CHaO CHzO 2CHaO +CH3OH 2CH30 +CHaOOCHa CHo02 CHaO +CHsOOH CHzO CHpO I M +CHaOI M I +CHzO HI CHaO CH3
+
+ + +
+
+ + +
+ +
+ CH3OOH -+ CH20 + Hz0
(a) (b) (c) (d) (e) (f) (9) (h)
(i) (j1
surface
surface + HzO + CHsOH, adsorbed oxides of I (m) CHJO + H I +CHaOH + I (P) CHI + CH20 + HO (4) HO + CHsO +CHzO + H20 (r) HO f CHsOO HzO + CHzOO (+ COz, CO, HCOOH) (s) HO f CHzO +H20 + CHO (t) CHIO + CHzO +CHsOH + CHO (u) CHIO + +CH,O + HOz CHsOz + HOz + CH30zH + (w) CHO f Oz+ OH + COz HOz + CO CHO + Oz+ (Y) 0 '2110 + + hI+ M + €1-C ( 0 0 + C&OI
0 2 ----f
----f
0 2
(VI
0 2
(X)
0 2
HCOO.. . --+ HCOOH (z) HCOO f CHsO +CHaOH COz (f')
+
The first thirteen steps were used in explaining the data on the photochemical oxidation of ethyl iodide in article I of this series,' and these steps will be re-examined in the present case. Also in ref. 1 there were given key references to the literature of methyl radical oxidation. This mechanism (1) J. Heicklen and H. S.Johnston, J . A m . Chem. Soc., 84, Nov. 20 (1962).
+
represents addition of molecular oxygen to the free radical t o form peroxy radicals and the subsequent products of these radicals. The direct bimolecular reaction of oxygen with methyl radicals t o produce hydroxyl radicals, step q, was proposed long ago by Bates and Spence2 and directly demonstrated by McKellar and Norrish3 a t 140'. Steps r to t and in part x to z indicate the products expected to follow from step q. The other steps are proposed on a more or less ad hoc basis t o discuss various minor products or secondary effects in the data., and these steps will be discussed later. Experimental The apparatus and experimental procedures have been described in article I of this series.' Eight series of runs were made with variation of the oxygen pressure and methyl iodide pressure (and thus the absorbed intensity). The incident intensity and spectral distribution of the radiation were the same for all runs. The incident radiation passed through a Corning 9-54 glass before entering the reaction cell to remove all radiation below 2200 A. The various products were analyzed in series, with some drift in the instrument and some variation in reproducing the initial chemical composition. Duplicate series were run in some cases. The final values of initial rates, half-lives and steady-state pressures of products (relative to diffusion through the pinhole) are estimated to have a standard deviation of about 10 to l5Y0 except for the unfavorable cases (e.g., water). Though this error is large, it must be remembered that the steady-state pressure of products was be, the order of magnitude of 0.01 to 30 tween 0.01 and 2 0 ~ or millionths of an atmosphere. Matheson tank oxygen was used, and impurities were nitrogen. Eastman Kodak Co. 0.3 % argon and 0.7 white-label methyl iodide was used. The cracking pattern of a methyl iodide-oxygen mixtures is given in Table I . Impurity peaks occur a t m / e 41, 43 and 254 in amounts of less than one part in 1000 compared to the 142 peak. The 254 peak belongs to molecular iodine which is present in one part of Iz to 4000 f 700 parts of methyl iodide. The peaks a t m/e 121 and 135 have not been identified, but they are probably unstable ion peaks rather than impurities. As the runs progressed, products appeared a t a large number of mass numbers. The mass spectra and sensitivity relative to oxygen were determined from the literature, from calibrations performed in this Laboratory, and estimated by analogy for the unstable products. Only in the case of CHJOOH did the cracking pattern have to be eqtimated. Only two cracking peaks of this coinpound were of (2) J. R. Bates and K. Spence, i b i d . , 63, 1689 (1931); l'rans. Foro(1931). (3) J . F. hlcKellar and R. G . W. Norrish, Proc R o y . SOC.(L@~zdo?t). 8263, 51 (19G1). d a y Soc., 27, 468
Nov. 5, 1962
PHOTOCHEMICAL
OXIDATION O F hfETIIYL IODIDE
TABLE I MASSSPECTRUM OF CHaI-02; CHII = 3.0 MM.; 0 m/c
Rel. height
Identification
m/e
Remarks
Rel. height
2
4031
= 10.0 MM.
Identification
Remarks
12 1.19 C+ 63'/a 0.04 I++ 13 2.82 CH + 70 .I4 ... In CHBI 14 7.2 CHp+ 71 .36 CHI++ CHI 96 .02 ... Background in mass spec. 81 15 O+ 16 ... 97 . 01 , . . Background in mass spec. OH + HzO in mass spec. 121 .32 ... 17 ... Unstable ion in CHaI 18 0.07 HzO' H20 in mass spec. 127 40.5 I+ h-2 Air in mass spec. 135 0.80 ... Unstable ion in CHsI 28 3.3 .. Unstable ion in 0 2 139 5.21 CI+ 301/a 1.81" 170 0 2 140 Shoulder CHI + 32 34 1.09 .. 141 Shoulder CH21+ Ar+ In 0 2 142 100.0 CHJ' 40 0.68 .. In CHsI 254 0.03' Iz+ Is in CH3I 41 .03 .. In C&I 43 .06 coz + In 0 2 44 .04 45 0-0.05 0 The ratio of m / e 301/3:32 diminishes rapidly with Oppressure and is essentially unobservable below 0 2 pressures of 1 mm. ; The ratio of CHsI and IZ pressures is 4000 i 700. +
+
+
*
Series [Ozl. mm. [CHoI], mm. Ri X 106/[CHJ], set.-' HzO CHoO CHaOH
TABLE I1 INITIAL RATESOF FORMATION OF PRODUCTS 1 2 3 4 5 ti 9.4 2.0 9.5 9.4 2.9 0.90 0.19 0.19 0.75 2.8 2.9 2.8
.. ..
..
7 0.27 2.9
0.080 3 ,0
8 9 4.0 .. .. 61 66 29.3 24.1 10.8 45 40 19.0 18 5.5 73 eo2 -7 .. 1.0 are given in Table IT.
+ k[L1/
(4)
4034
JULIAN
HEICKLEN AND HAROLD S. JOHNSTON
Vol. s4
TABLE V MASSBALANCE DATA n 2 3 4 1 5 Series G 8 9.4 2.0 9.5 9.4 2.9 0.90 0.27 0.030 [Ozl. mm. 0.19 2.9 0.19 0.75 2.8 2.8 2.9 [CHJ], mm. 3.0 139 118 145 -142 156 -62 -44 24 Ri (21) X 10fi/[CH31],set.? 162 140 ... 133 145 65 48 19 Ri (ZC) X 106/[CHJ], set.-' r107 i l G9 31 -24 -7 > 105 >58 h 64 1.6 -80 83 ... 69 31 26 12.7 17 24 15.8 32 10'Pm( 21)/ [CHJ ] 1.32 5.6 3.3 14.4 20 31 18.8 6.4 4.4 2.4 39 1 0 3 ~ ~ ~ ( 2[CHd c)/I 32 79 50 15 11.1 6.5 10fiRas( ZI)/[CH31],sec.-l 99 2.8 140 66 97 59 17 12.0 106R,,( ZC)/ [CH,I], set.-' 6.1 45 18.1 10.4 >13.7 8.8 21.0 >8.4 3.8 >1.8 d 18.5 10.4 7.6 1.85 >9.6 25.6 4.1 3.0 0.91 0.86 -0.42 0.123 -0.31 ... -0.94 1.05 +i(C), eq. 6 0.90 .77 1.01 .94 ,156 .92 .40 .29 %(I), eq. 7 ,29 0.91 .ll ,078 .44 .63 .38 .040 +sB(C),eq. 6 .10 .072 .042 0.64 .018 .21 .51 .32 %(I), eq. 7 106[R,(C:120) 4- 3X,(C02) Ri(HCOOH)][CH31], 1O6[2II,(HzO) Ri(CH30H) RR,(CH~OOH)]/[CH3], see.-'. sec. -l. lo3[P,J CHrO), 3P,,( C02) 4- Pea(HCOOH) / lo3[ape,(HzO) P,, ( CHIOH) f Pas(CH3COOH)]/ [CHf] : [CHJ]. a vs. * gives hydrogen carbon balance for initial rates; os. gives same information under steady-state condltions .
+
+
+
+
The quantum yields are plotted in Fig. 2 against the ratio [CH31]/[02]for all series for both initial rates and steady-state (with respect to leak through pin-hole) rates. The initial quantum 1
I
tc
t
Christie5 evaluated the ratio of rate constants, k b / k d , 0.53 X i o 4 cc./mole a t room temperature. The relative eficiency of CH.1 and O2 as foreign gases can be estimated from the boiling point correlation found for reaction c by Russell and Simoiis,fi and the equivalent pressure of methyl iodide is given in Table VI by [MIc = [ C H J ]
0
rl
+
+ 0.037[02]
(9)
I n these experiments the pressures of oxygen and methyl iodide were known by synthesis, and steadystate pressure of iodine was measured in the mass spectrometer. Thus every term in eq. 8 is known except the ratio of rate constants, k , / k d . This ratio was assumed to be zero, and curve A in Fig. 3
Fig. 2.-Relative quantum yield as a function of ratio of reactant pressures: 0, based on initial rate of formation of all products containing iodine; 0 , based on steady-state rate of all products containing iodine; A, initial rates of carbon-containing compounds; A, steady-state rates of carbon-containing compounds.
yields based on carbon are in excellent agreement with those based on iodine, and this result indicates that non-observed products (CHa, CO, HI) are negligible (perhaps excepting series 8) or that CO and HI, for example, are produced a t equal rates. The initial-rate quantum yield decreases with decreasing oxygen and is far below unity for the last three series. The steady-state quantum yields are far below those for the initial rate, as one expects from the competition of step d with steps b and q. The quantum yield a t steady state based on all carbon compounds is substantially greater than that based on all iodine compounds. Thus a t the steady state a significant amount of iodine is not observed; either H I or (more probably) the purple deposit is noted on the gold foil. Thus the analysis of steady-state data is based on the yield of carbon, not iodine. Equation 5 can be re-written as
Fig. 3.-Relative quantum yield based on steady-state rate of formation of all products containing carbon, eq. 8, for various assumed values of kq/kd: A, 0 ; B, 0.5 X D, 2 X The irregularity in the curves C, 1 X arises from the effect of the independently varied total effective pressure, M.
was prepared. I t is seen that the calculated curve is in serious disagreement with experiment. Next i t was assumed that k q / k d is: 0.5 x lo-', curve I3; 1.0 X curve C ; 2.0 X curve D. ( 5 ) M. I. Christie, Proc. R o y . SOC.(London), '244A,411 (1058). ( 6 ) K. E. Russell and J. Simons, ibid , 2 1 7 8 , 271 (1'353).
Nov. 5, 1962
PHOTOCIIEMICAL
OXIDATION O F METHYL IODIDE
TABLE VI INTERPRETATION OF INITIAL RATEDATA 1 3 2 Series 4 5 6 9.4 2.0 9.5 9.4 2.9 0.90 1 PzI, mm. 2 [CHsI], mm. 0.19 0.19 0.75 2.8 2.8 2.9 3 Rt( CHsOH)/Rt( CHaOOCH3) 8.5 10.7 8.7 9.8 8.9 9.3 4 of X 1O6/[CH3I], sec.-l 110 65 57 26 74 64 5 l.h.s.,“ eq. 16 0.18 0.32 0.29 0.147 0.43 0.23 6 [ 0 2 ] 1 ’ 2 of (mm.-sec.)*’2 570 1280 230 2000 710 108 9 6 5 5.5 7 l.h.s., eq. 18, mrn.-l .. .. 12.7 3.4 1.0 0.32 8 [Oil/ [CHaI 1 50 10.5 1.10 3.15 3.01 2.84 9 [MI1,.mm., eq. 9 0.54 0.26 0.158 0.092 0.061 10 l.h.s., eq. 19, rnm.-1’2 0.33 .. a 1.h.s. means left-hand side.
Curve A predicts quantum yields too low everywhere, and curve D gives quantum yields too high in all cases. The scatter of data is such that i t is difficult to choose between curve B and C. Each gives a fairly good representation of the data. Thus one concludes that reaction q certainly occurs in this system and that the value of k,/kd is between 0.5 X and 1.0 X From the distribution of products (see below) the lower figure is to be preferred. The quantum yield based on initial rates falls off a t low oxygen, Fig. 2 , for series 6, 7 and 8. For these runs, the yield of products is small, the background noise in the instrument is as great as ever, and a re-examination of the data shows the “initial rates” not to be truly initial, after all. The true initial rate is Ri = lim dx/dt as t
-*
0
(10)
where x is the pressure of a product. The small yields of products and the noise level are such as not to support the detailed analysis of eq. 10. The observed “initial rates” are ‘IRi”
= (x
- %)At - to)
(11)
4035
7 0.27 2.9 522 23 0.147 33
8 0.030 3.0
..
..
..
6.5 0.12 6.8
0.093 0.010 2.91 3 . 0 0
in this system. Hydroxyl radical is particularly reactive, and step t, abstraction of hydrogen from formaldehyde, may be important as well as disproportionation reactions such as r. On the other hand, methanol is a very important product in this system, reaction q does not lead to methsno!, and reaction b and its sequels also occur. The simultaneous occurrence of reaction b and reaction q, with the sequence of products formed in each case, is consistent with fact 2 above. The third fact is most easily explained by adding steps v and w. The HOz radical might also donate its hydrogen to CHsO, I, OH or another HO However, an easily abstracted hydrogen will go to the most abundant, not the most active radical; thus i t seems reasonable that w is the most important reaction of HOz. At first thought it might appear that the H 0 2 radical might also produce H20. However, this possibility must be ruled out for the following reasons: (1) the relative rate of water formation does not increase with the oxygen pressure; (2) to form water the H01 radical would have to produce H202 as an intermediate. No H~02was found. Even if HzO2 were present, water could not be an initial product from this source unless the heterogeneous decay rate of HzOz was so fast that i t reached its steady state value within about 10 or 20 seconds. Because of the good mass balances found for the initial rates, reactions j and thus p, are regarded as unimportant in this system, unlike the case for ethyl iodide. The free radicals reach their steady state very fast (estimated half-lives of 10-4 to second) compared to the steady state of products relative to diffusion through the pinhole (200-3000 seconds). Thus the steady-state analysis with respect to radicals can be applied t o the initial-rate data of the products. As in article I, the rate constant for reaction b, for example, will be written as k b ; but the rate of the reaction will be abbreviated by use of the symbol b, that is:
.
In the last three series, significant amounts of iodine had built up during the time over which eq. 11 was applied. The data in series 6, 7 and 8 are average rates over an early portion of the reaction; the distribution of products inside each series is a valid datum, but no comparisons of initial rates between series 6-8 and series 1-5 may be made. The constancy of quantum yield over the wide variation of conditions in series 1-5 indicates that trends from one series to another may be considered here. Initial Reaction.-The mechanism proposed for the initial reaction in the photo-oxidation of ethyl iodide does not account for all the products. It predicts that CHzO equals the sum of CHaOH and CH300H with no initial H ? 0 or COP. In fact the following important differences are observed in the CHd-02 system: (1) H20 and CO2 are definitely initial products; ( 2 ) Ri(CH30H) Ri b = kb[CHal[021[MI (12) (CHnOOH) < Ri(CH20) < Ri(CHa0H) RiI n these terms the initial rates of formation of (CH200H) 2Ri(H?O) in all cases: (3) the relative amount of CHzOOH increases markedly with products are the oxygen pressure. Ri(CH20) = f h q 4-r -/- v Water is formed from HO radicals produced b y R,(CH,OH) = f reaction q, and thus the initial products are in Ri(CH1OOH) = h w agreement (at least qualitatively) with the conclusion reached above that reaction q is important Ri(Ht0) = q
+
+ +
+ +
+
RL(CH3OOCHaj = g R,(CH,OI) = i
12,
,
,
,
i
i
j
“
‘
i
’
I
RdI2) = c R,(C04) s s
Other minor initial products, such as HCOOH, are omitted, since these may be rapidly-appearing secondary products (when present in very small amounts it is extremely difficult to establish -
7
i
Fig. 5.-Test
of eq. 18 and the evaluation of the ratio 0 2 and
Kh/K, from the intercept and the relative efficiency of CH31 as c a t a l > ~ tfor s reaction b. ,
2TC
W 2 EGC
600
1021, e’’,
*
1000 ‘ 2 0 3 ‘4OC 6 m ;
rr
’
H q - 5 ec
I ,I
Fig. 4.-Test of eq. 16 and the evaluation of the ratios of rate constants; slope, k y / k r 1 / 2 ; intercept, k h / ( k f k o ) 1 / 2 : 0, series 1-5; 0 , series 6-8; A, data of C&I-02 system; lower curve, CH3I-02; upper curve, C2HiI-02.
more, this value is close to the value of 12 1 found for the ethyl system.’ The competition between reactions f and v gives the relationship
where e is given in eq 13. Ignoring the term in CO which could not be measured, e can be computed and is listed as entry 4 of Table VI. The left-hand side of 16 is listed as entry 5 and the ratio [ 0 2 ] / e ’ / z R , ( I ) = 0 = a - 2c - i (13) is listed as entry (iof Table 19. Equation 16 preR,(CHa)= 0 = a - b - q dicts that a plot of X,(CH?OOH) (R,(CHIOH)e]’/z R,(CH,OO) = 0 = b - 2e - h - w t v s u s [OZ]‘e”2 should be linear Figure 4 is such a plot, and the expectation is verified. The interRi(CH30) = 0 = 2e - 2(f g) - h - i - r - v cept gives k h ’(kfke)’/2 = 0 14 and the slope gives R,(HO) = 0 = q - r - s k , ’kf = 1.44 x 10-4 (mm -set)-'/? or 0.59 (cc./ R((H02) = 0 = v - w mole-see.)"?. (The C:H51 data of article I are The rates of reaction in terms of products and rate shown on the same plot; in this case the data are constants are badly scattered owing t o the difficulty in analyzing the CzH500H.) The correlation shown in Fig. 4 a = R,(ZC) = Ri(Z1) (14) is the principal evidence in favor of step v, the abb = R (ZC) - R,(HsO) straction of a hydrogen atom from a free radical c = &(I*) by molecular oxygen. e = R,(CH30H) R,(CHIOOCH3) + 1/Z[R,(CH300H) In the section above on over-all quantum yield, Ri(CH3OI) Ri(H20) - Ri(C0) - Ri(C0~)l the relative efficiency of oxygen to methyl iodide f = Ri(CH30H) as an 34 gas for reaction b was taken to be 0 08’7, from the relative efficiencies for reaction c as found g = Ri( CHaOOCH3) by Russell and Simons. The data obtained here h = !kh/k,1’2kr1’2)e1!*(R;(CHIOH)} l’2 give a direct estimate of the relative eficiency of i = Ri(CH3OI) oxygen and methyl iodide for reaction b and also q = R,(H?O) for reaction i. Instead of eq. 9, one can write
whether a product is truly “initial”). Making the steady-state assumption of zero net rate for formation of each radical, one finds
+
+
+
+
R,(CII,3OOH) - h Ri(C0) r = Ri(H20) - Ri(C02) - R , ( C O )
v = w
s = Ri(CO2)
[M]b =
+
The mechanism predicts the constancy of several ratios of constants. The division of products between f and g should be constant if g is second order, not dependent on [MI. The ratio R,(CHsOH)/R,(CHsOOCH3)
[CHJ] { l $. Olb [02]/[CH31])
(17)
Competition between reactions b and q gives the re1at’ion
- kf/kg
(15)
is entry 8 of Table VI. Except for the unreliable value for series 7, the ratio varies only from 5.5 to 10.7 indicating that f and g are the only sources of CHaOH and CH?OOCH3, respectively. Further-
Entry 7 of Table V I lists for the left-hand side of eq. 15. These values are plotted against [O?]‘ [ C H J ] in Fig. 5 , The ratio of slope t o intercept for step b gives a value of C.OG for ob,in fair agreement with 0.03T from Russell and Simons. The intercept yields a value of 5.0 for k b / k q or 1.0 X 10s cc./mole. By combining this value
4037 TABLE VI1 STERPREI’ATlOh’ OF STEAI)Y-STATE 1)A.I.A E n t r y no.
1 2 3
Series [Or], mm. [CHJ], mm. k, X lo3 set.-'
5
k , X lo3, set.-'
1 9.4 0.19 5
19
2 2.0 0.19 4.5
3 9.5 0.75 0
..
5 2.9 2.9 7
4 9.4 2.8 2.6
21
17
24
12
12
30
6 0.90 2.8 4
TABLE VI11 RATESWITH CORRECTED ISITIAL RATES 4 3 5 6 9.5 9.4 2.9 0.90 0.75 2.8 2.9 2.8
COMPARISOS O F STEADY STATE
1 9.4 0.19
2 2 0 0 19
3.0 3.1 33 28 26 32 41 22 26 -60 22 19 (6.2)“ 1.64 -6 0 42 . . 0.13 .. .. 0.14 (6.6) (0 7) (6.2) (2,6) -25 4 6 11.1 6.6 6.1 2 2 2.5 3.9 -6 2 1 1.95 3.2 (0.69) (0.83) (0.58) -5 3.5 3.2 50 16 25 39 40 -60 31 23 Note: Parentheses indicate that lifetime used for the computation was ..
.. SO ..
5.4 12.6 40 47 37 28 (3.8)
