Daytime Atmospheric O ( 9 )BY RICHARD D. CADLENational Centre for Atmospheric Research, Boulder, Colorado, U.S.A.Received 7th October, 1963Atmospheric O(1D) concentrations have been estimated for the altitude range 10-240 km. Theestimates for the 100-240 km region are based on calculations of the 6300 8, line intensity by Brandtwhile those for the region below 100 km are based on calculations of the rates of photolysis byultra-violet radiation of 0 3 . The results suggest that in addition to the strong (50 kR) red day air-glow above 100 km, there may be a weaker glow (2 kR) in the 10-100 kni region. The results alsosuggest that the O(1D) concentrations are sufficiently high that the reaction 0(1D)+N2 -+N20+hmay contribute significantly to the N20 content of the atmosphere and the reaction 0(1D)+H20 -+20H may occur at an important rate in the stratosphere.Electronically excited atomic oxygen in the 1D state in the atmosphere is ofinterest for at least two reasons.One is that it is responsible for the red linesA = 6300 A and A = 6364 A (3P2 - ID2 and 3P1- 1Dz) in the airglow and aurora.The other is the possibility, which is discussed later, that it competes with groundstate (3P) atomic oxygen in certain chemical reactions.line, the more intense of the two, in the night airglowis 50- 100 R.l Dissociative recombination,The intensity of the 6300O,+ + e-, 0 ( 3 ~ ) + O('Dj, (1)O(lS)-+O(lD) + 5577 (2)(3)may be the dominant mechanism,2.3 although other processes may contribute :0 + 0 + 0 - 0 2 + O(W)N + 0 3 -+NO2 + O(lO),0 3 +ihv+O2 + O(lD),or the photolysis of ozone,(4)( 5 )in which the ultra-violet radiation is produced by processes such as0 + 0 - 0 2 + hv.(6)The intensity of the 6300A line in the twilight airglow (referred to the zenith)averages about 1 kR.1 That in the day airglow has been calculated by Brandt4to be about 50 kR referred to the zenith. Brandt's calculations suggest that thered lines in the day airglow result from emissions by O(1O) produced both byultra-violet dissociation of 0 2 in the Schumann-Runge region :and by dissociative recombination (eqn. (1)).The concentrations of O(1D) in the atmosphere are several orders of magnitudesmaller than the concentrations of O(3P). However, because of the difference inmultiplicity, certain reactions of O(3P) violate the spin conservation rule 5 , 6 whereasthe corresponding reactions of O(1D) do not.0 2 + JZV+ O(3P) + O(1D) (7)Examples areand0 + N2+ N20 + i h ~0 + CO-+C02 + hv.6R .D. CADLE 67Unfortunately, almost no information is available concerning rate constants forreactions involving O(1D). However, Laidler 5 has pointed out that the transmis-sion coefficient for a reaction involving a change in multiplicity is probably alwaysless than 10-4. Furthermore, the excitation energy for O(lD), about 45 kcal mole-1,can reduce or eliminate any activation energy. Thus certain atmospheric reactionsinvolving O(1D) may occur at velocities comparable with those involving O(3P)in spite of the difference in concentration.The purposes of this paper are (i) to estimate the concentrations of O(1D) in theatmosphere between 10 and 240 km altitude, (ii) to compare the concentrations ofO(1D) with the concentrations of O(3P), and (iii) to estimate the contributions ofO(1D) produced by photolysis of ozone below 100 km to the day airglow.METHOD OF APPROACHBrandt 4 reported estimated rates of emission of 6300 A radiation resulting fromultra-violet dissociation and dissociative recombination in the altitude range 100-240 km.These rates are readily converted to O(1D) concentrations by dividing bythe transition probability, 0.0069 sec-1.The photolysis of ozone 7 by the strong absorption in the 2200-3140 A regionalmost certainly produces O(lD), probably by0, + hv+O:,followed by :OJ-,02('Ag)+ O('0j (-91 kcal)0; -+.02( 'X,") + O('D) ( - 107-5 kcal)or Of+02(3X,)+ O('D> (-69.8 kcal) (13)(12)and probably largely by reaction (11). It may also be produced by absorptionin the 3140-35OOA region by (13). Thus ozone photolysis almost certainly producesconsiderable O(1D) between 10 and 100 km in the atmosphere. The O(1D) in thisregion is destroyed primarily by the following processes :O(lD)-,O(3P) + hv. (16)Thus the steady-state concentrations of O(1D) can be estimated for any given altitudeby the equation,where f is the absorption rate in photons/molecule sec, k14 and kls are rate con-stants for the corresponding reactions, A4 is the transition probability for (16),and the brackets indicate concentrations.Following Brandt, k15 is taken to be10-12 cm3 sec-1 and [MI] is [02]. The rate constant for reaction (14) is about4.5 x 10-34 cm6 sec-1 (Benson and Axworthy 8 ) and k14 is probably no larger thanthis. A4 is 0-0091 sec-1. Thus the first and the last terms in the denominatorcan be ignored and the equation becomes[O(lD)] = 1012f[03][02]-1.Values off were calculated from the absorption coefficients of ozone, the spectraldistribution of the intensity of solar radiant energy outside the earth's atmosphere,68 DAYTIME ATMOSPHERIC O(1D)and the scale heights for 0 3 . The last were calculated from the theoretical valuesof Dutsch.10 These two equations as written imply that the photochemical yield 4is unity. Volman 11 has reported results which suggest that at atmospheric pressureand very low concentrations of ozone in air, 4 may be 0.05 or slightly less.How-ever, his results can be extrapolated to the lower pressures existing at high altitudesusing the equationwhere kll is the rate constant for 11, kg is the rate constant for collisional deactiv-ation of 0; and [MI is the concentration of 0 2 plus Nz. At 30 km, 4 is estimatedto be 0.8, so in view of other uncertainties in the calculation, 4 is assumed to be4 = kl 1/@11+ k,L-Ml),unity.RESULTSThe resulting values off are shown in fig. 1 along with those obtained by Dutschfor the absorption by 0 3 of radiation up to 11,500A and producing both O(1D)and O(3P). Comparison shows that above 30 km most of the atomic oxygen fromozone photolysis, when first produced, is in the 1D state.I I I0-2 -3 -4log f (sec-1)FIG.1.-Values off for the photolysis of 0 3 leading to O(1D) (this paper), and to O(lD)+O(3P)(Diitsch 10).The estimated concentrations of O(1D) are shown in fig. 2 along with total atomicoxygen concentrations.9 Values for [03] were again the theoretical of Dutsch 10for 40 km and lower, and were those given in the Handbook of Geophysics9 foraltitudes above 40 km. The peak concentration for O(1D) produced by ozonephotolysis occurs at about 50 km rather than at the ozone peak, because of boththe rapid increase in f and decrease in 0 2 with increasing altitude.As expected, the concentrations of O(3P) exceed those of O(1D) by severalorders of magnitude and rate constants for reactions involving O(1D) would havR.D. CADLE 69to be correspondingly greater than those for the corresponding reactions involvingO(3P) in order for the former to be of importance to atmospheric chemistry. Asmentioned above, reactions involving O(1D) might be many orders of magnitudefaster than those involving O(3P). However, simple reactions involving atomsseldom have rate constants exceeding 10-12 cm3 sec-1 for second-order reactionsand 10-32 cm6 sec-1 for third-order reactions. Thus the concentrations of O(1D)are so small that many atmospheric reactions, especially third-order reactions, inwhich it might be involved are so slow that they are not of importance.log [O], CM-3Geophysics.9FIG. 2.-Concentrations of (10) and total 0.The latter are taken from the Handbook ofA possible exception is the reaction with N2 to form N20. The activation energyfor this reaction involving O(3P) is estimated to be 13 kcal mole-1 from the activ-ation energy and endothermicity for the decomposition of N20 (53 kcal mole-1and 40 kcal mole-1). Thus the excitation energy for O(1D) may largely or entirelyeliminate the activation energy. This reaction has been investigated by Grothand Schierholz.12, 13 They irradiated mixtures of nitrogen and oxygen (419 mmN2, 7 mm 02) with the resonance wavelengths of xenon at 1470 and 1295 A. Theresulting photolysis produces O(3P) and O(1D) with kinetic energies of both exceedingthe activation energy of the reaction. The N20 produced was removed from thegas mixture by condensation in a liquid-oxygen trap and the N20 concentrationswere determined with a mass spectrometer.Under these conditions about 10-3 ofthe atomic oxygen formed reacted with N2 to form N2O. However, it was notpossible to distinguish the reactions involving O(3P) from those involving O(1D).The following was suggested as a possible mechanism conttibuting to the N20formation :O+ Q2 + M-+O;+M (1 7)(1 8) 0; + Nz-+NzO + 0 70 DAYTIME ATMOSPHERIC o(l0)However, (18) involves a change in multiplicity unless 0; is a triplet. Further-more, the activation energy for (18) (when ground-state 0 3 is involved) is greaterthan the endothermicity of the decomposition of 0 3 into O(3P) and 0 2 , and if 0;is a triplet with this much excitation energy it would probably decompose rapidly.We can make a rough estimate of the rate constant for the reactionO(1D) + N2 + N2-+N20 + N2 (19)from the work of Groth and Schierholz by assuming (a) that O(1D) reacted at leastas rapidly as O(3P) with N2, (b) that the rate constant for the reaction of O(1G)with 0 2 to forin O3 is no greater than that for the corresponding reaction of O(3P)and is about 4.5 x 10-34 cm6 sec-1, ( c ) that the rate constant for the collisional de-activztions of O(1D) is 10-12 cm3 sec-1, and (d) that reactions (19), (14) and (15)were the main reactions leading to the removal of O(1D).Then,and k19 = 10-36 cm6 sec-1.Similarly, if the reaction is second order,O(~D)+N~+N~O+/ZV, (20)k20 = 10-17 cm3 sec-1.Rates of formation of N20 and estimates 9 of N20 concentrations at variousaltitudes are shown in tab!e 1.They suggest that if the reaction is second orderit may be an important source of atmospheric N20, since N20 is quite inert andmust have long residence times at least below the ionosphere.TABLE END RATES OF FORMATION OF N20 AT SEVERAL ALTITUDESaltitude(km)1020406090120formation rc< tecm-3 sec-1for thirdorderrezction0103102110-410-4for secondorderrenction01041041041010N 10 concentrations(Jcm -3101110121010109107-(a) as taken from Handbook of Geophysics.9It is of interest to compare the integrated rate cf photochemical N20 formation,as estimated from the order-of-magnitude results of table 1, with estimates byGoody and Walshaw 14 of the total rate of photochemical destruction of N20(8 x 1010 cm-2 sec-1) and of the average world wide N20 production by bacterialaction in the soil (1-6 x 1010 cm-2 sec-1). The integrated rate is 1010-1011 CM-2 sec-1if the reaction is second order, and 109 cm-2 sec-1 if the reaction is third order.Another reaction of interest is the oxidation of methane.In the ionospherethe methane undergoes photolysis as the primzry oxidation setp. However, in thestratosphere it can react with atomic oxygen and ozone :O(3P) + CH4+products (21)O(1D) + CH4-+products (22)Q3 + CH4-+products. (23R. D. CADLE 71The rate constants for (21) and (22) are not known. However, that for the reactionof O(3P) with n-butane has been found by Elias and Schiff 15 to bek = 5 x 10-11 exp ( - 4200/RT)or about 2 x 10-15 cm3 sec-1 at 25 km.The rate constant for (21) may not bemuch different, in which case, assuming a methane concentration at 25 km of1012 cm-3,(d[CI&]/d~)(21], 25 kmN lo6 SeC-' Cfn-3.The value of 1012 cm-3 for the methane concentration is based on the assumptionof a constant methane+air mixing ratio up to at least 25 km. However, the ex-change time for the atmospheric layer at 25 km is greater than two years,l6 so inview of the probable high rate of (21), the concentration of methane may be about1010 cm-3 at this altitude. Harteck and Kopsch 17, 18 found an activation energyof 7 kcal mole-1 for (21). Even if the activation energy is this high, the mesosphereis probably partially depleted of CH4 by this reaction.The pre-exponential factor, 5 x 10-11, is so large that spin-conservation rulesmust be followed and the products must be a triplet and a singlet, or two doublets.Perhaps the reaction isIt is very unlikely that the pre-exponential factor for (22) will be greater than5 x 10-11 and it will be much smaller unless the spin-conservation rule is obeyed,in which case the reaction might beThe activation energy for (25) might be smaller than for (24).If it is zero,(d[CH4]jdt)(25), 25 k m - 5 x lo4 cm-3 sec-l.The reaction of O(1D) with CH4 may produce CH2, as in reaction (29, but in theexcited singlet rather than the ground triplet state.The rate constant for reaction (23) has recently been determined by Dillemuth,Skidmore and Schubert 19 to beandThus reaction (23) is not an important mechanism for methane oxidation in theatmosphere.O(3P) + CH4+CH2 + H2O.(24)O(1D) + CH4+CH3 +OH. (25)k = 27 x 10-13 exp (- 15,30O)/RT cm3 sec-1(d[CH4]/dt)(23), 2 5 kmN ~ n i - ~ sec-'.Yet another reaction which may be important is0(1D)+H20-+20H+27 kcal. (26)The corresponding reaction of O(3P) is endothermic by 18 kcal mole-1 and is probablynot important. The concentrations of water vapour (about 1014 cm-3) and O(1D)(about 104 cm-3) at 25 km are such that if the pre-exponential factor in the usualexpression for the rate constant is about 10-12 cm3 sec-1 and the activation energyis low, reaction (26) must occur at an important rate.However, the rate constantfor (26) is not known.The concentrations of O(1D) produced by ozone photolysis below 100 km aresufficiently large to produce a weak day airglow. Integrating the concentrationbetween 20 and 100 km and multiplying by the transition probability gives anestimated intensity of 2 kR for the 6300 A line. This can be compared with theintensities mentioned above of 50 kR for the day airglow originating above 100 km72 DAYTIME ATMOSPHERIC o(l0)and 1 kR for the twilight airglow. It might be possible to detect such an atmosphericglow from a rocket at an altitudsl of about 100 km, shortly before sunset, by spectro-scopically analyzing the light received from the direction away from the sun andslightly above the horizon.The 6300 A line has been observed in the day airglow 2 0 s 21 and while the estim-ated intensities are somewhat conflicting they do not markedly disagree with Brandt’scalculated value.It must be emphasized that most of the above estimates are based on a collisionaldeactivation coefficient for O(1D) of 10-12 cm3 sec-1 and the assumption that oxygenis much more efficient than nitrogen in producing the deactivation. The figure10-12, used by Brandt,4 was originally taken from the work of Seaton22 who con-sidered it to be a minimum effective value for the aurora and airglow.The resultsof the calculations could easily be incorrect by an order of magnitude or more.The calculations must be repeated when more accurate values for the deactivationcoefficient are obtained; in the meantime the above calculations indicate what thesituation may be.1 Chamberlain, Physics of the Aurora and Airglow (Academic Press, New York, 1961), p.571.2 Bates, The Threshold of Space (ed. Zelikoff, Pergamon, New York, 1957), p. 14.3 Chamberlain, Astrophys. J. 1958, 127, 54.4Brandt, Astrophys. J., 1958, 128, 718.5 Laidler, Chemical Kinetics (McGraw-Hill, New York, 1950), p. 382.6 Wigner and Witmer, 2. Physik., 1928, 51, 859.7 Leighton, The Photochemistry of Air Pollution (Academic Press, New York, 1961), p. 50.8 Benson and Axworthy, J. Chem. Physics, 1957,26, 1718.9 Geophysics Research Directorate, U. S. Air Force, Handbook of Geophysics, revised edition10 Dutsch, Chemical Reactions in the Lower and Upper Atmosphere (ed. Stanford Research11 Volman, J. Amer. Chem. SOC., 1951,73, 1018.12 Groth and Schierholz, J. Chem. Physics, 1957,27,973.13 Groth and Schierholz, Ber., 1957,90,987.14 Goody and Walshaw, Quart. J. Roy. Meteorol. Soc., 1953, 79,496.15 Elias and Schiff, Can. J. Chem., 1960, 98, 1657.16 Martell and Drevinsky, Science, 1960, 132, 1523.17 Harteck and Kopsch, 2. Elektrochem., 1930, 36, 714.18 Harteck and Kopsch, 2. physik. Chem. B, 1931,12, 327.19 Dillemuth, Skidmore and Schubert, J. Physic. Chem., 1960, 64, 1496.20 Noxan and Goody, J. Atm. Sciences, 1962, 19, 342.21 Wallace, J. Geophys. Res., 1963, 68, 1559.22 Seaton, Astrophys. J., 1958, 127, 67.(Macmillan, New York, 1960).Institute, Interscience, New York, 1961), p. 167.Martell, 1963, personal communication