J. CHEM. SOC. FARADAY TRANS., 1994, 90(4), 549-552 Reactions of N(2 ’0) and N(2 ‘P) with 0, Yoshitaka Shihira Department of Energy Sciences, Tokyo Institute of Technology, Ookayama , Meguro-ku, Tokyo 152, Japan Teruaki Suzuki, Shin-ichi Unayama, Hironobu Umemoto* and Shigeru Tsunashima Department of Applied Physics, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo 152, Japan The rate constants for the reactions of N(2 ,D) + 0, and N(2 ’P) + 0, have been measured by employing a pulse radiolysis-resonance absorption technique at temperatures between 210 and 295 K. The rate constants were expressed by the following Arrhenius equations: kN(zD)+Oz= 9.4 x exp(-210/~) and kN(zP)+02= 3.1 x lo-’, exp(-6O/T) in units of cm3 s-’. The results for N(’D) + 0, were compared with the results of quasiclassical trajectory calculations on the basis of modified LEPS potential-energy surfaces.It is suggested that the main exit channel for the N(’D) + 0, reaction is the production of O(’D) under the present conditions. The reactions of metastable atomic nitrogen, N(’D) and They suggested that reaction (1) ought to follow a TI’’ tem-N(’P), with 0, are of great interest, from both an atmo-perature dependence, or possibly a somewhat stronger depen- spheric and reaction dynamics point of view. N(,D) and dence because of the presence of an energy barrier. On the N(’P) are important constituents in the upper atmosphere of other hand, the reaction of N(’P) may have a weaker or even the earth. The rate constants for the deactivation of these a negative temperature dependence, since the initial approach species at room temperature have been measured by many for insertion favours less 0, rotational motion and the investigator^.'-^ The contributions of these reactions to the increased lifetime of the collision complex at lower tem-formation of NO in the upper atmosphere have been dis- peratures raises the chance of a non-adiabatic transition.It is, cussed since a large amount of NO was observed in rocket therefore, necessary to measure the temperature dependence experiments.’ Thermodynamically possible reactive channels of the rate constants quantitatively and to check the model for N(,D) are as follows: proposed by Rawlins et al. Recently, we have succeeded in measuring the temperature N(,D) + 0, -+ NO(X ,lI)+ O(’D) + 175 kJ mol-’ (1) dependence of the rate constants for the deactivation of N(,D) + 0, --+ NO(X ’n) + O(3P)+ 365 kJ mol-’ (2) N(*D) and N(2P) with H, and D,.14 The rate constants could well be represented by Arrhenius equations.The Arrhe- Both these processes are symmetry-allowed. O(‘D) atoms nius parameters for N(2D)+ H, and D, could be reproduced formed in reaction (1) have been regarded as the major by transition-state theoretical calculations as well as quasi- source of the emission at 630 nm in the airglow and aurora.’ ’ classical trajectory calculations. Possible exit channels were Recent laboratory studies have provided important impli- also discussed. In the present work, the temperature depen- cations for the exit channels of these reaction^.'^.'^ Rawlins dence of the rate constants for the deactivation of N(’D) and et a!.’ measured the vibrational and rotational-state dis-N(,P) by 0, were measured.The experimental results were tributions of NO(X ,II) formed in the reactions of N(’D) and compared with those of quasiclassical trajectory calculations N(’P) with 0, near 100 K. They assumed that rotationally on the basis of modified LEPS potential-energy surfaces. excited NO results from reactions of N(’P) with 0,, while rotationally thermal NO results from reactions of N(’D). On Experimentalthe basis of this assumption, they concluded that N(’D) reacts with 0, uia a direct abstraction mechanism, along the The experimental apparatus and the procedure were the same ’A‘ surface which leads to O(’D) formation.They also con- as those described previou~ly.’~ A mixture of N, and 0, in a cluded that O(3P)is produced as result of a transition from stainless-steel vessel was irradiated by a pulsed electron beam the ,A’ surface to the ,A’ or ’A” surface which correlates with from a Febetron 706 apparatus (Hewlett Packard) to produce O(3P). It was also proposed that the reaction of N(,P) and metastable atomic nitrogen. The temporal variation of the 0, proceeds, in contrast, through a long-lived complex, uia concentration of N(,D) or N(,P) was traced by means of the an insertion mechanism on a highly attractive potential absorption of atomic lines of nitrogen. The wavelength used surface.for the detection of N(’D) was 149 nm corresponding to the On the other hand, quantitative experimental information 3,P+-2 ’D transition, while that for N(2P)was 174 nm cor- on the temperature dependences of the rate constants is not responding to the 3 ’P +2 ,P transition. These atomic lines available. Slanger et al. have reported the temperature depen- were derived from a cw microwave discharge in a flow of dence of the rate constant for the deactivation of N(,D) by N,-He. Transmitted light was detected with a photomulti- 02,1 but their result at room temperature has not been sup- plier tube (Hamamatsu, R976) through a VUV monochro-ported by recent measurements.’ Rawlins et al. predicted the mator (Shimadzu, SGV-50). The photomultiplier signal was trend of the temperature dependences of the rate constants.amplified and processed with a wave memory (NF Circuit 5 50 12 c ln m 0 4 0 0, pressure/Pa Fig. 1 Pseudo-first-order decay rates of N(’D) as a function of 0, pressure at 245 K Table 1 Rate constants obtained in the present work ~~ rate constant reaction temperature/K cm3 s-’ N(~D)+ 0, 295 4.57 f0.22 270 4.45 f0.16 245 3.72 f0.12 230 3.81 f0.20 210 3.48 0.15 N(,P) + 0, 300 2.53 f0.11 270 2.53 f0.09 245 2.09 f 0.07 227 2.46 f 0.13 212 2.34 f 0.10 T/K 250 200 1 0-133. 0 4, 0 5b0 103 KIT Fig. 2 Arrhenius plots for the N(,D) + 0, (0)and N(’P) + 0, (a)reactions Table 2 Arrhenius parameters for the N(,D) + 0, and N(’P) + 0, reactions reaction A/10-’2 an3s-l EJkJ mol-’ N(’D) + 0, N(’P) + 0, 9.4 & 1.5 3.1 f0.9 1.8 f0.3 0.5 f0.4 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Design Block, WM-852) and a computer (NEC, PC-9801). The temperature of the reaction system was controlled by introducing cold nitrogen gas from boiling liquid nitrogen to copper tubing surrounding the vessel. For the decay rate measurement of N(’P), the typical pressure of N, was kept at 100 kPa, while for the measurement of N(,D) the gas mixture was diluted with He because N(,D) is deactivated efficiently by N, .15*16 Typical partial pressures of 0, were 0-12 Pa. The sample gases were purified before being introduced into the reaction vessel. 0, (Toyo Sanso) was passed through a cold trap filled with glass beads. N, (Nihon Sanso) was passed through a column containing reduced copper chips at 590 K and a trap filled with glass beads at 77 K.He (Japan Helium Center) was passed through a column of copper chips heated at 590 K and a trap filled with molecular sieve 13X at 77 K. Experimental Results The time variation of the concentration of N(,D) and N(2P) showed an exponential decay. In Fig. 1, the pseudo-first- order decay rates of N(’D) at 245 K are plotted as a function of 0, pressure. Similar linear plots could be obtained at other temperatures and also for N(’P). The rate constants can be obtained from the slopes of such plots. Table 1 sum-marizes the rate constants obtained. The error limit is one standard deviation.The temperature dependence of the rate constants is shown in Fig. 2. The Arrhenius parameters, as determined by a non-linear least-squares method, are listed in Table 2. Both these reactions are characterized by small pre- exponential factors as well as small activation energies. Quasiclassical Trajectory Calculation Three-dimensional (3D) quasiclassical trajectory (QCT) cal- culations were carried out to evaluate the thermally averaged rate constants as well as the nascent vibrational-state dis- tribution of NO. The computational procedure is similar to those for N(,D) + H, and D, described previo~sly.’~ In order to carry out the QCT calculations, the exit channel must be specified. In the reactions of N(,D), both reactions (1) and (2) are energetically and symmetrically allowed. Then, we carried out the calculations by assuming that the reaction proceeds via reaction (1) or reaction (2).It may be considered that both channels are open and com- peting. However, the small pre-exponential factor observed in the present work strongly suggests that most of the exit chan- nels are closed and only one main channel is open. The potential-energy surfaces were calculated by using the modi- fied LEPS method employed in our previous work on N(,D) + H, and D,. In reaction (l),NO(X ,II) + O(’D) does not correlate with N(4S)+ O(3P)+ O(’D), but correlates with N(4S) + O(3P)+ O(3P). The reactant, N(’D) + 0, , correlates adiabatically with N(4S) + O(3P) + O(3P), not with N(,D) + O(3P)+ O(’P).Therefore both the diatomic potential curves for N-0 and 0-0 must be constructed in a manner similar to that employed previously for H--H.I4 For the N-0 curve, the equilibrium internuclear distance was assumed to be the same as that for a free NO molecule, 0.1 15 nm. The disso- ciation energy was calculated to be 448 kJ mol-’ by subtrac- ting the energy difference between O(’D) and O(3P)from the dissociation energy of NO. The value of be was calculated to be 32.7 nm-by using the modified D, and spectroscopically determined 0,.The potential parameters for 0-0, the dis- sociation energy, the equilibrium internuclear distance and the Morse parameter Be, were also determined in a similar J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 manner to be 0.121 nm, 273 kJ mol-' and 36.0 nm-', respec- tively. QCT calculations were performed for various sets of Sat0 parameters, ANOand Am. When AN0 and Aoo were chosen to be 0.161 and -0.020, respectively, the experimental results for the absolute values of the rate constants as well as the temperature dependence could be reproduced satisfactorily. Here, it was assumed that only one ,A' surface leads to the production of O('D). In other words, the apparent rate con- stants were divided by the ratio of the degeneracy factors, 15. This surface has a small, early barrier in the entrance channel. The barrier heights at various bearing angles are summarized in Table 3. The calculated results for the rate constants are shown in Fig.3 together with the experimental results. The vibrational-state distribution of NO was also cal- culated. The result is illustrated in Fig. 4 together with the estimation by Rawlins et al.' Canonical variational transition-state theoretical calcu-lations (CVTST) were also carried out by assuming collinear geometries. The procedure is the same as that described pre- vi~usly.'~In the case of N(2D)+ H, and D,, fair agreement was obtained between the QCT and CVTST results. In the present case, however, the rate constants obtained by the QCT calculation could not be reproduced by the CVTST cal- culation. For example, in reaction (l), the rate constant obtained by CVTST was around twice as large as that for QCT, when the above potential-energy surface (PES) was used.This discrepancy can be ascribed to the very low bending frequency at the transition state. The bending vibra- tional potential is too loose to be represented by a mixed harmonic-quartic potential.' ' Table 3 Saddle-point geometries and bearing angles barrier heights at various (?/degrees rltMlnm r&,/nm V/kJ mol-' 0 0.262 0.121 0.577 25 0.253 0.121 0.614 50 0.223 0.121 0.918 60 0.198 0.122 1.64 70 0.158 0.123 7.66 80 0.134 0.130 46.6 8,Bearing angle between the 0-0 axis and the vector connecting N and the centre of mass of 0,. riM,Distance from N to the centre of mass of 0,. r&,, ,0-0 internuclear distance. V, Barrier height at the saddle point.TIK ~ , ,30,0 , , , , 25; , , , , 27,0-11 -13.0' ' " . " ' " ' 10-13 3. 0 4. 0 5. 0 lo3 KIT Fig. 3 Comparison of the calculated and the experimental rate con- stants for N('D) +0,.(---) Results of QCT calculations. 55 1 0 10 I/' Fig. 4 Comparison of the calculated and the experimentally esti- mated vibrational-state distributions of NO(X *II)at 100 K for N('D) + 0,. 0,Experimental estimate by Rawlins et al.I3 a,QCT calculation for the O('D) channel. A,QCT calculation for the O(3P) channel It was also possible to reproduce the experimental results for the temperature dependence of the rate constants by assuming reaction (2). A satisfactory QCT result was obtained when AN0 and Am were chosen to be 0.180 and -0.385, respectively.However, reaction (2) seems to be less probable because the vibrational distribution of the product NO estimated by Rawlins et al. could not be reproduced at all, as is shown in Fig. 4. We also tried to reproduce the experimentally obtained rate constants for the reactions of N(2P)by assuming various exit channels. However, it was hard to find a LEPS surface which is consistent with the present experimental results. The calculated rate constants were found to be much larger than the experimental ones. The reaction of N('P) with 0, may involve a non-adiabatic process, as has been proposed by Rawlins et ~1.'~ Discussion Comparison with Previous Results Table 4 compares the present results for the rate constants obtained at room temperature with the literature values.'-' For both reactions of N('D) and N(2P),the present results Table 4 Comparison of the rate constants at room temperature reaction k/10-* cm3 s-' technique"*b ref.N('D) + 0, 4.57 f0.22 7.4 PR-RA FP-CL this work 1 5.2 & 0.4 FP-RA 2 5.3 k0.5 DF-RF 3 6.1 DF-EPR 4 6.6 k 1.0 DF-RA 6 4.6 f0.5 DF-RF 7 N(2P)+0, 2.53 f0.11 2.6 k0.2 PR-RA FP-RA this work 2 3.5 k0.13 DF-RF 3 2.51 0.14 1.8 *0.2 PR-RA DF-MPI 5 8 2.2 f0.4 DF-RF 9 Methods for production of metastable nitrogen atoms, PR, pulse radiolysis; FP, flash photolysis; DF, discharge flow. Methods for detection of metastable nitrogen atoms : RA, resonance absorption; CL, chemiluminescence; EPR, electron paramagnetic resonance; RF, resonance fluorescence; MPI, multiphoton ionization.agree well with those of the most recent and reliable measure- ments by Piper et aL7*' The present result for N(,P) also agrees well with our previous value.' Slanger et al.' measured the rate constants for the reaction of N(,D) with 0, over the temperature range 237-365 K. They concluded that the rate constants have a T112depen-dence. The Arrhenius fit to their results yields an activation energy of 0.23 kJ mol-'. This value is much smaller than the present result, 1.8 kJ mol-'. In their flash photolysis study, the relative number density of N(,D) was monitored by the emission of NO(B211-X211) resulting from the reaction of N(,D) with N,O.This measurement must have been affected by the lack of sensitivity. Note that their result at room tem- perature has not been reproduced by more recent and direct measurements. The NO(B 2rI-X 'll) emission mechanism may also be affected by the change in temperature. Exit Channel for the Reaction N('D) + 0, The main exit channel for the reaction N('D) + 0, has been disputed. Link and Swaminathan suggested in their review" that the O(3P)channel is favoured over the O('D) channel by at least 9 : 1 at thermospheric temperatures (300-1000 K). On the other hand, Rawlins et af. suggested that the adiabatic O('D) formation, reaction (l), is the major channel at 100 K by analysing the product state distributions.' The present results for the temperature dependence of the rate constants are consistent with either exit channel.However, judging from the result of the QCT calculations on the vibrational- state distribution of NO, the conclusion of Rawlins et al. is preferred. Rawlins et al. estimated the initial vibrational-state dis- tribution of NO formed in the reaction of N('D) with 0, by using a surprisal analysis. The distribution was fairly excited and peaked at u = 6. Fig. 4 shows a comparison of their result and that of the present QCT calculations. Fair agree- ment between the experimental and calculated results was obtained when reaction (1) was assumed. On the other hand, the calculated vibrational-state distribution is much hotter than the experimental one when reaction (2) was assumed.Both of the potential surfaces employed for reactions (1) and (2) had small, early barriers and their characters were similar except for the exothermicity. Therefore, the difference in the vibrational-state distributions for reactions (1) and (2) must reflect the difference in the exothermicity. Reaction (2) is more exothermic and more energy is partitioned into the product vibrational motion. The discrepancy between the present result and the propo- sal of Link and Swaminathan that the main product is O(3P) can be settled if we consider that the main exit channel J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 changes above 300 K. There must be a large activation energy for the production of O(3P) and this channel is closed at low temperatures.Since the activation energy obtained in the present work is small, it is reasonable to assume that the main exit channel is the formation of O('D) at the present temperatures, 210-295 K. It must be recalled, however, that the present assumption does not exclude the possibility of the production of O(3P)through a non-adiabatic transition after passing through the barrier. In conclusion, the present results are generally consistent with the model presented by Rawlins et al. that the adiabatic O('D) formation, reaction (l), is the major channel for the reaction N('D) + 0, and O(3P) formation, reaction (2), occurs only through a non-adiabatic transition in the 'exit' channel of reaction (1). The reaction of N('P) with 0, may involve a non-adiabatic process.For further discussion, tra- jectory calculations which take surface-hopping into con-sideration will be profitable. Rate constant measurements at high temperatures will also be helpful. References 1 T. G. Slanger, B. J. Wood and G. Black, J. Geophys. Res., 1971, 76, 8430. 2 D. Husain, S. K. Mitra and A. N. Young, J. Chem. SOC.,Faraday Trans. 2, 1974,70, 1721. 3 M. P. Iannuzzi and F. Kaufman, J. Chem. Phys., 1980,73,4701. 4 B. Fell, I. V. Rivas and D. L. McFadden, J. Phys. Chem., 1981, 85, 224. 5 H. Umemoto, K. Sugiyama, S. Tsunashima and S. Sato, Bull. Chem. SOC.Jpn., 1985,58, 3076. 6 P. D. Whitefield and F. E. Hovis, Chem. Phys. Lett., 1987, 135, 454. 7 L. G. Piper, M. E. Donahue and W. T. Rawlins, J. Phys. Chem., 1987,91,3883. 8 C. M. Phillips, J. I. Steinfeld and S. M. Miller J. Phys. Chem., 1987,91, 5001. 9 L. G. Piper, J. Chem. Phys., 1993,98,8560. 10 C. A. Barth, J. Geophys. Rex, 1964,69, 3301. 11 R. Link and P. K. Swaminathan, Planet. Space Sci., 1992, 40, 699. 12 J. P. Kennealy, F. P. Del Greco, G. E. Caledonia and B. D. Green, J. Chem. Phys., 1978, 69, 1574. 13 W. T. Rawlins, M. E. Fraser and S. M. Miller, J. Phys. Chem., 1989,93, 1097. 14 T. Suzuki, Y. Shihira, T. Sato, H. Umemoto and S. Tsunashima, J. Chem. SOC.,Faraday Trans., 1993,89,995. 15 K. Schofield,J. Phys. Chem. Ref. Data, 1979,8,723. 16 K. Sugawara, Y. Ishikawa and S. Sato, Bull. Chem. Soc. Jpn., 1980,53,3 159. 17 B. C. Garrett and D. G. Truhlar, J. Phys. Chem., 1979,83, 1915. Paper 3/05960J; Received 5th October, 1993