Ion and Charge Exchange Reactions involving AtmosphericGasesBY J. SAYERS AND D. SMITHElectron Physics Dept., The University, Birmingham 15Received 15th January, 1964In order to study reactions between ions and neutral particles, a series of experiments have beencarried out which utilize time resolving mass-spectrometric monitoring of the afterglows in atmo-spheric gases produced by radio-frequency discharges. The following reactions have been investig-ated : (i) the charge exchange rates of the O+ ion with oxygen and nitrogen ; (ii) the charge exchangerates of the He+ ion with oxygen and nitrogen. Rate coefficients and cross-sections are presentedfor some of these reactions, together with suggestions concerning the mechanisms of the reactions.Since measurements of the rates of reaction between ionized and neutral atomsof atmospheric gases were required at thermal energies, the afterglow method wasused. Microwave methods could not be used in the present studies since the chargetransfer reactions investigated produce no obvious effect on the electron loss process.The method is essentially that of Dickinson and Sayers 1 and consists of ionizingthe gases under investigation with a high power radiofrequency discharge pulse andobserving the ion density decay rate in the afterglow. The method assumes that thecurrent flowing to the mass spectrometer collector is proportional to the instan-taneous undisturbed ion density in the afterglow plasma.The discharge plasma isproduced in a large glass envelope.In the afterglow, the ions and electronsrapidly reach thermal equilibrium with the gas. In this way, ions are producedhaving a Maxwellian energy distribution and their reaction with neutral species canbe studied at thermal energies. Similar work has been carried out by Fite and hisco-workers.2, 3The great advantage of the mass-spectrometric method is that positive identifica-tion of the reactants and products in a given reaction can be obtained and also themass spectrometer can be used as a monitor of any impurities present in the after-glow. These measurements are now possible due to the development of mass spectro-meters with short ion path-lengths which reduce considerably the chance of reactionstaking place between the ions and neutral species in the mass spectrometer itself.In all the measurements carried out in the various gas mixtures, neutral heliumwas the dominant species present and He+ was the dominant ion formed in theionizing pulse.In all cases, the helium was at a sufficiently high pressure so as toact as an effective buffer to the diffusion of ions and electrons to the walls of thedischarge vessel. In one series of measurements described, the rate of decay ofthe He+ ion in the presence of neutral oxygen and then nitrogen was studied. Themost important product of the reactions was in the first case the Of ion and in thesecond case the Nf ion. The rate of decay of the O+ ion in the presence of mole-cular oxygen and then molecular nitrogen is also described.EXPERIMENTALThe discharge vessel consisted of a Kodial glass cylinder of diameter 14-5 cm and overalllength about 30 cm.The electrode system consisted of two parallel flat nickel discs of16168 ION AND CHARGE EXCHANGE REACTIONS11 cm diam. inside the vessel placed about 16 cm apart and two external sleeves of copperabout 5cm wide placed about 7cm apart. The mass spectrometer was contained in aside-arm of the main discharge vessel, the ions formed in the discharge tube entering themass spectrometer through a small circular orifice of 0.18 mm diam. at the centre of aNilo-K disc of 2.5 cm diam. which was effectively sealed to the wall of the discharge vessel.Particular attention was paid to obtaining good vacuum conditions. Kodial glass wasused in preference to Pyrex glass: the latter is reputed to give off chloride ions whenbombarded in a gas discharge. Bakable metal taps were used throughout the system tothe exclusion of all grease joints and rubber gaskets.The discharge vessel was baked to350°C using an oven and the rest of the tubulation, gas train, etc., was outgassed usingheating tapes. Where possible, metal parts were outgassed by eddy current heating. Thisallowed the residual pressure in the system to be reduced to 10-7 mm Hg. Cold traps wereincluded above mercury diffusion pumps and in the gas circulating system. The massspectrometer was differentially pumped, the gas being returned to the main dischargechamber. Hence, during operation, the system was isolated from the evacuating pumps.The gases used in the experiments were spectroscopically pure gases obtained from theBritish Oxygen Co.Ltd., and the glass containers were sealed to the system via metal taps.Pressure measurements were made with a McLeod gauge.The ionizing pulses were of 10 psec duration at a frequency of 7 Mc/sec and a repetitionfrequency of 50 pulses/sec synchronized to the a.c. mains supply. The power in the pulsecould be varied continuously from the minimum required to initiate the discharge to about200 kVA. A constant discharge power of about 25 kVA was usually used in these ex-periments. Ion currents flowing to the mass spectrometer collector were usually - 10-9 Aand in order to measure these currents as a function of time they were passed throughthe grid resistor of a wide-band amplifier, the resulting time-varying amplified voltage beingdisplayed on an oscilloscope screen.To obtain the gas mixture in the required proportions, the gas required at the lowestconcentration, usually oxygen or nitrogen at a pressure of - 10-3 mm Hg, was first intro-duced and its pressure measured accurately with the McLeod gauge.Helium gas was thenquickly introduced to raise the total pressure to approximately 0.5 mm Hg pressure.Measurements of decay rates of various ion species were obtained by photographing theoscilloscope display together with a calibration time trace. From the photographs, theconcentration of the tuned ion species as a function of time in the afterglow was obtained.In some cases, it was found necessary to use a cine camera to record the oscilloscope traces,since a rapid movement of the oscilloscope trace resulted during the first few ionizing pulses.RESULTSOXYGENPreliminary measurements of the decay rate of the Of ion density in oxygenwere made by Dickinson and Sayers.1 The discharge tube was not baked, the onlyattempt at cleaning it was by running a discharge in helium for several hours.Fromthe semi-logarithmic plots of ion density as a function of time in the afterglow,a decay rate 13. can be obtained for different partial pressures of oxygen. A plotof A as a function of oxygen partial pressure should produce a straight line, theslope of which provides a value for the rate coefficient k of, in this case, the reactionThe value given by Dickinson and Sayers iso++02+o,’+o.k = 2-5+0*4 x 10-11 cm3 sec-1 at 293°K.The experiment was repeated in the present apparatus without baking and thevalue obtained for the rate coefficient wask = 2.6-&0-4 x 10-11 cm3 sec-1 at 293°KJ .SAYERS AND D. SMITH 169It was noticed, however, that under these conditions an ion of mass 12, probablycarbon, was present in the afterglow as an impurity.The experiment was repeated using the improved vacuum techniques describedin the previous section. The decay of the 0' ion was much slower than beforeand the graph of decay constant against partial pressure of oxygen showed lessscatter of the points about a straight line(see fig. 1). Also, ionized carbon was not0 4 6 8 10 I2oxygen partial pressure in mm Hg x 103FIG.1.-Decay constants for the Of ion density in oxygen afterglows at 293°K.observed in the afterglow. The reaction rate coefficient determined from theseresults isk = (1.64+0-05) x 10-11 cm3 sec-1 at 293°K.The corresponding cross-section isCT = (2*15+0.06) x 10-16 cm2.There seems little doubt that the lower value obtained is due to the more stringentvacuum techniques adopted in this work.The temperature dependence of the rate coefficient for the O+ ion reaction withoxygen has also been investigated. Prolonged baking was required in order to ob-tain measurements at the higher temperatures in order to reduce the impurity level.Measurements were also made below room temperature by surrounding the dis-charge vessel with solid carbon dioxide.Rate coefficients and correspondingcross-sections for the various temperatures are given in table 1. Over this restrictedrange of temperature, the rate coefficient appears to be approximately proportionalto the inverse square root of the temperature.TABLE 1temp. rate coefficient."K cm3 sec-1 x 1011452 1-35 -f0*05388 1-72 f0.06293 1.64 f0-05210 2.w *@locross-section,cmzx 10161.32 f0.051-95 f0.072.1 5 f0-063-70 f0-2170 ION AND CHARGE EXCHANGE REACTIONSThe charge exchange reaction occurring here is0++0~-+02++0~and it has been suggested4 that the dominant mechanism is ion-atom interchange.The observed increasing cross-section as the temperature decreases is in qualitativeagreement with the theory of this process.However, on the basis of the Gioumousisand Stevenson equation 5 the theoretical rate coefficient for the reaction is 9 x 10-9cm3 sec-1 which is an order of magnitude greater than the present results.By studying the emission of the oxygen red line (6300A) from the night-time I:layer, which he assumed was emitted from decaying excited oxygen atoms formedin the above charge exchange reaction, Nakamura 6 has calculated a value for therate coefficient ask = 1.3 x 10-11 cm3 sec-1at the temperature in the F layer ( N 1400°K). This value is in close agreement-perhaps fortuitously-with the present work. The phenomena of clean-up ofoxygen was always noticeable in the afterglow to some extent. Hence the photo-graph of the oscilloscope trace had to be taken as soon as possible after the initiationof the discharge.0XYGEN-I-NITROGEN MIXTURESThe reaction between the ionized oxygen atom and the nitrogen molecule is ofespecial interest in the upper atmosphere for it is known that the concentration ofmolecular nitrogen is considerable.The method of examining the reaction is toadd nitrogen to a mixture of oxygen and helium, such as that used in the previousexperiment. Any change in the decay rate of the Of ion must be due to the addedcomponent. An effective total rate coefficient kT may be given by the relationwhere ko is the rate coefficient for ion loss due to the oxygen molecule, kN is therate coefficient for ion loss by the nitrogen molecule, and a is the ratio of the nitrogenpartial pressure to the oxygen partial pressure.The measurements were made in two series, in one of which the nitrogen-oxygenratio was approximately 1 : 1 and in the other series the ratio was approximately2 : 1.It was not consistently possible to obtain nitrogen and oxygen pressuresthat had exact 1 : 1 or 2 : 1 ratios, but these ratios were maintained to within 10 %.The results observed for the decay rate ;1 as a function of oxygen or nitrogenpartial pressure were somewhat scattered, and to obtain reasonable accuracy inthe h a 1 value of the rate coefficient kT, many more measurements were taken thanwith pure oxygen. The values of the effective total rate coefficient kT, obtained for thetwo different mixtures at 293"K, together with the value for the rate coefficient ofthe O+ ion in pure oxygen, are shown in fig.2. From this graph the rate coefficientfor the loss of oxygen by reaction with molecular nitrogen at 293°K iscorresponding to a cross-section ofThe addition of a third gas to the discharge increases considerably the possible ion-exchange reactions. However, most reactions can be excluded in favour of thereaction,kT = ko+akNkN = (2-7 0-2) x 10- l1 cm3 sec- ',0 = (3*6&0.3)x cm2.O++NpNO++N+l.l eVJ . SAYERS A N D D. SMITH 171and it is assumed that this is the most important reaction mechanism. The resultobtained for this reaction can be compared with the value of the charge exchangecross-section measured by Hasted 7 using beam techniques. The value he obtainedwasCJ = 2.8 x 10-16 cm2at an energy corresponding to 32,000"K, and this value tended to zero as the energywas reduced.The value obtained for the cross-section in the present work, althoughcarried out at a lower energy, is larger than the above value.The results described in this and the previous section have been described inmore detail by Batey.8ratio of nitrogen to oxygen partial pressuresFIG. 2.-Variation of the effective rate coefficient for the reactions of the O+ ion in Nz+ 0 2 mixtureswith the composition of the mixture at 293°K.HELIUM -k NITROGEN MIXTURESThe partial pressures of nitrogen were in the range 10-4 to lO-3mmHg andhelium was added to make the pressure up to about 0.6 mm Hg. In fig. 3 the decayconstant ;1 for the He+ ion density is shown as a function of the nitrogen partialpressure for two different discharge powers.The curve is expected to be a straightline with a small intercept on the decay constant axis corresponding to decay ofthe He+ ion by diffusion and conversion to the molecular ion only. However, thecurves in fig. 3 show intercepts on the nitrogen partial pressure axis.In order to explain this it has been suggested that a significant amount of dis-sociation of molecular nitrogen, which is assumed to be the neutral reacting species,occurs in the ionizing pulse. It must be further assumed that the number of dis-sociations is independent of the nitrogen partial pressure so that the partial pressureof molecular nitrogen is decreased by a constant difference from the measuredvalue.The value of this constant difference would decrease with a decrease in th172 ION AND CHARGE EXCHANGE REACTIONSdischarge power, and so the forms of the curves in fig. 3 are explained. This as-sumption is not very satisfactory and the effect could be due to the phenomena ofclean-up of nitrogen in the discharge.3 0 -x0- I I I I I I I I I I 0 I 2 3 4 5 6 7 8 9 10nitrogen partial pressure in mm Hg x 104FIG. 3.-Decay constant for the He+ ion density in helium+nitrogen afterglows as a function ofthe nitrogen partial pressure for two discharge powers at 293°K.On the assumption that the neutral reacting species was N2, the rate coefficientat various temperatures-determined from straight lines drawn to the results asindicated in fig.3-and the corresponding cross-sections are given in table 2.TABLE 2temp. rate coefficient, cross-section,O K c m 3 sec-1 x 109 ~ m 2 x 1014503 1 -03 0.59408 1 -40 0.89293 1 -45 1 -09195 1 *75 1 -60Similar measurements have been made by Fite and his co-workers2 who findcross-sections of - 10-15 cm2. The present results are not sufficiently accurate todefine the form of the dependence of the rate coefficient on temperature. However,since the cross-section decreases with increasing temperature, it may be deducedthat if the mechanism of the charge exchange process is charge transfer the energydefect for the transfer is very small and the process may be accidentally resonant.Bates and Patterson9 have suggested that the complex ion HeN+ is unlikely tobe formed, and so the mechanism of the process is unlikely to be an ion-atominterchange reaction. In the present experiments the dominant nitrogen ion inthe afterglow was found to be the atomic ion N+, contrary to the observations ofFite who found the N i ion to be dominant.The mechanism of the exchangereaction in the present experiments therefore appear to be dissociative charge transfer :He+ + N p H e + N+ + N + 0.3 eV,although the energy defect is rather larger than would be expected on the basis oJ . SAYERS AND D. SMITH 173the experimentally determined variation of the cross-section with temperature.Some evidence was found of the presence in the afterglow of a very small densityof a positive ion of mass 18 a m.u., but it is possible that this was the ion HzO+released from the walls of the discharge vessel rather than the ion HeN+.HELIUM -k OXYGEN MIXTURESThe first measurements in helium+oxygen mixtures were made by a similarmethod to that used for the experiments in helium+nitrogen mixtures. Oxygenpartial pressures between 10-4 to lO-3mmHg were used, and helium was addedto make the total pressure 0.6 mm Hg.However, it was noticed that the spectro-meter signals representing the decay of the He+ ion density changed with time quiterapidly after initiation of the discharge, and that the value obtained for ;1 was de-pendent on the time between the initiation of the discharge and the exposure ofthe recording film. At a given measured pressure the values obtained for ;1 variedby up to a factor of three depending on t h s time.In order to investigate this effectfurther, the signals were photographed with a cine camera. The decay constantsdetermined from various frames of a typical cine film are shown in fig. 4 as a function'Ol-----II I I I I4 8 12 I6 2 0 24I tl3m frame number (time after initiation of discharge)FIG. 4.-Typical variation of A for the He+ ion density in heliumfoxygen afterglows with timeafter initiation of the discharge. Initial partial pressure of oxygen is 3 x 10-4 mm Hg.of the frame number. The frames were numbered starting from " 1 " at the firstafterglow to be recorded. The time after initiation of the discharge is also indicatedin fig. 4.The shape of the curve in fig.4 is largely explained by the phenomenon of clean-up. During an ionizing pulse, and in the afterglow following it, a proportion ofthe oxygen ions diffusing to the walls of the discharge vessel becomes permanentlyattached there. This process occurs during every ionizing pulse, but the numberof available sites for oxygen on the walls becomes smaller as the number of occupiedsites increases, and so the clean-up rate may be expected to decrease with time.The decay constant for the He+ ion density by charge exchange with oxygen may,therefore, be expected to decrease rapidly at first, and then more slowly as the rateof clean-up of oxygen decreases. The diffusion of oxygen from other parts of th1 74 ION AND CHARGE EXCHANGE REACTIONSapparatus into the discharge vessel would tend to reduce the rate of decrease of theoxygen partial pressure in the discharge vessel.The initial slow rate of decreaseof the decay constant is difficult to explain. It seems that before the oxygen may beadsorbed on to the walls, the walls must first be prepared by bombardment withoxygen ions. A possible reason for this is that the sites for oxygen must first bemade available by the removal of other less active atoms, such as helium, which mayinitially occupy the sites.If the curve of the decay constant against frame number is extrapolated to zeroframe number, the decay constant indicated should be that which would be ob-served if the clean-up were not important. In fig. 5 the values of the extrapolatedoxygen partial pressure in mm Hg x 104FIG.5.Variation of the extrapolated h for the He+ ion density in helium+oxygen afterglowswith the partial pressure of oxygen at 308°K.decay constant at 308°K have been plotted against the partial pressure of oxygen.As the point at zero oxygen partial pressure, which corresponds to the decay of theHe+ ion density in pure helium, is more accurately known than all the other pointsthe best fit line to the experimental points has been drawn to pass exactly throughthis point. The points lying below the line, and marked with dots, represent decayconstants that have been determined from a particular frame number of the cinefilms. They show how clean-up may affect the value obtained for the decay con-stant if the measurements are not made immediately after initiating the discharge.It would be possible to fit lines to the results modified by clean-up in the same waythat lines were fitted to the results of the helium+nitrogen experiments, i.e., withan intercept on the partial pressure axis (see fig.3). However, the intercept in theseresults is smaller than that in the helium+nitrogen results. As oxygen is a morereactive gas than nitrogen, the clean-up of oxygen was expected to be faster thanthat of nitrogen, and so a larger intercept was expected in the helium+oxygenresults on the basis of the suggested clean-up mechanism. A reliable predictionof the variation of the intercept with the gas concerned cannot be made on thebasis of the dissociation mechanism suggested in order to explain the intercept,because the cross-sections for the dissociating reactions are not known.The useof the cine camera technique in further helium + nitrogen experiments is, thereforeJ . SAYERS AND D. SMITH 175required before the cause of the intercept may be satisfactorily explained. Therate coefficient determined from the line drawn in fig. 5 on the assumption that theneutral reacting spscies is the 0 2 molecule isk = (1.05 f. 0.08) x 10-9 cm3 sec-1,and the corresponding cross-section isB = (7.7+0-6) x 10-15 cm2.The errors in these measurements are probably due largely to inaccuracies in themeasurement of pressure. Fite and his co-workers2 have recently made similarmeasurements and found a rate coefficient of 5 x 10-10 cm3 sec-1, and a corres-ponding cross-section of about 5 x 10-15 cm2.In the present experiments, as in those of Fite and his co-workers, the dominantoxygen ion in the early afterglow was the atomic ion Of.This suggests that theprocess occurring is either dissociative charge transfer,He+ + 0 p H e + O+ + 0 + 5.9 eV,where some of the energy released is taken into excited states of the products, orthe ion-atom interchange reaction suggested by Bates and Patterson,9He++02+(HeO+)’+O7followed by a radiative dissociation,(HeO+)’+He+ O++hv.Evidence was found to suggest the presence of a very small density of a positiveion at mass 20 a.m.u. in the very early afterglow. This signal appeared only whenoxygen was admitted to the discharge vessel, and so was possibly the ion HeOf.In fact, the rate coefficient for the He++02 reaction determined by the presentexperiments is in agreement with the value indicated by Bates and Patterson forthe ion-molecule reaction (on the basis of the Gioumousis and Stevenson equation).If the ion-molecule reaction is the dominant charge exchange process, the HeOfion must dissociate very rapidly after formation because the mass 20a.m.u. ionwas present only in very small quantities in the very early afterglow. Fite andhis co-workers found no trace of the HeOf ion in heliumfoxygen afterglows, andthey suggest that dissociative charge transfer is the dominant process.Design and construction of the apparatus used in the initial stages of the work,together with the results concerning the O+ ion, are due to Dr. G. R. Court andP. H. Batey.This research was supported in part by the Geophysics Research DirectorateAFCRL of the Air Research and Development Command, U.S.A.F.1 Dickinson and Sayers, Proc. Physic. SOC., 1960, 76, 137.ZFite, Smith, Stebbings and Rutherford, J. Geophysic Res., 1963, 68, 3225.3 Fite, Rutherford, Snow and Van Lint, Disc. Faraday SOC., 1962,33, 264.4 Bates, Proc. Physic. SOC. A , 1955, 68, 344.5 Gioumousis and Stevenson, J. Chem. Physics, 1958,29,294.sNakamura, J. Geomag. Geo.-elec., 1961,12, 114.7 Hasted, Proc. Roy. SOC. A , 1951,205,421.8 Batey, Ph. D. Thesis (University of Birmingham, England, 1962).9 Bates and Patterson, Planet. Space. Sci., 1962, 9, 599