Laboratory Studies of the Chemiluminescence from theReaction of Atomic Oxygen with Nitric Oxideunder Upper Atmosphere Conditions *BY G. DOHERTY~ AND NEVILLE JONATHANGeophysics Corporation of America, Bedford, Massachusetts, U.S.A.Received 13th January, 1964Light emission from the reaction of atomic oxygen with nitric oxide has been observed over thepressure range 0-85-400 microns Hg. By calculation of the partial pressures of atomic oxygen andnitric oxide it is concluded that the light emission shows a first-order dependence with respect tothe atomic oxygen and nitric oxide partial pressures but is independent of the total pressure.The results are interpreted as being consistent with a previously postulated three-body mechanism.In addition, they confirm that a previously determined overall rate constant for light emission isapplicable under upper atmosphere conditions.This rate constant is used to show that the chemi-luminescent reaction of atomic oxygen with nitric oxide, which has been postulated as the cause ofthe night airglow continuum, could indeed be responsible provided that the nitric oxide concentra-tion is of the order of 10s molecules/cm3 over the altitude range 90-110 km.In early studies of the continuum in the night airglow Rayleigh demonstratedthat its spectral distribution could not be the result of scattered sunlight.1 Recenttheories have independently suggested that the continuum is the result of the chemi-luminescence associated with the nitric oxide + atomic oxygen reaction?-5 RecentlyShefov 6 and Yarin 7 have shown that the nocturnal continuum distribution is closeto that of a laboratory-produced airglow.Atomic oxygen is formed above 70km by photodissociation of oxygen mole-cules.Wallace 8 gives estimates of the nighttime densities which vary from approxim-ately 1010 atoms/cm3 at 70 km to 1012 atoms/cm3 at 100 km for a 12-h day followedby a 12-h night. The specific amount of nitric oxide present is unknown thoughtheoretical arguments have been put forward for its existence.9 An upper limitof 108 molecules/cm3 has been placed as a result of experiments using a rocket todetermine the ultra-violet absorption spectrum in the altitude range 60-87 km.10Studies of the precise altitude of the continuum have been hampered by the lowlight intensity levels but rocket experiments have suggested the strongest emissionlies between 90 and 112 km.11, 12The essential mechanism for the nitric oxide + atomic oxygen light emission is0 + NO+N02 +hv,followed by the much faster non-light-emitting reaction,Extensive laboratory studies have been made of this reaction in the pressure range1-5 mm Hg.Most experiments have indicated that in the 1 mm pressure rangethe chemiluminescence is the result of a three-body reaction although the lightemission is second order, dependent only on atomic oxygen and nitric oxide con-centrations.13-15 The rate constant for overall light emission has been found by* This work was supported in part by the National Aeronautics and Space Administration.j- present address : Department of Law, Boston College, Boston, Massachusetts.730 +NO2402 + NO74 NO 4- 0 C HEM1 L UMI NES CEN C EFontijn and Schiff to be 1-7 x 10-17 cm3 molecule-1 sec-1.16 However, these workersdoubt the applicability of their value under upper atmosphere pressure conditions.The present study was, therefore, undertaken to extend laboratory measurementsto pressure regions similar to those which exist in the region of the airglow.Theexperimental data may then be used to calculate whether there is sufficient atomicoxygen and nitric oxide present in the atmosphere to produce the observed nightairglow continuuni.EXPERIMENTALThe reaction cell used has been developed along the lines of multipass cells which arecommon in absorption spectroscopy. It is shown diagrammatically in fig.1. It consistedof a 50-l., 3-necked Pyrex flask covered on the outside by aincrease the light-gathering power. The cell was continuallylayer of magnesium oxide toevacuated through the centreMAGNESIUMFIG. 1NITRIC OXIDE I ~ ~ ~ 7TO PUMPSOXYGEN.--Schematic form of reaction cell for observation of chemiluminescence.arm by a 4-in. oil-diffusion pump which had a maximum pumping efficiency in the 1-100micron pressure region when backed by a large mechanical pump. A liquid-nitrogen trapwas maintained between the diffusion pump and the cell at all times. The cell pressurewas measured by two independent McLeod gauges. The reactant gases were either letICOLD fTRAP - I "I - -CISCUAffiEFIG.2.-Apparatus for observation of chemiluminescence.into the cell through the two side arms and mixed near the top centre of the vessel or theywere premixed at the point of entry into the vessel. There was no significant differencein the observed light emission whichever method was used and so the atomic oxygen andnitric oxide were generally mixed at the point of entry. The atomic oxygen concentratioG . DOHERTY AND N . JONATHAN 75was determined in the reaction vessel at the end of each run by titration with nitrogendioxide added through the second side arm. The resulting light emission was plotted asa function of NO2 flowrate.A block diagram of the complete apparatus is shown in fig. 2. Oxygen (MathesonExtra Dry Grade, 99.6 % minimum purity) was passed, via a flowmeter, needle valve andcold trap, through a microwave discharge.The partially dissociated oxygen entered thereaction vessel after .passage through a Wood's light trap. Known flow rates of nitricoxide were added to the atomic oxygen via a stainless steel needle valve and a calibratedsilicon oil capillary flowmeter. 'I he nitric oxide was purified by passage through a columnof Ascarite and a dry-icefacetone trap placed on the low pressure side of the valve. Across check on the purity of the nitric oxide and on the accuracy of the nitrogen dioxidetitration technique was obtained from the fact that the maximum light intensity from thetitration was equal to that attained with a PNO/PO ratio of 1/4. The kinetics of this havebeen studied in some detail 13 but may be explained simply in this case as being due to thefact that the initial reactionis very much faster than bothandO+NO2+02+NO0 + NO+N02 +hvO+NO+M+N02+M.Hence, during the titration, maximum light emission occurs when half the steady-stateatomic oxygen has been destroyed.At this point the atomic oxygen remaining is equalto the nitric oxide which has been formed and the resulting light emission is equal to theproduct of these partial pressures. At pressures of the order of 10 microns it is consideredmore accurate to use the position of maximum light emission rather than the end pointto calcu'ate the 0 atom concentration.17A Perkin-Elmer 1 12G single-beam, double-pass grating spectrometer equipped withan E.M.I.9558 B photomultiplier was used to record the light emission spectrum between3000 and 7000A. The overall light emission was measured by mounting an R.C.A.1P28 photomultiplier, in a light-tight housing, directly on to the aperture of the reactioncell. The resulting d.c. signal was fed into a Victoreen micro-ammeter. Checks weremade to ensure that the photomultiplier and micro-ammeter gave linear readings overthe range of light emission encountered.RESULTSSPECTRUM OF THE LIGHT EMISSIONThe spectrum of the chemiluminescence was obtained between 3500 and 7500Aat total pressures of 100 and 400 microns Hg. Under our conditions of low resolu-tion, the spectrum appcared as the usual airglow " continuum '' with a maximumintensity at approximately 6500A.The light emission at pressures below 100 pwas too low to allow spectroscopic measurements to bt: made. However, crudespectroscopic measurements of the light emission were made by fitting the IP28photomultiplier (S 5 cathode) and E.M.I. 9558 B photomultiplier (S 20 cathode)with various combinations of Corning filters so that approximately lOOOw of theradiation were observed by the photomultiplier at any time. By comparison ofthe readings obtained with and without the filters at 10 p and 100 p it was concludedthat there was no obvious shift in the spectral distribution at the lower pressure.DEPENDENCE OF THE LIGHT EMISSION ON CHANGES I N CONCENTRA-TION OF SPECIES AND VARIATION OF THE PRESSUREThe atomic and molecular oxygen partial pressures were kept constant whilstusing a steady oxygen flowrate of 26 cm3/min before partial dissociation.Smal76 NO 4- 0 CHEMI LUMI NES CEN CEamounts of nitric oxide were added, via a calibrated flowmeter, to the oxygen streamat the point of entry into the reaction vessel. At pressures below 30 microns theresidence time was short enough so that there was virtually no loss of atomic oxygenin the reaction vessel. At pressures above this, the principal losses were due to thethree-body recombination reactionkiO+NO+M-+N02+M,whch has a rate constant of approximately 6-0 x 10-32 cm6 molecule-2 sec-1.133 18-20The partial pressure of nitric oxide remained unchanged since that which was usedup in forming nitrogen dioxide was regenerated immediately by the very fast secondaryreactionkiwhere ki = 2.6+0-6 x 10-12 cm3 molecule-1 sec-1.Hence, according to steady-state reactor theory, the difference between initial and steady-state concentrationsof atomic oxygen at any pressure isO+NO,+NO+02,where [O,] and [O] are the initial and steady-state concentrations of atomic oxygen,[NO] is the initial (and steady-state) concentration of nitric oxide, [MI is the totalconcentration. All these concentrations can be expressed in units of molecules(or atoms)/cm3. The residence time t of the species in the reaction vessel is equalto V/F, where Y is the volume of gas reduced to s.t.p. and P is the total flowratein cm3 s.t.p./sec. Hence, in table 1 it was calculated that at 43.5p, with a nitricoxide partial pressure of 2 p and a residence time of 6-6 sec, there was a 7 % lossin the atomic oxygen concentration.At higher pressures the corrections becamegreater and difficult to apply.TABLE 1.-~N"ENSlc1[y AS A FUNCTION OF THE ATOMIC OXYGEN AND NITRIC OXIDE PARTIALPRESSURES OVER A TOTAL PRESSURE RANGE OF 3 TO 8 0 pintensity I0total pressure atomic oxygen nitric oxidepressure concentration nit$yo$ide (A) ( x 105 A)( P Hg) (p Hg) ( X 106 A)(P Hg) run no.1 3.02 9.53 19.54 30.05 43.543-56 80.080.080.00.130.440.901-351 -93(1 -79)"3-38(3-01)*3.38(2.70)"3-38(2.43)*0.360.902.13.32.02.01.02.03.00.8252-35-147.8210.514.714.3 513.02 . 9 7 ~ 10-71-08 x 10-52 .5 8 ~ 10-52.1 x 10-52-07 x 10-61 . 4 7 ~ 10-5zwx 10-53 . 9 ~ 10-50.630-520.570.580-540*59*0*49*0.53"0.53 ** These values are corrected for loss of atomic oxygen by three-body recombination, as explainedin text.The pressure in the reaction cell was varied from 3 to 80 ,u by partially closingthe valve between the pump and the cell. At the end of each series of additions ofnitric oxide, the atomic oxygen flowrate was measured by titration with nitrogendioxide. Hence, knowing the flowrates of molecular oxygen, atomic oxygen anG. DOHERTY AND N. JONATHAN 77nitric oxide, and the total pressure, the partial pressure of each constituent wascalculated. The light emission was plotted as a function of the nitric oxide partialpressure at each different total pressure.Below 40p, straight line plots were ob-tained but above this point a slight curvature appeared which became severe at80 p and higher. This was due to losses in the atomic oxygen partial pressure as aresult of the three-body non-radiative recombination process. The results of atypical low pressure run are given in fig. 3. It may be seen that the light emission2.6 &0 0.2 0.4 0.6 0.8 1.0 1.2partial pressure nitric oxide (p Hg)FIG. 3.-Light emission as a function of nitric oxide partial pressure at total pressure of 9.5 p.is directly proportional to the nitric oxide partial pressure. If one defines I0 as thelight emission per micron of nitric oxide and atomic oxygen such thatintensity = 10[0][NO],then an essentially constant value is obtained for 10.A typical set of results for aseries of runs over the pressure range 3 to 8 0 p is given in table 1 and fig. 4. Itmay be seen from table 1 that the light emission is directly proportional to theatomic oxygen and nitric oxide partial pressures and independent of the total pressure(and hence third body) over the range 3 to 80 p.The direct relationship between the atomic oxygen partial pressure and the lightemission was further investigated under conditions of constant total pressure and afixed partial pressure of nitric oxide. The atomic oxygen partial pressure wasvaried by changing the power of the microwave discharge and by moving a copperwire into the oxygen stream immediately after the discharge.After each set ofruns the atomic oxygen partial pressure was measured as already explained. Theresults are shown in fig. 5. It can again be seen that the light emission is directlyproportional to the atomic oxygen concentration and I0 has a constant value ingood agreement with the other data78 NOfO CHEMILUMINESCENCE16 0 -14.0 -12.0 -n ,6: 10.0-c. $2XWY0.0 -.I.E 6.0-4.0 -2 . 0 -0 I 2 3partial pressure atomic oxygen ( p Hg)FIG. 4.-Light emission as a function of atomic oxygen partial pressure over total pressure range3-80 p.0 points at which atomic oxygen loss is negligible. @ points corrected for atomic oxygen loss. r 2.2 7 , I , , I , , , , , , , , , I02.00 01 0 . 2 0.3 0.4partial pressure atomic oxygen (p Hg)FIG.5.-Light emission as a function of atomic oxygen partial pressure at total pressure of 9.5 pG. DOHERTY AND N . JONATHAN 79Attempts were also made to study the light emission at pressures below 3 pby cutting down the flowrate of oxygen to 9-10ml/min. In this way the lightemission was studied over the pressure range 0.85 to 7.5 p. These results are givenin table 2. It may be seen that the value of I0 is again constant but is somewhatlower than before. However, 10 is still independent of the amount of third bodypresent otherwise I0 would show a pressure dependence. At these low pressures,a small error in the measurement of either pressure or flowrates may result in this10 difference. Hence, since the constancy of 10 still holds, we attribute this lower10 to an inherent experimental error.TABLE VALUES OF I0 AT OXYGEN FLOWRATES OF 9-10cm3/minintensity 10total pressure pressure atomic pressurebefore NO addition oxygen nitric oxide oc Hl3) (P Hg) (c' Hg) (A) ( X 105 A)0.85 0.04 0.2 owx 10-7 0.421.0 0.065 0.2 0 .5 2 ~ 10-7 0.405.0 0.234 2.0 0 . 1 8 ~ 10-5 0.387.5 0.442 2.0 0.33 x 10-5 0.37EFFECT OF ADDED GASESThe pressure of the reaction vessel was maintained at 3, 30 and 100 p with con-stant flows of atomic and molecular oxygen and nitric oxide. Small amounts ofa foreign gas were added in quantities not exceeding 10 :< of the total flow, Thegases used were dry air, oxygen, nitrogen, and carbon dioxide. The additions didnot cause any noticeable change in the light emission indicating that there was little,if any, quenching by the added gases.Attempts were made to conduct similarexperiments with carbon monoxide but these results showed a slight but definiteincrease in the light emission, probably as a result of the chemiluminescent reaction0 + co+co2 + liv.DISCUSSIONIt is believed that the experimental data described above are in complete agree-ment with the three-body mechanism already postulated to account for the chemi-luminescent reaction between atomic oxygen and nitric 0xide.14 It is not proposedto reproduce the earlier deduced potential energy diagram of the system, but inorder to facilitate comparison between the original derivation of the mechanism andits description here, the original numbers, etc., are maintained.The complete reaction mechanism for light emission can be written 14 asNO + 0 + M+NO;(C) + MNO;(C) + M+NO + o + MNO:( C) -+ N O:(B) (4)NO;( B) + N 0; ( C)NO:+NO2 + hvNO;+M-+NO2+Mwhere NO*,(B) and NO,*(C) are two different excited electronic states.Using thisscheme and neglecting the possibility of appreciable light emission from NOg(C), thisstate having the longer lifetime,zl the steady-state treatment for light emission yields 180 NO4-0 CHEMILUMINESCBNCEAt low pressures it is anticipated that quenching is an efficient process withalmost every collision effective, which means that k7~[M] and k,c[M] are approxim-ately 1 x 106 sec-1 at 300 p. k 6 ~ is given as 4 x 106 sec-1.21Hence, simplifying with k6B% k7~[M],At low pressure k6~%(k3[M]+k7~[M]) and since I is independent of pressureHence,then k4 < (k3[M] +k7c[M]) which means also that k6+ ks.The above equation is consistent with the light emission showing a pressureindependence in spite of the reaction being a three-body process.Such a reactionscheme would also exhrbit the dependence of the light emission on the nature ofthe third body as found by Clyne and Thrush,ls since k7c[M] varies according tothe nature of M. Fontijn and Schiff 16 pointed out that the pressure independencemust be lost at sufficiently low pressures since the condition (k3+k7c)[M]%k4 nolonger holds. It was calculated that this condition should occur at pressures ofseveral microns and hence they cast doubt on the applicability of their rate constantto upper atmosphere calculations.Insertion of the suggested reasonable valuesof k3 M 10-10 cm3 molecule-1 sec-1 and k7cz 10-10 cm3 molecule-1 sec-1 shows thatat a total pressure of 1 p, k4 must be considerably less than 6.6 x 103 sec-1 for thelight emission to remain pressure independent. In view of our results it is there-fore suggested that a more reasonable upper limit for k4 is lO3sec-1 rather thanthe previously suggested value of lO4sec-1. The three-body mechanism is thenconsistent with the experimental data even down to a pressure of 0.85p. Also,our observation that there was virtually no quenching by the added gases can beaccounted for by this mechanism.The experimental data outlined here are also consistent with a direct two-bodyrecombination reaction mechanism in which there is virtually no quenching of theexcited species.Such a scheme has been postulated as a result of experiments similarin some ways to ours but without direct measurement of the reactant concentra-tions.22 Nevertheless, arguments against this have already been given 14 and inaddition it is impossible to explain the different third-body efficiencies 15 in termsof such a scheme.The importance of our data in so far as upper atmosphere reactions are con-cerned is that the previously determined value for the rate constant for light emission1.7 x 10-17 cm3 molecule-1 sec-1 is applicable to overall system pressures at leastas low as 0.85 p. Such a pressure corresponds to an altitude of 93.5 km.23 Therate constant was therefore used to calculate the possible photon flux of the nightairglow continuum using the formulad[NO;]/dt = 1.7 x 10-”[O][NO].Values for the oxygen atom concentrations at various altitudes were taken fromthe calculations by Wallace for a 12-h day and night cycle8 and the nitric oxideconcentration was given the allowed upper limit 10 of 108 molecules/cm3 over theentire 70-1 10 km range.Temperature effects were neglected since the light emissionhas only a slight negative temperature dependence.15 The results are presented intable 3. The total photon flux was calculated to be approximately 3 x 109 photonsG. DOHERTY AND N. JONATHAN 81cm2 sec. The field data obtained by Shefov 6 indicate a total integrated emissionof approximately 3300 rayleighs which corresponds to a flux of 3.3 x 109 photons/cm2 sec.Consequently, the calculated and field data are in agreement but no greataccuracy should be expected when one considers the uncertainty in the nitric oxideconcentrations. The distribution of photons in the range 70-110 km is in apparentagreement with the airglow continuum intensities as measured by rocket experi-ments. This must be the case since the atomic oxygen concentrations follow thisdistribution and a uniform concentration of nitric oxide has been assumed.TABLE 3.-POSSIBLE PHOTON FLUX FROM THE MGHT AIRGLOW CONTINUUM ASSUMING THECAUSE TO BE THE of- NO REACTIONtotal particle atomic oxygen nitric oxideconcentration concentration concent rat ion flux altitude(km) (molecuIes/cm3) (atoms/cm3) (mo~ecu~es/cm~) bhotomlcrn2 set)70 2.1 x 1015 1 x 101080 4.0 x 1014 2 x 101090 5 .9 ~ 1013 1 x 1011100 7 . 8 ~ 1012 1 x 1012110 1.2x 1012 1 x 101226x 1061 x 108IX 1091 . 7 ~ 109Taken together, the experimental and field data indicate that chemiluminescencefrom the nitric oxide+atomic oxygen reaction may be a prime contributor to thenight airglow continuum. This can only be the case if the nitric oxide concentra-tion is of the order of 108 molecules/cm3 in the 90-1 10 km range.We wish to thank Prof. A. Dalgarno for several valuable discussions during thecourse of this work and Dr. Peter West for helpful comments during the preparationof this manuscript.1 Rayleigh, Nature, 1920, 106, 8.2 Nicolet, The Threshold of Space, ed. Zelikoff (Pergamon Press, Oxford, 1957), p. 40.3 Krassovsky, Dokl. Acad. Nauk, U.S.S.R., 1958, 78, 669.4 Bates, The Earth as a Planet, ed. Kuiper (Univ. of Chicago Press, Chicago, 1954), p. 576.5 Broida and Gaydon, Proc. Roy. Soc. A, 1954,222, 181.6 Shefov, Spectral, Electrophotometrical and Radar Researches of Aurorae and Airglo w (NASA7 Yarin, Spectral, Electrophotometrical and Radar Researches of Aurorae and Airglow (NASA8 Wallace, J. Atmos. Sci., 1962, 19, 1.9 Bates, J. Geophysic. Res., 1950, 55, 306 ; Proc. Physic. Soc. B, 1951, 64, 80510 Jursa, Tanaka and LeBlanc, Planet. Space Sci., 1959, 1, 161.11 Koomen, Scolnik and Tousey, J. Geophysic. Res., 1956, 61, 304.12 Heppner and Meredith, J. Geophysic. Res., 1958, 63, 51.13 Kaufman, Proc. Roy. SOC. A , 1958,247, 123.14 Broida, Schiff and Sugden, Trans. Furaday Soc., 1961, 57,259.15 Clyne and Thrush, Proc. Roy. Soc. A , 1962,269, 404.16 Fontijn and Schiff, Chemical Reactions in the Upper and Lower Atmosphere (Interscience17 Jonathan, to be published.18 Harteck, Reeves and Mannella, J. Chem. Physics, 1959, 29, 1333.19 Ogryzlo and Schiff, Can. J . Chem., 1959, 37, 1690.20 Ford and Endow, J . Chem. Physics, 1957,27, 1156.21 Neuberger and Duncan, J. Chein. Physics, 1954,22, 1693.22 Harteck and Chase, to be published.23 Minzner, Champion and Pond, The ARDC Atmosphere (no. 115 Air Force Surveys in Geo-Tech. trans. F-106, 1962, section IV, no. 5, 1961), p. 45.Tech. trans. F-106, 1962, section IVY no. 5, 1961), p. 39.Publications, New York, 1961), p. 239.physics, 1959)