PHOTOCIIEMISTRY OF NITROGEN DIOXIDE
Nov. 5, 19G2
DEACTIVATION
OF
TABLE I1 EXCITEDtert-BuToxY RADICALSBY SEVERAL GASES
T = 26', X = 2537 A. Deactivating gas
1020 k i / k z , ml./rnolecule
( CHa)aCONO"
5.6
NO
5.0
Rel. deact. efficiency on coll./coll. basis
CzHs
2.4 1.1 At 79", ki/ks = 1.2 X 10-20ml./molecule.
sz
1.0 1.1 0.46 0.23
The range of efficiencies from a diatomic molecule to a polyatomic molecule is quite small. Such a narrow range is reminiscent of that found for deactivation of molecules excited to very high vibrational levels. 23 Nitric oxide has an unexpectedly high efficiency for a diatomic molecule. This indicates that 4 is probably not a simple deactivation step. It does not seem possible to choose among several suggestions as to the exact nature of this reaction. If every collision ketween the excited radical and a tert-butyl nitrite molecule leads to deactivation, an upper limit of kz of 8 X l o Q sec.-' may be computed. For comparison, the rate constant for decomposition of the thermalized tert-butoxy radical a t 26' is probably not greater than l o 3sec.-l. From the values of k3/kz a t 26' and 7 9 O , Ez - E a is found to be 6.1 kcal./mole.
4011
collision rate. If all other sources of temperature dependence of kJ are ignored, E? is about 6.5 kcal./ mole, with an uncertainty of several kcal. The activation energy required for decomposition of thermalized tert-butoxy is twice this The constant k? was estimated to be 5 8 X l o Q set.-'. If E2 is 6.5 kcal., A2 5 1014a7 set.-'. If reaction 3 goes in one collision in one hundred, AS is reduced to a value near the middle of the range expected for a simple unimolecular reaction. (e) Effect of Wave Length.-The IoTered quantum yield of acetone, -0.2, when 3130 A,. radiation is used, compared with -0.8 a t X 2537 A, demonstrates the strong effect of wave length. Thompson and Dainton7 found that prolonged illumination of tert-butyl nitrite vapor in the banded region led to no net decomposition. This suggests that the excited radical effect must be indeed small under these conditions. ADDEDIN PRooF.-The ratio k l / k z for moderation of excited radicals by $thane has been found t o be 6 X 10%nil./ molecule a t 3130 A., a value substantially greater than t h a t found a t 2537 A. (Table 11). The effect of a change in wave length is therefore to change both the number and energy of the excited radicals formed in the primary process.
The apparent activation energy of deactivation steps like 3 is often taken to be zero. This will not generally be the case, since k3 depends upon the
Conclusions.-The present results indicate that an excited radical effect must be considered in the vapor-phase photolysis of alkyl nitrites following absorption of light in the continuum. The apparent occurrence of deactivation processes suggests that the effect may be much less important in the liquid phase. It is not possible to decide with certainty in which degrees of freedom the excitation lies. Relative deactivation efficiencies given in Table I1 are consistent with quenching of vibrational energy, but the possibility of electronic excitation of the radicals may not be excluded.
(23) A. F. Trotman-Dickenson, "Gas London, 1955, p. 84.
(24) D. H. Volman and W. M. Graven, J. Am. Chcm. SOC.,1 5 , 3111 (1953); G. R. McMillan, ibid., 82, 2422 (ISGO).
+
(CH,),CO* --+ CHSCOCHa CH, (CHS)aCO* (CHi)aCOKO + (CHa)sCO (CHa)XONO
+
+
[DEPARTMENT OF
Kinetics,"
(2) (3)
Butterworths,
CHEMISTRY, UNIVERSITY O F CALIFORNIA, LOS i h G E L G S
24,
CAI,IFORNIA]
The Photochemistry of Nitrogen Dioxide at 3130 and 4050
A.
BY FRANCIS E. BLACET, THOMAS C. HALLAND PHILIPA. LEIGHTON RECEIVED FEBRUARY 3, 1962 I n mixtures of nitrogen dioxide and 0'8-enriched oxygen, scrambling in t h e 0 2 and transfer of OE8to the NO2 are produced by irradiation a t 3130 A., but not a t 4050 A. Added gases are more effective in quenching the nitrogen dioxide photolysis a t 4050 A. than a t 3130 A. At both wave lengths nitrogen dioxide reacts with isobutane t o form t-nitroisobutane. The quantum yield of this reaction with 6 mm. NOS and l o a t m . i-C~Hiois =0.1 a t 3130 1.and ~ 0 . 3 a5t 4050 A,, while t h a t of the photolysis with 6 mm. NOz alone is 0.96 a t 3130 A. and 0.36 a t 405p A. An activated molecule mechanism is capable of explaining these and other observations on nitrogen dioxide a t 4050 A.
Introduction At 3130 and 3660 A. the diffuse absorption spectrum, absence of fluorescence, and photochemical behavior of nitrogen dioxide long have been regarded as evidence of its photodissociation with a primary yield of close to unity. At 4050 A. the absorption spectrum shows discrete structure and the molecule fluoresces, yet photochemical reactions also occur, with some characteristics in common and some distinctly different from those
at shorter wave lengths. Both dissociation and activated molecule reactions have been postulatedo1,2 as primary photochemical processes a t 4050 A., but evidence regarding the relative importance of these and the photophysical processes which also occur has been indeterminate, and hitherto no complete explanation of the photochemistry of (1) W. A. Noyes, Jr., and P. A. Leighton, "The Photochemistry of Gases," Reinhold Publ. Corp., New York, N . Y., 1941. (2) S. Sat0 and R. J. Cvetanovi6, Can. J . Chem., 36, 279 (1958).
4012
FRANCIS E. BLACET,THOMAS C. HALLA N D PHILIPA. LEIGHTON
nitrogen dioxide a t this wave length has been proposed. Experimental data bearing on the phoJochemistry of nitrogen dioxide a t 3130 and 4050 A. were presented by one of the authors (T. C. H.) in a doctoral dissertation at the University of California, Los Bngeles, in 1953. With the increasing importance of nitrogen dioxide photochemistry in air pollution, some of these data have been rather widely quoted.3-7 They are here described in detail, and the mechanisms which they suggest are applied to a number of hitherto uncoordinated observations. Experimental Photolyses were conducted in a 3 X 20 cni. cylindrical fused silica cell connected by graded seals and stopcocks t o the storage bulbs, manometers, traps, and Toepler and diffusion pumps required for introducing, measuring, and removing the desired gases. The photolysis system, which was mercury free and had a volume of 210 ml., consisted of the cell, a trap, and a solenoid circulating pump, and was isolated from the gas handling system during runs by stopcocks so placed as to minimize blind end tubes. The system was enclosed in a n insulated box for temperature control, and most runs were conducted a t 25'. For higher temperatures the box was heated by electric coils and a circulating fan. The cell was illuminated along its axis by a collimated and filtered light beam from a Hanovia Type A medium pressure mercury arc, which u-as housed in a separate box. Two filters were used: ( 1 ) a 3-cm. layer of 46 g . of SiS04'6H20 and 14 g. of CoS01.7H20 in 100 cc. of H20 plus a 1-cm. layer of 5 g. of K H phthalate/l. H20, and ( 2 ) a 3-cm. layer of 0.02 iV I1 in CC14 plus a 1-cm. layer of 0.05y0quinine hydrochloride in H2O. With the first filter the phthalate solution was constantly renewed by circulation from a 5-1. reservoir, and the decay in transmission was less than ly0over a 5-hr. irradiation. From the transmission curves of the filters, nieasured on a Cary spectrophotometer, and the line intensities of the arc, it was estimated t h a i 94% of the radiation transmitted by filter l was a t 3130 ,A. and 88% of t h a t transmitted by filter 2 was a t 4017-4078 A . Energy flux measurements before, during, and after each photolysis were made with a phototube, calibrated in situ against a thermopile-galvanonieter system which in turn !vas calibrated against standard carbon lamps. By means of ausiliary illumination from a sinall incandescent lamp, the phototube also served as a photometer for measuring the SO2 concentration in the cell before and after each run. Sitrogen dioxide was purified by the method of Harris, et ~ l . and , ~ stored a t -80" as white solid n T 2 0 4 . Carbon dioxide, difluorodichloromethane, and the hydrocarbons used as added gases were purified by freezing, pumping, and volatilizing the middle third of each sample. In the standard photolysis procedure, after the system had been thoroughly outgassed and the mercury arc brought to operating equilibrium, stopcocks leading to the Toepler pump and manometer were closed to prevent contact between liquid mercury and I i 0 2 . The N204 storage bulb then was warmed, NO2 was introduced to the desired pressure as measured by the photometer, and any foreign gas t o be used was added. The system contents then were condensed a t liquid nitrogen temperature in the trap, the system pumped, the contents volatilized and recondensed, ( 3 ) H. W. Ford, Can. J. Chem., 38, 1780 (1960). (4) H. W .Ford and H. Endow, J . Ckem. P h y s . , 27, 1138, 1 2 i 7
(1957). ( 5 ) P . A. Leighton. "Photochemistry of Air Pollution," Academic Press, Iric., New York. N. Y , 1061. ( 6 ) P. A . T.cighton and W. A . Perkins, J r . , "Solar Kadiatioii, . 4 1 ~ sorption Rates, and Photochemical Primary Processes in Urban Air." Air Pollution Foundation, Report No. 14, Los Angeles, California, I!).%. 17) P A. Leighton and I%'. A Prrkins, Jr , "Photochcmical Srconila r y Itcactions in Urban Air," Air Pollution Foundation, ICepiirt Si,. 2.i, San hfarino, California, 19.78. (8) L. Harris, G W. King, \V S. Henrdict and R. \V. 1l. P e a r w , J . Chenz. P h y s . , 8, 7fi5 (1040).
Vol. s4
and the system repumped. hText the radiation flux was measured, the trap contents revolatilized, the solenoid pump activated, the NO2 pressure measured by photometry and the irradiation started. Transmitted radiation was measured a t the beginning and end of the run. After the run, the remaining NO2 pressure was read by photometry, the condensable products again frozen out, and the radiation flux remeasured. The non-condensable products then were collected by Toepler pump in the thimble of a Blacet-Leighton microgas analysis apparatus, the volume measured, and the sample analyzed by transfer of a small portion to a modified Westinghouse Type LV mass spectrometer. The oxygen content was determined from the mass 32 peak height relative t o t h a t from pure air, and from this and the radiation measurements the quantum yield was calculated by the standard m e t h ~ d . ~ I n the runs with 018-enriched Oa, the sample was stored over mercury in a fused silica thimble of the microgas analysis apparatus, from which the desired amount could be drawn through a capillary for introduction into either the photolysis system or the mass spectrometer. As originally received, most of the 0 ' 8 in the sample was present as 01W6. It was equiiibrated before use by irradiation, in the thimble, by 2537 A. from a resonance lamp. A 1-hr. irradiation gave complete equilibration but resulted i n a slight sacrifice in 0'8-content, probably due to outgassing of 0 ' 6 0 1 6 from the walls. -111 quantum yield runs were conducted at an initial NO2 pressure of = 6 mm., and irradiation times, ranging from a few minutes to an hour or more, were chosen to keep the amount of KO2 decomposition small, generally =575 of that initially present.
Results The results from the irradiation of mixtures of NO2 and 018-enriched0 2 are summarized in Table I. In these experiments, the initial XOz pressures TABLE I SUMMART OF ISOTOPE EXCHATGE DATA Irradiated a t 3130 A. 4047
A. Initial NO2 pressure, nim. 10 4.6 Ol8-enriched O2 pressure, mni. 0 . 3 4 4 0.281 Photoproduced O2 pressure, r n m . 0.257 0.291 0 ' 8 in O2 beforc irradiatiim 24.5 24.3 0'6 in O2 :titer irraiiiation" 11,O 2 3 . 1 Estiniatetl atoin 0 1 s in S O SO2 L X 0.0 after irradiation Ratio 033fi/0234 before irradiatioii 0 . 139 , lXl Ratio 023fi,'0234 after irradiation .Oi6 . If8 Corrected for photoproduced 0 2 (mass 38).
+
were adjusted to give approximately the same amounts of photodecomposition a t the two wave lengths. Xfter irradiation the mixture was refrigerated a t - 190' for 30 min. to remove the oxides of nitrogen, and the amount and distribution of the 0ls in the oxygen was determined by mass spectrometry. The experiments were conducted at %', and a dark run of 12 hr. %t115' showed no exchange. The results at 3130 A. show two effects, a loss of 01* in the oxygen and a reduction in the 0 2 3 6 / 0 2 3 4 ratio after irradiation. The loss of OIs presumably was due to its transfer to the remaining NO and NOa, and confirming this, mass spectrographic exaiiiiiiattiun UT the reiiiaiiiiiig N O r sliowecl ;t s i i i a l l peak a t iiiass 48 accompanying the major peak a t mass 46. Both of these effects would be expected to follow photodissociation of the NO2 to yield single OJti atoms. I n contrast, irradiation a t 4030 A. produced no significant change in the 1>9)P .4.1,eigiitiin a n d I' (19:jZ).
B Iil;rcet, .I
.,liii
Clicin. SOL.,5 4 , 316,:
Nov. 5, 1962
PIIOTOCIIE~IISTHY OF NITROGEN DIOXIDE
20
01 0
l i
I
1000 1500 2000 Pco,,mm Fig. 1.-Effect of carbon dioxide on quantum yield of oxygen formation: 0 , this work; A, Holmes and Daniels.IO
500
0 2 3 6 / 0 2 3 4 ratio and no significant transfer of 0 l8 from the oxygen to the nitrogen dioxide. The quantum yields obtained for oxygen formation from pure NOz a t 25' and =56 mm. pressure wherea@o,= 0.96 a t 3130 A. and %o, = 0.36 a t 4050 A. One run a t 147' and 4050 A. gave @o, = 0.37, indicating that the yield a t this wave length is temperature independent or nearly so. The effects of added carbon dioxide and difluorodichloromethane on the quantum yields are shown in Fig. 1 and 2. In neither case were any observable products of interaction with the added gases obtained, and a t both wave lengths the yields were approximately inversely proportional to the added gas pressure. The straight lines in the figures give the values in Table I1 for b in l / @ 0 2 = a b ( M ) . The effects of etha5e and propane on @on were measured a t 3130 A. with similar results, and the observed coefficients are included in Table 11. All runs were a t PXO,= 6 mm. Both in nitrogen dioxide alone and in the pre2ence of added gases the quantum yield a t 3130 A. decreased with decreasing NO2 pressure. Thus in NO2 alone @oP decreased from 0.96 a t PNO,= 6 mm. to 0.72 a t PKO,= 1.7 mm., and in the presence of 1 atrn. of propane it decreased from 0.34 at PSOZ = 6 mm. to ~ 0 . 1 a5t PSO,= 1.5 mm. The effects of $ecreasing NOz pressure were not studied a t 4050 A.
+
QUEXCHIXG
TABLE I1 COEFFICIENTS FOR EFFECTS OF ADDED GASES Q'ave length,
b, I. mole-'
3130 4050 3130
8.2 43 100 2800 33 46
coz coz
A.
CFZCB CF2CIz Ci"
4050 3130
CsHs i-CdHio
3130 3130
165
In the case of isobutane a photochemical interaction occurred, both a t 3130 and 4050 A. A condensable product was obtained which remained (10) H
H. Holmes and 1'
Daniels, J
A m . Chem SOC.,56, 630
400
0
800 PCF~CIZ, mm
1600
1200
of difluorodichloromethane on quantum yield of oxygen formation.
Fig 2 -Effect
in the trap when the nitrogen oxides and unreacted hydrocarbon had been pumped off a t -80'. Application of the techniques of Rock'' and Thomas and Sigfried12 suggested that the bulk of this condensate was tertiary nitroisobutane, and comparison of its mass spectrogram with those of a number of possible compounds showed a close resemblance only with that of t-nitroisobutane. The same product was obtained thermally by heating a mixture of 6 mm. of NO2 and 1 atm. of isobutane a t 130' for 1 hr. In both the photochemical and thermal reactions, water also was identified as a product. The composition of the condensable product, as shown by its mass spectrogram, was essentially independent of the NO2 and isobutane pressure over a wide range. The quantum yields in Table I11 were based on the height of the mass 56 peak in the spectrogram of the condensate, relative to that from a known quantity of t-nitroisobutane added to a NO1-isobutane mixture in the photolysis systein, then condensed and separated by the same procedure. The peak heights so obtained averaged only 0.76 of those given by introducing the same quantity of t-nitroisobutane directly into the spectrometer system, showing that the procedure used did not give complete recovery. The ovariability in yield shown by the runs a t 3130 A. and = l atm. of isobutane (Table 111) probably also is due QUANTUM
ox h, Added gas, M ,
(1934).
4013
TABLE 111 YIELDOF I-ISONITROBUTANE FORMATION T = 2 5 O , P N O ,= 6 4~ 0 3 mni.
Wave Ipngth,
Isobutane pressure,
A
mm
*t--C.rHghOz
3130
756 756 768 774 1206 384 756
0 12 05
4050
.os .14 .08 .I6 .35
to incomplete recovery, and for this reason the results are only semi-quantitative. Tohemost that may be said is that the yield a t 3130 A. is approxi(11) S hf Rock, Aizal C h e i n , 23, 261 (1951)
(12) B W Thomas and W D Sigfried, t b d , 21, 1022 (1949)
FRANCIS E. BLACET, THOMAS C. HALLAND PHILIPA. LEIGHTON
4014 0.5 [
PNO, = 1.7 mm. quite precisely. This explana-
I
10-1
1
10
100
1000
PNOr.
Fig. 3.-Quantum yields .Of oxygen formation in nitrogen dioxide photolysis a t 4050 A,: 0,this work; A, Dickinson and Baxter”; 0 , N ~ r r i s h . ~ ‘
tion supports a primary yield of unity and requires that, for M 7 NO2, ks/ks < 0.04. At 4050 A. the exchange experiments indicate that NO2 is not photodissociated in significant amounts. If so, its photoreactions a t this wave length must involve activated molecules, and these processes must account for the experimental observations on the quantum yield of NO? photolysis, on its fluorescence, and on its reactions with hydrocarbons. A mechanism which appears to accomplish this for the photolysis and fluorescence is KO2 h~ +NOz‘ I* (9)
+ +
+
~ Z J
+
0 N0zNO 0 2 0 NOS+NOa’ NOJ’ +NO2 0 NOj’ NO2 +NO2 0 NOZ’ M +T i 0 8 M NO NO, +2N02
+
+ +
+
+ + + 2
kz ka k4 N O kb k6
(2) (3)
(5) (6) (7)
Introducing Ford and Endow’s value4 of kz = 2.1 X l o 9 1. mole-‘ sec.-l and Ford’s estimate of k3 = 5 X lo”, the observed reduction in @OZ with decreasing NO2 pressure in the presence of 1 atm. of C3Hs gives k4/ka = 0.01 mole l.-l, and the data in Table I1 yield the ratios CFzClz 0.038
C2H4
C ~ H B i-CdHia
0.013
0.017
0.06
The Ford mechanism predicts a +ox of ynity for the nitrogen dioxide photolysis a t 3130 A. in the absence of added gases. A value of less than unity might be due to a primary dissociation yield of less than unity, to the participation of NOs as i V in reaction 6, to other reactions such as NO NOa’ + 2N02, or to the wall removal of 0 atoms or NOe’. Of these possibilities, only wall removal will explain the observed decrease in @o, with decreasing NO? pressure in the absence of added gases. Insertion of a wall removal term k,(N02)-] in eq. 8, with k , = 2.45 X l o 3 set.-*, accounts for the observed values of (PO, = 0.96 a t P N O= ~ 6 mm. and 0.72 a t
+
N0z’
kii
+NOz
----f
kv kis kis kn
(10) (11)
(12) (13) (14) (15) (16) (17)
For the quantum yield of oxygen formation this l Fives ~ ~ ~ ~ In pure NO2 the mechanism thus predicts that a t pressures such that kl0(NO2) > 12,s i k14 the quantum yield will approach independence of the pressure with a limiting value of 1/@0,? 1 kll/k!o. A fit with reported values a t 4050 A. is shown in Fig. 3, where the line is from eq. 18 with (M) = 0 and
+
(4)
For the quantum yield of oxygen formation this gives
M COz ke/ks = 0.003
kio kii kiz
+ h~ X02‘ +KOz” NOZ” +NO2 + h” KO*” + KO2 +NOz + NOz 902” +M NO2 + M
Discussion The isotope exchange experiments give direct evidence of the photodissociation of nitrogen dioxide a t 3130 but not a t 4050 A., supporting the earlier assumptions that this is the ~ a s e . Reactions following dissociation have been discussed in detail by Ford,3 with the conclusion that the mechanism which best conforms with present understanding, under conditions such that the reactions of 0 atoms with NO and 02 are negligible, is NO2 +NO 0 I (1)
+ + NO2 ---+ 2 N 0 + Oz NO,’ + NOz +NO2 + NOz NOz’ + M +NO2 + M XOz’
mately 0.1 a t isobutane pressures of 750-1200 mm., while the yield a t 4050 A. may be higher than that a t 3130 A. and may be proportional to the isobutane pressure.
+
L-d. s4
At low nitrogen dioxide pressures the mechanism predicts that @oxwill fall off and approach proportionality to the pressure, but it is apparent from Fig. 3 that no observations are available in the proper range to test this. Values of the ratio (k13 k14)/k10 much larger than 3 X mole 1.-l produce significant departures from the observed yields a t 3-11 mm., but smaller values of this ratio, with some adjustment in kll/klo, give equally good fits. A value of (k13 k14)/k10 much smaller than 3 X may, however, be questioned on the basis of absolute rates. Thus, from the absorption coefficients of nitrogen dioxide, Neuberger and Duncan15 have estimated a lifetime relative to radiative transition of 2.6 X lo-’ sec. for NO2’, which gives k13 = 3 X lo6 set.-'. Even with (kls k14)/k10 = 3 X this requires that klo > 10” 1. mole-’ sec-1, and to explain the observed values of @ol,kll must be larger still. With added carbon dioxide and difluorodichloromethane, the values of b in Table 11, combined with eq. 18 and 19 for PNo2= 6 mm. yield the ratios
+
+
+
ill
=
coz
k12/kin = 0.014
CFzCl2 0.9
(13) R . G. Dickinson and W. P. Baxtet, J . A m . Chcm. SOC.,5 0 , 774 (1928).
(14) R. G . W.Norrish, J . Chem. Soc., 1158 (1929). (15) D . Neuberger and A. B. F. Duncan, J . Chcm. P h y s . , 22, 1693 (1954).
PHOTOCHEMISTRY OF NITROGEN DIOXIDE
Nov. 5, 1962
Compared with these ratios, that of kll/klo = 1.7 indicates that reaction 11 is faster than would be expected if only simple deactivation were involved. An alternate path which might account for this is NO$’
+ KO2 +NO3 + NO
(114
followed by (7). This conforms with Ford’s suggestion that 0, NOz’, and NOa’ may each attack NO2 by two paths, to form either a peroxy or a non-peroxy activated complex. The peroxy complex in each case in postulated to yield OONO, which decomposes to 0 2 NO, while the nonperoxy complex yields NO,. Table IV lists the reactions involved and the peroxy/non-peroxy rate constant ratios indicated by Ford’s estimates and this work. The apparent increase in this ratio in going from 0 atoms to N03’ is noteworthy.
+
TABLE IV REACTIOKS OF 0, hTO,’, AND KO,’
Reactants
0
+ KO’ + NO,
NO*’ N03’
+ No2
WITH
Reaction nu. of process forming Peroxy Non-peroxy complex complex
3 lla
2 10 5
6(M = KO?)
4015 ki3
’*’
=
=
+ kic + (kio + kii)(NOz) + k i ~ ( M ) (22) + + (kio + KII)(NOZ)+ kdM)I [k16 + k16(N02) + k17(M)1 (23) k13
kl4kl5
kl4
These equations show that, with increasing (NO2) or increasing (M), the fluorescence from N01” will be quenched more rapidly than that from NOz’. The predicted quenching curves for pure NO2 in Fig. 4 are based on the rate constant ratios in (19) and are independent of the absolute values of af0. They suggest that, if NO2 does fluoresce from both states a t 4050 A. a t pressures above = 0.1 mm. most of that observed will be from N02’.
NO2 Rate constant mtio Peroxy Non-peroxy
0.004 0.6 >25
The greater effectsoof added gases on $0, a t 4050 relative to 3130 A. are reflected in the larger ratios of klz/klo relative to ke/k6. The difference thus may be assigned to the postulate that, relative to their collisional deactivations, the reaction of NOz’ with NOz to form a peroxy complex is less efficient than that of NOg’ with NOZ. The radiationless transfer (14) was proposed by Neuberger and Duncan16 to account for their observations on the fluorescence of nitrogen dioxide. On excitation by blue light, a t NOz pressures of 0.6 to 12 p , the fluorescence gave lifetime traces yielding the relation, with PNO,in mm.
0 10-4
io-* 10-1 1 10 100 PNOZ, mm. Fig. 4.-Predicted quenching of nitrogen dioxide fluorescence at 4050 A. 10-3
Norrish16reported a weakening of the fluorescence in the region of 6250-6550 A., while that in the range was little affected, in going from of 5600-6050 4350 to 4050 A. as the exciting wave length. This change in distribution was not confirmed by Neul/tf = 2.25 X l o 4 2 X lo6 P sec.-l (20) berger and Duncan. Some weakeFing of fluoresExciting bands centered at 3950, 4300, and 4650 cence in going from 4350 to 4050 A. might be exA. gave approximately the same values of tfo. pected from eq. 22 and 23, since klo is essentially On excitation by 5460 A., on the other hand, the zero a t ths longer wave length. If the value of klo fluorescence gave no observable lifetime trace. a t 4050 A. is that given by (19) and the other This led Neuberger and Duncan to postulate that constants do not change with wave length, a f ( 4 0 6 0 ) / the first excited state NOz’ is short lived and the @fcraeo) would be 0.88 a t PNO,= 0.1 mm. and 0.84 fluorescence excited by 5460 A. comes only from a t P N O ,= 10 mm. An increase in k14 with dethis state, while the second state N02” is relatively creasing wave length would increase af”but delong lived and produced the traces observed at crease af!and in the case, if fluorescence occurs 3950-4650 A. The firstostate, NOz’, also might from both states, the over-all effect observed would fluoresce a t 3950-4850 A. but the experimental depend on the pressure. method used would permit no decision on this. In mixturesaof Nos, Nz04, and N z O ~ exposed to For the fluorescent lifetime of N02” the mech- light a t 4050 A. and shorter wave lengths the only anism gives observed change is the decomposition Nz06 + 2N02 1/2 0 2 , accompanied by the association l / t P = k16 k d K O d kl (M) (21) From Neuberger and Duncan’s data kl6 = 2.25 X 2N02 e N204. The concentration of the light l o 4 set.-' and k16 = 3.7 X 1Olo 1. mole-’ set.-'. absorber NOz increases as the reaction proceeds. The absence of any observed wave length varia- Using a constant pressure of 51.5 mm. of hi206 tion in tf over the range 3950-4650 A. suggests that, a t Oo, Holmes and DanielslO estimated the quantum over this range a t least, kls and kl6 are independent yields of Nz06decomposition from the total product pressure and the equilibrium constant P2r;o,/ of the exciting wave length. The quantum yields of fluorescence derived P N ~ o=, 13.53, with P in mm. At the two wave from the mechanism are (16) G . W. Norrish, J. Chcm. Soc., lGll (1929).
8.
+
+
+
+
I’RAKCIS E. BLACET,THOMAS C. HALLAND PHILIP,i. LtiIcrrrox
4016 2.0
Vol. s4
1
/
3
1.2
0.2 1.0 0 Fig. 5.-Quantum
100 200 300 Product pressure, tnm. yields in the photolysis of nitrogen pentoxide at 3660 d.
I
I
50
100 Product pressure, mm.
0
150
200
lengths studied in$he NO:! photodissociation region, 3130 and 3660 A., under these conditions it is Fig. 6.-Quantum yields in the photolysis of nitrogeti pentoxide a t 4050 A . necessary in estimating the yields relative to KO2 absorption to correct for absorption by Xz04, an experimental determination of the intercept a t and a t 3130 A. this correction is quite severe. P = 0 is required for any definite specification. Using Holmes and DanielslO absorption coefficients The photoreaction of nitrogen dioxide with for ??Oz and N204, the estimated yields a t 3130 A. isobutane at 3130 8.probably involves abstraction are somewhat larger, while using Hall and Blacet’s of the tertiary by oxygen atoms and coefficients,l7 they are smaller than those a t 3660 hydroxyl radicals,hydrogen and combination the resultant ,4. For this reason it is believed that there is no isobutyl radicals with NO2 to giveofnitroisobutane significant difference in the yields relative to SO2 0 $. i-C4H10 +OH i-GHg kzi ( 2 7 ) absorptiFn a t the two wave lengths, and those OH + i-CaHlo +H20 z-CaHg (28) a t 3660 A. are probably the more reliable. These values are plotted as circles in Fig. 5 i-C4Ho + SO1 +L-CIHOSO~ (29) against the mean product pressure during each time Together with reactions 1-6 these give, as an upper interval over which @ was measured. They may limit since other reactions of OH and i-C4H9 be explained by reactions 1-7 together with h - 2 0 ~ radicals are not taken into account -t N O a KO2. The line in Fig. 5 is estimated = 2@0,, kz, ka, and kJka as from eq. 8 with before, and ks,’k5 = 0.003 for ;LI = EO2 and 02, 0.01 for bl = Nz06,and 0.06 for M = NzO,. Similarly, Holmes and Daniels’ results a t 4050 shown as circles in Fig. 6, may be explained by Based on CvetanoviE’s resultsl8 the value of kz7 reactions 9-14,together with N205 NO? NOz may be estimated a t = 5 X lo7 1. mole-’ set.-'. and (T). The solid line in Fig. 6 is from eq. 18, With this and the estimates of kz, ka, kd Ik5, and kfi ‘k5 with the ratios in (19) and kl:!.’klo = 0.14 for 14 = derived from the NOs photolysis, for 11 = iN?05,0 . i for 11 = N z 0 4 and , 0.014 for M = 02. C4H10 and 6 mm. of SO2 eq. 30 gives @ . c ~ H ~ s o ~ These explanations may be questioned on a t 20.11 and 0.12 for 760 and 1200 mtn. of iGH1c. least two points. First, they require the assump- These are consonant with the experimental contion that k6 and klz are smaller for M = K.06 clusion that @ c 4 H r ~ 0=0.1 2 in this prespre range. than fFr M = S204.Second, the mechanism a t The approach of the yield a t 3130 A. to inde4050 A. predicts a quantum yield of zero a t zero pendence of isobutane pressure, which the mechproduct pressure, whereas Holmes and Daniels’ anism predicts and the experimental data do not results suggest a finite value a t zero pressure. deny, results from the coincidence that a t the These points suggest the possibility of direct re- pressures used kz(N02)< k2;(C4Hlo)< k:!(NOz) and actions of SO:{’and SOZ’ with iY?O6, such as k5(?JOa) < k e ( J f j < k 4 . At higher NO? pressures, or either higher or lower i-CIH1,)pressures, a greater so:j’ f x205+ so::+ 0, + ” S O . (2-4) change in yield with isobutane pressure is to be sos’+ sroi +s o + or + “so2 ( 2 3 ) expected. Xt 4030 the explanation of the nitrogen diFor example,oadding(23) to the preceding mechoxide-isobutane reaction must be sought in terms anism for 4050 A. gives of activated molecules. postulated mechanism is
+ +
+
-
A.
+
x.
A%
The dashed line in Fig. 6 is from this equation with k:!6/klQ= 1 and kl2/kl0 = 3.2 for A 1 = N206, 1 2 for A i = N204 It should be emphasized that these values are only illustrative, other combindtions give equally good or even better fits, and ( 1 7 , I C 11.111 mcl I
I., Blacct
J C h i n Ph3s 2 0 , 1745 (1‘332)
NO*’
+ Z-CIHIO HKOr + z ’ - C ~ H ~ i-C4H9 + NO? +i-CaHoKOn 2HXOz +HzO + S O f S O ? ----f
k31
(:31) 129) (32)
Combining these with reactions 9-14 introducing the rate constant ratios in ( I n ) , and neglecting 1
I h , 1:.
J. Cvetanovi6, ibid , 30,
Ill
I
I!j,i!J)
Nov. 5, 1962
PHOTO-OXIDATION OF AZOETHANE
the resultant nsmall term containing gives, for 4050 A.
k13
+
k14
4017
where again the major products were identical but the product yields a t 4050 A. appeared to equal or exceed those at shorter wave lengths, may be explained b y postulating that 0 atoms and NOZ’ react with 1-butene to yield the same first product
+
In this case the variation in quantum yield with 3130 A.: 0 C4H8 +C4HsO kta (3.1) isobutane pressure, or with the (NO~),/(C~HIO) 4050 d . : XOz’ -I- CiHs +NO kgj (35) ratio, is more straightforward than that predicted a t 3130 A. The dependence on pressure indicated The resulting quantum yield equations will reby the yields in Table I11 suggests that, when M = semble (30) and (33), with appropriate substitui-C4H10> k12,1k1? is small, and the absolute yields tions and elimination of the factor 2 in (30). suggest that klo:’kdl = lo2, but the experimental Cvetanovik’s value of k34 is 1.1 x IO9 1. mole-’ values are so rough that this cannot be stated with set.-' and the pressures used by Sat0 and Cvetanocertainty. To a first approximation, the yield a t vic were 3 mm. of NO2 and 30 mm. of i-C4H*. 4050 A. will exceed that a t 3130 A. if 2.7klo,’k31 < For these values eq. 30 pr!perly revised gives a34 = k3k~(M)/2k27k4,and in the investigated pressure 0.45. The yield a t 4050 A. and the same pressures range this would appear to be the case. The will exceed this value if (klz 0.27klo)/k36 < 1.2. postulated mechanisms thus permit both a higher Acknowledgment.-The authors are grateful to yield and a greater dependence on isobutane pres- the Air Pollution Control District, Los Angeles sure a t 4050 relative to 3130 fi. County, and to the National Institutes of Health, The observations of Sat0 and CvetanoviP on U. S.Public Health Service, for funds granted in the photooxidation of 1-butene by nitrogen dioxide, partial support of this research.
+
+
[ C O S T R I B U T I O N FROM THE D I V I S I O N O F P U R E CHEMISTRY, S A T I O N A L
RESEARCH COUNCIL,
OTTAIVA, C A N . ]
The Photo-oxidation of Azoethanel BY H. CERFONTAIN~ AND K. 0. KUTSCHKE RECEIVED JUNE 13, 1962 The photo-oxidation of azoethane has been studied over the temperature range 28’ t o lF2’ at various oxygen and azoethane pressures and over a wide range of intensities, Sitrogen, acetaldehyde and ethanol are major products a t all temperatures. The amount of nitrogen formed is greater than in t h e photolysis of azoethane in the absence of oxygen and increases with decreasing oxygen pressure within given limits, especially a t high temperatures, where nitrous oxide and ethylene become important products. A mechanism for the formation of ethylene, nitrous oxide and additional nitrogen is proposed. At temperatures above 100’ acetaldehyde is produced, in part, by oxidation of ethoxy radicals. At high conversion several C1-products are formed. The acetyl radical is a n intermediate in the formation of these products from acetaldehyde. I t was not possible t o determine a rate constant for the reaction of ethyl radicals with oxygen.
Introduction In the photolysis of azoethane no evidence was found for the existence of C?HhNzradicals3 Xzoethane was therefore expected to be a convenient source of ethyl radicals for a study of the reaction between these radicals and molecular oxygen. The oxidation of ethyl radicals has been the subject of several investigations. Jones and Bates4 have studied the photo-oxidation of ethyl iodide. They followed the course of the reaction manometrically, and concluded that acetaldehyde and ethanol are th.e main products. Finkelstein and Soyes” and, later, JolleyGhave made an extensive study of the photo-oxidation of diethyl ketone. I t seems well established that ethyl, propionyl and pentanonyl radicals are involved in the oxidation. Carbon monoxide, carbon dioxide and acetaldehyde were the major products detected. (1) Issued as i T R C . No. 7079. (2) National Research Council Postdoctoral E’ellow 1855-1957; present address: Laboratory for Organic Chemistry, University of Amsterdam, Nieuwe Achtergracht 129, Amsterdam, Netherlands. (3) H. Cerfontain and K. 0. Kutschke, Can. J . Chem., 36, 344 (1 958). (4) L. T. Jones and J. I t . Bates, J . A m . Chem. Soc., 5 6 , 2283 (lY34). ( 5 ) A. Finkelstein and W. A. S o y e s , Jr., Disc. Fuvaday SOL.,14, 82 (193).
(6) J 1s. Jolley, J . A m . Chena. SOL, 79, 1337 (1327).
No analysis was made for ethanol. Tracer experiments showed that carbon monoxide was formed solely from the carbonyl group and that a certain fraction of the carbon dioxide, especially a t higher temperature, originated from the a-ethyl carbon atom. Ethyl hydroperoxide is the major product in the mercury-photosensitized oxidation of ethane under some conditions7 The results show that ethyl hydroperoxide is not formed by hydrogen abstraction of the C2H602radical from the parent compound. Turner, Callear and CvetanoviC., who studied the oxidation of ethyl radicals produced from mercury-photosensitized hydrogenation of ethylene arrived a t the same conclusion.& The main reaction producing ethyl hydroperoxide was thought to be C2HjOz
+ HOz +CzHsOzH + 0 2 Experimental
Azoethane, prepared by Merck, Sharpe and Dohrm, Montreal, was redistilled and degassed several times and the middle fraction collected and stored as a liquid in a blackened tube behind a mercury cut-off. T h e oxygen was .. .
. - .. . . .
(7) J k S. Watson and E%.d e G. Darwent, J . P h y s . C h e m . , 61, 577 (1967). ( 8 ) A. H . Turner, A. 13. Callear a n d K . J. CvetanoviC, unpul,lished work.
