Journal of Atmospheric Chemistry 1 (1983), 53-63. © 1983 by D. Reidel Publishing Company.

0167-7764/83/0011-0053501.65

53

Kinetics of the Reaction of Nitrogen Dioxide with Ozone R. A. COX and G. B. COKER Environmental and Medical Sciences Division, AERE, Harwell, Oxon., England (Received: 13 December 1982) Abstract. The kinetics of the reaction of NO 2 with 03 have been investigated at 296 K, using UV absorption spectroscopy to monitor decay of NO 2 or O 3 and infrared laser absorption spectroscopy to monitor formation of the reaction product N20 s . The results both for the rate eoeffig/iont at 296 K (k I = 3.5 x 10 -t7 ems molecule -t s -t ) and the reaction stoiehiomel~y (ANO2/AO3 = 1.85 -+ 0.09) are in good agreement with previous studies, confirming that the two step mechanism involving formation of symmetrical NO 3 as an intermediate is predominant. NO2 +O 3 NO 3+NO 2 +M

, NO 3 +O 2 ' N 20 s +M

(1) (2)

A possible minor role for the unsymmetrical ONOO species is suggested to account for the lowerthan-expected stoichiometry factor. The importance of this reaction in the oxidation of atmospheric NO2 is discussed. Key words. NO 2 oxidation, ozone reactions, nitrogen pentoxide, nitrate radical, kinetic spectroscopy, atmospheric NO 2.

1. Introduction The reaction o f NO2 with Oa is of interest in atmospheric chemistry since it provides a potentially important non-photochemical oxidation process for atmospheric NO2. In laboratory studies, the formation o f the major product, nitrogen pentoxide, has been satisfactorily accounted for by a two-step mechanism involving intermediate formation o f the symmetrical NOz radical. NO2 + 03

" NO3 + 02

(1)

NO3 + NO2 + M

' N20 s + M

(2), ( - 2 )

At ambient temperatures the equilibrium in reaction (2) favours the forward reaction which occurs rapidly (k2 = 3 . 7 x 10 -3o cm 6 molecule -2 s -1) and reaction (1) is rate determining. A number o f kinetic studies o f reaction (1) have been reported in which the decay o f reactants has been monitored using absorption spectroscopy (Graham and Johnston, 1974; Becket et al., 1974), mass spectrometry (Davis et al., 1974; Huie and Herron, 1974) and chemiluminescence. (Wu et al., 1973). The results show excellent agreement for the rate constant k, and its temperature dependence. In two studies

R.A. COX ET AL.

54

Closed cycle cooler ~~.~!

_

,aser source [@-] J

/,4~:ltiporf inter a ,d exit (Not snownl Manifold

To 9as handling system

1 ]

I )

I

I

1841 Teflon coated pyrex celt Deut__~rium

Lamp

I//////////////7777"?11 I ~ . ,

lPhotomultiplier /

Flu°rescen'" b'acklight'" F"--lq [Ljl

I

spectrometer* IR Signal

Idriver~ I

Data-

] I Signa: I ProcessincJ I

~ / ' tTiming/Circuitry

_ . 075m monochromator

I

L

"* S=Tunable diode laser, M:Mode sorting monochromator,D=HgCd Te d~tector, LCM =Laser control module Fig. 1. Schematic Diagram of UV/IR Laser Spectrometer.

(Graham and Johnston, 1974; Wu et al., 1973) the stoichiometry for reactant consumption a = ANO2/AO3 was determined and found to be somewhat less than the value of 2 expected on the basis of reactions (1) + (2). Alternative minor processes have been suggested to account for this, but the evidence for their occurrence is not unequivocal. In the present work, time-resolved absorption spectroscopy has been employed to monitor N2Os, using a diode laser infrared source, and NO2 and 03 using conventional UV techniques, in the reaction of NO2 with Oa. kl Was determined at 296 K for a range of reactant conditions. The results are discussed in terms of the various elementary reactions involved and the significance of this reaction for the atmospheric oxidation of NO2 is also considered. 2. E x p e r i m e n t a l The kinetics studies were conducted using infrared and ultraviolet absorption spectroscopy to monitor ~ecies concentration in a 18.4 1 long path pyrex cell which could be evacuated to < 1 mtorr. The apparatus is illustrated schematically in Figure 1. The infra-red source is a tunable diode laser (Spectra Physics LS3) coupled to the cell, which contained White reflective optics giving multiple traversals (up to 108) of the 1 m base path, for the infrared monitoring beam. The beam was chopped mechanically before dispersion on a monochromator (for selecting lasing modes) and detection on a liquid N2-cooled Hg Cd Te detector. Output from the lock-in amplifier (Brookdeal 9503) was displayed on a chart recorder. The laser frequency could be selected by current/temperature tuning using the laser control module. The UV beam from a deuterium lamp traversed the cell twice

KINETICS OF THE REACTIONOF NITROGENDIOXIDE WITHOZONE

55

before dispersion on a monochromator (Spex) and detection on a photomultiplier in the conventional fashion. Gases were introduced to the cell from a high vacuum line and pressures measured on an MKS Baratron. A pure sample of NO2 was prepared by the reaction of NO with 02 and pure N2Os was prepared by the reaction of NO2 with O3 using the method described by Davidson et al. (1978). O3 was used as a 5% mixture in 02 prepared as required by passing a flow of pure O2 through a silent electrical discharge. Concentrations of NO2 and O3 were determined in the cell by UV absorption at 350 and 255 nm, respectively, using published absorption cross-sections (Baulch et al., 1980). N2Os was monitored by IR absorption at 745.84 cm -1. Absorption by N2Os in the region 740-746 cm -1 exhibited no resolvable structure, even with the narrow bandwidth (~-5 × 10 -3 cm -1) diode laser source. Presumably, the rotational features were fully broadened at the temperature (296 K) and total pressure (10 torr, mainly N2) employed, since strong attenuation of the radiation by N2Os was observed. The absorption followed the Beer-Lambert law and the absorption cross-section at 745.84 cm -~ , 2.5 x 10 -t a cm 2 molecule-i, was determined relative to the consensus value (Baulch et al., 1980) of 3.3 × 10 -1 s cm 2 molecule at 215 nm in the UV, so that absolute N2Os, concentrations could be determined. NO2 could also be monitored in the IR using a characteristic absorption line close to the N2Os monitoring frequency. The accuracy of the concentration measurements of O3 were within +5% and for NO2 and N2Os -+10% approximately. In the experimental runs, the reactant in excess was added to the evacuated cell first and the pressure made up to 10 torr with N2. The second reactant was then added and the absorption-time behaviour recorded in both IR and UV until the reaction was complete. Finally, the absorption due to the excess reactant was measured. The experiments were conducted at room temperature, 296 + 2 K.

3. Results Experiments were first conducted to determine the stability of the reactants and products with respect to the cell walls. NO2 appeared to be quite stable but 03 decayed slowly with a first-order decay coefficient of 2.4 x 10 -4 s -1. A sample of pure N2Os added to the cell decayed more rapidly, ka = 11.3 x 10-4 s -1 with the appearance of a complex system of absorption lines at 850.4 cm -1 ascribed to HNO3. This is faster than the overall homogeneous decomposition rate for pure N2Os at this temperature and 10 torr pressure, 1.3 x 10 -s s-1, calculated from the results of Connel and Johnston (1979) and is believed to result from a heterogeneous reaction of N2Os with adsorbed water vapour. It was necessary to take this decay into account in the determination of kl from N2Os formation. Kinetic experiments to determine kl were conducted both with NO2 and with 03 as excess reactant in the concentration range ( 2 - 2 0 ) x 1014 molecules cm -3. The initial concentration of the minor reactant was in the range (0.5-1.0) x 1014. In most cases, the concentrations were such that pseudo first-order kinetic behaviour resulted, as is illustrated in the logarithmic decay plots presented in Figure 2. For decay of O3 or

56

R.A. COX ET AL

\

rj

0

\

\=,,,

\

[NO2] : 3.08 (11.)

[03] 5.2 (15) =

'31 = 2.1.8 (I/.)

I 0

1[NO2= 1.49 (151 I

I

I

I

I

I

I

I

20

1.0

60

80

100

120

1l,0

160

Time (s) Fig. 2. Logarithmic concentration-time plots for O= in excess NO= (circles) and NO= in excess Oa (squares) in the NO2 + O= reaction. Filled points show data from formation of N20 s after correction for decay of this product by the method described in the text.

NO2 the ratio Ao/At, where A 0 and A t are the optical densities at time = 0 and t respectively is plotted. The slope gives k!( = kl [NO2] or a kl [03]) directly. For N2Os formation, analysis was complicated by the decay of this product. This led to a rise in N2Os absorption during reaction to a maximum, Am=x, followed by a decay. The time dependence of N2Os absorption is given by: _ k____f_ I

At

k a - k I A~o[exp(-klt) - exp(-kat) ]

(i)

57

KINETICS OF THE REACTION OF NITROGEN DIOXIDE WITH OZONE

/

65 0.10

%

0.05

I 5

[o3]

or

I 10 [NO 2]

molecule

I 15 cm - 3 x 10 -l&

Fig. 3. Plot of first order rate coefficient against concentration of excess reactant (NO~ or 0 3) in the NO~ + 0 3 reaction.

where k I and k a are the first-order rate coefficients def'med above and A~o is the final concentration of N2Os if no loss by decay occurred. Then a plot A ~ ( = Amax) - A t vs t can be used to determine k I. If k d 4= O, then Areax < A~o and k t as determined from this plot is overestimated. A method of successive approximations can be used to determine k z. The time for the maximum absorption A max can be obtained from the derivative of(i)

tm~x = ( k ~ - kz) -~ ln(k,~/k ~)

(ii)

58

R.A. COX ET AL.

An approximate value of k ~r is determined using the measured Amax for Aoo, and a new value of A~o calculated from (ii) and (iii). A ~ = Ama x exp(k a tmax). Three iterations were required to obtain successive plots of constant slope. Overall correc. tions to Ao~ ranged from 10 to 40%. Figure 2 shows that the corrected first-order plots based on N205 formation corresponded well with those based on reactant decay, except for a small deviation at large extents of reaction particularly at low excess reactant. This possibly indicates a change in mechanism at large extents of reaction. Table I shows the values of k I from Oa decay and N~Os formation as a function of [NO2]. The agreement Table I. Pseudo first order rate constants k 1 from O3 decay and N203 formation. T = 296 K [NO~]av melee cm-3 X 10-14 1.98 3.08 4.98 5.86 8.31 10.6 15.7

kI(s -t x 10a) From 0 3

From N~O5

8.15 13.6 19.8 25.2 31.5 37.5 55.5

8.56 13.6 18.7 23.5 32.0 36.5 55.4

is good, providing support for the two-step mechanism for N2Os formation involving the rapid reaction of NO3 with NO2. Figure 3 shows a plot of k z vs [NO~]. Linear regression analysis gave a value of: kl - (3.45 _+0.12) x 10 -17 cm ~ molecule -1 s-I(296 K) from the slope of this plot. The intercept (2.1 + 0.5)x 10 -3 s -x , was significantly greater than expected from the observed decay of [03] in the absence of [NO2]. Also shown in Figure 3, is the data obtained from experiments with excess 03 present where k t, as determined both from NO2 decay and N205 formation, is plotted against [03]. In this case, the slope can be equated to o~kl. Linear regression analysis gave: akl = (6.3 -+0.2)x 10 -17 em 3 molecule -1 s -I The intercept in this case was also slightly positive, (5.0-+ 1.7)x 10 -3. These results therefore, yield, a value for the stoichiometdc ratio of et = 1.83 -+ 0.09. The value of a was also determined from measurements of the overall change in concentration of NO2 and Oa during the course of the reaction. The mean value obtained from experiments with NO2 in excess was a = 1.88 +--0.31 and with O3 in excess a--- 1.85 -+0.40. In the latter experiments there appeared to be a significant decrease in ot with increasing [O3],

KINETICSOF THE REACTIONOF NITROGENDIOXIDEWITHOZONE

59

particularly at [03] > 1 × 10 Is cm -3. The overall yield ofN:Os, ~ (N2Os), corrected for loss by heterogeneous decay, was compared with the amount of NO2 reacted. The average value of the ratio ANO2/¢ (N2Os) was 1.96 -+0.29 for all experiments, showing that N2Os was the only stable N-containing product of the reaction of NO2 with 03.

4. Discussion The value of kl obtained in this kinetic study is in excellent agreement with the consensus value at room temperature (298 K) based on previous studies, i.e., 3.2 x 10 -17 cm 2 molecule -1 s -1 (Baulch et al., 1980). The present work is the first to employ measurements of the rate of formation of the product N2Os to determine kl, and the results confirm that N2Os is the major N-containing product, which is formed in a rapid reaction of the initial product of the NO2 + 03 reaction, i.e., NOa + NO2 + M

--N205 + M

(2)

The present results also confirm that the overall stoichiometry for reactant decay ANO2/ AO3 is significantly less than the value of 2 expected from the simple mechanism involving reactions (1) + (2). Wu et al. (1973) have reported values of 1.68 -+0.15 (NO2 in excess) and 1.88-+0.15 (Oa excess) and Graham and Johnston (1974) give 1.89-+0.08. Wu et al. (1973) have suggested that the low value resulted from the side reactions: NO2 + NO3 NO2 + 03

~NO + NO2 + 02 • NO + 202

(3) (4)

with subsequent reaction of NO with 03: NO + 03

"NO2 + 02.

(5)

Graham and Johnston (1974) failed to observe chemiluminescence from reaction (5) and argued that NO could not have been formed. However, NO also reacts rapidly with NO3, k6 - 2 x 10 -11 cm a molecule -1 s -1 : NO + NO3

" 2NO2

(6)

and it can be readily shown that the ratio Rs/R6 TM ksk2 [M]/k6k~. For their experimental conditions only ~-20% of the NO would have reacted via (5) and from their sensitivity limits as much as 1% of the initial NO2 could have formed NO via reactions (3) or (4). This would lead to a stoichiometry of ~ = 1.94, i.e., significantly higher than the observed values. An alternative explanation is the formation of an unsymmetrical NO3 molecule as an intermediate in reactions (3) or (4): NO2 + NO3 (or 03)

• ONOO + NO2 (or 02)

(8)

A reaction of ONOO with either O3 or NOa would lead to the same effect on reaction stoichiometry as does NO formation. A value of a = 1.85 requires ONOO is formed in about 3~aaf the reactive collisions between NO2 and O3 or NO2 and NO3.

60

R.A. COX ET AL.

NO2 + O3-----*'ONOO + O2 NO2 + NO3

(la)

~ ONOO + NO2

(2a)

Involvement of unsymmetrical NO3 was proposed originally by Ogg et al. (1950), to explain isotope exchange experiments for 02 and NO2 in the presence of NO. More recently, Audley et al. (1980) have suggested the reaction ONO0 + CO ~

C02 + NO~

(9)

to account for C02 formation in the thermal decomposition of N20s in the presence of CO, with reaction (2a) as the source of ONO0. Also Bhatia and Hall (1980) proposed that the IR absorption at 1837 crn -a observed in matrix isolated products of the reaction of NO with 02 and 03 is due to the peroxynitrate ONOO species. However, positive identification of the unsymmetrical NO3 intermediate in the NO2--O3--N2Os system is required in order to clarify the reaction mechanism, and provide firm quantitative information on the kinetic parameters of ONO0 reactions. 5. Oxidation of Atmospheric NO2 The kinetic data for reaction (1) allow the forward rate of NO2 reaction with ozone to be calculated for any atmospheric conditions. Typically Oa concentration in surface air is 30 ppb giving at lifetime for NO2 of 11 h at 25°C. However, in order to evaluate the overall rate of oxidation of NO2 to HNO3, it is necessary to consider (a) alternate reactions of NO3 (b) the rate of reaction of N2Os with H20 and (c) the reversibility of step (2). The reactions occurring are summarised in Scheme 1. (2) +NO 2

(1)

NO3

NO2 + 03 +NO

2NO2 NO2 + O

1

+ (hv)

(-2)

~)

N2is

~+H20

(1) 2HNOa(g)

dry deposition (ground or aerosols) Scheme1.

In daylight the lifetime of NO3 with respect to photolysis in the visible spectrum is TM 10 s whilst reaction of NO3 with NO and NO2 have characteristic time constants of 2 and 11 s respectively for an NOx (NO + NO2) concentration of 3 ppb. Therefore, the reactions converting NO3 back to NO2 are dominant in daytime, thereby reducing drastically the overall rate of NO2 oxidation. In situations of reduced sunlight intensity, i.e., in winter at middle or high latitudes, in clouds, and also at night, reaction (2) becomes relatively more important and the fate of N205 needs to be considered. The kinetics of the dissociation of N2Os via the reverse reaction (-2) are reasonably

KINETICS OF THE REACTION OF NITROGEN DIOXIDE WITH OZONE

61

well established, (Connel and Johnston, 1979). At 1 atm pressure and 283 K the lifetime of N2Os with respect to dissociation is approximately 130 s. The rate of reaction of N2Os with water vapour is not well defined. Laboratory studies (Morris and Niki, 1973; Cox, 1974) have shown that the homogeneous reaction is extremely slow, but that heterogeneous formation of HNO3 occurs readily when H20 and aerosol particles are present with N2Os. The evidence suggests that the presence of liquid water, either in the form of cloud and fog droplets or on aerosol surfaces, is necessary for conversion of N2Os to HNO3 in the atmosphere. The rates of transfer of gaseous N2Os to droplets or aerosols can be estimated using the theory developed of Fuchs (see Fuchs and Sutigin, 1971) as modified by Chamberlain et al. (1960), i.e., 1 dc c dt

where

= k a = 47rrnD/

D

r

_.--ZD + - r+A 7tlV

C = gas phase concentration, D = diffusion coefficient (cm 2 s-X), N = ( R T / 2 1 r M ) 1A,

r n A 7

= = = =

radius of particle (cm), number of particles (or droplets) per cubic cm, mean free path of gas molecule (cm), accomodation coefficient.

The key unknown quantity is the accommodation coefficient which could realistically take any value between l0 -~ and 1 for a water soluble reactive species. The current best estimate for the attachment of H20 to water droplets is ~ = 0.03 (Pruppachker and Klett, 1978) and a value of the order of 10 -2 would seem reasonable for N2Os. For cloud droplets of radius 12/~m attachment of N2Os (M= 108) is calculated to occur at a specific rate of 5.6 x 10 -4 n s-1 . For typical values of n = 100 we obtain a lifetime of 180 s for N2Os. Attachment to small aerosol particles, (n = l04 cm -3, r = 0.3/~m), with the same accommodation coefficient occurs at a much slower rate, the lifetime being ~-25 rain. Even slower rates are expected for removal of N2Os by deposition at the ground under a 100 m inversion using a deposition velocity of 1 cm s-1 . Comparison of these values with the N2Os dissociation rate leads to the conclusion that in a droplet free atmosphere with low aerosol population, the equilibrium between N2Os and NO3 in reaction (2) can be maintained, and the total concentration of NO3 + N2Os will increase with time. This is supported by direct atmospheric observations of NO3 under conditions of low humidity (Plattet al., 1981). Under these circumstances the overall rate of HNO3 formation from NO2 will be determined by the rate at which N2Os reacts heterogeneously with water vapour. On the other hand, if the N2Os + H20 reaction occurs rapidly, e.g., in the presence of droplets, only low steady-state concentrations of NO3 and N2Os will be present. Indirect support for this comes from the direct observations of NO3 radicals in the atmosphere, (Platt e t a l . , 1981). NO3 was generally below limits of detection under humid conditions and

62

R.A. COX ET AL.

on the appearance of fog NO3 concentration was observed to decline rapidly. Under these conditions the overall rate of conversion of NO2 to HN03 by reaction with 03 is determined by kl and the competition between reaction (2) and the alternate loss processes for NO3. Clearly the overall rate of NO2 oxidation is therefore a complex function of atmospheric conditions and a detailed model is required to determine quantitatively the oxidation rate. Some of the kinetic parameters necessary for this model are now available but others, in particular the accommodation coefficients for attachment of N20s molecules onto atmospheric particles, require further experimental study. Furthermore available observational evidence (Platt et al., 1981 ; Noxon et al., 1980) strongly indicates that our knowledge of NO3 chemistry is incomplete and that additional removal reactions are necessary to account for NO3 behaviour.

Acknowledgement This work was conducted as part of a program of Air Pollution Research funded by the Department of the Environment.

References Audley, G. J., Baulch, D. L., Campbell, I. M., and Hamill, L.T., 1980, Evidence for a new intermediate in IN205 decomposition, J.C.S. Chem. Comm. 433-434. Baulch, D. L., Cox, R. A., Hampson, R. F., Jr., Kerr, J. A., Troe, J., and Watson, R. T., 1980, Evaluated kinetic and photochemical data for atmospheric chemistry, J. Phys. Chem. Re./. Data 9, 295471. Becket, K. H., Schurath, V., and Seitz, H., 1974, Ozone-olefin reactions in the gas phase, 1. rate constants and activation energies, Int. J. Chem. Kinet. 6, 725 739. Bhatia, S. C. and Hall, J. H. Jr., 1980, A matrix isolation infrared spectroscopic study of lhe reactions of nitric oxide with oxygen and ozone, J. Phys. Chem. 84, 3255. Chamberlain, A. C., Eggleton, A. E. J., Megaw, W. J., and Morris, J. B., 1960, Behaviour of iodine vapour in air, Disc'. ki~rad. Soc. No. 30, 162. Cox, R. A., 1974, Particle formation from homogeneous reactions of sulphur dioxide and nitrogen dioxide, Tellus 26, 235 240. Connell, P. and Johnston, H. S., 1979, Thermal dissociation of N205 in N 2, Geophys. Res. Lett. 6, 553 556. Davidson, J. A., Viggiano, A. A., Howard, C. J., Dotan, I., Fehsenfeld, F. C., Albritton, D. L., and Ferguson, E. E., 1978, Rate constants for the reactions of O: +, NO2 +, NO +, H3 O+, CO3 , NO2 and halide ions with N20 5 at 300K, J. Chem. Phys. 68, 2085 2087. Davis, D. D., Prusazcyk, J., Dwyer, M., and Kim, P., 1974, A stopped flow-time of flight mass spectrometry kinetics study: Reaction of ozone with nitrogen dioxide and sulphur dioxide, J. Phys. Chem. 78, 1775-1779. Fuchs, N. A. and Sutugin, A. G., 1971, High-dispersed aerosols, in btt. Reviews q/Aerosol Physics and Chemisto', vol. 2 (eds. G. M. Hidy, and J. R. Brock), Pergamon, New York, pp. 1 60. Graham, R. A. and Johnston, H. S., 1974, Kinetics of the gas phase reaction between ozone and nitrogen dioxide, J. Chem. Phys. 60, 4628-4629. Huie, R. E. and Herron, J. T., 1974, The rate constant for the reaction 03 + NO2 = 02 + NO3 over the temperature range 259 362K, Chem. Phys. Letl. 27, 411-414. Morris, E. D. and Niki, H. J., 1973, The reaction of dinitrogen pentoxide with water, J. Phys. Chem. 7"/. 1929-1932.

KINETICS OF THE REACTION OF NITROGEN DIOXIDE WITH OZONE

63

Noxon, J. F., Norton, R. B. and Marovich, E., 1980, NO 3 in the troposphere, Geophys. Res. Lett. 7, 125 128. Ogg, R. A. Jr., Richardson, W. S., and Wilson, M. K., 1950, Experimental evidence for the quasiunimolecular dissociation of nitrogen pentoxide, J. Chem. Phys. 18, 573. Platt, U., Perner, D., Schroder, J., Kessler, C., and Toenissen, A., 1981, The diurnal variation of NO 3, J. Geop173"s. Res. 86, C12, 11965 11970. Pruppacher, H. R. and Klett, J. D., 1978, Microphysi~'s ¢~l Clouds and Precipitation, D. Reidel, Dordrecht. Wu, C. H., Morris, E. D. Jr.. and Niki, H., 1973, The reaction of nitrogen dioxide with ozone, J. Phys. Chem. 77, 2507 2511.