Formation and Reactions of the Excited O2(A3C;) MoleculesBY P. HARTECK AND R. R. REEVES, JR.Dept. of Chemistry, Rensselaer Polytechnic Institute, Troy, New YorkReceived 20fh January, 1964The emission of the Herzberg bands can be readily observed in the laboratory by catalyzingthe recombination of oxygen atoms on nickel. In a clean system the ozone concentration mayincrease due to the reaction with oxygen (O2A3Z;+O2-+O3$O). Addition of a few % hydrogenresults in a strong OH emission (2Z++2n) apparently due to the reaction 03,43Z;+H2+OH+OH2Ci. Addition of NO give a characteristic reddish hue to the emission. The surface catalysisalso was apparent in several glass systems.The night sky radiation includes emission of thc Herzberg oxygen band system(O2A3E; +02X3E-;) which is generally attributed to the production of the OzA3E;molecule by oxygen atom recombination processes.1 This emission has been ob-served in the laboratory,z and recently by use of a surface catalyst to obtain arelatively intense emission.3 In this paper we describe some observation of theHerzberg emission when enhanced by a surface catalyst.In addition, the effectsof NO, H2 and Nz on the emission are discussed.EXPERIMENTALTwo similar experimental arrangements were used to study the surface-catalyzedexcitation of the Herzberg emission of the oxygen molecules using nickel as the catalyst.One of these consisted of a glow discharge, producing oxygen atom which were pumpedby a 1397 Welch pump into a 45 mm quartz reaction tube containing the nickel catalyst.FIG.1 .-Schematic of apparatus.Spectra of the Herzberg emission from pure oxygen and with the addition of nitric oxideor hydrogen were recorded using a 0.5 m Jarrell-Ash scanning monochromator. Theother was specifically used to detect the formation of any ozone in the system by absorption8P. HARTECK AND R . R . REEVES JR. 83of the mercury 2537A line. The 2537A line was monitored by a photomultiplier tubewhich was in a bridge circuit.A third experimental arrangement shown in fig. 1 was also used which consisted ofa 200 1. reaction vessel which could be readily operated at pressures from 5 to 20 microns.This vessel was evacuated by a 6-in. glass pipe system connected to a CVC MHG-900mercury diffusion pump and a Welch vacuum pump (1398).The vessel could be pumpedat a rate of about 200 l./sec. The pressure was monitored by an NRC Alphatron 530pressure gauge and the light emission was observed with an EM1 6256B photomultipliertube using various filters to obtain the desired wavelength region for observation. D.c.amplifiers were used to produce a signal level which could be plotted on an oscilloscopeequipped with an attached Polaroid camera for direct recording of results.The following technique was used to add a gas stream with a well-defined 0-atom con-centration. The oxygen atoms were produced in a glow discharge and pumped by a 1397Welch pump through a long 45 mm Pyrex tube, where the pressure was generally in theregion from 500 microns to 1 mm. The light intensity in the 45 mm tube at the pointat which the sidestream was removed was monitored by a 931 A photomultiplier tube.The change in intensity with addition of various reactants including NO, H2 and N2, wasalso studied.A sidestream was taken from this tube to feed the 200 1. vessel with the samerelative oxygen-atom concentration.RESULTSUsing a stream of oxygen with an atom concentration of approximately 20 %, theHerzberg bands were observed by eye as a blue luminosity being emitted from the nickelcatalyst surface to a distance of several mm from the surface at pressures from 100 to 300microns. Addition of molecular hydrogen to the stream resulted in emission of the (0,O)and (1,O) OH bands (A2P-tX2II) in the ultra-violet. Addition of appreciable amountsof hydrogen resulted in a marked decrease in the Herzberg emission with correspondingincreased intensity in the OH emission.Further addition of molecular hydrogen reducedthe Herzberg emission to an unobservable level and finally OH bands also disappearedapparently because of interference with the surface catalysis process (spectra shown in fig. 2).When small amounts of nitric oxide were also added to the oxygen-atom stream, inaddition to the oxygen afterglow, a red coloration was observed in the same region wherethe Herzberg bands are emitted. This emission was similar to the emission of the oxygenafterglow from NO and 0-atoms, but shifted to the red. Large amounts of NO quenchedthe Herzberg emission.The addition of molecular nitrogen had no specific effect on the Herzberg emission.These studies were made using a nickel catalyst.It is possible to obtain this type of cata-lysis under normal conditions of fairly clean nickel or cobalt and a reasonably clean oxygenatom stream. To obtain a blue emission readily visible to the eye requires very clean oxygenand a conditioned catalyst. It was not always easy to reproduce the best conditions.Experimentally it was observed that a partial pressure of about 1 p ozone was obtainedwhen excited oxygen molecules were produced by surface catalysis in the 100-300 p pressureregion. Various experimental conditions were used : changing pumping speed, geometricconfiguration of the nickel catalyst, partial pressure of the oxygen atoms, and the totalpressure.Despite this large number of variations the ozone concentration remained inthe region of 1 p and substantially higher concentrations could not be obtained. Withoutthe catalyst, which could be removed by a magnet, the partial pressure of the ozone droppedto less than 0-1 p. Thus, about 10 % of the oxygen atoms must have been converted toozone via the excited OzA3X+, Herzberg level.A weak surface catalysis could also be observed in the large 200 1. glass reaction vesselafter it had been operating for some weeks. This was attributed to undefined surfaceconditioning. A typical surface catalyzed emission is shown in fig. 3. This photographwas made with NO present and has the red coloration mentioned above. This catalysiscovers a surface area of approximately 300 cm2.The Herzberg emission in the blue wasnot intense enough to be visible to the eye, but was readily observed by the photomultiplierFIG. 3.-Reaction vessel with surface catalysis evidcnt at upper left.[To j k e poge 84FIG. 4.-Photograph of Herzberg band emission84 EX c I TE D 02(~3z MOLE c u L E sFig. 4 is a photograph taken in the ultra-violet region only, where the emission of theHerzberg bands can be seen from various areas of a tube through which O-atoms areflowing. All these experiments were performed in a considerably lower pressure region,i.e. 5-20 p. Herzberg emission from the 200-1. vessel was measured as a function of oxygenatoms and nitric oxide concentrations. Generally the Herzberg emission increased pro-portional to the oxygen atoms and was reduced by the addition of NO.However, con-ditions were not stable enough to give reliable quantitative results. The addition ofN o emission observed H 2 z 1 0 '__- - - _ _ _FIG. 2.-Emission of Herzberg bands.hydrogen under these conditions yielded entirely different results compared to the experi-ments in the 100-3OOp region using the nickel catalyst. Apparently hydrogen enhancesthe surface catalysis effect at low pressure and the intensities of the Herzberg emission andthe reddish emission in the visible (assumed to be due to NO2 formation) were increasedin most cases by a factor of 10-100. In a freshly cleaned system no emission of the Herzbergsystem could be observed from the 200-1. flask at those very low pressures.DISCUSSIONBy the use of a catalyst such as nickel it is possible to produce the Herzbergbands with a strong emission intensity.This technique permits some experimentalobservations which are otherwise virtually impossible to obtain in the laboratory.Since the Herzberg emission is observed in the night-sky radiation, it is pertinent tothe chemistry of the upper atmosphere. It is significant that the Herzberg emissionproduced by surface catalysis extends only a few mm from the catalyst indicatinga lifetime of approximately lo-ssec. Since the radiation lifetime is much longer(- 1 sec), the lifetime must be limited by deactivation processes or chemical reactionsP. HARTECK AND R. R. REEVES JR. 85One possibility is the interaction of the excited species with a ground-state oxygenmolecule to produce ozone.The experimental results indicate that ozone is formed,but the maximum partial pressure is limited to about 1 p. This may be partly dueto ozone also quenching the excited oxygen molecules and becoming destroyed inthe process. Alternately, it also may be that the collision of two excited oxygenmolecules results in a deactivation of both by formation of two 0-atoms and theoxygen ground-state molecule. These reactions are given as reactions (l), (2) and(3) in table 1. Those believed to occur with NO and H2 are also listed in table 1.The ozone formation by reaction (1) may be considered analogous to the reactionof excited CO with CO : 49 5co"+co-,co2+c. (6)Studies of the absorption of the iodine line at 2062A by NO indicate that a similarreaction also occurs :NO* +NO+N02 + N (74+N20 + 0.(7b)Excited nitrogen molecules formed by surface catalysts, however, have been foundto be much less reactive than expected.5 The reaction with N2 is impossible sincethere is no indication of a molecule N3 with a reasonable binding energy.It is of basic interest to know what fraction of the 0-atoms result in excited02A3Z$ molecules using the nickel catalyst. An estimate can be made from theintensity of the emission of the Herzberg bands and the fraction of excited mole-cules that will emit radiation compared to the number formed. The diffusion ofthe excited species from the surface corresponds to a lifetime in the order of 10-5sec, limited by deactivation processes.The radiative lifetime is about 1 sec andtherefore the fraction emitting would be 10-5 of those formed. Since the intensityis in the order of 1012 light quanta/sec, then 1017 excited molecules would be prim-arily formed. Since about 3 x 1018 0-atoms/sec stream over the catalyst, roughlyone-tenth must form the 02A3Zfi molecule. This agrees favourably with themeasured ozone concentration corresponding to a 10 % conversion of atoms toozone, but the estimate is only an approximation.The rates of reaction for ozone formation (0 + 0 2 + M-+03 + M) and the ozonedestruction by 0-atoms (0 + 03+202) are of major importance for the upper atmo-sphere. The values given in the literature for these rates are not consistent,7and the steady-state of ozone in a mixture of oxygen and 0-atoms varies also.Afterthe observation of the Meinel bands in the night sky and the laboratory experiment86 EXCITED o~(A~c;) MOLECULESsimulating this emission,s it became evident that the H-atom-ozone reaction wasvery fast and that minor amounts of H-atoms could give misleading results in thestudy of the rate of the reaction between ozone and 0-atoms. Therefore, it seemedthat a system entirely free of hydrogen atoms should give reliable results. However,due to surface catalysis additional ozone can be formed making the ozone concen-tration higher than that corresponding to kinetic equilibrium. Therefore the errorsmay be in either direction and reliable results can be obtained only when both effectsare under control.The authors thank Mr. T. Rolfes and Mr. W. Chace for assistance in obtainingthe experimental results, and Dr. B. A. Thompson for her helpful suggestions. Thiswork was carried out under research grant no. AF-AFOSR-174-63 from the AirForce Office of Scientific Research.1 Barth, J. Geophys. Res., 1962, 67, 1628.2 Barth and Patapoff, Astrophys. J., 1962, 136, 1962.3 Mannella and Harteck, J. Chem. Physics, 1961, 34,2177.4 Groth, Pessara and Rommel, 2. physik. Chem., 1962, 32, 192.5 Harteck, Reeves and Thompson, Z. Natirrforsch., in press.6 Harteck and Safrany, unpublished results.7 Kaufman and Kelso, Chemical Reactions in the Lower and Upper Atmosphere (Interscience,8 Gamin and McKinley, J. Chem. Physics, 1956, 24, 1256.N.Y., 1961), p. 255