2434
J . Phys. Chem. 1991, 95, 2434-2431
expected as in the a-NpOD case.
is on the order of a few nanoseconds. Together, both calculations and the experimental results indicate a kinetic isotope effect of kH/kD 1 200.
Summary and Conclusions We were unable to measure a deuteron-transfer rate for a-and 8-NpOD.(ND3),. Based on the steady-state and time-resolved data, the deuteron-transfer rate was determined to be longer than the excited-state lifetime for both systems. The tunneling calculations for both the simple one-dimensional case and with inclusion of heavy atom motion also indicate that the tunneling rate
Acknowledgment. This work was supported by the National Science Foundation. The authors thank Professor E. R. Bemstein and his research group for their collaboration in obtaining the molecular jet mass spectra. The authors also thank Professor D. W. Pratt for a preprint of ref 14.
Thermal Decomposition of Peroxyacetyl Nitrate and Reactions of Acetyl Peroxy Radicals with NO and NOPover the Temperature Range 283-313 K Ernest0 C. Tuazon, William P. L. Carter,* and Roger Atkinson Statewide Air Pollution Research Center, University of California, Riverside, California 92521 (Received: August 27, 1990)
The thermal decomposition of peroxyacetyl nitrate (PAN) in NO-N02-air (or N2) mixtures has been studied at 740 Torr of total pressure over the temperature range 283-313 K. The experimental data obtained yield a rate constant k3 for the thermal decomposition of PAN of k3 = 2.52 X 10'6k'3S73/T s-' and a rate constant ratio for the reactions of the CH3C(0)OO' radical with NO, ( k , )and NO (k2)of k2/kl = 1.95 0.28, independent of temperature over the range 283-313 K. These data are compared and discussed with other literature data.
*
Introduction Peroxyacetyl nitrate, CH3C(0)OON02(PAN), is formed in the atmosphere in the degradation reactions of many organic compounds of both anthropogenic and biogenic origin.1.2 In the atmosphere, the formation of PAN acts as a reservoir of radicals and oxides of nitrogeq2 and its formation then influences the rate of formation of ozone and other manifestations of photochemical smog. Is3v4 PAN is formed from the reaction of the acetyl peroxy (CH3C(0)O0.) radical with NO,: CH3C(0)OO* + NO2
CH3(0)OON02
(1)
with the acetyl peroxy radical being formed from the oxidation of a variety of organic c0mpounds.I The rate of PAN formation and its effectiveness as a reservoir of the CH3C(0)OO' radical and NO, is controlled by the ratio of the rate constant of reaction 1 to that for the competing reaction of acetyl peroxy radicals with N O (reaction 2).Iq5 In addition, the net rate of PAN buildup CH,C(O)OO' + N O CH,C(O)O' NO2 (2)
+
CH,C(O)O'
fast
CH3'
rection is also in the falloff regime between second- and third-order kinetics at and below atmospheric pressure at room temperature,S,l3+15 However, prior to the present study and the parallel independent study of Kirchner et a1.,I6 the only information concerning the rate constant k2 or the rate constant ratio k2/klarises from relative rate s t ~ d i e s , ~ -with * J ~those of Cox et aI.I7 and Cox and Roffey* employing complex reaction schemes. On the basis of these data, the recent chemical mechanisms of Carter et a1.'* and Stockwell et aI.I9 assume that k 2 / k l = 1.5 at room temperature and atmospheric pressure and that this ratio is independent of tem-
+
+ C02
in the atmosphere is also influenced by its thermal decomposition:
-!..
CH3C(0)OON02 CH3C(0)OO' + NO2 (3) which occurs at significant rat= at room temperature and above, with the lifetime of PAN due to thermal decomposition (reaction 3) being -35 min at atmospheric pressure and 298 K.'" An accurate knowledge of the rate of PAN decomposition and the rate constant ratio k 2 / k lis important for the correct formulation of chemical mechanisms used in urban and regional airshed models3 The kinetics of the thermal decomposition of PAN have been studied numerous times,&" and this reaction is in the falloff regime at and below atmospheric pressure at room temperat ~ r e . ~ J *Absolute J~ rate constants for reaction 1 have been determined over the temperature range 248-393 K,I3-l5 and this *Author to whom correspondence should be addressed.
0022-3654/91/2095-2434%02.50/0
(1) Atkinson, R. Atmos. Enuiron. 1990, 24A, 1. (2) Roberts. J. M. Atmos. Enuiron. 1990. 24A. 243. (3j Dodge, M. C. J . Geophys. Res. 1989,'94, 5121. (4) Carter, W. P. L.; Atkinson. R. Enuiron. Sci. Technol. 1989,23, 864. ( 5 ) Atkinson. R.; Baulch, D. L.; Cox. R. A,; HamDson, R. F., Jr.; Kerr, J. A.; Troe, J. J. Phys. Chem. Ref. Data 1989, 18, 881. (6) Pate, C. T.; Atkinson, R.; Pitts, J. N., Jr. J . Enuiron. Sci. Health 1976, A l l , 19. (7) Hendry, D. G.; Kenley, R. A. J. Am. Chem. SOC.1977, 99, 3198. (8) Cox, R. A.; Roffey, M. J. Enuiron. Sci. Technol. 1977, 11, 900. (9) Schurath, U.; Wipprecht, V. Proc. 1st European Symp. PhysicoChemical Behavior of Atmospheric Pollutants, Commission of the European Communities, 1979, pp 157-166. (10) Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. In?. J . Chem. Kinet. 1985, 17, 525. ( 1 I ) Senum, G. I.; Fajer, R.; Gaffney, J. S. J. Phys. Chem. 1986,90,152. (1 2) Reimer, A.; Zabel, F. 9th Int. Symp. Gas Kinetics. Bordeaux, France, July 20-25, 1986. (13) Bridier, I.; Caralp, F.; Loirat, H.; Lesclaux, R.; Veyret, B.; Becker, K. H.;Reimer, A.; Zabel, F. J . Phys. Chem., in press. (14) Addison, M. C.; Burrows, J. P.; Cox, R. A.; Patrick, R. Chem. Phys. Lett. 1980, 73, 283. (15) Basco, N.; Parmar, S . S.In?. J. Chem. Kinet. 1987.19, 115. (16) Kirchner, F.; Zabel,F.; Becker, K. H., Ber. Bunsen-Ges Phys. Chem. 1990, 94, 1379. (17) Cox, R. A.; Derwent. R. G.;Holt. P. M.; Kerr, J. A. J . Chem. SOC., Faraday Tram. 1 1976, 72, 2061. (18) Carter, W. P. L.; Lurmann, F. W.; Atkinson, R.; Lloyd, A. C. Development and testing of a surrogate species chemical reaction mechanism; EPA-600/3-86-031, 1986. (19) Stockwell, W . R.; Middleton, P.; Chang, J. S.; Tang, X . J . Geophys. Res. 1990, 95, 16343.
0 1991 American Chemical Society
The Journal of Physical Chemistry, Vol. 95, No. 6, 1991 2435
Peroxyacetyl Nitrate and Acetyl Peroxy Radicals perature. In contrast, to fit model simulations to the results of some outdoor environmental chamber experiments, Gery et a1.20 derived a large temperature dependence of e-5250/T for k2/kl for use in the latest carbon bond mechanism. Dodge3 showed that this large temperature dependence in the k2/kl ratio in the carbon bond mechanism of Gery et causes that mechanism to predict significantly less ozone formation at temperaures below -298 K than do the mechanisms of Carter et aI.'* and Stockwell et aI.l9 This arises because the expression of Gery et aL20 predicts that as the temperature decreases, the reaction of the acetyl peroxy radicals with NO2 becomes increasingly more important relative to its reaction with NO, while the other two mechanisms predict that these two reactions are much less sensitive to temperature. Since the reaction with NO2, forming PAN, is a radical termination process, while the reaction with N O is radical propagating, the carbon bond mechanism predicts lower overall radical levels (and thus slower ozone formation rates) at lower temperatures than do the other two mechanisms. In this work, we have studied the decay rates of PAN in the presence of NO and NO2 over the temperature range 283-3 13 K and determined the rate constant, k3, for the thermal decomposition of PAN and the rate constant ratio k2/kl.
Experimental Section The dark decay rates of PAN were monitored in the presence of N O and NO2 at a total pressure of 740 Torr of air or N 2 over the temperature range 283-313 K. Experiments were carried out in a 5800-L evacuable, Teflon-coated, environmental chamber2' fitted with a multiple reflection optical system interfaced to a Nicolet 7 199 Fourier transform infrared (FT-IR) absorption spectrometer. The chamber was temperature controlled by using an ethylene glycol heating/cooling system and maintenance of the temperature within h0.6 K was monitored by several thermocouples placed inside the chamber. The concentrations of PAN, NO, NO2, and other reaction products were monitored by FT-IR absorption spectroscopy. For each experiment, PAN was introduced into the chamber, and its decay was followed for 30-1 50 min, depending on the temperature, after which N O was added. At each temperature, the rate of decay of PAN alone was established to be totally negligible with respect to the rate of decay in the presence of NO. The initial concentrations were (in molecules cm-3 units) PAN (1.15-1.37) X I O l 4 and N O (2.41-2.94) X lOI4. NO2 was not added initially but was formed during the dark decay of PAN and was present prior to N O addition at concentrations of (0.12-1.45) X 10') molecules ~ m - ~Synthetic . air (80%N 2 + 20%0,) was used as the diluent gas, except for one experiment at 301.5 K for which N 2 was used as the diluent gas. After the addition of NO, the PAN decays were monitored for from -9 min at 313 K, with 60%loss of PAN and spectra recorded every 28 s, to -3 h at 283 K, where a 40% loss of PAN was observed and spectra were recorded every 8-1 1 min. PAN was prepared by photolysis of a C2HSONO-N2-02 mixture in a specially designed reactor22and purified by preparative-scale gas chromatography. The only impurities detected by infrared spectroscopy were traces of methyl nitrite and water. The ethyl nitrite used in the feed gas mixture was distilled from a commercial sample containing 15% (by weight) C 2 H 5 0 N 0in ethanol (Aldrich Chemical Co.). N O (299.0%) and NO2 (199.5%) were obtained from the Matheson Gas Co. The spectra of the reaction mixtures were analyzed by successive subtraction of absorption features with the use of calibrated reference spectra. The interactive subtraction routine yielded factors of the concentrations represented by the reference spectra. The initial PAN spectrum for each experiment served as reference ~
~~
____
____
"0
loo0
2000
3000
TIME (s)
Figure 1. Measured time-concentration profiles for PAN, NO,and NO2 in air diluent at 301.5 K.
and the subtraction/analysis of PAN was based on its absorption band at 1841.5 cm-I, with an absorpti~ity'~*~~* of 10.2 cm-l atm-' (base 10). N O and NO2 analyses were based on their finestructured absorption bands centered at 1876.0 and 1617.5 cm-l, respectively. Reference spectra of N O and NO2 for the different temperatures were generated by flushing known partial pressures (measured at room temperature with an MKS Baratron, 100-Torr sensor) of the gases in calibrated 2-L and 5-L glass bulbs into the 5800-L chamber filled with 1 atm of N2. In the case of NO2, the partial pressures in the glass bulbs were c ~ r r e c t e d ~ for ~ - ~the ' presence of N204, with the correction for the range of partial pressures (up to 4.5 Torr) employed amounting to 14%. The above calibration for NO2 was also verified by introducing into the chamber known weights of N 0 2 / N 2 0 4in a small vial. Subtraction of absorptions by some of the reaction products facilitated the analysis of NO2. The reaction products observed were C H 3 0 N 0 2 , C H 3 0 N 0 , CH3N02, HCHO, HONO, and HN03. The room-temperature reference spectra that had been previously compiled for these compounds were sufficient to facilitate processing of the spectra. Systematic errors in the measurement were estimated to be within 24% for PAN and within f 8 % for N O and NO2.
Results In the presence of N O and NO2, the decomposition of PAN is governed by reactions (1-3) above. Thus, at a constant total pressure -d[PAN] k,k3[PAN] [NO]
-
dt k 2 [ W + kl[N021 -d In [PAN] kzkdNO1
-
(1) (11)
dt k 2 W I + kl"21 wher the rate constants kl and k3 are the bimolecular rate const ts for the total pressure (and [MI) conditions employed. Rearrangement of eq I1 leads to
b
and plots of (4 In [PAN]/dt)-' against [NO,]/[NO] should yield
~~
(20) Gery, M. W.; Whitten, G. Z.; Killus, J. P.; Dodge, M. C. J . Geophys. Res. 1989,94, 12925. (21) Winer, A. M.;Graham, R. A.; Doyle, G.J.; Bekowies, P. J.; McAfee, J. M.; Pitts, J. N., Jr. A d a Enuiron. Sei. Technol. 1980, IO, 461. (22) Stephens, E. R.; Burleson, F. R.; Cardiff, E. A. J. Air Polluf. Control Assoc. lW5,15. 87.
(23) Stephens, E. R. Anal. Chem. 1964,36,928. (24) Tsalkani, N.; Toupance, G. Afmos. Enuiron. 1989,23, 1849. (25) Verhook, F. H.; Daniels, F. J. Am. Chem. SM. 1931. 53, 1250. (26) Dunn, M. G.; Wark, K., Jr.; Agnew, J. T. J . Chem. Phys. 1%2,37, 2445. (27) Harris, L.; Churney, K. L. J . Chem. Phys. 1967,47,1703.
Tuazon et al.
2436 The Journal of Physical Chemistry, Vol. 95, No. 6, 1991
TABLE k Rate Comtuts k3 a d Rate Constant Ratios k d k , Obtained from the Decays of PAN in N+NOZ-Alr (or N2) Mixat 740 Torr of Total Pressure
T,"K
1@k3: s-I 0.413 0.025 0.628 0.038 1.09 0.09 2.99 t 0.12 6.99 0.63 7.11 0.22 14.5 1.9 30.0 2.7 41.0 3.3
283.4 285.7 289.5 295.9 301.5 301.5c 307.3 31 1.2 312.9
k2/klb 1.98 0.16 1.76 0.13 1.96 0.22 1.95 0.10 2.1 1 0.22 1.72 0.14 2.30 0.30 1.92 t 0.22 1.86 0.15
** * * * *
* * *
In air diluent, except as indicated. Indicated statistical errors are 1 least-squares standard deviation. N2 diluent gas. 1x10*
-..
4ooo
0
I2000
eo00
5
TIME (s)
Figure 2. Plots of In [PAN] against reaction time for the thermal decomposition of PAN in NO-N02-air mixtures at 740 Torr of total pressure and 289.5 and 301.5 K.
'r
301.5
t 0
2x0-5 3.0 2
4
6
8
CN0,I 1 CNOI
Figure 3. Plot of cq I11 for the thermal decomposition of PAN in an
NO-N02-air mixture at 301.5 K. straight line plots with intercepts of I l k 3 and slopes of k l / k 2 k 3 . Figure 1 shows a typical set of timeconcentration profiles for PAN, NO, and NO2 during the decomposition of PAN in an N 0 - N 0 2 a i r mixture at 301.5 K. In all experiments carried out in air diluent, the N O and NO2 concentrations, and hence the [NO]/ [NO,] concentration ratios, changed markedly during the experiments due to the thermal oxidation of N O to N02,5 and hence eq 11 predicts that the PAN decays would not be exponential. For the experiment carried out in N 2 diluent, the conversion of N O to NO2 was significantly slower than in the analogous experiment in air diluent, with the N 0 2 / N 0 concentration ratio varying by a factor of 7.6 in N2 diluent compared to a factor of 18 in air diluent over the same time interval. As shown in Figure 2 for the decays of PAN in NO-N02-air mixtures at 289.5 and 301.5 K, the PAN decay rates, -d In [PAN]/dt, were gentle curves. For each experiment, k3 and k2k3/klwere obtained from plots of (4In [PAN]/dt)-l against [NO,]/[NO], as indicated by eq 111. For each time t, the value of d In [PAN]/dr was approximated by {d In [PAN]/dtJ, = (In [PAN],, - In [PAN],,)/(t2 - t , ) (IV) where tl is the time of the measurement made immediately prior to time t, and r2 is the measurement immediately following. (In most cases, t - rl = r2 - t . ) Similar results were obtained by fitting the PAN concentration-time data to smooth polynomial curves and using these equations for the curves to derive values of d In [PAN]/dt. However, the present method (eq IV) was preferred, since it gave a more accurate indication of the magnitude of the scatter of the data.
I
I
32
3.4
36
IOOOIT (K)
Figure 4. Arrhenius plot of the PAN thermal decomposition rate constant k3 at 740 Torr of total pressure of air or N2 diluent. (---) Recommendation at 740 Torr of total pressure from the IUPAC data evaluation panel.s
A plot of eq I11 for the 301.5 K data shown in Figures 1 and 2 is shown in Figure 3, and a reasonable straight line plot is evident. The values of k3 and k 2 / k lobtained from the intercepts and s l o p of these plots, respectively, by least-squares analyses are given in Table I. As seen from Table I, the data obtained in N 2 or air diluents a t 301.5 K were indistinguishable within the two least-squares standard deviations. With the Arrhenius expression, a unit-weighted least-squares analysis of the rate constants k3 for the thermal decomposition of PAN at 740 Torr of total pressure of air or N2 leads to k3 = ( 2 %+4.00 , ~ ) x 1016e-(13573&284)/3s-l where the indicated errors are 1 least-squares standard deviation. The Arrhenius plot for k3 is shown in Figure 4. The Arrhenius plot for the rate constant ratio k 2 / k l is shown in Figure 5. Within the experimental errors, this rate constant ratio k z / k l is independent of temperature over the range (283-31 3 K) studied (Table I), and hence a unit-weighted average leads to k,/kI = 1.95 f 0.28 where the indicated 1 least-squares standard deviation includes both random errors and the systematic uncertainties in the N O and NOz measurements. Discussion The rate constants k3 determined in this work for the thermal decomposition of PAN at a total pressure of air or N2 diluent of 740 Torr are in good agreement with the recent recommendation
The Journal of Physical Chemistry, Vol. 95, No. 6, 1991 2431
Peroxyacetyl Nitrate and Acetyl Peroxy Radicals
3.0r ==
N '
t
t
0"
30
32
2.5
2x)
et a1.I' at 296 K and the rate constant ratios determined over the temperature range 296-328 K by Cox and Roffey? which range from 1.2 to 3.1 with an average value of 1.9 f 0.7 and no obvious dependence on the temperature within the large uncertainties. An average of the present data and those of Kirchner et a1.I6 leads to a rate constant ratio kJkl = 2.2 at 740-750 Torr of total pressure, independent of temperature over the range 283-321 K. These two sets of data indicate that the temperature dependence of this rate constant ratio is