Chemical Reactions Contributing to the Nightglow BY D. R. BAITS Dept. of Applied Mathematics, Queen’s University, Belfast Received 8th January, 1964 Recent work on the chemical reactions contributing to the nightglow is briefly surveyed. The view is expressed that in spite of difficulty in interpreting certain laboratory measurements the Chapman and the Barth mechanisms should continue to be regarded as possible sources of X 5577 ; and that three-body recombination of oxygen atoms should continue to be regarded as a possible source of the Herzberg and Atmospheric bands. The position regarding the D-doublet of sodium remains undecided. To avoid unnecessary repetition I shall, in general, confine myself to the research that has been carried out since 1960 when I last reviewed the subject.1 FORBIDDEN GREEN LINE OF ATOMIC OXYGEN Chapman’s suggestion 2 that the presence of A 5577 of 01 in the nightglow spectrum is due to was widely accepted in 1960, not because of any compelling positive evidence, but because of the failure to find a plausible alternative excitation mechanism.To account for the observed mean intensity of 250 rayleighs * it was necessary to suppose that the rate coefficient kl was at least about 2 x 10-35 cm6 sec-1. An attempt at determining kl has been made by Young and Clark 3 who measured the absolute intensity, I(5577), of the green light emitted from a 72-1. bulb con- taining a known quantity of atomic oxygen produced by Assuming that the loss of O(1S) atoms by non-radiative processes is inappreciable they deduced that kl is less than 2 x lO-36cm6 sec-1 and hence that the Chapman process (1) is not responsible for the A 5577 line of the nightglow.However, Young and Sharpless 4 stated later that such a conclusion is premature. Thus finding that where n ( 0 ) is the number density of 0 atoms they were led to believe that some deactivation process like is, in fact, important. They interpreted their measurements as indicating that kl/k4 is 2.6 x 10-22 cm3. If kl were 2 x 10-35cm6 sec-1 or more (as required by Chapman’s theory) k4 would then be at least about 1 x 10-13 cm3 sec-1. Barth and Hildebrandts have also sought to determine kl in the laboratory. With the aid of an electron paramagnetic resonance spectrometer they found n(0) to be 3 x 1015 cm-3 in a mixture of atomic and molecular oxygen (with n(02) ap- proximately 3 x 1016 cm-3) flowing through a tube of 3.5 cm diam.and 48.6 cm length. From their inability to detect A5577 and the calibrated sensitivity of their * A rayleigh corresponds to an emission rate of 106 photons (cm2 column)-1 (sec)-1. 21 0 + 0 + 0+02 + O(1S) (1) NO +N-+N2+ 0. (2) I (5577)ocn(O)2 (3) ops) + o(~P)+o(~D) + ope) (4)22 THE NIGHTGLOW equipment they deduced that the photon emission rate was less than 2 x 107 cm-3 sec-1. To obtain information on deactivation they carried out a series of sub- sidiary measurements on I(5577) from a mixture of atomic oxygen and nitrogen. They inferred from these that collisions with the walls of their tube were unim- portant compared with and that k5 is about 5 x 10-15 cm3 sec-1.Since the radiative transition probability A(1S- 10) is 1.3 sec-1 and A(1S- 3P) is much lower they concluded that O(1s)+o2-O(3P or 1D)+O2, (5) This limit is 250 times smaller than the minimum value needed for Chapmans' theory. In view of the results obtained by himself and Hildebrandt, Barth6 has put forward the hypothesis that the excitation of A 5577 in the upper atmosphere is due to the sequence O+O+M+O~(X,U)+M, (7) 02(X,U) + 0 - 0 2 + O(lS), (8) in which M is any species and (x,v) signifies that the 0 2 molecule is electronically and vibrationally excited. This two-stage mechanism does not differ fundamentally from the single stage mechanism of Chapman. It is not possible to predict with assurance which is the more effective. The internal energy of an O~(X,II) molecule may be reduced below that required for (8) by wall deactivation (rate Wzv), by the emission of a photon (transition probability Asv) or by atom-atom interchange and other collision processes which I shall simply write taking kg to be the mean rate coefficient.It may be seen that the yield of O(1S) atoms from the (7,8) sequence is O~(X,U) + M + 0 2 + M, (9) A7,8n(0)3 cm-3 sec-1, (10) where A,,* is an effective three-body rate coefficient given by k,k,n(M) W""+A""+ k,n(M,' A7,8 = If the sequence is to account for the 25577 line of the nightglow without incon- sistency with the null observation in the experiment of Barth and Hildedrandt 5 it is necessary that 1 7 3 should decrease by a factor of 250 or more in passing from conditions at an altitude of about 100 km to conditions in the tube used.Denoting the number denisties concerned by I atmos and I lab respectively the requirement may be written : 250n(M I lab){AXv+ k,n(M I atmos)) < Wxv. n(M I atmos) Since n(M I lab) and n(M I atmos) are about 3 x 1016 cm-3 and 1.5 x 1013 cm-3 respectively and since Wxv would not exceed 1 x 104 sec-1 even if' every encounter with the sides of the tube brought about deactivation, requirement (12) leads to and It is very doubtful if these inequalities are in fact satisfied. kg c 1 -3 x 10-15 cm3 sec-1, Am < 2 x 10-2 sec-1. (13) (14)D. R . BATES 23 Other hypotheses regarding the excitation of 15577 have been advanced.4~ 7-9 Some of them have been examined critically by Dalgarno.10 None are at all at- tractive.Further laboratory work on process (1) and processes (7) and (8) is desir- able. These remain the two most promising possibilities. HERZBERG AND ATMOSPHERIC BANDS OF OXYGEN It is natural to attribute the Herzberg (A3Zz-,X3Z;) and Atmospheric (blZz-’X3Zg) bands of the nightglow to oxygen atoms recombining to form electronically excited oxygen molecules 0 + 0 + M-, 0,(A3Zi) + M, (1 5) The observed intensities are much less than the intensities which would be expected to result from processes (15) and (16) so it has been suggested there must be rapid deactivation, probably by the atom-atom interchange process 02(a) + 0-0 + 02(p) (17) whiGh may remove either electronic or vibrational energy.1 Barth and Patapoff 11 have reported that they have confirmed in the laboratory that three-body recombination of oxygen atoms does indeed give rise to the Herzberg bands.However, they apparently did not determine the absolute intensity. The spectral range which they covered did not include the Atmospheric bands. Young and Sharpless 4 9 9 have also investigated the emission of the oxygen bands from an afterglow. According to them oxygen atom recombination takes place almost entirely into the ground electronic state of the molecule * so that processes (15) and (16) fail to account for the nightglow intensities (about 1.5 kilo- rayleighs in the Herzberg bands and about 15 kilorayleighs in the Atmospheric bands). It is difficult to see how the ground electronic state could be specially favoured in recombination. The theoretical work of Benson and Fueno 12 has been cited in this connection but it does nothing to make the results of Young and Sharpless understandable. The highest vibrational levels of 02(A3Z:U+) and 02(blX,+) concerned in the night- glow are the seventh and zeroth which have excitation energies with respect to the normal molecule of 4.9 eV and 1.6 eV.The excitation energy of the Herzberg bands, like that of A 5577, is so high that the only obvious mechanism for supplying it is three-body recombination of oxygen atoms. In contrast, there are many mechanisms which can provide the low excitation energy of the Atmospheric bands. However, the great intensity of the emission restricts the possibilities drastically. Perhaps first consideration should be given to the processes involving hydrogen and its oxides.Reactants pass through these processes extremely rapidly as is evident from the observed intensity of the Meinel bands which requires 139 14 that vibrationally excited hydroxyl radicles are formed at a rate of some 1012 (cm2 column)-1 (sec)-1. Catalytic recombination of oxygen atoms by hydrogen is comparable in importance to direct recombination and con- ceivably contributes to the Atmospheric bands by some process like HO2+O+OH+O2(b1X~). (18) * It will be recalled that Barth’s hypothesis on the excitation of h 5577 requires recombination into excited electronic states.24 THE NIGHTGLOW However, it is scarcely worth pursuing this at present since rocket experiments (cf. Packer 15) have shown that the distribution with altitude of the emission of the Atmospheric bands is effectively the same as that of the Herzberg bands and that of 15577 but is markedly different from that of the Meinel bands.B-DOUBLET OF ATOMIC SODIUM Chapman 16 has tentatively suggested that the B-doublet may be excited by NaO + O+Na(2P) + 02, NaH + OjNa(2P) + OH. (19) (20) Na+ 0 p N a O + 0 2 (21) and Bates and Nicolet 17 that it may be excited by The sodium compounds might perhaps be formed by and followed by or Na + 0 2 + M +Na02 + M Na02+O-+NaO+02 NaOz+H+NaH+02. As Dalgarno 10 has pointed out the objection that process (19) is endothermic is not conclusive since the NaO molecules resulting from processes (21) and (23) are probably vibrationally excited. Potter and Del Duca 18 have attempted to determine the altitudes of the emissive layers due to processes (19) and (20) from a theoretical study of the photo-chemistry of sodium and its compounds.They obtained much lower altitudes than the rocket measurements 15 indicate and concluded that neither in (19) nor (20) is the chemi- luminescent process operative. However, the uncertainties are necessarily very great since the rate coefficients of some of the processes are unknown and since the adopted distribution with altitude of n ( 0 ) and n(H) may be considerably in error. It would be unwise to abandon either of the two hypotheses under discussion because of the lack of accord between the calculated and observed altitudes. Another hypothesis on the D-doublet has been advanced by Potter and Del Duca.18 This is that vibrationally excited oxygen molecules transfer some of their energy to free sodium atoms Taking n(Na) to be 1 x 104 cm-3 through a layer 10 km thick it may be seen that the process would yield the winter intensity of some 200 rayleighs if n(02, u>l2) were 1 x 107 cm-3 and k25 were 2 x 10-9 cm3 sec-1.Potter and Del Ducca estimated the rate of formation of 02(u>12) as 5 x 105 cm-3 sec-1. At an altitude of 90 km, n(02) is about 1.3 x 1013 cm-3 and n(0) is about 1 x 1012 cm-3. Hence in order that n(O2, u >12) should be as high as re- quired, the rate coefficients for the deactivation processes and would have to satisfy the inequalities, 02(v > 12) + Na-02 + Na(2P). (25) Oz(u > 12) + 0- 0 + 02(u < 12), 02(U >, 12) 4- 02(U = 0) -+ 02(u < 12) -I- 02(U < 1 2), (26) (27) (28) k26 < 5 x 10-14 cm3 sec-1, k27<4 x 10-15 cm3 sec-1.D. R.BATES 25 These inequalities are quite severe since process (26) can take place simply by atom- atom interchange (which, contrary to the assertion of Potter and Del Duca, can change the vibrational quantum number by more than unity), and since process (27) may be in very close energy balance. In justification of their adoption of an extremely high value of k25 Potter and Del Duca recalled that the emission of the D-doublet is strongly quenched in mole- cular nitrogen, bromine and (presumably) oxygen, the rate coefficients of the col- lision processes involved being of the order 10-10 to 10-9 cm3 sec-1. The process in oxygen would be the inverse of (25), as assumed by Potter and Del Duca, if the electronic energy were converted entirely into vibrational energy.However, some is probably converted into translational energy 10 since the process apparently takes place through the crossing of potential energy surfaces.19 This makes it likely that the value assigned to kzs is much too high. This research has been sponsored by Cambridge Research Laboratories, OAR, through the European Office, Aerospace Research, United States Air Force, under Grant No. AF-EOAR 63-85. 1 Bates, Physics of the Upper Atmosphere, ed. Ratcliffe (Academic Press, New York, 1960), 2 Chapman, Proc. Roy. SOC. A, 1931, 132, 353. 3 Young and Clark, Physic. Rev. Letters, 1960, 5, 320. 4 Young and Sharpless, Reactions of Nitrogen and Oxygen Atoms with Applications to the Upper Atmosphere Nightglow (Technical Report of the Stanford Research Institute, California, 1962), chap. 4, and J. Chem. Physics. chap. 5. 5 Barth and Hildebrandt, J. Geophys. Res., 1961, 66, 985. 6 Barth, Science, 1961, 134, 1426; J. Geophys. Res., 1962, 67, 1628. 7 Krassovsky, Ann. de Geophys., 1958, 14, 395 ; Planet. Space Sci., 1963, 10, 7. 8 Young and Clark, J. Chem. Physics, 1960,32, 607. 9 Young and Sharpless, Reactions of Nitrogen and Oxygen Atoms with Applications to the Upper Atmosphere Nightglow (Technical Report of the Stanford Research Institute, California, 1962), chap. 5 ; J. Geophys. Res., 1962, 67, 3871. 10 Dalgarno, Planet. Space Sci., 1963, 10, 19. 11 Barth and Patapoff, Astrophys. J., 1962, 136, 1144. 12 Benson and Fueno, J. Chem. Physics, 1962,36,1597. 13 Chamberlain and Smith, J. Geophys. Res., 1959, 64, 611. 14 Ferguson and Parkinson, Planet. Space Sci., 1963, 11, 149. 15 Packer, Ann. de Geophys., 1961, 17, 149. 16 Chapman, Astrophys. J., 1939, 90, 309. 17 Bates and Nicolet, J, Geophys. Res., 1950, 55, 235. 18 Potter and Del Duca, J. Geophys. Res., 1960, 65, 3915. 19 Laidler, J. Chem Physics, 1942, 10, 34, 43.