摘要:

Reaction Rates Determined From Ionospheric ObservationsBY R. C. WHITTEN AND I. G. POPOFFStanford Research Institute, Menlo Park, CaliforniaReceived 3rd January, 1964Many investigations have employed ionospheric observations to compute relaxation parameters.We use some recent D- and E-region measurements for the same purpose. Solar flare X-ray dataobtained during the Naval Research Laboratory Sun Flare I1 programme, together with an 18-Mcsudden cosmic noise absorption record, are used to compute an effective dissociative recombinationcoefficient corresponding to -80 km altitude. It is found to be about 5 x 10-7 cm3 sec-1 to withina factor of about 2.Relative ion concentration data obtained in the altitude region 110 to 200 km are used to com-pute rate constants for several charge rearrangement processes which are believed to be the dominantones in the E-region.Several of them agree quite well with laboratory measurements, but one doesnot. Possible reasons for the disagreement are discussed.It is the purpose of this paper to report values for several ionospheric " relaxationparameters " which were deduced from observations of ionospheric phenomena.Such investigations have been carried out over a period of many years but themethods employed have varied. We utilize some recent observations of D-regionand E-region phenomena, respectively to obtain an effective recombination co-efficient and charge rearrangement rate constants.The sunlit region of the lower ionosphere is largely responsible for the absorptionof energy from high frequency radio waves which propagate through it. The ex-planation of this effect lies in the large neutral particle densities and thus large elec-tron collision frequencies ; according to magneto-ionic theory the signal attenu-ation (in db) is proportional to the product of collision frequency and the electrondensity integrated over the ray path.Even for relatively small electron densitiesthe absorption can be significant.Under certain conditions of enhanced radiation emission by the sun, the rateof production of free electrons at D-layer altitudes can rise by several orders ofmagnitude in a very short time. Solar X-rays in the 2-8 A range which are respon-si ble for the so-called sudden ionospheric disturbances are of particular interest.If one can perform simultaneous measurements of the intensity and spectral formof the ionizing radiation, as well as of the electron density, the rate at which theelectrons combine with the positive ions is easily obtained.This programme andthe resulting computation of the effective recombination coefficient will be discussed.Various processes other than recombination also have important bearing on decayof the plasma. At E-region altitudes these are effectively restricted to charge-transfer reactions and ion-atom interchange reactions. With the aid of recentmeasurements of the relative concentrations of various species of ions we shallderive estimates of the rate constants corresponding to the most important processesof these types.EFFECTIVE DISSOCIATIVE RECOMBINATION COEFFICIENT I N THE D-REGIONIn 1959 the Rocket Astronomy Group at the U.S.Naval Research Laboratory(NRL) conducted a series of observations (Sun Flare 11) 1 of solar X-rays with the18186 IONOSPHERIC OBSERVATIONSaid of rocket-borne detectors. Ideally this apparatus was capable of yielding thecoarse features of the spectrum from about 0.2A to 60A. In order to observe theX-radiation emitted by a flare, a patrol of the solar disc was conducted while therocket probe was held in readiness for immediate launching at Point Arguello,California. When the beginning of a Class 2f flare was observed in Ha at about22 : 45 UT on 31st August, 1959, the probe was launched and reached maximumaltitude very shortly after peak X-ray intensity.The spectrum constructed fromX-ray intensity data obtained at 22 : 45 UT is shown in fig. 1. The broken lines10-51 1wavelength, AFIG. 1.-X-ray spectrum of the flare observed at22 : 50 UT on 31st August, 1959. The solid curverepresents the most probable spectrum based onextrapolated data. The broken lines represent theestimated upper and lower limits of intensity.60 70 80 90altitude, kmFIG. 2.Ionization rate profile obtained from thesolid curve in fig. 1.represent the estimated upper and lower limits of intensity. The scintillationcounter measurements (in the energy range 15-80 keV) indicated that the very hardportion of the spectrum was softening as it decreased in intensity. The reader isto be cautioned against inferring from fig.1 that the spectrum was continuous inits entirety; the range A>1*5A may, in fact, have contained much line structure,but the resolution of the sensors was insufficient to observe it.When X-rays are absorbed in the upper atmosphere, they produce (a) primaryionization by photoelectric " stripping " of electrons from neutral molecules, anR. C . WHITTEN AND I . G . POPOFF 187(b) secondary ionization by “ stripped ” electrons. A photon of wavelength 2 Awill produce about 200 free electrons by this process. The ionization rate 4 can bewritten :where a&) is the absorption cross-section of X-rays in air as a function of wave-length; dfOo/dd is the spectral irradiance at some reference altitude 20; no is theneutral particle number density2 at altitude 20; H is the scale height (in general,a function of z) ; x is the zenith angle ; and Wis the mean energy required to producean ion pair in air (-32 eV).The ionization rate profile corresponding to the X-rayspectrum presented in fig. 1 is shown in fig. 2.The ionization rate is related to the electron density Ne by the equation :N dAdt l + A l + A dt’--- ( a D + A a i ) N : - - 2 - dNt? --in which A is the negative ion-to-electron density ratio, and a D and cci are the dis-sociative and ion-ion recombination coefficients, respectively. In the sunlit D-region the third term on the right-hand-side of (2) is very small compared withdN,/dT and thus is neglected. According to recent estimates 39 4 of 3, and a( we canalso neglect these quantities at altitudes above 75 km.Hence, for the upper D-region, (2) simplifies toto yield C ~ D if the electron density can be obtained.During the course of the flare observation, cosmic noise intensity I at 18 Mcwas recorded at Los Angeles (see fig. 3). It is related to the absorption coefficientk by the equality :dN,/dt M q - etDN:, (3)Idiln = exp [- 2JD (kd - kn)dzl, (4)where d and n refer to the disturbed and normal D-region and of indicates integra-tion over that segment of the ray path which lies in the D-region ; at high frequencies,k is related to the electron density Ne and collision frequency by 5 9 6where e is the electronic charge in e.s.u., m is the electron mass, c is the velocityof light in free space, f is the wave frequency, and v is the most probable electroncollision frequency.7 At an altitude of 80 km v w 6 x 105 sec-1.In order to compute the electron density profile at the time of observation ofthe flare X-rays, several trial X-ray pulses with different time characteristics wereassumed and the corresponding electron density profiles were computed by numericalintegration of (3), using assumed values of a~ close to our final result.Thesevalues of Ne were then substituted into (4) and (5) in order to obtain the computedcosmic noise absorption at various times after commencement of the effect. Theresulting relative intensity curves were compared with the shape of the rise segmentof the riorneter record (fig. 3); that trial X-ray pulse shape which yielded the bestfit to the experimental data was accepted as an adequate approximation to theactual pulse. The corresponding value of a~ is the effective dissociative recom-bination coefficient for the absorbing region : b~ = 5 x 10-7 cm3 sec-1 to withina factor of two.The principal uncertainty lies in the measured solar X-rayspectrum (fig. 1). We have tacitly assumed that the shape of the X-ray spectrum iskw 5N,e2v/4xmc f 2, (5188 IONOSPHERIC OBSERVATIONStime-independent. This is not completely valid but is believed to be justified for ourpurpose. The absorption coefficient profile is shown in fig. 4. Clearly maximumabsorption occurs between 80 and 82 km.6 .5 -4 -3 -2--NRL OATA-I I 1FIG.3.-Intensity at Los Angeles of 18 Mc cosmic noise during the occurrence of a class 2+ solarflare on 31st August, 1963.70 80 90altitude, kmFIG. 4.-The 18 Mc absorption coefficient profile.LeLevier 8 recently reported the computation of an effective D-layer dissociativerecombination coefficient which was obtained from a riometer record from MidwayIsland taken during the high altitude nuclear detonation of 9th July, 1962. Thevalue so obtained was between 3 x 10-7 and 7 x 10-7 cm3 sec-1. Recent laboratorymeasurements of dissociative recombination coefficients, as well as the effectiverecombination coefficients obtained from D-region measurements, are tabulated intable 1R . C. WHITTEN A N D 1. G. POPOFF 189Apparently the values of ao(0;) and aD(Ni) are functions of time, the latestdata being consistently smaller than that which it supersedes.This probably re-flects the fact that fewer contaminants are present in the more recent experiments.Except in a crude way it is impossible at the present time to relate the laboratoryvalues to ionospheric processes because of the potential importance of excited statesof the ions. In any event the effective values are not too different from those ofTABLE 1aD (effective) "D(o+) UD(N+) ag(NO+)(61113 sec-1) (cm3 sec-1) (cm3 sec-1) (cm3 sec-1)3 to 7x 10-Ta ( 1 . 7 ~ ) ~ 10-7 c (3 L-~)x 10-7 c 1.3 x 10-65x 10-7b (3.8 f 1) x 10-7 d (5.9 rt 1) x 10-7 d 3~ 10-7f2 . 4 ~ 10-7 ga LeLevier ; 8 b this report ; C Biondi ; 9 d Kasner, Rogers and Biondi ; 10 e Gunton and Inn ; 11f Doering and Mahan ; 12 B Stebbings and Van Lint.13~ ( 0 : ) and ao(NO+) ; as we see later, N; is not likely to be an important ion in theD-layer.The measurements of ao(N0f) are not very reliable ; Gunton 14 improvedhis technique for measuring this quantity and has recently obtained a much smallervalue than reported in table 1. Using data from the diurnal variation of theE-region critical frequency, Bowhill 1 5 has derived a recombination coefficientwhich apparently corresponds to NO+ : ED = 5 x 10-9 cm3 sec-1. The small mag-nitude of this result probably reflects the elevated temperature of the E region.CHARGE TRANSFER AND ION-ATOM INTERCHANGE IN THE E-REGIONCharge-transfer reactions and ion-atom interchange reactions are very importantin controlling ionospheric relaxation rates by " shuffling " the identities of theions.The nature of the ionic species in turn has an important bearing on the rateof electron recombination and hence on the electron density profile. The followingreactions are probably the dominant ones in the shuffling process in the E-region :o+ + 0 2 - + 0 + 0; (7)NZ + O+N2 + 0' (8)N l + 02-+N2 + 0;o+ +N~+N+NO+.Reaction (9) is quite fast, particularly at low altitudes. This supports our statementthat recombination of N: is unimportant in the D-region.It is possible to make estimates of the rate constants characteristic of reactions(7)-(10) if one knows reasonably well the neutral particle densities,16 the ion produc-tion rates,16 the dissociative recombination coefficients,l7 and the relative ion con-centrations.18-20 They may be obtained at a given altitude by inserting the valuesof the known parameters into the equilibrium equations 190 IONOSPHERIC OBSERVATIONSsubscripts), the square brackets represent concentrations of the indicated species, anda%6*C represent the dissociative recombination coefficients of Ni, 0; and NO+,espectively. We require that[el = [Nl]+[O,']+[O']+[NO+], (15)q (electron) = qa+qb+qcif electric charge is to be conserved.The values of the various known parametersat 150 km are presented in table 2.TABLE 2neutral particle 16 concentra-ion production rates 16 at 150dissociative recombination co-efficients 17 at 150 km (cm3relative ion concentrations 19at 150 km (cm-3) ; NASA4-09, 1047 EST, 20th April,1960, Wallops Island)tions at 150 km (cm-3) "21 = 2 x 1010, [02] = 1 x 109, [O] = 3 x 1010km (cm-3 sec-1)SW-1)qa = 2~ 103, qb = 1 .6 ~ 102, qc = 1 . 7 ~ 103Q,(N;) = 2~ 10-7, aD(o;) = 9~ 10-8, aD~o+) = 4~ 10-8[N;]/[e]-~0,01 ; [O:]/[e] = 0.3[NO+]/[e] = 0-6; [O+]/[e] = 0.1In order to arrive at estimates of the rate constants, k7, kg, kg and klo, it wasfirst necessary to compute the electron density. This was done by setting the sumof the right-hand members of (11)-(14) equal to the total ion production rate, ex-pressing the various ion concentrations in terms of the electron density and solvingfor [el. The rocket data employed were taken from flight NASA 4.09 (solar zenithangle = 28") of 20th April, 1960. Values of the rate constants were computed bysubstituting the values of the various parameters from table 2 into the equilibriumequations (11)-(14). The values which best fit the data are :k7-2 x 10-11 cm3 sec-1;kg - 2 x 10-10 cm3 sec-1;kg-2 x 10-11 cm3 sec-1;klo-2 x 10-12 cm3 sec-1.Of these, k7, kg and klo have recently been measured in the laboratory by variousworkers. The results are :k7 Dickinson and Sayers 21Sayers 22Langstroth and Hasted 23Fite et al.24kg Fite et al.24klo Talrose et aZ.25Sayers 22Langstroth and Hasted 232-5 x 10-11 cm3 sec-1;1.6 x 10-11 cm3 sec-1;1.8 x 10-12 cm3 sec-1;(1 to 15) x 10-11 cm3 sec-1;2 x 10-10 cm3 sec-1;<6.7 x 10-12 cm3 sec-1;2.5 x 10-11 cm3 sec-1;4.7 x 10-12 cm3 sec-1.Our estimates of k, and kg are in good agreement with the laboratory measure-ments.In fact, the agreement is perhaps misleading, since our estimates are to orderof magnitude only. On the other hand, our estimates of klo is in rather poor agree-ment with one or two of the laboratory measurements. Since the estimated un-certainty in our computation is only about half an order of magnitude, the obviouR. C. WHITTEN AND I . G . POPOFF 191conclusions are (if the laboratory measurements are correct) that we are using toosmall a value for the dissociative recombination coefficient of NO+ or that anothercharge rearrangement process for conversion of NO+ into a different ion is effective.However, it is believed highly improbable that ao(NO+) is in error by more thanhalf an order of magnitude.17 Nor can we find an exothermic charge rearrangementprocess which will remove NO+.Two other possible explanations of the discrep-ancy are temperature dependence of the rate constant, and the existence of differentexcited states of NO+ in the ionosphere from those in the laboratory experiments.If a temperature dependence does exist as a result of an activation energy, it is inthe wrong direction to provide satisfactory explanation. The E-region at 150 kmis characterized by a temperature of about 1000"K, whereas the laboratory experi-ments were carried out at room temperature. Hence one would expect klo to belarger in the 3-region than in the laboratory.This research was supported by the Advanced Research Projects Agency throughContracts AF 19(601)-199A and AF 19(604)-8355.1 Chubb, Friedman and Kreplin, J.Geophys. Research, 1960, 65, 1831.2 US. Staniiurd Atmosphere, 1962 (U.S. Government Printing Office, Washington, 1962).3 Nicolet and Aikin, J. Geophys. Research, 1960, 65, 1464.4Dalgarn0, Ann. Geophys., 1961, 17, 16.5 Sen and Wyller, Physic. Rev. Letters, 1960, 4, 355.6 Molmud, Physic. Rev., 1959, 114, 29.7 Phelps and Pack, Physic. Rev. Letters, 1959, 3, 340.8 LeLevier, J. Geophys. Research, 1964,69, 481.9 Biondi, Ado. Electronics, 1963, 18, 67. (Academic Press, New York, 1963).10 Kasner, Rogers and Biondi, Physic. Reu. Letter$, 1961, 7, 321.11 Gunton and Inn, J. Chem. Physics, 1961, 35, 1896 (L).12 Doering and Mahan, J. Chem. Physics, 1962,36, 669.13 Stebbings and Van Lint, Measurements of Atmospheric Reactions, General Atomic, 21 st Nov.,14 private communication.15 Bowhill, J. Atmos. Terr. Physics, 1961, 20, 19.16 Norton, Van Zandt and Denison, Proc. Int. Conf. Ionosphere, London, 1962, (The Physical17 Whitten and Poppoff, to be published.18Taylor and Brinton, J. Geophys. Research, 1961, 66,2587.19 Meadows and Townsend, in Space Reseurch I (North-Holland Publishing Co., Amsterdam,20 Istomin and Pokhunkov, in Space Research I11 (North-Holland Publishing Co., 1963).21 Dickinson and Sayers, Proc. Physic. SOC., 1960, 76, 137.22 Sayers, in Atomic and Molecular Processes, ed. by Bates (Academic Press, New York, 1962).23 Langstroth and Hasted, Disc. Furday SOC., 1962, 33, 298.24 Fite, Rutherford, Snow and Van Lint, Disc. Furaday Soc., 1962, 33, 264.25 Talrose, Markin and W i n , Disc. Faraday SOC., 1962, 33, 257.1963 (unpublished report).Society, London, 1963).1960)

ISSN:0366-9033    

DOI:10.1039/DF9643700185

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

Negative Ions in Afterglows in Atmospheric Gases *BY WADE L. F I T E ~ AND J. A. RUTHERFORDGeneral Dynamics/General Atomic, San Diego, CaliforniaReceived 8th January, 1964Mass spectrometric monitoring of ions in afterglows in atmospheric gases has been appliedto the study of negative ions. Nitrogen+oxygen+hydrogen mixtures as well as the " pure " gasesand ordinary air have been examined in a bakeable afterglow mass spectrometer. For pressuresof the order of 1 torr and after times of the order of milliseconds, minute impurities produce negativeion spectra which bear little resemblance to those which could be expected. Especially significantare the ions COT (mass 60) and NO3HT (mass 64). It seems likely that many of the atmosphericphenomena usually attributed to 0 5 and NO5 involve these and other heavier ions.The mass spectrometric monitoring of afterglows in gas discharges has proveduseful for determining the ion species present in afterglows and also for ascertainingrates at which ion-molecule collision processes such as charge exchange and ion-atom interchange proceed.Following the first application of this method by Phelpsand Brown 1 who studied afterglows in helium, the technique has been appliedprincipally to the study of processes believed operative in the upper atmosphere?-6The studies, however, have been restricted to collisions in which positive ionsparticipate.Of no less interest for atmospheric purposes are processes involving negativeions. It is to this subject that the present work is addressed.The findings suggestthat the catalogue of negative ions in the atmosphere is probably far from completeand that perhaps the principal negative ions present in much of the atmosphereare not in fact those which are commonly presumed. They also suggest that manyof the laboratory experiments involving negative ion processes, where the experi-ments were done without benefit of mass spectrometric monitoring, are very likelyin need of considerable re-interpretation.EXPERIMENTALAPPARATUSThe basic apparatus for after glow mass spectrometry consists of a source chamber inwhich an afterglow can be excited and which is connected to an ion mass analyzer througha small aperture. In the source chamber, which is nothing more than a confined volume,a mixture of gases is placed and a short pulse of power is applied in order to produce aweakly ionized plasma.After cessation of the power pulse, ions escape from the decayingplasma through the aperture, and are accelerated, focused and mass analyzed. By tuningthe analyzer to accept ions of a given elm ratio and by displaying the output of the iondetector on an oscilloscope, the time history of each ion species in the afterglow is presented.From these histories, the ion-molecule collision processes occurring can be identified fromthe build-up of one type of ion at the expense of another and rate coefficients can be derived.Alternatively, rather than using fast detection circuitry and displaying the time histories,*Portions of this research were supported by the Defense Atomic Support Agency underi present address : Department of Physics, University of Pittsburgh, Pittsburgh 13, Pennsylvania.Contract DA49-146-XZ-041.19W. L .FITE AND J . A. RUTHERFORD 193each ion current can be integrated over the duration of the afterglow, which is particularlyconvenient in displaying the entire ion spectrum on, say, an X - Y recorder.In the present experiments, the apparatus used is that shown schematically in fig. 1.The entire apparatus, including the source chamber, was constructed of stainless steel,with gold O-rings and was thoroughly bakeable. Mercury diffusion pumps were used.Provision was made to produce the afterglow using pulses either of radio-frequency power,provided at internal electrodes (not shown), or of 40 MeV electrons from the General Atomicelectron linear accelerator (LINAC). A pusher electrode was placed in the source chamberto provide a weak drift field to move the ions toward the pinhole aperture.(Use of a driftfield of the order of 1 V/cm was helpful in positive ion studies, but for negative ion experi-ments, a drift field was of no consequence, and the pusher electrode was normally con-nected to the rest of the source chamber.)PUSHER 7 ELECTRODE iION LENSESSLITMAGNETPOLE FACEJEDIFFUSION' 1 I LiNAioEAMLlOUlO NITROOEN TRAPS AND I tINLETPlRANlBAKEABEVALVEPUMPSFIG. 1 .-Schematic diagram of apparatus for mass-spectrometric studies of afterglows.After emergence of the ions through the pinhole aperture, they were quickly acceler-ated, usually to 400eV, and focused in the first of two differentially pumped vacuumchambers of the mass spectrometer proper.After passage through the slit separatingthe two mass spectrometer chambers, the ions were magnetically analyzed in a 60" sectorfield. The ions were further accelerated to several keV before impinging on the first dynodeof an electron multiplier. The output of the multiplier fed into a pre-amplifier to reducethe impedance of the signal before display on the oscilloscope. With the circuitry used,the time constant was less than 1 psec.Not shown in fig. 1 is a small electron gun mounted in the ion focusing lens system.Using this gun, electron impact ionization of the gas flowing from the source chamberenabled making measurements of the relative concentrations of the gas mixtures at anytime; the total gas pressure in the source chamber was measured by a Pirani gauge.Thegases were admitted to the source chamber through bakeable variable leak valves andstainless steel lines. The gases used, with the exception of air, were normally those soldas spectroscopically pufe; hydrogen was admitted to the system through a palladiumleak. Every reasonable effort was made to limit the admission of impurities into the gasesused.In the experiments summarized here, the excitation of the afterglows was done by pulsesof 30 mc/sec radio-frequency power. The pulse durations were varied from 1 to 10 psecfor the most part, but occasionally longer pulses were used.One characteristic of this instrument which proved disadvantageous in the presentstudies is illustrated in fig.2. This figure shows an overlay of oscilloscope traces for 0;and 0 2 issuing from an afterglow in pure 0 2 at a pressure of about 1 ton. At early times,the positive ion was readily extractable from the afterglow, but an abrupt cut-off of positive194 NEGATIVE IONS IN AFTERGLOWSion current occutred at just under 1 msec. Simultaneously negative ion signals, whichcould not be obtained early in the afterglow, abruptly became observable and the negativeion signals continued throughout the remainder of the afterglow. When more than onepositive ion was present, all would cut off together and all negative ion signals would becomeobservable at the same time.It is presumed that a reversal of the electrical polarizationof the entire plasma during the decay was responsible for the observed behaviour. Whileit was found that use of the pusher electrode in the source chamber could shift the cut-offpoint slightly, it was not found possible to obtain negative ion signals throughout theentire period of the afterglow.0 0.2 0.4 06 0.8 1.0 1.2 1.4 1.6time (msec)FIG. 2.-Oscilloscope traces of 0; and 0; ions from an afterglow in 0 2 .Since one of the two methods used to obtain rate coefficients depends on knowledgeof the ions early in the afterglow (see ref. (3)) and this knowledge was not available innegative ion observations, principal interest in the present experiments was centred onthe species present rather than the reaction rate coefficient values.GENERAL CONSIDERATIONSInterest in the mere identification of the negative ions in laboratory afterglowsis predicated from results such as are shown in fig.3. This figure is an overlay ofoscilloscope traces taken on the principal negative ions issuing from an afterglowin " pure oxygen " at pressures of less than about 0.5 torr. While the 0- and 0 5ions are expected to be present, the negative ion at mass 60, which is seen to dominatethe spectrum at late times in the afterglow, is not. The mass 60 ion must arise froman impurity in the source chamber that was present despite the use of a baked systemand considerable care in the gas handling and, as judged by means of electron impactionization, occurred in a fraction of no more than about 10-3.If such minor impurities are capable of producing the dominating heavy negativeions at later times in laboratory afterglows, and possibly at the expense of the lighternegative ions, it is reasonable to expect that heavier negative ions may be abundantlypresent in the atmosphere.It becomes of interest to examine the heavy negativeions and attempt to identify them in order to guide the assessment of their possibleroles in atmospheric phenomenaW. L. FITE AND J . A. RUTHERFORD 195RESULTSOXYGENFig. 4 shows the time-integrated negative ion signals present in an afterglowin oxygen at a pressure of 1 torr in the baked stainless steel afterglow mass spectro-meter. In this figure, it is seen that the principal negative ions are 0-, 03, 03and mass 60.Increasing the gain of the instrument gives the negative ion spectrum0 I 2 3 4 5 6 7 8time (msec)FIG. 3.-Oscilloscope traces of negative ions in an 0 2 afterglow ; P = 0.4 torr.32FIG. 4.-Negative ions in an afterglow in 0 2 ; P = 1 torr; low gain.shown in fig. 5. The OH- peak at mass 17 indicates that traces of either water orhydrocarbons were in the instrument, and the peak at 46 (NO-,) indicates somenitrogen impurity in the gas. The peak at 21.3 has been found to accompany thepresence of 05 in afterglows and seems to be an Aston peak caused by breakupin flight of 03 into O-,+O. This breakup can be collision induced according tostudies of this peak as a function of pressure in the mass spectrometer tube, but itG196 NEGATIVE IONS I N AFTERGLOWSalso seems to occur spontaneously in part.The mass 64 ion is presumed to be 01and the existence of the mass 76 ion suggests that the mass ion 60 ion can attachan additional oxygen atom.In drift tube experiment, Pack and Phelps 7 have concluded that COj is readilyformed, which would suggest that the mass 60 ion is COT. However, N2O3 isalso a possible identification of the ion in the oxygen afterglow. Additional experi-ments to attempt to identify the mass 60 ion were clearly required.6C723;36-. -FIG. 5.-Negative ions in an afterglow in 0 2 ; P = 1 tom; high gain.c 0 2 4 - 0 2 MIXTURESIn order to explore the possibility of the mass 60 ion being COT, carbon waspurposely introduced into the oxygen afterglow.While either CO or C02 couldbe used, it was presumed that the power pulse producing the afterglow wouldprobably make ample quantities of both these gases from either, and there seemedto be little preference for choice of the introduced gas. C02 was used as a matterof convenience.Fig. 6 shows the time-integrated negative ion spectra found when the mixturewas approximately of equal parts and the total pressure was about 0.8 torr. As isevident, the addition of C02 quenches 0- and 0, in the afterglow and mass 60becomes by far the dominant ion in the spectrum, although masses 76 and 78 arealso seen to go off scale in fig. 6.The presence of mass 46, however, indicated that in these experiments nitrogenwas not entirely eliminated, and the possibility that mass 60 could be N203 remained.Going in the opposite direction and trying to enhance N207, if it was the mass 60ion, was clearly suggested.N2+02 MIXTURESWhen mixtures of spectroscopically pure N2 and 0 2 were used in the afterglow,mass 60 was not found to be enhanced, which presumablyeliminates N203 as themass 60 ion.However, it was found that another heavy ion at mass 64 was worthW. L. FITE AND J . A. RUTHERFORD 197of some attention. Fig. 7 shows the time-integrated spectrum of negative ions froma 4 : 1 N2+ 0 2 mixture at a pressure of 0.25 torr, and shows the almost completeabsence of mass 60. This figure also shows the pattern at masses 61-66 repeatedFIG. 6.-Negative ions in the afterglow of a 5 : 4 mixture of COZ and 0 2 , at a total pressure of 0.8ton ; masses 60, 76 and 78 are off scale.4 664HEIGHT OF 46 PEAK WHEN GAINIS REDUCED BY A FACTOR OF 2.5FIG.7.-Negative ions in 4 : 1 N2 + 0 2 mixture.at masses 79-84, suggesting that the addition of a water molecule was involved.This, in turn, suggests that the mass 64 ion is an N03H-, i.e., an NO: ion with awater molecule clustered to it198 NEGATIVE IONS I N AFTERGLOWSAIRThe negative ions which are customarily considered to exist in ordinary air are0 2 and NOT. In light of the findings that heavier negative ions are quite prominentand sometimes dominant under the best conditions of gas purity that could be achieved46FIG. &-Negative ions in air afterglow ; P = 0.41 torr.82 L 62 LFIG.9.-Negative ions in air afterglow ; P = 0.7 torr.in these experiments, it became of interest to examine the afterglow negative ionspectra in ordinary laboratory air with its multitude of impurities.Fig. 8-11 show the time-integrated spectra of negative ions from room tem-perature air at pressures of 0.41, 0.7, 1.2 and 1.5 torr, respectively. A strikinW. L . FITE AND J . A . RUTHERFORD 199feature of this sequence of plots is the growth of the mass 64 peak, relative to themass 46 peak, with increasing pressure. The relative growth of the mass 82 peakis perhaps even more striking.At higher mass numbers, a cluster of ions resembling that at masses 79-83 wasfound with the major peak at 96, and strong single peaks at approximately masses108 and 122.Only occasional small peaks were observed at even higher massesup to about mass 280, where the observations stopped.64FIG. 10.-Negative ions in air afterglow ; P = 1.2 torr.64FIG. 11 .-Negative ions in air afterglow ; P = 1.5 tom.With regard to the transient character of these ions in the afterglow, the decayconstants of the different peaks at 1.2 torr pressure are as follows :mass decay time (exponential)46 0.2 msec62 2.564 0.779 1.580 2.082 0-200 NEGATIVE IONS I N AFTERGLOWSThere is little quantitative information to be deduced from these decay times becausedifferences in rate of diffusion of the various ions to the walls, as well as ion destruct-tion-creation processes, could lead to different decay constants.The importantexperimental point is, however, that NO: is a relatively short-lived ion when im-purities exist that permit the formation of heavier ions.AIR 4- D2 MIXTURESThe conjecture that the mass 64 ion is an NO: ion with an H20 moleculeclustered to it has already been put forth. In order to check this identification,it is of interest to consider afterglows in air to which D2 has been added. If themass 64 ion does contain two H atoms, then with the deuterium present, the peakshould split into a triplet. Fig. 12 shows that this is the case. This X - Y plotFIG. 12.-Negative ions in 10 : 1 air+Dz afterglow.was taken at a pressure of 2 torr with a 10 % D2 in air, which apparently scrambledthe H and D atoms in the 64-65-66 triplet in roughly equal numbers. Correlationof the pattern in the neighbourhood of mass 80, in fig.12, with the pattern withoutthe D2 present, as in fig. 11, is not so obvious.HYDROGENAdding D2 to the air in the afterglow chamber, although promising to help inthe identification of the mass 64 ion, also gave the possibility of generating wholenew groups of negative ions. It was therefore important to examine the negativeions in afterglows in hydrogen with the inexcludable impurities in the instrument.When this was done, major negative ions were seen at masses 17, 25 and 61.Other negative ions, including mass 1 and a number of ions between masses 31and 46 were observed in lesser quantities. In order to determine the number ofhydrogen atoms in each of the major ions, mixtures of Hz and Dz were also usedand a mass spectrum of such a mixture is shown in fig. 13.From the splittinginto doublets it is clear that each of the three major ions contains a single hydrogenatom, leaving little doubt that mass 17 is OH-, mass 25 is C2H- and mass 61 isprobably CO3H-W. L. FITE AND J . A. RUTHERFORD 20 1DISCUSSIONSeveral points are indicated from results of the type presented here. First,negative ion spectra in afterglows produced in the laboratory are complex andquite minute impurities in a weakly ionized plasma can produce a great variety ofions. Within milliseconds at pressures of -1 torr, reactions occur to produce ionsup to masses slightly over 100 in copious amounts and up to at least 280 in lesseramounts.45 3532FIG.13.-Negative ions in H2fD2 afterglow.Secondly, it is exceedingly difficult to control the impurities which can affectlaboratory afterglow experiments on negative ions. In the present experiments thecarbon was inexcludable and this impurity produced major negative ions in bothhydrogen and oxygen afterglows. It is believed that the carbon probably came fromthe stainless steel of the afterglow source chamber itself, being liberated perhapseither by sputtering during the power pulse or by chemical reactions with freeradicals or excited molecules present in the early afterglow.Thirdly, in view of the care taken to control impurities in these experiments,it seems extremely likely that many other experiments, performed in numerouslaboratories, on attachment, detachment, negative ion mobility, etc., may havebeen subject to considerably more impurity difficulties than is realized.The easeof formation of COT, apparently at the expense of 0- and/or 0 2 when a carbonimpurity is present, suggests that experiments on oxygen negative ions are par-ticularly subject to misinterpretation; perhaps, however, they are no more so thanexperiments on NO, performed under less stringent conditions than the mostthorough baking and trapping to exclude water vapour which apparently clustersto form the ion NO3H3.Fourthly, the clustering of water molecules to NO3 appears to occur very readilyfor one and two molecules. Addition of more than two water molecules did notseem to occur in the present experiment.Fifthly, if and when it becomes possible to probe the atmosphere for negativeions in a rocket experiment, it will be highly important to have the range of themass spectrometer extend to at least mass 76 and highly desirable to be able to covermass 122, at least at altitudes where three-body collisions can occur to produce thenegative ions observed in these experiments.Sixthly, in view of the ease of formation of heavy negative ions in air, apparentlyat the expense of lighter negative ions, it seems likely that a number of importantatmospheric effects involving photodetachment, mutual neutralization, etc., operatethrough these heavier ions rather than the lighter ions normally considered202 NEGATIVE IONS IN AFTERGLOWS1 Phelps and Brown, Physic. Reu., 1952, 86, 102.2 Dickinson and Sayers, Proc. Physic. SOC., 1960,76, 137.3 Fite, Rutherford, Snow and Van Lint, Disc. Faraday SOC., 1962, 33,264.4 Langstroth and Hasted, Disc. Faraday SOC., 1962, 33, 298.5 Fite, Smith, Stebbings and Rutherford, J. Geophys. Res., 1963, 68, 3225.6 Sayers and Smith, Atomic CoZlision Processes, ed. McDowell (North Holland Publ. Co.7 Pack and Phelps, Bull. Amer. Physic. SOC., 1962, 7, 636.Amsterdam. 1964) pp. 871-876

ISSN:0366-9033    

DOI:10.1039/DF9643700192

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

Electron Affinities of the Oxides of NitrogenBY A. L. FARRAGHER," F. M. PAGE * AND R. C. WHEELER tReceived 15th January, 1964The electron affinities of NO and NO2 have been measured by the magnetron technique. Thevalues of 21 and 92 kcal/mole respectively were both derived from direct electron-capture processes.Evidence is also presented for the existence of an excited state of the NO5 ion lying 92 kcal abovethe ground state. The value of 92 kcal for the electron affinity of NO2 is higher than previousestimates, but is consistent with the properties of this ion and reflects its importance in the upperatmosphere.The magnetron method has been used extensively to measure the electronaffinities of radicals produced by the pyrolysis of suitable substrates 1 9 2 at a heatedfilament which becomes the source of thermionic electrons and negative ions.Thedemonstration that the dominant negative ions in the upper atmosphere werederived from the oxides of nitrogen, coupled with the stability of the salts of theseions, prompted an investigation into the oxides, NO and NO2, from which, beingstable free radicals, the ions might be formed by direct capture. Attempts werealso made to measure the electron affinity of NO3, produced by pyrolysis of thevolatile anhydrous copper nitrate,3 but the absence of essential thermochemicaldata about the bond energies in this compound prevented the interpretation of theresults.EXPERIMENTALThe NO was supplied by Matheson Company Inc. and used unpurified. NO;! wasprepared either by heating lead nitrate in a stream of oxygen, following the directions ofDodd and Robinson 4 or by oxidation of NO.The NO2 was always purified by fractionaldistillation before use.The apparatus used was identical to that described previously 1 except for the minormodifications of using solid molybdenum anodes and maintaining both grids at the samepotential. With this afrangement it was possible to reduce the background ion currentto 10-5 of the thermionic (electron) current in the absence of a gas. Filaments of bothplatinum and platinum-rhodium alloy were used and the temperatures, in the presence of5 x 10-3 mm of gas, were measured with either Leeds and Northrup or Cambridge dis-appearing filament pyrometers. The temperatures corresponding to certain fixed valuesof the filament current were measured repeatedly by either instrument and found to bereproducible to better than f5"K.The electron affinity was derived from a plot of the logarithmic ratio of the electronto ion currents against the reciprocal of the absolute temperature of the filament, the slopebeing equal to the apparent electron affinity E'/R.lRESULTSNITROGEN DIOXIDEThe graphs characteristically consisted of three parts : (i) a low-temperatureportion having a slope corresponding to an apparent electron afinity of 96 kcal/mole ;(ii) an intermediate region with an apparent electron affinity of 5 kcal/mole; (iii) a* College of Advanced Technology, Birmingham.2031- Queens University, Kingston, Ontario204 ELECTRON AFFINITIEShigh-temperature portion with E’ = -32 kcal/mole.Although all the graphs wereof the same form, the individual portions were not all shown well together.LOW-TEMPERATURE REGIONThe 20 runs which showed this portion well enough for a reasonable line to bedrawn gave a mean value of 94_+ 10 kcal/mole. The most reliable experimentallines from this set had values of 92.8, 99.7, 97.0, 94.3, 90.5, 102 and 97.0. HenceE’ = 96.2_+3-7 kcal/mole (fig. 1).I6.0 7.01041~FIG. 1.This value is attributed to the process NOz+e-,NO;. The specific heat cor-rection is estimated as -3/2RT, and since for this process the apparent is also thetrue electron affinity 1 thenENO~ = 92+ 3.7 kcal/mole at 0°K.The scatter in these results is undesirably large largely because the measured ioncurrents in this region were 1012-10-13 A, i.e., close to the limit of detection of theapparatus (lO-14A) and comparable with currents induced by the vibration ofthe apparatus in the field of the solenoid.The slope is also strongly affected bysmall errors in the measured temperatures.HIGH-TEMPERATURE REGIONThe most reliable straight lines had slopes corresponding to -E‘ being 31, 31,34, 33, 33, 34 and 29 kcal/mole (fig. 2). Hence E‘ = -32.1 & 1 kcal/mole. Thisis attributed to the process NOZ+e-+NO+O-. The specific heat correction isestimated as -5/2RT so that E’ = -40-8 kcal/mole at 0°K. The ON-0 bondenergy is 72 kcal5 so that Eo = 72 - 40.8 = 3 1.2 & 2 kcal/mole. This is in good argee-ment with the value of 34 kcal deduced from measurements on N20, H20 and 02.6INTERMEDIATE REGIONThe results in this region had a mean slope of 50& 5 kcal/mole (20 runs), thebest straight lines having values 3.2, 6-4, 6.9, 1-8, 1-8, 5.0, 4.6, 7.7 and 7-7 kcal/mole.Hence E’ = 5-0 12.2 kcal/mole (fig.3a)A . L . FARRAGHER, F. M. PAGE AND R . C. WHEELER 205There are several explanations of this result. The most obvious is that sincethis region corresponds to a maximum, the apparent slope has no real meaning.This is unlikely to be true since the only results which were considered had flatportions which extended over too great a temperature range to be mere experimentalerror. They had, moreover, no sign of a maximum and so the reasonable assumptionis that they represent a real process.I I6.0 7.01 0 4 / ~FIG.2.16.0 7.01W/TFIG. 3.The reactions occurring on either side of this can both be ascribed to the inter-action of NO2 with the filament and so this process also probably involves N02.However, this might be due to electron capture by NO, formed by the gas phasedecomposition of N02. In order to eliminate this possibility the electron a&tyof NO was measured. The results showed that with NO at the same pressure asthe NOz, the ion currents were ten times less than those found in this region and ha206 ELECTRON AFFINITIESa slope of 26 kcal/mole. These are also plotted in fig. 3b for comparison. It there-fore seems lughly unlikely that this result is due to NO.We are therefore left with a reaction involving NO2 and yielding a positive electronaffinity.Fission of the molecule is impossible since the only reasonable products,NO- or 0-, would both give negative slopes. The only possible reaction wouldtherefore seem to beThe specific heat correction is estimated as - 3/2RT and soNO2 + e+[NOJ*.= 0 & 2.2 kcal/mole at 0°K.NITRIC OXIDEAs previously mentioned, the ion currents obtained from NO were much lowerthan those from NO2. The results obtained were 27.4, 26.1, 21.1, 28.4 and 25.0kcal/mole, or E' = 25.6+2-5 kcal/mole. The specific heat correction is estimatedas -3/2RT. HenceEN, = 20.6 & 2.5 kcal/mole at 0°K.DISCUSSIONPrevious estimates of the electron affinity of NO2 vary greatly. X-ray diffractionhas established7 the structure of the ion as angular with an 0-N-0 angle of 125"and an N-0 distance of 1.13 A.The axial ratio is therefore about 2 : 1 and theassumption of spherical ions will not lead to a large error. From the knowndensities of sodium, potassium and silver nitrites the crystal radius of the ion isestimated as 2.05, 1.98 and 1.96 A, giving a mean value of 2.0 A. This is greaterthan the " thermochemical radius " of 1.55 A proposed by Yatskimirskii 8 and usedby him in conjunction with the lattice energy formula of Kapustinskyg to derivevalues for the lattice energies of the alkali metal nitrites. Based on this workPritchard 10 proposed an electron affinity of 37.5 kcal/mole for NO2. If the radiusof 2-0A is substituted in Kapustinsky's formula the electron affinity becomes some20 kcal higher. The uncertainty in these methods is, however, high.Smith and Seman11 claim to have observed photodetachment of the electronat 3.0 eV thus placing an upper limit of 69 kcal/mole on the electron affinity.Con-versely, Curran 12 has observed charge transfer between the ions 0-, SF;, SF;,Cl- and NO2. This places a lower limit of 87 kcal/mole on the electron affinity.Such a high value would be in agreement with the ease of forming NO; by electricaldischarges in air, but is to be contrasted to the difficulty of forming it by directelectron capture or discharges in NO2.13 The electron capture cross-section ofNO2 has been estimated by Fox 14 as 10-20 cm2, approximately one-hundredththat of oxygen.There have been few previous estimates of the electron affinity of NO, duelargely to the impossibility of performing lattice energy calculations.Pritchard 11has suggested a small positive value and the value given here is in accord with thisestimate.From the observations cited above the formation of NO; is favoured by thepossible occurrence of bimolecular collision processes, but direct electron captureis inefficient. This may be explained if the thermoneutral process leads to the forma-tion of an excited ion. The promotion energy observed here corresponds closelyto the weak absorption band occurring at 3100A in the nitrite ion, alkaline nitritesolutions, nitromethane, nitrosyl chloride, nitric acid and the nitrates. This haA .L. FARRAGHER, F. M. PAGE A N D R . C. WHEELER 207been attributed to an no,n* transition.15 The occurrence of this transition at awavelength corresponding so closely to the electron affinity of NO? supports theidentification of the thermoneutral process.Of processes leading to the formation of the ground-state nitrite ion, those whichinvolve direct electron capture or simple bond formation, e.g.,NO,+e+NO,, (0NO + 0-3N02, (ii)are exothermic by at least 92 kcal, and the resultant ions will break down, so thatthe processes will be very inefficient. The intervention of a chaperon moleculewill increase the efficiency but the corresponding processes leading to the formationof the excited ions will always be favoured. Since the transition to the ground state,and electron detachment, are in almost exact resonance, the excited ions will alsobreak down very rapidly.The overall process will therefore always be of very lowefficiency.Any processes involving formation of the ground state of the ion by a bimolecularmechanism involving bond breaking would be expected to be much more efficientand two possible examples are(iii) NO2 + X-+NO, + X,NO+O,+NO, +O. (iv)In both cases AH will be less than excitation energy of the ion and there is alsothe possibility of kinetic energy transfer to the eliminated X. Reaction (iii) cor-responds to the charge transfer observed by Curran, and the value of 92 kcal pro-posed here for the electron affinity of NO2 supports Curran’s lower limit of 87.Reaction (iv) is of interest since Branscombe 13 has noted that the NO, formed indischarges through air seems to be formed at the expense of 0, ions.In the light of these observations it seems probable that the NO, ions foundin the E layer during daytime 16 are formed by processes (iii) and (iv).Since96.5 % of the negative ions are NO, we can assume the overall equilibrium givenby reaction (i) and write :5040V 5T 2log,&, = -- +-log,, T-5-9;ifT = 280°K and K, = [NO2][e]/[N0;][NOi]/[e]-lO-14 17 and taking [NO21 as 107 cm--3,18 use of the Saha equation givesV = 0.48 eV.It is clear therefore that the NO2 and NO; are not in thermodynamic equilibriumwith the electrons. Alternatively, assuming processes (iii) or (iv), then, since theentropy of reaction will be negligible,log KPf exp (- AHIRT),Substitution again leads to the conclusion that NO, is not in thermodynamicequilibrium with the oxygen ions.The NO, ion is probably present at a steady-state concentration determined bythe rate of photodetachment, but in the absence of any data on the rate of radiativeattachment we cannot estimate the position of equilibrium.It is probable, however,that the rate of photodetachment from the excited state will be prohibitively largeand that the NO; in the upper atmosphere is in the ground-state condition208 ELECTRON AFFINITIESThe authors thank Prof. C. C. Addison for the gift of anhydrous copper nitrate,the Royal Society for a grant for the purchase of apparatus, and the U.S. Armyfor support under contract DA-91-591-EUC-6-2808.1 Page, Trans. Faraday SOC., 1960, 56, 1742.2 Gaines and Page, Trans. Faraahy Soc., 1963,59, 1266.3 Addison and Hathaway, J. Chem. SOC., 1958, 3099.4 Dodd and Robinson, Experimentd Inorganic Chemistry (Elsevier).5 Cottrell, The Strengths of ChemicaZ Bonds (Butterworths, London, 1954).6 Page, Trans. Faraday Soc., 1961,57, 1254.7 Langseth and Walles, 2. physik. Chem. B, 1934, 27, 209.8 Yatsimirskii, Izvest. Akad. Nauk. S.S.S.R., Odtel. Khim. Nauk., 1947, 453, 411.9 Kapustinsky, Acta physicochim., 1943, 18, 370.10 Pritchard, Chem. Rev., 1953, 52, 529.11 Smith and Seman, quoted in Branscombe, Proc. 5th Int. Conf: Ionization Phenomena in Gases,12 Curran, Physic. Rev., 1962, 125, 910.13 Branscornbe, The Threshold of Space (Pergamon Press, 1957), p. 101.14 Fox, J. Chem. Physics, 1960, 32,285.15 Trawick and Eberhardt, J. Chem. Physics, 1954, 22, 1462.16 Johnson and Heppner, Trans. Amer. Geophys. Union, 1956, 37, 350.17 Webber, Proc. 5th Nat. Con$ Ionosphere (London, July, 1962), p. 63.18 Harteck, The Threshold of Space (Pergamon Press, 1959, p. 37.1961, p. 9

ISSN:0366-9033    

DOI:10.1039/DF9643700203

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

GENERAL DISCUSSIONDr. M. Ackermann (Centre National de Recherches de I’Espace, Bruxelles) said :Prof. Nicolet has shown the importance of the activation energy E3 of the reaction0 + 0 3 -I- M-dO2on the aeronomic photochemistry of oxygen. The value of the rate constantk2 = 2.2 x 10-35 exp (1900/RT) cm6 sec-1 given by Dr. Thrush for the reactionM being argon, in his comment about the data obtained by Kaufman and Kelso 1leads to the following implication.Benson and Axworthy 2 and Castellano and Schumacker 3 have measured thedependence of k ~ / k 3 on temperature. Their determination of E3 depends on thevalue of E2. The values of (E3-E2) equal to 6.60 kcal/mole in ref. (2) and to(3)0+02+M+03+M (2)concentration (cm-3)FIG. 1.5.44 kcaljmole in ref. (3) imply E324 kcal/mole if E2 = - 1.9 kcal/mole.Takinginto account the relative efficiencies given in ref. (3) for different third-bodies inreaction (2) and considering the data of ref. ( 1 ) and ref. (4), the following set ofvalues can be assumed :kl = 3 x 10-33 cm6 sec-1for the reaction 0 + 0 + M+O2 + M, k2 = 3 x 10-35 exp (950/T) cm6 sec-1, k3 =7 x 10-12 exp (- 2000/T) cm3 sec-1.1 Kaufman and Kelso, this Discussion.2 Benson and Axworthy, J. Chem. Physics, 1957,26, 1718.3 Castellano and Schumacher, 2. physik. Chem., 1962,34, 198.4 Mathias and Schiff, this Discussion.20210 GENERAL DISCUSSIONThe resulting mesopheric concentration of 0 and 0 3 are shown by the linesdrawn between stars in the figure with the curves given by Prof. Nicolet in fig. 4of his paper.The important influence of the partition of between reactions (2) and(3) on the distribution of ozone with altitude is evident.Prof. F.S. Dainton (University of Leeds) said: Amongst the reactions shownby Prof. Groth was that between an oxygen atom in its 1D state and methane whichwas writtenO(1D) + CHq-CH2 + H20and Prof. Nicolet assigned an energy of activation of 7-8 kcal to this reaction. Onthe other hand, Prof. Norrish and his colleagues have provided convincing evidencethat oxygen atoms in the 1D state produced in the photolysis of gaseous ozoneattack substances like ammonia, water, and methane to detach a hydrogen atomand produce one or more hydroxyl radicals. I would like to raise a general questionas to the mechanism of these reactions.It seems that the oxygen atom in its 1Dstate ought to be able to undergo ready insertion between CH, OH, and NH bonds,thereby forming initially a highly excited adduct which might conceivably decom-pose after a very short space of time which would be longer the more polyatomicthe adduct molecule. There is clear evidence of the existence of this insertionreaction when oxygen atoms in the 1D state are generated in liquid water. Someyears ago Prof. Taube showed that oxygen atoms emerging from the photolysisof dissolved ozone in liquid water produced hydrogen peroxide. More recently,Mr. Fowles in my laboratory has been studying the photolysis of nitrous oxide inliquid water using 1849 A light. In this latter case both the kinetics of the reactionand also experiments in which N20 fully labelled with 1 8 0 indicate that the oxygenatom which is detached from the nitrous oxide in the primary act enters a watermolecule in its immediate vicinity with 100 % efficiency. Furthermore, there isno evidence for hydroxyl radical production from this process and one can onlyconclude that the highly energized H202, which in the gas phase flies apart intotwo hydroxyl radicals, is immediately deactivated in liquid water.This wouldgive a lifetime for the excited H202 molecule which would be greater than about10-10 sec, but less than the collision interval in the gas-phase experiments. Also,if the aqueous reaction O(1D) + H20 were to give two hydroxyl radicals initiallythen some of these would certainly escape since in the photolysis of hydrogenperoxide using either 3130 A or 2537 A light, 50 % of the hydroxyl radicals whichare formed in this primary dissociation escape from the solvent cage into the bulkof the solution.It would be very interesting to carry out experiments on the photo-lysis of N20 using 1849A light at pressures of substrates such as water, methane,etc., which would bridge the gap between the pressures used in Prof. Norrish’s kindof experiment and those concentrations which are obtained in the liquid state.Dr. J. B. Hasted (University College, London) said: The origin of the discrep-ancies between both Of02 and O+N2 rate constants deduced from afterglow mass-spectrometer measurements by Sayers and co-workers, by Fite and co-workers,and by Langstroth and Hasted, is presumably to be sought in the differences ofmethod of excitation, which may produce different excitation of the neutral reactants.Similar effects are noted in the kinetics of neutral species. The observations of Sayersand Smith with helium + nitrogen mixtures certainly indicate the relevance of excitationpower.However, we have been unable to detect significant variation of Of decay ratewith power of applied breakdown pulse either in02 + He, or in02 + N2 + He. The rela-tive efficiencies of r.f. breakdown, “ d.c. pulse ” breakdown and Linac electron pulse tGENERAL DISCUSSION 21 1produce ions without exciting either them or the neutral species, are insufficiently under-stood, so that the safest course would be to inject a mass-analyzed pulse of ions into thegas.The difficulties encountered in injection into a mobility tube need not inhibit suchan approach.With regard to the form of the cross-section function for charge transfer plusion-atom interchange processes, such as O+Nz, my early beam measurement is notparticularly relevant to the thermal energy processes, since the latter are believedto be dominated by ion orbiting in the inverse fourth power field. Whilst non-resonant ion-atom charge transfer processes decrease with decreasing impactvelocity in the non-orbiting region, the evidence in the orbiting region indicates thatfor a two nucleus process there is no enhancement but that for complex mole-cular systems there is.For ion-atom interchange processes almost all the avail-able evidence indicates cross-sections falling with increasing impact velocity in theorbiting region. Thus where there are two competing processes it is plausible toassume that the ion-atom interchange may dominate at low impact velocities whilstthe charge transfer must dominate at sufficiently high energies ; in the light of theseconsiderations the agreement between the O+H charge transfer rate derived byHanson et al. and the cross-section measurements of Fite et al. may be fortuitous.For orbiting collisions the energy liberated is divided between the products inthe ratio of their masses. It is clear that the escape of He+ from the atmospherecould not be effected by a near resonance charge transfer process, but this does notrule out the competitive role of such a process.The late appearance of negative ions in Fite and Rutherford’s afterglows mightbe caused in part by plasma polarization and in part by slow rates for the relevantattachment processes.We have observed negative ions in the earlier afterglowin halogens, but with the aid of a few volts positive bias on the radiofrequencymass-spectrometer probe inserted into the plasma.A theory of the metallic probe in an electronegative plasma has been developedby Boyd and Thompson,l who find that for negative ion-electron ratios greaterthan about 2, the current of positive ions to a negatively biassed probe is proportionalto the square root of the ion density; but for ratios much less than 2 the current isdirectly proportional to the ion density.We have observed positive ion decays tochange their logarithmic slope by a factor of two after a certain delay time; suchobservations might be profitably employed as a check on the moment at whichnegative ion densities become appreciable. For a probe at wall potential no suchcritical time is observed, and even in an electronegative gas a positively biassedprobe should collect negative ion currents proportional to the negative ion density.Dr. D. Smith (University of Birmingham) said: It seems likely that the methodused to ionize the gas will determine the number of excited species present in theafterglow and it seems probable that the r.f. ionizing pulse will be particularlyefficient in this respect.However, in order to influence the reaction rates the ex-cited species would have to last well into the afterglow. Even then, it is not easyto predict what their effect would be in the reaction rates. At the moment wehave no information on this aspect of the work but it is hoped that the programmeof work that we are engaged on will help to clarify the situation and perhaps explainthe discrepancies between our results and those of Langstroth and Hasted.Dr. J. B. Hasted (University College, London) said: We had hoped to be ableto report an interchange process between O+ and N2 in our drift tube. However,the small size of cross-section, coupled with unsuitability of rare gases as buffers,1 Boyd and Thompson, Proc. Roy.Soc. A, 1959,252,102212 GENERAL DISCUSSIONnecessitate the injection of O+ into pure N2. The dominant ion is identified asN4+ at low Xj’p : during the thermalization process, O+ is converted by charge transferto Nl. Identical peak patterns are obtained by the injection into nitrogen of Of,N; and N+. Previous nitrogen mobility studies 1, 2 yield two peak patterns. Atlow and moderate X/’ there is interconversion between N i and N4+, and recentlyan interconversion between N i and N+ (large mobility) at high X/p has been reported.The mobilities of N4f and N i are almost identical. We observe the first of theseprocesses ( N i +Nz-+N:), 5 = 3 x 10-16 cm2 at 0.24 eV. But at moderate X/pwe also report a third peak (fig. 1) of mobility midway between that of N i and N i ,Xlp, V cm-1 torr-1FIG.1.-Mobilities in nitrogen as a function of X / p . Broken line indicates the N$ data of Varney.tentatively identified as vibrationally excited N;, perhaps o = 1. The mobility isapparently much too low to arise from a collision between unlike systems. Inter-conversion rate constants are deduced as follows :N,’(u = O)+N,+Nl(u = 1)5 = 2 x cm’,N;(U = I)+N,+N,I(o = 0)5 = 1.5 x 10-16 cm2,although there may be included a contribution from the dissociation of N4+ :N l +N,+N;(V = 1)o = 4 x 10-l~ cm’,all at a mean energy of 0.24 eV.1 Varney, Proc. Sixth Con$ Ionization Phenomena in Gases (Paris, 1963).2 McAfee, Proc. Sixth ConJ Ionization Phenomena in Gases (Paris, 1963)GENERAL DISCUSSION 21 3The vibrational excitation and deactivation mean cross-sections are, unexpectedly,much larger than those for neutral molecules; this may be related to the fact that,unlike the latter, they arise during inverse fourth-power orbiting collisions, whichallow a longer time for the interconversion of kinetic and internal energy.Sincethe onset energy for the excitation process is double the excitation energy (-0.25 eV)a large onset cross-section is implied by our measurement. For this interconversionthe measured mean cross-section rises with increasing Xlp, as expected. The con-tribution of non-resonant charge transfer processes to vibrational activation anddeactivation of molecular ions, noted by Bates, should not be ignored.Prof. F. S. Dainton (University of Leeds) said: Dr.Hasted said that the ratesof the vibrational energy transfer reactionsN,’(v = 1)+N2(u = O)+N,’(u = O)+N,(u = l),andNZ+(u = O)+N,(u = l)+NZ+(u = 1)+N2(u = 0),were about 1000 times larger than the bimolecular collision rates between neutralmolecules and ascribed this to the possible existence of a sticky collision complex.I would like to ask whether the sticky collision complex is the N4+ ion which wasshown in some of Dr. Hasted’s curves ?If this is the case, is there a complete “ scrambling ” of the four nitrogen atomsin this tetratomic molecule ion and would it be possible to identify such scramblingby repeating Dr. Hasted’s experiments with mixtures of isotopically pure 14N2and lsN2 or do other processes occur in his experiments whch would lead to thisscrambling and thereby obscure the results ?Finally, in view of the fact that the vibrational quanta will be different in themolecule and in the ion, what is considered to be the mechanism of the exchangeif it does not involve a tetratomic complex in which energy flow can occur quitefreely between the vibrational modes ?Dr.J. B. Hasted (University CoZZege, London) said : In reply to Prof. F. S. Dainton,It is very probable that the activation and deactivation processes proceed by anN i complex formed by a classical inward-spiralling process, in which energy canexchange between the vibrational modes. The answer to the question, whether theN i we observe is in fact this complex, is unfortunately not self-evident.The ob-served N i must have a lifetime in excess of 1 msec, the period spent in the drift tube.The time necessary for the distribution of energy among the degrees of freedom ispresumably much smaller. Any complex taking part in our observed process,must be destroyed in the drift tube, but it is not established that this takes placewithout collision. However, the N;(u = 0) is formed either from the NZ we ob-serve or the Ni(u = 1) via a shorter-lived complex.Unfortunately the mass-sensitivity of peak identification is not sufficiently goodin a mobility experiment to allow the distinction between 14N and 15N. In mass-spectrometer source experiments, such as those carried out by Saporoschenko, usemight be made of this facility.Dr. F. A.Baker and Dr. L. A. PCtermann (Znstitut BatteZZe, Genieve) said: Duringthis discussion we have learnt of unusual catalytic effects occurring at nickel sur-faces and of the important role that may be played by certain gaseous impuritiesin attempts to reproduce “ atmospheric ” chemical reactions in different laboratories.1 Saporoschenko, Physic. Reu., 1958,111, 1550.NZ(u = O)+N~-+N:(V = 1)+N2214 GENERAL DISCUSSIONA large amount of work has been done to establish the nature, magnitudes andsources of various gaseous “ residuals ” that have been encountered and subsequentlyminimized in attempts to produce ultra-high vacua (UHV). In our experience, usingsilicone oil diffusion pumps in conjunction with activated alumina traps over thetotal pressure range from 10-8 to 10-11 torr, hydrogen and carbon monoxide arefrequently the most abundant residual gases.Where higher pressures have beenstudied, it is usual to find some carbon dioxide and, with little or no degassing bybaking, water vapour and methane.Working in the above stated pressure range we find that either a “thermalflash ” to approximately 250”C, or pulsed bombardment by electrons at low power(5 mA, 100 V) readily releases some 10-6 torr l./cm2 of hydrogen andcarbonmonoxidefrom the surface of commonly used metals (molybdenum, Kovar, nickel and variousstainless steels).Our pumping speeds are 0.8 l./sec for H2 and 0.2 l./sec for CO. If, e.g., theelectron bombardment is regularly pulsed on for 10 sec and off for 60 sec, the de-sorption peak pressure amplitude from a 20cmz metal surface becomes quasi-stable in the 10-7 torr region.Let us assume that a system of 2-3 1.volume, subjected to fairly stringent de-gassing procedure, provides an ultimate pressure of 5 10-8 torr. If “ spectro-scopically ” pure work gas is then allowed to flow through this volume to providea working pressure of 10-1 torr, the impurity level is 5 0.1 p.p.m. However, slightheating or bombardment of surfaces as discussed above may produce impurities upto 10p.p.m. cm-2. This level would increase with time without a continuous flowof gas. Such gas sources appear to be replenished from the bulk material andare consequently difficult to exhaust.We are also aware that low energy ions are very much more efficient than electronsin desorbing gas.So, we would ask if sufficient attention is being given by specialistsin this field to the results obtained by UHV specialists which provide direct in-formation on the nature, magnitudes and sources of impurities.Dr. M. A. A. Clyne, Dr. J. C. McKenney and Dr. B. A. Thrush (CambridgeUniversity) said : We have recently measured the rate constant of the reactionover the temperature range 193-373°K in a flow system. The concentrations ofmetastable oxygen molecules were limited to a very low value by passing argoncontaining extremely low concentrations of molecular oxygen through the r.f.discharge used. In this way, reactions of metastable oxygen molecules, which complicatethe kinetics, were obviated. Great care was taken to eliminate hydrogenous im-purities.The rate constant of reaction (1) was determined from the accelerationof the oxygen atom decay when molecular oxygen was introduced downstreamfrom the discharge. Oxygen atom concentrations which were typically 10-3- 10-4mole % were measured by air-afterglow emission using a E.M.I. type 9558B photo-multiplier cell. No emission by electronically excited oxygen molecules could bedetected under these conditions.We obtain k p = 2-2 x 10-35 exp (+ 19OO/RT) cm6 molecule-2 sec-1 in excellentagreement with the value of kk = 1.7 x 10-35 exp (+ 1700/RT) which Jones andDavidson 1 deduce from their data on the decomposition of ozone in shock-tubesand Benson and Axworthy’s work on its thermal decomposition. TakingkpZ/kp = 1.7, our data give kp2 = 9.1 x 10-34 cm6 molecule-2 sec-1 at 300°K inagreement with Kaufman and Kelso.1 Jones and Davidson, J .Amer. Chem. SOC., 1962, 84, 2868.0+02+Ar = O3fA.r (1GENERAL DISCUSSION 21 5Prof. W. L. Fite (University of Pittsburgh, Pennsylvania) said: In the detectionof excited 0 2 using a mass spectrometer, the process e+O*, +O-+O appears tobe much more satisfactory than using positive ion formation below the groundstateionization threshold.For example, fig. 1 shows curves 1 for 0- formation in electron collisions withnormal 0 2 molecules and with molecules issuing from a gas discharge. While thestates of excitation of the discharged 0 2 are not known, the shift in the dissociativeattachment cross-section curve is spectacular and 0- formation might be well em-ployed as a detector of 0; in experiments of the type reported by Mathias et al.8 I4-X--X- e + O2 - 0-e + o~*--o-0 2 4 6 8 10 12 14 16 18electron energy (eV)FIG. 1 .4 - production curves for electron collisions with groundstate and excited 0 2 .Even more interesting is the observation that an only slightly less spectacular shiftof the dissociative attachment cross-section curve occurs on heating the 0 2 from300 to 2100"K, suggesting that if electronic excitation is required to produce theshift, then the lAg state must be very effective. However, vibrational excitationof the ground electronic state may also contribute to the shifting of the 0- productioncurve.Dr. A. B. Callear (University of Cambridge) said: In collaboration with Greenand Williams,:! we are presently investigating the optical absorption spectrum ofprocesses occurring during and following a microwave discharge in various gases.The method might be regarded as a variant of flash photolysis, with a powerfulmicrowave-pulse generator in place of the photolytic flash.Our results so farare in some respects very sinlilar to those of flash photolysis; the fast electronsare subject to the optical selection rules. One important difference is that theexcitation is not limited by the quartz cut-off at about 2000 A.The peak electric field in our cavity is 1400 V/cm, with gas pressures of a fewhundred mm. Thus the electron temperature may be similar to that in the dis-charge used by Bader and Ogryzlo.3 Although we have not yet attempted anyexperiments with oxygen, it may be of value to speculate on the mechanism.1 Fite and Brackmann, Proc.Sixth Int. Con$ Ionization Phenomena in Gmes (Paris, 1963),2 Callear, Green and Williams, Nature, 1964, 201,70.3 Bader and Ogryzlo, this Discussion.vol. I, p. 28216 GENERAL DISCUSSIONThe main primary process in a micro-wave discharge in oxygen should be ex-citation to the upper state of the Schumann-Runge bands.O2x3z; + ~--QB~E; + e.The B3Z; state will decompose to 1D and 3P oxygen atoms if it has the requiredenergy. Otherwise it may predissociate,Apparently the cross-section for the banded region is small compared to the cross-section for the continuum.1 Thus there should be produced a high yield of O(1D)as suggested by Prof.Norrish.2 However, the O(1D) should react rapidly with0 2 , a possible scheme being0(1D)+02X3E;-,0,aXA,+03P.There should also be substantial vibrational excitation in the discharge, and theabsence of emission from blE:(u = 1) shows that a fast relaxing process is occurring,and the vibrational temperature of the system may be ambient. The reaction ofatomic oxygen with 0 2 would cause rapid vibrational relaxation.0+02(21 = n)-,[o3]-+0+02(11 = m).Prof. H. I. SchB (McGiZZ University) said: There appears to be a discrepancybetween the two sets of experiments reported by Dr. Warneck. In his flow experi-ments, in which 0 2 was added downstream of the photolysis cell, the 0 3 quantumyield gave values of y = k4/k3[M] of about 41, while the literature data quotedgave y = 8 for O(3P) atoms. This was taken as evidence that k3 is much lowerfor O(1D) atoms than for O(3P) atoms.However, in the static circulating system the ratio of 0 2 to CO quantum yields,R = ~(OZ)/~~(CO) was found to be about 0.3, whereas, if reaction (3) were unim-portant, R should be 0.5.Dr. Warneck attributes the 0 2 deficiency to the formationof 03. However, the maximum amount of 0 3 which could be formed, consistentwith his mechanism, would be the steady-state value given byCombining this with the mass balance relationshipyields a maximum value forif y = 41. Thus R must lie in the range 0.50>R>0.48 to be consistent with theresults of the flow experiments.Even with y = 8, R could not be less than 0.42.If the value of k3/k4 reported by Mathias and Schiff (in this Discussion) for O(3P)atoms were used, a considerably lower limit for R could be obtained.However, if the missing 0 2 were due to 0 3 formation, then the ratio of [02]/[CO]should increase with irradiation time, due to the build-up of 0 2 and the dependenceof reaction (3) on [02]. Thus the straight line plots of fig. 3 should not have beenobtained. Similarly, in fig. 2, the quantum yield of 0 2 should decrease with pressuredue to the dependence of (3) on [MI. Thus it is not likely that the missing 0 2 ispresent as 0 3 , unless there is an appreciable amount of 02(1Ag) present and the1 Lassette, Silverman and Krasnow, J. Chem. Physics, 1964, 40, 1261.2 Norrish, this Discussion.[COI = 3[031+2[021R = 7/(3 +27) = 0.4GENERAL DISCUSSION 217reaction 02(lAg) + O(1D) + M+03 + M is very fast.Similar 0 2 deficiencies havebeen found in our laboratory and in the work of Mahan.We have also obtained evidence that the reverse actionO(lD)+CO+(M)+COz+(M) (5)may also be important. Thus, we found a decrease in ~ ( O Z ) when a small amountof CO was added. Also, at higher temperatures (110OC) plots such as fig. 3 werenot linear but had slopes which decreased with irradiation time. The inclusionof this reaction (5) implies a CO quantum yield less than unity for which we havealso obtained some evidence.Prof. F. S . Dainton (University of Leeds) said: Would Dr. Warneck care tocomment on the lack of material balance which he finds in those of his experimentswhich contain no oxygen initially and which has been observed by other workersincluding Mahan and ourselves.It has consequences for the kinetics of the re-action which Dr. Schiff mentions. Perhaps I may add that I think that this is agenuine observation in the sense that it is not due to any loss of ozone by reactionwith the walls or tap grease, etc., since in experiments on the radiolysis of liquidcarbon dioxide carried out by Dr. R. L. S. Willix in our laboratories, it has beenpossible to identify very small amounts of ozone and to get a material balance in asystem which decomposes exclusively to carbon monoxide, oxygen and ozone,whereas Dr. Baulch working on the vacuum u.-v.photolysis of carbon dioxidehas not been able to find any traces of ozone even using similar analytical methodson similar quantities of material to those used by Dr. Willix. There may be a genuinedifference here in that in the radiation chemical case oxygen atoms are producedin the 3P state whereas in the vacuum u.-v. experiments the oxygen atoms are pro-duced in the 1D condition. Perhaps oxygen in the 1D state, whichis known to be able toundergo the insertion reaction and addition reaction to double bonds, can alsoundergo addition to a carbonyl group in carbon dioxide to form the species co3first proposed by Katakis and Taube which would be iso-electronic with cyclo-propanone.Dr. Peter Warneck (Geophysics Corp. of America, Bedford, Mass.) said: TheCO2 photolysis experiments employing the circulating system were primarily de-signed to determine CO quantum yields.The 0 2 quantum yields also obtainedwere not considered too reliable because it was thought that the small amounts ofozone simultaneously produced by reaction (3) might react with metal surfacesand/or stopcock grease, thereby resulting in a loss of oxygen, and precluding asteady-state treatment for ozone. The unpublished results mentioned by Prof.Dainton, however, may open up new aspects since they suggest that the oxygendeficiency is a genuine finding. If this is accepted, the photolysis of pure CO2will require additional study and interpretation, although the ratio determined inthe present study probably remains unaffected because of the fairly high concentra-tion of oxygen admixed in that case.The formation of appreciable amounts ofozone via reactions of the type,02(1A~)+O(lD)+M403+M,can be precluded because of the rather small upper limit for the ozone quantumyield found with the flow system using pure C02.Note added in proof:Since this Discussion was held we have obtained additional experimental resultswhich are in good agreement with the findings mentioned by Prof. Dainton. I218 GENERAL DISCUSSIONcould be verified that the extent of ozone formation in the closed system is neglig-ible, the upper limit for the 0 3 quantum yield being Q(O3) si 0.03. The [02]/[CO]ratio was found to be dependent on the light intensity, on the ratio of carbon dioxideto admixed rare gas concentration, and on the time interval after which an irrad-iated sample was subjected to analysis.The results appear to be consistent withthe notion that 1D oxygen atoms can react with C02 to form an unstable additionproduct. An estimate for the rate constant associated with this process indicatesthat its influence is negligible under the experimental conditions employed in thedetermination of y.Dr. R. D. Cadle (National Center for Atmospheric Research) said: In reply toDr. Dalgarno, the intensity of the 6300A line in the day airglow is still uncertain.Values reported in the literature vary from about 2 to 40 kR referred to the zenith.One recent paper 1 suggests a value of 30-40 kR, nearly that calculated by Brandt 2and used as the basis of some of the calculations in my paper.Part of the widevariation in the reported values may represent a diurnal or seasonal effect. How-ever, in view of the difficulties in measuring the intensity of this line, this is far fromcertain.A direct comparison of the value of Brandt with the measured intensity of the6300 A line cannot be used to obtain the collisional deactivation coefficient of O(lD),since Brandt's results depend only in part on this coefficient; nonetheless, theintensity of the 6300A line in the day airglow, when established, can be used to seta lower limit for this coefficient.If the collisional deactivation coefficient does turn out to be much greater thanthat used in my paper, the corrected values for O(1D) concentrations at altitudesbelow about 140 km can readily be calculated from the values in fig.2.Dr. E. A. Ogryzlo (University of British Columbia) said: There is some experi-mental evidence which indicates that the formation of excited molecules in highfrequency discharges may indeed proceed via the reaction :02(3~;) + e*+ 0 ( 3 ~ > + 0(9) + e (1)(2) o(9) + 02(3~;)-, 02(lAg) + o ( ~ P )orA large cross-section for reaction (1) at an electron energy of 8 eV is suggestedby the work of Schulz,3 and rapid deactivation of O(1D) by oxygen according toreaction (2) was first proposed by Bates and Dalgarno.4In a typical flow system operating at pressures of about 1 mm Hg, and at adistance of several cm from the discharge there is no evidence for any O(1D)atoms 5-7 and there is considerable evidence for the presence of 02(lAg).8s 9 How-ever, at low pressures and immediately after the discharge, Nutt and Biddlestone 10have reported evidence for the presence of O(1D) and the absence of 02(lAg).These two observations are understandable if at low pressures and at short re-0 2 ( l q >1 Jarrett and Hoey, Planet.Space Sci., 1963, 11, 1251.2 Brandt, Astraphys. J., 1958, 127, 54.3 Schulz, Physic. Rev., 1959, 116, 1141.4 Bates and Dalgarno, J. Atm. Terr. Physics, 1953, 4, 112.5 Harteck and Kopsch, Z. physic. Chem. By 1931, 12, 327.6 Kurt and Phipps, Physic. Rev., 1929, 34, 1957.7 Rauson and Beringer, Physic. Rev., 1952, 88, 677.8 Foner and Hundson, J. Chem. Physics, 1956, 25, 601.9 Bader and Ogryzlo, this Discussion.10 Nutt and Biddlestone, Trans.Faraday SOC., 1962,58,1368GENERAL DISCUSSION 219action times, only reaction (1) has occurred, while under the usual high pressureoperating conditions used in kinetic studies O(1D) has been removed by reaction(2) to form the observed 02(1Ag).The absence of vibrationally excited 0 2 in the discharge products may indeedbe associated with a rapid deactivation by oxygen atoms. This would be con-sistent with our recent observation1 that when 02(1Ag) and 02(1X:g+) are producedin the reaction Cl2 + H202-+2HCl+ 02*, emission from vibrationally excited 02(1 Ag)and 02(1XcB+) is observed. It is possible, however, that a great deal of direct vibra-tional excitation does not occur in electrically discharged oxygen since the cross-section for vibrational excitation of 0 2 by 0.2-4-5 eV electrons appears to be quitesmall compared with that for N22 (where the vibrationally excited moleculescan easily be observed in a flow system3).With nitrogen we have then a specificmechanism for creating vibrationally excited molecules, but not for adding to thetranslational energy of the molecules. Vibrational relaxation would therefore beslow. On the other hand, in oxygen there is a “primary step” for creating trans-lational energy (absorption into the Schumann-Runge continuum) but appar-ently none that is very effective in creating vibrational excitation. Under thesecircumstances the relaxation of a small amount of vibrational excitation wouldbe rapid.Dr.F. Kaufmann (Aberdeen Proving Ground, Md.) said: Doherty and Jonathanfind the intensity of the O+NO chemiluminescence to be first order in [O] and in[NO], but independent of total pressure down to 0.85 microns Hg. They explainthis result using the mechanism proposed by Broida, Schiff, and Sugden 4 and therebyobtain an upper limit of 103 sec-1 for the rate constant, k4, for crossing from NO,*(C)to NO,*(B). Such a value of k4 invalidates the basic assumption of the mechanism,viz., that radiation from state C can be neglected compared to that from B. Itis assumed in the paper that C is the state observed by Neuberger and Duncan 5whose radiative rate constant k 6 ~ is 2.3 x 104 sec-1, i.e., 170 times slower thank 6 ~ = 4 X 106 sec-1 which was calculated from the integrated absorption coefficientof normal NO2 and therefore represents an upper limit, because transitions to otherexcited states may be partly responsible for this absorption.The ratio of the steady-state concentrations of states B and C in the proposedmechanism is given bywhich, at low pressure, is “k6~/k4 = 4000.It is therefore clear that radiationfrom state C, far from being negligible, is at least 4000/170 = 23 times more in-tense than that from B which invalidates the initial assumption. Moreover, radi-ation from C should show the expected fall-off with decreasing pressure such thatI0 = I[O]-1[NO]-1 would be down to one-half its high pressure limit when thepressure is 7 microns if the proposed rate constants are used.The inapplicability of the proposed mechanism, the apparent pressure inde-pendence of Io, but its dependence on the nature of M (Mr.Kelso and I have founda range of a factor of 2.2 in I0 for different M) together make this chemiluminescencea puzzling phenomenon.1 Brown and Ogryzlo, Proc. Chern. SOC., April, 1964.2 Schulz, Physic. Rev., 1959, 116, 1141.3 Morgan, Phillips and Schiff, Disc. Faraduy SOC., 1962, 33, 118.4 Broida, Schiff, and Sugden, Trans. Faraday SOC., 1961,57,259.5 Neuberger and Duncan, J. Chern. Physics, 1954, 22, 1693220 GENERAL DISCUSSIONProf. P. Harteck (Rensselaer Poly. Inst., Troy) said : I would like to add a fewcomments to the paper of Jonathan and Doherty. Our results are in full agree-ment with their showing that the light intensity observed is proportional to thenitric oxide and oxygen atom concentrations even in this low pressure region.Theoverall mechanism is a second order reaction for the light emission. There can bemade many arguments as to why a simple two-body reaction is not likely, but athree-body mechanism at these low pressures which is followed by a quenchingreaction to result in an observed two-body mechanism seems impossible. A three-body recombination rate coefficient for the reaction NO + 0 + M+N02 + M hasbeen measured by ourselves and others as 6 to 7 x 10-32 (particles/cm3)-2 sec-1,which, incidentally, is a very high rate of recombination. The results of our low-pressure studies show the emission intensity follows the second-order reaction rateand the intensity at 10 p total pressure can be calculated from the absolute intensitymeasurements of Schiff. In order to obtain the necessary intensity of photonemission with a primary three-body mechanism, each three-body collision reactionwould have to result in a light emission with no quenching.The reaction wouldhave to appear third order unless a substantial part of the excited molecules werequenched, which would also show a marked decrease in intensity. Of course, thesecomparisons are made at 10 p and the results have been shown to be valid to evenlower pressures. We therefore conclude that the light emission is due to the simpletwo-body reaction.From my own experiments, I realize how difficult are the experiments of theauthors.I have shown in my paper that’even in the cleanest systems certain con-ditions of the surface can produce the excited oxygen molecule and may lead toerroneous results. I would like to ask the authors what precautions were taken toinsure a clean system with a minimum of interfering effects.Dr. D. B. Hartley and Dr. B. A. Thrush (Cambridge University) said: Experi-ments in this laboratory on the air-afterglow reaction and on the related chemi-luminescent NO+O3 reaction 1 indicate that the same electronic state of NO2 isresponsible for almost all the emission by NO2 observed in these reactions and forthe bulk of the visible and near ultra-violet absorption spectrum of N02. Theapparent discrepancy between the radiative life measured by Neuberger andDuncan 2 (4.4 x 10-5 sec) and the value estimated from the integrated absorptioncoefficient (-2 x 10-7 sec) can be explained partly by the expected dependence ofthe transition moment on bond angle but mainly by free internal conversion betweenthe excited state and the ground state. The latter process ensures that moleculeswith high energies are predominantly in the ground electronic state which has amuch higher density of levels at these energies.The absolute magnitude of the quantity 10 in the relationand the fact that it has a negative temperature coefficient similar to that for thereaction 3both support the view that a third body is involved in the excitation of the air-afterglow, as does our recent observation of charges in its intensity distribution forM = H20, a species which gives a high value of kl and a low value of 10 comparedwith oxygen or argon.1 Clyne, Thrush and Wayne, Trans.Faraday SOC., 1964,60,359.2 Neuberger and Duncan, J. Chem. Physics, 1954,22, 1693.3 Clyne and Thrush, Puoc. Roy. SOC. A, 1962,269,404.O+NO+M = N02+GENERAL DISCUSSION 221We have shown that the air-afterglow has a similar intensity distribution to theNO2 fluorescence excited by HgA = 4358 A radiation which produes NO2 mole-cules with energies just below the predissociation limit. Baxter's data1 on thequenching of this fluorescence can therefore be used to determine the third-orderrate constant for recombination into the state which emits the air-afterglow. Thevalue obtained is within a factor of two of the total rate of recombination.Asthe rate of recombination into a shallow electronically-excited state is believed tobe much lower than into a ground state, we consider that there is strong evidencethat the predominant mechanism of the air-afterglow consists of three-body recom-bination into the high vibrational levels of the electronic ground state of NO2followed by radiationless crossing into the excited state from which emission occurs.Direct application of Baxter's data might suggest that 10 should fall to half itshigh-pressure value at about 0.05 mm pressure for M=02. If allowance is madefor collisional redissociation of newly-formed NO2 molecules from levels just belowtheir dissociation limit, this pressure would be significantly reduced, perhaps by afactor of ten.Dr.N. Jonathan (Geophysics Corp. of America, Bedford, Mass.) said: I agreewith the idea of Dr. Hartley and Thrush that the chemiluminescence arises fromone excited state only. We have recently looked at the light emitted up to SOOOAwith an E.M.I. 9558B photomultiplier and find the same pressure independence forI0 in the relation I = Io[O][NO]. The results we have presented are not in agree-ment with their suggestion that 10 becomes pressure dependent at 50 microns or less.It remains possible that there is a cross-over point above the maximum pressureof 80 microns at which we were able to work. This seems unlikely for several reasons :(i) The light emission below the cross-over point should vary directly as the pressure ;(ii) by using higher flow-rates we have made approximate measurements up to 150microns and failed to observe an appreciable pressure effect; (iii) we have recentlymeasured the ratio of (I*)No to (10)co with a 1P28 photomultiplier in the 20-60micron region and have obtained results which are in reasonably good agreementwith those of Clyne and Thrush 2 in the 1 mm pressure region.As Dr.Kaufman has pointed out, the Broida, Schiff and Sugden mechanism isnot entirely satisfactory, since radiation from the NO,*(C) state should not beneglected. This is also true in the 1 mm pressure region considered by the origin-ators 3 since altering k4 by an order of magnitude (as we have done) does not makeNO,*(C) negligible.However, as these authors pointed out, the inclusion of radi-ation from the NO,*(C) state does not basically alter the scheme since at low pressuresthe equation reduces toThe radiation from NO,*(C) could be ignored if k6c-102sec-l. This value isperhaps not unreasonable if one adopts Dr. Thrush's suggestion that NO,*(C) isthe highly vibrationally excited ground electronic state. However such a schemeleads to a value for kl > 10-30 cm6 molecule-2 sec-1 and must also be regarded asunlikely in view of the experimental average value of 6 x 10-32 cm 6 molecule-2 sec-1for the total recombination rate constant.This mechanism is subject to many variations since one can only guess at mostof the rate constants involved. One is reluctant to talk in terms of a three-bodymechanism at pressures as low as 1 micron but the experimental data seem to1 Baxter, J.Amer. Chem. Soc., 1930, 52, 39202 Clyne and Thrush, Proc. Roy. Sac. A, 1962,269,404.3 Broida, Schiff and Sugden, Trans. Faraday SOC., 1961,57,259222 GENERAL DISCUSSIONindicate this is the case. We have recently obtained some results for lo with argon,nitrogen and oxygen as the respective third bodies at these low pressures. Thevalues are in very good agreement with Dr. Kaufman’s unpublished data obtainedin the 1 mm range. This would appear to be yet another argument for the three-body mechanism.In conclusion, I must go back to the original purpose of the experimental work.It would appear that the rate constant for light emission obtained in the 1 mmpressure range may be used to calculate the possible intensity of the night airglowcontinuum.Provided that the nitric oxide concentration is of the order of 108molecules/cm3 then the reaction of atomic oxygen with nitric oxide is a possiblecause.Dr. B. P. Levitt (Imperial College) said: I should like to support Dr. Thrush’ssuggestion that a single radiative electronic state of NO2 is responsible for the airafterglow with a radiative lifetime of about 2 x 10-7 sec (the value calculated fromthe absorption spectrum 1) rather than the higher value of 4.4 x 10-5 sec found 1from measurements of fluorescence excited by 4358 A radiation. This latter valueis used by Doherty and Jonathan. Clyne, Thrush and Wayne have recently shown 2that this state also gives rise to the banded emission observed from NO2 heatedto about 1000°K from the fluorescence excited by 4358A radiation and from the0 3 + NO chemiluminescence.The (nearly) continuous emission from NO2 heated in the shock tube to about2000°K arises from the same state as the air-afterglow, the principal evidence beingthat it shows the same short wavelength cutoff just below 4000A.3The absolute intensity of emission from NO2 per unit wavelength interval hasbeen measured at 2000°K. It rises from zero at 4000 A in a complex, but essentiallyexponential manner to about 1 x 1028 photons sec-1 mole-1 cm-1 at 5800 A, corres-ponding to emission from successively lower vibrational levels of NO;.It thenrises linearly to 6 x 1028 at 9400 A, the limit of observation.The activation energyof emission falls from a maximum near 4000A until 5800A and then remainsconstant : the bulk of the emission which lies below 5800 A thus comes from thelowest vibrational levels of NO;, the variation with wavelength being due to Frank-Condon factors and the relative shapes of the two potential energy surfaces. Forthe 0 3 +NO chemiluminescence which also arises 2 from the lowest levels of NO,*Thrush has suggested that the emission rises to a maximum at 11,000 A with anequal number of quanta emitted at longer wavelengths. If this distribution holdsfor the shock tube measurements, the intensity plot can be integrated to give a meanradiative lifetime of 2.6 x 10-7 sec, in fortuitously good agreement with Neubergerand Duncan’s value of 2.6 x 10-7 sec estimated from the absorption spectrum.Dr.B. A. Thrush (Cambridge University) said: Could the red emission observedby Dr. Harteck and Dr. Reeves near the surfaces when nitric oxide is added be due toelectronically excited NO2 formed in the reaction, NO+O3 = NO*, + 0 2 . Thisreaction is very rapid and we have recently shown that it yields electronicallyexcited molecules with comparatively high efficiency particularly at elevatedtemperatures.Prof. P. Harteck (Rensselaer Poly. Inst., Troy) said: In reply to the commentof Dr. Thrush, there was a decrease in the emission of the Herzberg bands withaddition of nitric oxide. From this we concluded that the excited oxygen molecule1 Neuberger and Duncan, J.Chem. Physics, 1954,22, 1693.2 Clyne, Thrush and Wayne, Trans. Farahy SOC., 1964, 60, 359.3 Levitt, Trans. Faraday SOC., 1962, 58, 1789; 1963, 59, 59FIG. 1GENERAL DISCUSSION 223was reacting with the nitric oxide before emission took place, resulting in an excitedNO2 molecule. This is not conclusive, however, and further study is necessary toverify the mechanism responsible for this red emission. From the energetic stand-point both are practically equal as can be easily seen.Dr. N. H. Sagert and Dr. B. A. Thrush (Cambridge University) said: In recentstudies with Dr. D. W. Setser of the reaction between oxygen atoms and cyanogen,lit has been established that added nitric oxide is excited in collisions with meta-stable A3Z:+, nitrogen molecules to levels v = 0 and l of the A2Z+ state (y-bands)and very slightly to level v = 0 of the B2II state.This has led us to suggest that anitrogen molecule in the A3Z: state might be the intermediary in the very efficientconversion by nitrogen of NO from the C2II state to the A2C.f state. In the associ-ation reactionthe very rapid 2 secondary steplimits the steady-state nitric oxide concentration to a very low valueN+O+M = NO+M (1)N+NO = NzfO (2)which would severely reduce the extent to which nitrogen could transfer the energyfrom a newly formed nitric oxide molecule in the C2n state to another in its groundstate. If such a process were important in the excitation of the NO y-bands instudies of the nitric oxide afterglow such as those made by Young and Sharpless 3in which oxygen atoms are produced by partially titrating active nitrogen withnitric oxide, the ratio of the intensity of the 8’ = 0 and 1 y-bands to the other bandsystems of nitric oxide should increase with the amount of nitric oxide added owingto the increased steady-state concentration of nitric oxide predicted by eqn.(3).No such dependence was detected by these workers.If, however, excess nitric oxide is added to active nitrogen, N, 0 and NO co-exist in significant concentrations in the mixing zone. Under these conditions weobserve a strong enhancement of the v‘ = 0 and 1 y-bands of NO. This is shownin spectrum (b) which may be compared with the excitation of added NO in theO+C2N2 reaction which appears in spectrum (a) together with CN emission atlonger wavelengths.By contrast, spectrum (c) shows the normal N+O afterglowobserved when less than the stoichiometric quantity of nitric oxide is added toactive nitrogen, where the NO /3-bands are much more intense than the y-bands.Experiments are in progress to determine whether metastable A3Zi nitrogen mole-cules in active nitrogen contribute significantly to the NO y-band emission underthese conditions.Dr. A. B. Callear and Dr. I. W. M. Smith (University of Cambridge) said: Prof.Gaydon has suggested that the rotational distribution of the 6 bands may show someevidence of heterogeneous predissociation. If the predissociating state is &A state,the fluorescence spectrum should show abnormally strong emission from stateswith low J values.It would, in fact, be very difficult to obtain sufficient intensitywith the required dispersion, by means of optical excitation. The most intensesource of 6 emission is from the afterglow of an electric discharge. If the pre-dissociation is heterogeneous, the pre-association mechanism would populate states1 Setser and Thrush, Nature, 1963,200, 864.2 Clyne and Thrush, Proc. Roy. Soc. A , 1961,261,259.3 Young and Sharpless, Disc. Faradzy SOC., 1962,33,228224 GENERAL DISCUSSIONwith large J and transitions from states with small J should be abnormally weak.This experiment is a practical possibility.It is stated in our paper that the simplest and most reasonable mechanism fordissociation of C2IT(u = 0), is via the a4n state. At low pressures, the D2Z+ statedecomposes less rapidly so that the E bands are stronger than the S bands in emission.We believe that the change of A in the D2Z++a4ll transition is sufficient restrictionto account for this phenomenon. The D2Z+ state may decompose at a rate ofabout 108 sec-1 compared to 6-6 x 108 sec-1 for C2n(v = 0).Dr.Thrush’s experiments appear to show that N2A3Z; will transfer energy tonitric oxide to produce the A2Z+ state. The quenching of NO C2II by N:! may in-volve N2 as an energy carrier (though the p bands were not observed) and we havedesigned experiments to test this possibility.Prof. W. L. Fite (University of Pittsburgh, Pennsylvania) said: In using noblegases as carriers in flow tube experiments on reaction rates, where discharges areinvolved and where oxygen is present, I would like to call attention to experimentson oxygen lasers performed by Bennett et aZ.1 In particular, it was found that anAr+O2 gas mixture is much more effective than a Kr+02 mixture in producingthe 8446 A line (33P-33S transition) of atomic oxygen. Since population of theupper state was believed to arise through electron-impact excitation of O(210)and/or 0(21S), the authors concluded that collisions of metastable argon atomswith oxygen molecules were very effective for producing metastable oxygen atomsthrough transfer of excitation followed by dissociation.If this interpretation is correct, it would appear that argon is perhaps not thebest choice for a carrier gas in any experiment where the presence of metastableoxygen atoms would render ambiguous the interpretation of results on reactionrates.Conversely, if some flow tube experiments could be performed using bothKr and Ar as carriers, then perhaps the comparison of reaction rate results couldshed further light on the operation of oxygen lasers.Dr. Peter Warneck (Geophysics Corp. of America, Bedford, Mass.) said: I wouldlike to ask Dr. Berkner what he thinks about the possibility that layers of substanceswith saciently high absorbing power in the ultra-violet spectral region may havedeveloped already in the primitive atmosphere owing to photochemical activity,thereby providing a shield against u.-v. radiation in a similar manner as the ozonelayer in the present oxygenic atmosphere. From the experiments performed byMiller we know, e.g., that aldehydes could have assumed this function. Theimplications upon the evolutionary development of life on dry land are evident.Prof. P. Harteck (Rensselaer Poly. Imt., Troy) said : We (R. Brown, S. Dondesand P. Harteck) have recently found that the number of atoms produced by ionizingradiation could be determined by the isotopic exchange in 14N14N + 15N15N mixturesto form 14N15N. The results of the exchange show a G value (number of atoms formedper 100 eV) for N atom formation of about 11.5 when the oxygen concentration isless than 0.015 %. The observed results are in good agreement with our theoreticalvalue, G(N) = 11.7 derived from consideration of ion recombination followed bydissociation, and excitation followed by dissociation. In experiments with oxygenpresent at 1 and 10 % concentration, the G(N) value is reduced to 6 and belowindicating that ion recombination to form nitrogen atoms does not occur becauseN ; transfers its charge to 0 2 .must be excluded. These results agree with our earlier nitrogen bation studies.Similarly it can be shown that the reaction :N l + 02+NO+ + NO1 Bennett, Faust, McFarlane and Patel, Physic. Rev. Letters, 1962, 8,470GENERAL DISCUSSION 225Ionizing radiation used in these experiments consisted of " pile " radiation and thekinetic energy of 235U fission fragments. Gas pressures were up to 20 atm.We (S. Dondes, P. Harteck and C. Kunz) have recently found the presence ofthe 5577A Auroral Green (01) line in the emission taken from the Po-210 alphairradiation of nitrogen gas at 1 atm containing 1/10,000 of oxygen. The 5577Aline disappears when the electric field is applied to remove the ions from the system.From all the possible reactions, the only reaction which fits the observed results isN++02-)NO++O(lS)+6.6 eV.These results are in full agreement with the findings of Volpi,l who has shown thisreaction to be very fast (1 x 10-10 cm3 sec-1). A full report of these experimentshas now been published.21 Volpi, J. Chem. Physics, 1963, 39, 518.2 2 . Naturforsch., 1964, 19a, 6

ISSN:0366-9033    

DOI:10.1039/DF9643700209

出版商:RSC    

年代:1964

数据来源: RSC

摘要:

GENERAL DISCUSSION 225AUTHOR INDEX *Ackermann, M., 209.Bader, L. W., 46.Baker, F. A., 213.Bates, D. R., 21.Becker, R. A,, 149.Berkner, L. V., 122.Bloomfield, C. H., 176.Cadle, R. D., 66, 218.Callear, A. B., 96, 215, 223.Ching, B. K., 149.Clyne, M. A. A., 214.Cook, G. R., 149.Dainton, F. S., 210, 213, 217.Dalgarno, A., 142.Doherty, G., 73.Farragher, A. L., 203.Fite, W. L., 192, 215, 224.Harteck, P., 82, 220, 222, 224.Hartley, D. B., 220.Hasted, J. B., 176, 210, 211, 213.Huffman, R. E., 159.Johnson, G. R. A., 87.Jonathan, N., 73, 221.Kaufmann, F., 26, 219.Kelso, J . R., 26.Larkin, P. S., 112.Larrabee, J. C., 159.Levitt, B. P., 222.Marshall, L. C., 122.Mathias, A., 38.McKenney, J. C., 214.Nicolet, M., 7.Ogryzlo, E. A., 46, 218.Page, F. M., 203.Popoff, I. G., 185.Pktermann, L. H., 213.Reeves, R. R., Jr., 82.Rutherford, J. A., 192.Sagert, N. H., 223.Sayers, J., 167.SchifF, H. I., 38, 216.Smith, D., 167, 21 1.Smith, I. W. M., 96, 223.Tanaka, Y., 159.Thrush, B. A., 112, 214, 220, 222, 223.Warman, J. M., 87.Warneck, P., 57, 217, 224.Wheeler, R. C., 203.Whitten, R. C., 185.Young, R. A., 118.* The references in heavy type indicate papers submitted for discussion

ISSN:0366-9033    

DOI:10.1039/DF9643700225

出版商:RSC    

年代:1964

数据来源: RSC