THE STUDY OF COUJSION-INDUCED PHENOWA WITH A DEMPSTER-TYPE MASS SPECTROMETER THESIS FOR THE DEGREE 0F PI. D. MICHIGAN STATE UNIVERSITY HAROLD H; HARRIS 1.967.- 0-169 This is to certify that the thesis entitled THE STUDY OF COLLISION-INDUCED PHENOMENA WITH A DEMPSTER-TYPE MASS SPECTROMETER presented by Harold H. Harris has been accepted towards fulfillment of the requirements for Pho Do degree in Chemistry widow Maj r professor Datelflflfly 2,767 hug—~— LIBRAR Y 1' Michigan State .C I ' ”t University —. '1'"- :w.’ ‘i . WMY!"1.~;\ .lfil. l.. r. in. II . 'D'. '4 ' .ouV. flu , ‘t' . {El I‘ll .t ‘ 5'0 ABSTRACT THE STUDY OF COLLISION-INDUCED PHENOMENA WITH A DEMPSTER~TYPE MASS SPECTROMETER by Harold H. Harris In spite of the large number of Dempster-type analyt- ical mass spectrometers in use, this instrument has not been used in the study of collision-induced phenomena. The reason for this is that there is only a small volume in the spectrometer in which reactions can occur and have the products appear in the mass spectrum. When the analyzer pressure is high enough so that a measurable num- ber of reactions occur, it is also high enough to attenu- ate the ion beams. Whenever such attenuation exceeds a few percent, the equations usually used to calculate the reaction cross sections are not valid. In the present work, new equations are developed for use in the measurement of both total collision cross sec- tions and reaction cross sections in a Dempster mass spec- trometer. Experimental studies are made of the following reactions, and their reaction cross sections are deter- mined at 1000 volts ion accelerating potential: The for (6) was the Harold H. Harris .2 Ar+2 + Ar —->- Ar+ + Ar+ 0': 3.6x10'l6 cm2 00* + 002 -—-> c“ + o + 002 o a 6.6x10‘17 cm2 N2+ + N2 —-> N+ + N + N2 0- 3.5x10‘17 cm2 N2O+ + N20 —->-No+ + N + N20 0 - 6.3xlo‘17 cm2 co+2 + co -—> c” + (0*) + (co+) as 1.4::10’17 cm2 value of Kuprianov and co-workers (1) of 1.0 x lO-r7cm? the reaction co++co——>c++o+co used for calibration. In addition, following systems were determined: 00" in co 16 OT . 2.leO- cm 0+ in CO CIT - 5.0xlO"17 cm2 0" in co GT . 1.8xlo'l6 cm2 Ar+ in Ar (yT - 2.2)(10-15 cm2 Ar+2 in Ar (IT - 1.2xlO-15 cm2 REFERENCE 8. Kuprianov, M. Tikhomirov, V. Potapov, and P. Karpova, Soviet Physics-JETP, fig, 569 (1956). the total collision cross sections for THE STUDY OF COLLISION-INDUCED PHENOMENA WITH A DEMPSTER-TYPE MASS SPECTROMETER BY .4 Harold Hf) Harris A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1967 To my Parents and Mary 11 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. Morley Russell for his guidance and encouragement through the performance of this research and the prepara- tion of this thesis. Even though he was at another univer- sity during the time most of this investigation was done, he was never too far away to provide helpful advice in times of difficulty. Many thanks are also due Mr. John Bartelt for helpful discussions of the theoretical portions of this work. The mass spectrometer used in the present work was purchased in part with funds from the National Science Foundation. 111 II. III. IV. VI. VII. VIII. IX. TABLE OF CONTENTS INTRODUCTION. HISTORICAL. O O O O O O O O C O O O O O THEORETICAL . . . . . . . . . . . . . . A. The Measurement of Total Collision Cross Sections . . . . . . B. Measurement of Reaction Cross Sections EXPERIMENTAL. . . . . . . . . . . . . . RESULTS A. Carbon Monoxide. B. Argon. C. Carbon Dioxide D. Nitrogen . E. Nitrous Oxide. CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK. . . . . APPENDIX 1 — Ionization Gauge Calibration Curves. . . APPENDIX 2 - On the Theoretical Estimation of the Effective Collision Volume Length. REFERENCES. iv 11 15 22 3O 30 Al 50 56 6O 65 69 75 80 Table LIST OF TABLES Mass Spectra of the Gases Theoretical IL for Several Reactions. Figure SUD“) IO ll l2 13 14 15 16 LIST OF FIGURES The Dempster Mass Spectrometer. Auxiliary Vacuum System . Cooled McLeod Gauge Carbon Monoxide. CO+ Intensity vs. Pressure. Carbon Monoxide. O+ Intensity vs. Pressure Carbon Monoxide. 0* Intensity vs. Pressure Carbon Monoxide. Peak 5.1a vs. Pressure. Carbon Monoxide. 15.14/112.O vs. Pressure. Carbon Monoxide. Peak 10.28 vs. Pressure Carbon Monoxide. 110.28/112.O vs. Pressure Ar on. Ar Intensity vs. Pressure Ar on. Ar Intensity vs. Pressure. Argon. Peak 80 vs. Pressure. Argon. IBO/IAO vs. Pressure. Carbon Dioxide. C Intensity vs. Pressure Carbon Dioxide. Peak 5.1“ vs. Pressure. vi £353 13 23 26 3A 35 36 37 38 39 40 A6 in 49 53 54 LIST OF FIGURES - Continued Figure l7 l8 19 2O 21 22 23 24 25 26 27 28 29 Carbon Dioxide. 15.14/112.O vs. Pressure. Nitrogen. N Intensity vs. Pressure Nitrogen. Peak 7.0 vs. Pressure Nitrogen. I7.O/Ilh.o vs. Pressure Nitrous Oxide. NO+ Intensity vs. Pressure. Nitrous Oxide. Peak 20.45 vs. Pressure Nitrous Oxide. 120.45/I3O.O vs. Pressure Carbon Monoxide. Pressure Gauge Calibration. Argon. Pressure Gauge Calibration. Carbon Dioxide Pressure Gauge Calibration. Nitrogen. Pressure Gauge Calibration. Nitrous Oxide. Pressure Gauge Calibration. Geometric Estimation of 11. vii 58 59 62 53 6A 70 71 72 73 7A 76 I. INTRODUCTION Despite the fact that the Dempster, or 1800 deflection, mass spectrometer is very common in analytical mass spec- trometry, it has not been applied to the study of collision- induced phenomena. There are two reasons for this. First, since both the ion source and the analyzer are inside the magnetic field, it is comparatively difficult to achieve differential pumping. Second, the effective collision volume is both small and unknown. Its small size means that, at pressures high enough for collisional phenomena to occur in measurable amounts, the ion beam loses a large fraction of its intensity due to collisions with molecules. Under these conditions, the approximations which are usually made in order to interpret the results are no longer valid. The purpose of this work is to show that meaningful interpretations of collisional phenomena occurring in a 1800 spectrometer may be made. Several collision-induced bands have been studied, and the results are comparable to measurements obtained by other means. Scrupulous attention has been given to the measurement of the pressure in the effective collision volume, and the calibration of the pressure gauge. Unfortunately, the literature of mass spectrometry is not consistent concerning the nomenclature used to describe reactions which occur inside a mass spectrometer. In this thesis, 1. the definitions of these terms will be as follows: secondary process: any process occurring in the spectrometer which requires a collision. collision-induced phenomenon: a process which re- quires a collision in the spectrometer analyzer tube. ion—molecule reaction: a process which requires a collision in the spectrometer ion source. Aston band: a diffuse peak caused by either a metastable transition or a collision-induced reaction. II. HISTORICAL The first mass spectrometer (1,2) was built by the Nobel laureate, F. W. Aston. Ions were produced by a gas- eous discharge and separated into a homogeneous energy beam by deflection in an electrostatic field. Then a magnetic analyzer was used to further differentiate them according to their mass-to-charge ratio (m/q). The resolved ion beams darkened the photographic film used for a detector. Almost all of the ions fell on narrow lines, corresponding to the mass-to-charge ratios of the stable isotopes of the gas in the discharge tube. When the vacuum in the instrument was allowed to deteriorate, diffuse bands were sometimes observed in the spectra (1,2). Although he did not study these bands in detail, Aston recognized their probable origin as being ions which change their mass-to-charge ratio during a collision with a molecule, after electrostatic deflection, but before magnetic analysis. Aston observed a band at m/q 5.2 in the spectrum of carbon monoxide, and at m/q 0.5 in the hydrogen spectrum. He showed that the particle giving the 0.5 peak was a pro- ton, arising from the dissociation of H2+ after it leaves the electric field (2). The explanation of the band at 3.: pC'Ss's ~‘Ll A, " adv‘b 'h- “N c No 1.. m/q 5.2 in CO was also given by Aston, as being the dissoci- ation of 00+ to give 0* and o. If the kinetic energy of the CO+ ion is partitioned between the fragments according to their masses, he reasoned that the resultant band should appear at m/q 5.14 since the C+ ion would carry 3/7 of the energy of a primary carbon ion, with mass 12. A subsequent observation of the band by Bainbridge and Jordan (3) in their double-focusing instrument gave an m/q of 5.1A51;0.002, in agreement with theory. Although there are several ways of achieving double-focusing, a typical double-focusing instrument first separates the ion beam electrostatically, according to velocity, and then mag— netically, according to mass. The increased resolution of the arrangement makes more accurate mass determinations possible, and collision-induced bands appear sharper, fur- ther facilitating accurate determination of the apparent mass of Aston bands. Mattauch and Lichtblau (4) studied about 30 bands with their double—focusing spectrometer. Another of the early experimenters was Smyth (5,6), who observed secondary processes with a mass spectrometer. He studied the H3+ ion in the high pressure spectrum of .hydrogen, and he also observed the bands produced by nitro- tgen. An ion of mass ml and charge ql accelerated through a potential V will possess a kinetic energy given by 1 (l) ‘2'le 3" qlv where v is the velocity of the ion. If the ion then dis- sociates to produce a new ion of mass m2 and charge Q2, the kinetic energy of the second ion will be (2) T = firnev2 == (n12/ml)qlV assuming that the kinetic energy lost or gained during the dissociation process is small compared to the kinetic energy possessed by the first ion. The radius of curvature of an ion in the magnetic field H is m “(a-t) (mt where c is the velocity of light, q is the charge of the ion, and m is its mass. Using equation (3), we find the radius of the second, or "daughter”, ion to be 2vq *3 mgt (4) “‘75 In“ H q2 1 This is the same radius as that traversed by a primary ion of mass-to-charge ratio (m/q)*. where (5) (m/Y‘ {—7—7—(m2/q2)? q 3 "‘1 ql This equation was first derived by Dempster (7) and is widely used in the prediction of the positions of Aston bands. Although it is very useful, equation (5) does not determine the origin of an observed peak, since there is § only one observable, (m/Q) . Hipple and co-workers (8,9,10) were the first to recognize that the diffuse peaks in the mass spectra of hydrocarbons were due to ions which spontaneously dissoci- ate between the electrostatic field and the magnetic analyzer. Hipple also showed that in the Dempster, or 1800 deflection, spectrometer (the type used in this study) the bands produced by these "metastable“ ions occur at the same position they do in sector-field instruments (10). In addi- tion, Hipple’s spectrometer (9,10) was equipped with a means of measuring the energy of the ions causing a metastable peak. This was accomplished by varying the potential of the analyzer with respect to the final slit. The fraction of the original accelerating energy possessed by the frag— ment ions was obtained by varying the analyzer potential until the metastable peak disappeared. Then, from equation (2), m2/q2 V2 ‘6) 6751‘s where ml and m2 are the initial and fragment masses, q1 and q2 are their charges, respectively, V1 is the accelerating potential, and V2 is the retarding potential. This system provides a second equation for Aston bands, and makes the determination of their origins unambiguous. Variations of Hipple's experimental apparatus have been widely used in the study of both metastable ions and collision-induced Phenomena. Henglein and Ewald used a "parabolic" mass spectrometer (invented by Eisenhut and Schutze (11.12)) to study colli- sional dissociation of both positive and negative ions from a number of gases, including ethylene, carbon dioxide, hydrogen, methane, water, propane, nitrogen, and several isotopically labeled analogs of these substances (l3,lh). Their apparatus consisted of an ion source where the ions were formed by electron impact and accelerated electrosta- tically, a collision chamber which was maintained at high pressure, an analyzer which consisted of colinear magnetic and electrostatic fields, and a photographic plate detector. The ions in the beam leaving the collision chamber were de- flected by the magnetic field, in a direction perpendicular to their direction of motion, and perpendicular to the field. The magnitude of the deflection was dependent on the mass of the ions, as in the usual magnetic analysis spec- trometer. The electrostatic field deflected the ions per— pendicular to their direction of motion, but parallel to the field, and separated them according to their energy. The result was a set of spots on the film, the vertical position giving the relative energies of the product ions producing the darkening, and the horizontal position their mass-to- charge ratio. Probably the most detailed studies of Aston bands have been carried out in the laboratory of C. E. Melton, with a sector-type magnetic spectrometer specially designed for the purpose (15). The source and the analyzer are completely separated, except for the narrow slit which allows the ion beam to enter the analyzer. A separate vacuum system pumps each side of the slit. This apparatus allows the pressure in the analyzer to be varied by two orders of magnitude without seriously changing the source conditions. Because of this almost complete separation of source and analyzer, Melton was able to study the collision of ions of one gas on the molecules of another. He measured the reaction cross sections for the dissociation of carbon monoxide molecular ions in collision with a variety of simple gases (16), as a function of electron energy and as a function of ion kinetic energy. These results showed that the singly charged molecular ion converts its kinetic energy into vibrational energy when dissociating, while the doubly charged molecular ion converts its electronic energy into vibrational energy. Melton also reported a number of collision-induced dissoci— ations in the mass spectra of acetylene (l7), methane (18), n-butane and i-butane (19), and formic and deuteroformic acid (20). Larkin Kerwin and co-workers have used a sector-type magnetic spectrometer (21) to study phosphorus (22) and oxygen (23). In addition, they have published a review of techniques useful in the study of collision processes in a mass spectrometer (2A). There have been several modern reinvestigations of bands studied qualitatively by the early mass spectrome- trists, including the band at 0.5 in hydrogen (25,26), the bands in n-butane (27), and in methane (28). An interesting unresolved controversy appears in the literature concerning the mass spectrum of nitrous oxide. Begun and Landau (29,30) have evidence for some metastable contribution to the band at 20.45 mass units, corresponding to the transition: (7) N2O++M—->NO++O+M Friedman and Irsa (31) have also examined the band, and found no evidence for any metastable contribution. The band was also observed in the present work, and it will be discussed again later. A photoionization source provides reactant ions with a well—defined energy. Experiments performed with such an ion source can therefore yield more detailed information about the excited states of the reactant ion than the usual elec- tron impact source instrument. Excited states may be of importance to collision reactions. Some experiments using such a source have been done by Chupka and Refaey (32). Perhaps the most powerful tool available for the study of ion collision phenomena is the tandem mass spectrometer. This device consists of two spectrometers, the mass-analyzed ion beam of the first being used to bombard molecules and lO produce secondary ions in the source of the second spec- trometer. Such a system has been used by Lindholm (33-35), Fedorenko (36), Rourke and co-workers (37), and Abbe and Adloff (38-u0), among others. Since collision bands may occur whenever the instrument pressure is high, they often appear in field-emission spec- trometers (Al). The same phenomena which give rise to Aston bands in mass spectrometry may be studied in other ways, which do not involve mass analysis (42-46). A review of these techniques has recently been published by Hasted (A7). Collision-induced phenomena also occur in time-of— flight spectrometers, and some new instrumentation has been devised for their detection (48-50). The detection of neutral fragments is a particularly important advantage of the time-of-flight analysis method. III. THEORETICAL A. Measurement of Total Collision Cross Sections Consider a parallel beam of positive ions of homogene- ous velocity passing through a hypothetical gas of solid spherical molecules. If an ion passes within a distance r of a molecule, we shall consider a collision to have oc- curred. The cross-sectional area (IT presented by a mole- cule is therefore (8) CT = dIr If there are N molecules of the gas per cubic centimeter, the probability that an ion will collide while moving a small distance cfx cm through the gas will be N0idx. If we presume that any such collision removes an ion from the beam (or if we define a "collision" as an interaction that removes an ion from the beam), the amount of beam intensity lost in traversing a distance cfx, after starting with an intensity I will be: 0 (9) (TI - NOTIO x 0n integration, we find the intensity after moving a dis- tance L through the gas to be (51): 11 I2 (10) I .. Ioexp(-No;I.L) where (IT is called the total cross section for collision of the ion in the beam with molecules of the gas. Applying equation (10) to a Dempster-type mass spec- trometer, shown in Figure l, the ion current measured at the collector is given by (11) I - lidimexn-9,N,oTl)exp(~i2NQoT2)exp(-93~3o-T3> where i is the electron beam intensity,,Q is the length of the electron beam (cm), Q1 is the cross section for ioniza— tion (cm2), N is the source pressure (molecules/cm3), 7) l is a term that includes the ion collection efficiency and also takes into account ions which collide with the slit edges,fl1 is the distance (cm) from the electron beam to the first slit, is the average total collision cross section “I: (cme) for ions in the source, 22 is the distance (cm) be- tween the accelerating electrodes, N2 is the pressure (molecules/cm3) between the accelerating electrodes,o':T2 is the average total collision cross section (cmz) for ions be- tween the accelerating electrodes, ,€3 is the distance (cm) from the second accelerating electrode to the collector, N3 is the pressure (molecules/cm3) in the analyzer, and<3i3 is the total collision cross section (cmz) for ions in the analyzer. The total collision cross sections,<3&1, 0E2, and<3' are not identical because they are dependent on T3 l3 to purrcs :ri 7-xli1'. l your. 5,5'9‘ Entrance Slit Jon Porn .‘teflglh '1) Collector Exit Slut Figure No). The Dempster Mass Spectrometer. Electron Accelerating Beam Electrodes Repeller Electrodes l ' ‘ — Gos lnlet l Focuing Electrodes 1' I‘d thure I(bl. Detail of the Ion Source. energy various of the the use Where L 10h kine traveled (13) where T 0 LV ‘ it!“ n «e H. r4 A a. i 5 (h v Q 4 (D/ , T r ‘1 L (_ 14 energy (46), and the ions have different energies in the various portions of the spectrometer. The energy dependence of the total collision cross sections is also the reason for the use of the adjective "average" in the definitions of Oil andCTTq, since the ions are undergoing acceleration in the source: (7% is better described as ave ( > 1 L 12 (j z -¢;[ rave L o UT(T)d)? where L is the distance the ions are accelerated, T is the ion kinetic energy, and ,6 is the distance the ion has traveled from the first electrode. T is given by (13) T . T + qu 0 where T0 is the kinetic energy possessed by the ion as it enters the accelerating area, B is the field between the electrodes, and q is the charge on the ion. Since )(1 +92 is much smaller than93, equation (11) may be approximated, to good accuracy, as (14) I = iQQiN Nf?exp(- (BN 3 GT3 If the pressure in the spectrometer is uniform, (15) N = N Under these conditions, the maximum intensity occurs when (16) §fi=o=1Qq1Nq(..393qT )exp( m? 301.3) + ilqmexp(-NP30T3) 15 Solving (16) gives 1 (17) OT ”W- 3 3 This provides a convenient determination of the total collision cross section of an ion with its parent gas. Only the analyzer dimension 93 and the instrument pressure is required, with no standardization with an external system necessary. B. Measurement of Reaction Cross Sections The measured ion intensity for ions in an Aston band caused by the reaction (18) A+ + M«——>--B+ + c + M may be derived in much the same way we found the intensity of a primary ion, in equation (11). Consider as before a parallel beam of positive ions of homogeneous velocity passing through a gas. In the derivation of equation (11), we considered collisions only as processes which remove ions from the original beam. In fact, there may be several processes which can "destroy“ the ions. For instance, they may dissociate, they may exchange charge with the gas molecules, they may abstract a portion of the gas molecules, or they may simply be scattered out of the beam. Each of these processes has a characteristic cross section 01. 16 Clearly, (19) OT =i<71 13‘ The number of particles of the product of reaction 1 pro- duced will be given by (20) I = Io[l - exp(-,QO‘1N)] i where I is the intensity of the beam of product ions, IO 1 is the reactant ion beam intensity,.Q is the length of the reaction chamber,(ji is the cross section for reaction, and N is the pressure. Coggeshall (52) has shown that reac- tions that contribute to an Aston band in a Dempster-type mass spectrometer must occur in a small volume near the entrance to the analyzer (see also Appendix 2). Using Coggeshall's result and equations (10) and (20), we may write down the intensity of the Aston band caused by reaction (18) occurring in a 1800 deflection instrument as (21) I =ng N exp(-RNo )eXP(-)(N0’ ) AB 1A 173 1 1 TIA 2 2 T2A X [l "’ exP(-94NuO-AB)]exP(-23N3OTBB) where I is the intensity of the band, and the other sym— AB bols are as in equation (11), with the subscripts A and B indicating that the constant in question is for the reac- tant ion A+ or the product ion B+, respectively. .24 is 17 the effective length of the reaction chamber, and CAB is the reaction cross section (cm2). Normally, the product ion B+ may be formed as a pri- mary ion in the electron beam as well as by the collision— induced reaction. The ratio of the intensities measured for the Aston band and for the primary ion which is the product of the reaction is I (22) 513—- A ‘11 exp(‘fiNio'r )exp('92N2-C+ + 0+ + Ne This result is very confusing when viewed in light of the results of Melton and Wells. Melton and Wells studied the collision-induced dissociation of 00+2 into 0* and 0, with a series of molecular targets, and showed the cross section for neon to be over one thousand times smaller than that for CO. On this basis, we would have to conclude that Kuprianov would measure a cross section in excess of 2 x 10-13 cm2 for reaction (32)! 34 31 EEIVO. x ouammmua c m o - O ..l 'A‘o l .‘.“" ..ID! .I‘ ‘ 0mm VA'A‘ O z_0l x 0 31 ’O O Q (l U ‘0 . O (0.0) . 0 9381a .m> 35:25 +00 3:852 5280 e .. v 2:2... I. i .2... 35 A: I N. 0. SI EEvco. x 233E m m h 0 0 ea 233$ 2, 2.2.2:. +0 apiece: c0960 0 8:9... 0 o ee 8 36 BI EEV v0: scammed o. e._ m. o._ m o o N o . re as. 32:3 0 BE. moo 4 .N_.Hl Z r X Samoan. m> 2.22:. +0 0.. 2585.2 5260 .w. l 0 ~59“. T LON 6N f) > D) D CC CC C 78 5m < 37 BI EEV ¢o_x Sammoi w. v. N. o. m m c N P P p P F b b p .25 32:34 O .25 moo Q 00+ 0+ +0 Aloo++oo Summea .m> Sm xoma 02.882 8250 x. «Sod 38 n I“ '- Q 5 W » K9 3 N. 3.5 remind 0 ~35 moO Q Q \\ 0 0. SI EEV wo. x2335 m w v N b b 8335 m>o.m_~\¢_w~ 00+O++o§ou++oo 02.852 5900 m 839“. «3 c'u T (T 39 3‘: EE. V0.1 233i o a. m. e m w a .o 4 \ vN 4 fl 4 1? I 4 m .6 4 . 8 1 x 4 ‘ rm m an 4 v 4 '0 233$ .m> mmd. .68 d. +oo+o++otuoomfoo . 02.652 c0960 '0. 0 0.59m a 140 ‘ . ., . A . ...: .1.‘ ‘. .‘fi‘utr V. ... ‘a‘ 14.11 1.. ..‘uu-‘v L 3... .55 vo.x23uoi m._ m. N._ o._ .m m w N 4 4 4 4 4 Q 4 1 1\ 4 ‘ Summoi .2, o.~_~\ mud: +09. o++o 1100++Tou 4 3:652 593 1 o. 2.6: fi ' :3 I/QZ'OII 9 r t,opv. 2| “1 B. Argon The Aston peak technique can be of use in the measure- ment of charge transfer cross sections, but it is limited to reactions in which the ions accelerated in the source retain at least some or their charge. For instance, the following reactions could theoretically produce Aston bands: (3“) Ar+2 + Ar ——>- Ar+ + Ar+ (35) Ar+3 + Ar -—> Ar'.'2 + Ar+ (36) Ar+3 + Ar ——+~ Ar+ + Ar+2 +2 (37) Ar + Ar -+> Ar+3 + Ar- Reactions which neutralize the bombarding ion are not de- tected because the neutral fragments are not collected. A search was made for evidence of reactions (35) and (36), but neither of the corresponding bands was detected, presumably because of the very small cross section for triple ioniza- tion of argon (65). Reaction (37) has a very low cross sec- tion at low energies and is not observed in mass spectrom- eters (66). The band produced by reaction (3“) appears at m/q 80, and there is the possibility of interference with the ion AP2+, which may be produced by collision of an argon ion with a molecule in the source (67). However, the band dis- aDpears when the electron energy (uncalibrated) is lowered below about 50 volts, which is somewhat higher than the ap— pearance potential or Ar+2 (u3.38 e.v.) but much higher than .r .t-‘N ‘rjxr.v.--x N2 the appearance potential of Ar+ (15.76 e.v.) (68). An exact appearance potential measurement was not feasible because of the low intensity of the Aston band. The band intensity is also observed to decrease as the electron energy is in- creased above 130 volts, as does the intensity of the peak at m/q 20, due to Ar+2. The present experimental section is the only one not performed at 70 volts electron energy. It was done at 130 volts in order to maximize the band intensity. Figures 11 and 12 yield the total collision cross sec- tion for the doubly and singly charged argon cations. They are 1.22 x 10'15 and 2.20 x 10'15 cm2, respectively. Cramer (69) has measured the total collision cross section for Ar+ on Ar to be “.0 x 10'15 cm2 at an ion energy of 400 e.v. He found it to consist of a charge transfer cross section of 2.0 x 10'15 cm2 and an elastic scattering cross section of 2.0 x 10'15 cm2. Although this value is within a factor of approximately three of the results of the present work, even this agreement is largely fortuitous. The reason for this is the difference in experimental methods. Cramer's results were obtained using a total charge collection technique which collected nearly all of the "slow" ions produced when a beam of argon ions collided with a sample of argon gas. By contrast, the Aston peak technique, in measuring the disappearance of charge from the original beam, excludes not only ions that are scat- tered a very small angle out of their normal flight path, ..- ..-.-. A.“ ‘- ‘ A V. ....q. ._ 43 but also those which have lost a small fraction of their momentum. One would expect the results from two such dif— ferent methods to differ greatly, but a similar effect can be operative even when the apparatuses in use are very similar. A relatively small difference in experimental parameters could cause a large change in the "resolution" 5"" of the equipment. By "resolution” is meant the ability of the apparatus to detect a change in the velocity or momentum of an ion and register it, therefore, as "scattered." r Because of the differences in the resolution of various 3 j instruments, it is a common practice for each investigator to normalize his results to a commonly accepted value for the cross section of an easily measured reaction. The lack of this type of normalization in the comparison of the values in the present work to values in the literature will cause the differences between them to sometimes be very large. In order to avoid some of these difficulties, at least in the literature of the Aston band technique, it is here proposed that subsequent results be reported relative to the cross section for the collision-induced dissociation reaction (31) of carbon monoxide. The pressure dependence of the m/q 80 band is displayed in Figure 13. It exhibits the second-order behavior at low pressure which is common to all of the collision bands in the present work. Also of interest is the coincidence of the lines formed by points measured with the two inlet uu sywstems. This is further evidence that the spectrometer 113 near pressure equilibrium even when the gas is leaked flJTtO the source. The ratio of the Aston band at m/q 80 to the peak at mph; 40 is plotted in Figure 1A. The slope of the straight Itine, which is independent of the position of the gas inlet, yixelds a value of 3.56 x 10'16 cm2 for the cross section of reaction (3“). Since all of the experiments in the present vuark were performed at 1000 volts accelerating potential, true kinetic energy of the initial (doubly charged) ion is 2000 e.v. Using the Aston peak technique in a magnetic sector xnass spectrometer, Weiner, Hertel, and Koski (70) measured the cross section for the partial charge transfer from Ar++ to its parent gas to be 0.6 x 10'16 cm? at 2000 e.v. rising to 2.0 x 10'16 cm2 at 6000 e.v. Hasted and co—workers (71) also have measured this cross section, employing a collision chamber technique and collecting the slow ions produced when the doubly charged ions pass through the argon gas sample. However, this method does not discriminate between reaction (34) and the reaction (38) Ar+2 + Ar -e>Ar +’Ar+2 Since (38) is a resonant process, it is likely that it has a high cross section at low energy, and this may account for the high value of 2.& x 10‘15 t 10% cm2. A later “5 urement in the same laboratory (“5) was reported, and his case the results were separated into the contribu- LB from reactions (31%) and (38), and those were found -16 and 6.6 x 10'16 cm2 1 15%, respectively. be 2.0 x 10 a sum of the cross section (38) and half of (314) are 1 excellent agreement with the measurement of Flaks and olov‘ev (72) of 7.8 x 10':L6 cm2. (The reason for summing the cross sections in this way is that the slow ion collec- tion method measures the total cross section for charge production in the target gas, and this amounts to the cross section for production of dipositive ions plus half of the cross section for production of simple ions.) It is felt that the later paper by Hasted and co-workers (NS) and the result of Flaks and Solov'ev (72) represent the best ex- perimental values in the literature for the partial charge transfer reaction, and although there are a number of other papers on charge transfer reactions in argon, these will not be mentioned here. Citations of these papers may be found in references (#5) and (72). 46 3: 58:07 05305 o_ 3 N. o. m m e N 9 m... e... w... 2 x. ..h 3 ~ 9. 9532a .2. 3.2.2:. #4 3 .803 9 ._ 959E ....“ ..w 9 G , 9 o ; ,9 «a. o I 0 Q. o as .u 1 we 1" 0 n\\ u? 3: E Evvo. x 05305 e N. O. m w. .25 32.83 O .2... So AN 238.... .2, 5.22:. *4 .394 N. 950E m b )- ‘38 fl... fwd «sub 4 P4 b 148 3... E E. so. x 838i m m s. ._ . .25 egzxi O «0:: wow q 2332.". .9 cm goon. +3 ...a #2 ea .593 n_ 9.32.... .Nn 3x EEV¢o.xoSmmoi ‘49 .2... 32.34 O .2... 25 AN t? 0.532.; u> 0: \ Cow +.<+.+..< In ..4 4&4 .394 952... 'N 50 C. Carbon Dioxide The CO+ and 00+2 ions produced by the ionization and fragmentation of carbon dioxide should undergo the same collision—induced reactions as the same ions produced by the ionization of carbon monoxide. Therefore, one might expect all of the Aston bands which appear in the high pressure mass spectrum of C0 to also be found in the spectrum of 002. However, an examination of the analytical spectra of these gases (Table 1) shows why some of the bands are not present. The intensity of an Aston band is proportional to the abundance of the reactant ion in the source. The m/q 28 peak in carbon dioxide is only about 3% of the total ionization of the molecule (the sum of the ion intensities) whereas the same ion accounts for about 90% of the ionization in carbon monoxide. At a given pres— sure, then, the Aston peaks resulting from the dissociation of CO+ ions would be about 30 times less intense in CO2 than in CO. The m/q 14 peak, due to C0++ ions is absent from the spectrum of 002. Such an ion would have to be produced by the dissociation of the ion C02++. Since the two positive charges could be separated by the formation of two charged species, the doubly charged ion is not produced, at least not in measurable amounts. As a result, the Aston band at m/q 10.28 which was studied in carbon monoxide, was not visible in CO2. 51 The m/q 5.114 peak, due to the reaction (39) 00" + 002——> c+ + 0 + co 2 was not large enough for study with the auxiliary inlet, but the use of the gas inlet increased the intensity enough so that a quantitative measurement was possible. However, introduction of gas directly into the source causes the source pressure to be higher than the analyzer pressure, and therefore one of the conditions (equation 15) used for the derivation of the expression (equation 17) for the measurement of total collision cross sections was violated The development of the theory for the measurement of reac- tion cross sections does not require that the entire mass spectrometer be at one pressure, but only that the absorp- tion of the ion beams in the source be small. Using the same equations as previously, Figure 17 yields 6.57 X 10-17 2 cm for the cross section of reaction (39). It will be noticed that there is a non-zero intercept in Figure 17, indicating a metastable contribution to the peak. This was the only such behavior observed in the bands studied, and it is believed to be an artifact, arising from a systematic error in the calibration of the pressure gauge of approximately 4 x 10"5 mm Hg. The reason for the neces- sity of this supposition is the nature of the reactant ion, 00+. Being a simple diatomic ion, it would seem that there would be no energy levels capable of storing the dissociation 52 energy for as long as 10"6 seconds, the approximate time that the ion spends in the source. Although there are no previous measurements in the literature for the cross section of reaction (39), one might well be surprised to find that this cross section is more than six times larger than that for the same reac- tion when carbon monoxide is the target gas. Certainly the difference in molecular sizes cannot account for this large difference in cross section. It must be remembered, however, that the reactant ions in the two cases are not identical. One of them was produced by the direct electron impact ionization of carbon monoxide while the other was the result of the unimolecular dissociation of 002+. Potapov (73) has shown that such differences in the preparation of the reactant ions may affect the cross section for their collision-induced dissociation. He studied the reaction (to) 00+ + Ne -—+ 0* + 0 + Ne and found the cross section to be 1.75 times larger when the initial ion was produced by dissociation of CO + than 2 when it was produced by the ionization of CO. . ... ... ... r ... . . .... (En-.1133. raisin unlitlehk. ADI EEVVO_ x Summoun. o. ... N._ o. m w e N o 53 h P t p l .0 2:88.... .m> 5.22:. +0 8.55 895 .9. m. 2.6... 514 3: EE. sO. 08:53:. .0. c. N. O. m w v N b p p h b b p scammed .m> Eh xcwa Noo+ odb +Noo..oo ooioé cones”e w. seamed .0. VN_ S 55 3: EEvcoioeauaoea mm. m N. O. m w ¢ N O «V e «V q .o. 4 . 4M «.1 6N I N. . .9 Q r U ‘ .../... .fl .Onfi Q 0 «u 9 4 2:32... .2, o.N._\ in. N00... O++01|NOU++OU .On 02.6.0 .8960 Q N. 2:9... 56 D. Nitrogen The only Aston band which was detectable in the nitro— gen system arose at m/Q 7.0 from the reaction + + (141) N2 +N2——>N +N+N2 Since this Aston band coincides with the possible ion N+2, there was some concern that part of the measured band in- tensity could be contributed by the doubly charged atomic ion. Such a contribution would be first order in pressure, and would appear as a non-zero intercept in the graph of the ratio of the reaction peak to the product ion peak (Figure 20). Absence of such a contribution indicates 2 that the N... ion is not interfering with these measure- ments. The slope of the straight line in Figure 20 yields a cross section for reaction (41) of 3.52 x 10'17 cm2. Although there have been numerous studies of charge-transfer reactions in nitrogen (MO), (70), (7“), there apparently has been no previous study of the collision-induced dis- sociation of N2+. 57 3: EE. vo.xo=.mm2n. S N. o. m o v N h b P p p p 05¢.de .m> 332:. ...z :3on m. 2:2... 58 w. O. b 3: 56:0...2335 m w ¢ - h b 233i m> ON .301 N2+ z++z ... Nz+mz 509:2 m. 2:0: PN 59 «1 «V 31 66:0... 233$ o. m w P D r h-Q’ rN <4 2:82.". .2, oiuxowfi N2 + z++z ... Nz+mz 5352 om 23: «1 60 E. Nitrous Oxide The nitrous oxide system had been studied a number of times before the present work was begun, but only in a qualitative manner. In these earlier studies, the pressure in the collision volume was not measured directly, nor was the pressure gauge calibrated. Metastable contributions were reported by Begun and Landau (29), (30), with a spec- trometer without repeller electrodes, but were denied by Friedman and Irsa (31). No cross sections were reported. Although metastable transitions are not common in triatomic molecules, they have been reported in carbon dioxide (75) and in H28 (76). A recent paper by Newton and Sciamanna (77) confirms the existence of the metastable transition (42) N20+-——>-NO+ + N and measures the (unimolecular) half-life to be less than 0.2 x 10'6 sec. Since the residence time in the source for an ion of m/q 44 is about h.6 x 10-7 sec at 15 volts repelling potential in our spectrometer, one might expect to observe some metastable transitions. However, a single rate constant does not indicate what portion of the ions will decompose at this rate. Newton and Sciamanna also neglected to report the electron energy used in their study. They could not measure cross sections for the 1.4.5..__ n.- ---q.;——_ - 61 collision-induced reaction because the pressure in their analyzer was not known. In the present work, no metastable contribution to the peak at 20.u5 was observed (Figure 23), but this should not be considered a refutation of the work of Newton and Sciamanna, or of Begun and Landau, since their experimental conditions may have been much different from those in the present work. Since the nitrous oxide system was studied with the gas inlet exclusively, no measurements of total collision cross sections are possible, although the intensity of the + ion as a function of pressure is presented in Figure 16 NO 21. The slope of Figure 23 yields a value of 1.24 x 10- cm2 for the reaction (143) N20" + N20 -—-+ NO+ + N + N20 62 m,_ 3.... 6.5 «.0. x 05395 E m. o. m m L r 2335 .2. 3.32:. +02 326 32:2 .N 959... 5N 3: EEVvaoEmmoi 63 o. m w v N o . . - . r to 4M G .N 4‘ t \ IV 1\ . 7a.. .0 .V 4 and 4 .m 4 T 4 o. 838i .2 9.0m ..oom . o~z+oéz io~z++o~z é .oExo 32:2 um 2:2“; . i 614 o. x8335 v w P LV 0538.”. m> o.omH\mvo~H o~z+ o.+oz f o~z+ +o~z 09x0 39:2 mm 050.... Jr (b VI. CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK The primary advantage of the technique developed in this thesis is that quantitative measurements of the cross sections of collision-induced phenomena may be made without a specially built mass spectrometer. The required modifi- cations to an analytical instrument are slight and do not preclude analytical use of the instrument. The spectrometer used in the experimental part of the present work was con~ currently used for the analytical mass spectrometer require- ments of the rest of the chemistry department. Such prac- tical considerations, while not of scientific moment, often contribute strongly to the interest in a given field. It is hoped that many scientists with access to a Dempster—type analytical spectrometer will be encouraged to undertake measurements of this kind, knowing that such a project need not be expensive or elaborate. A possibly useful bonus to the present method was dis- covered as the mass spectral data were being reduced. If the Aston band intensity is divided by the intensity of an the resultant ratio will This ion other than the product ion, not, in general, be a linear function of pressure. behavior could be useful in the identification of the 65 66 _ origin of an Aston band. If one happened to choose a “wrong“ ion that had the same total collision cross section as the product ion, the result would, however, be a straight line. An important disadvantage to the technique of studying Aston bands in a Dempster spectrometer is that, because the source and the analyzer are both in the magnetic field, it is difficult to pump them separately. This means that one cannot study the collision of the ions of one gas on the molecules of another. It has occurred to the author that this difficulty could be circumvented by measuring the bands in a series of mixtures of the two gases and extrapo lating to the pure gas. Such a procedure would be very time consuming. The experimental cross sections reported in the pres- ent work should not be considered as significant contribu- tions to the literature of collision-induced dissociations, but only as examples of the reactions which may easily be studied using the method developed in the theoretical section. Since the technique of studying collision-induced Aston bands in a Dempster spectrometer is established, it is hoped that these bands will be thoroughly reinvestigated, as a function of both electron energy and ion energy, in order to bring to light more of the details of the dissoci- ation process. It is felt that the collision-induced band in nitrogen would be particularly fertile, since it has not 67 previously been observed and nothing is known about the mechanism of its formation. An explanation should be found for the unusually low value for its cross section. Additional theoretical work should be carried out to better define the length of the effective collision volume, both for collision-induced processes and for metastable transitions. The rudimentary calculations carried out in Appendix 2 are obviously inadequate, but it is expected that the recent work of Coggeshall (52) and Newton (78) will be extended to give a realistic estimate of the dis- tance an ion can travel into the analyzer before dis- sociating and still be collected. It is anticipated that such a realistic estimate will require inclusion of the effect of the kinetic energy released on dissociation. In the absence of a complete theoretical treatment of the size and shape of the collision volume, the dependence of a?“ on the radii of parent and daughter ions could be determined by a careful comparison of data obtained with a Dempster instrument with the results of measurements from a sector-field spectrometer having a well-defined collision chamber. Newton (78) has shown that peak shapes of metastable transitions can only be explained by assuming that kinetic energy is released in the reaction. Presumably, these kinetic energy effects are also present in collision- induced dissociations, and, in addition, there is the 68 probability that kinetic energy is transferred to the target ion. Since the mass spectrometer is essentially a momentum selector, these kinetic energy effects tend to widen the ion beam, and some ions are therefore lost by collision with slits. Of course, the number of ions lost is dependent on the size of the kinetic energy effects and the selectivity of the spectrometer. For this reason, very different cross section measurements may be obtained with different instruments. At the present time, neither the kinetic energy released by the reaction, nor that trans- ferred to the target ion, is known,so that the effects may not be taken into account when absolute cross sections are compared. Fortunately, absolute cross sections are seldom of importance to the chemist. He is usually more interested in how the cross section varies with ion accelerating energy and with electron energy than in its absolute value. Therefore, although the Dempster spectrometer method shares the problem of the unknown kinetic energy effect with the sector-field method, it should not be a bar to its use. VII. APPENDIX 1 The following pages present the results of the calibra- tion experiments as described in the Experimental section. The calibrations of the ion gauge for argon, carbon monoxide, and nitrogen were performed with the cold trap cooled with liquid nitrogen. Since carbon dioxide and nitrous oxide condense at liquid nitrogen temperature, the trap was cooled in an acetone-dry ice bath for those runs. 59 7O 3122.0. A... . $2532.. 3.... o! o a .w ’ b 5.65:5 0250 2335 02.832 8900 cm saga Pu (6H mmbomamuud canoe) 00' '3‘ N “mm 71 no: EEvaxoSmmoi to: o: o. m w v r h b > 5:05:50 3:00 05¢.de .803 8 2.5a 9.: 60 (r H W unbOIXOH‘SSBJd abnog uol i b ( 72 N. a: 9 55:0... 9532.. to... 32 o o v L b 3:05:00 330 9.335 8:85 3960 mm 83: N co ‘2 (OH ww)vo|xamssmd obnog uol $3 73 3: Sacco... 233E ca... 22 w. v. N. O. o 0 t» ’ ’ bN 8:05:00 @950 23de $0222 NN as”: m (6H ww)bO|XOJFSSOJd abnog uol fN. rw. 714 A :II'IV .... u .. . _ — n n”... . . . .. . . .I .. in. . . - .. , c |. . A . [.11 units ..‘v .71.: £L.h.§'. (... 3: servo... 95:9... no: a! o. c. T o. o o v u o ''''''' I! P D P P D P b I o J. M U 9 D n a.» .d .... n n ... ..x. fi~_mu:b w m H m... em. 8.85:3 3.50 sawed m 02.6 32:2 mm 2.6: row im VIII. APPENDIX 2 Coggeshall (52) has analyzed the path of a metastable ion in a Dempster mass spectrometer, for the purpose of explaining the peak shapes observed for metastable transi- tions. Although his analysis did not include kinetic energy effects, and he therefore did not obtain realistic peak shapes, his approach may lead to a first approximation to the length 94 of the effective collision volume for collision-induced reactions. Figure 29 pictures the path of an ion in the analyzer of a Dempster spectrometer. The ion enters through a slit of width 26: at the top of the figure. After travelling a distance in the analyzer, the ion may collide with a molecule and dissociate. When the ion dissociates, its radius changes from the initial value, R1, to the radius R2 characteristic of the product ion. The center of the cir- cle described by the initial ion is constrained to the line between the entrance and exit slits by the fact that the ions enter perpendicular to this line. It may be seen that the ion path which leads to the maximum distance of travel before dissociation is one-in which the ions pass very near the inside edges of the entrance and exit slits. This is the path illustrated in Figure 29. 75 76 T]: 2‘ 2R2 21L; 2*, 7 . ’— a Rl-ZPHJS FIgUfe 29. Geometric estimation of l4. 77 The distance Q4 may be obtained from a simple trigo- nometric argument. Consider the triangle formed by the center of the circle described by the initial ion, the exit slit, and the point where the ion dissociates. Applying the law of cosines to this triangle, the angle CK is found to be ( )Q -l 2125- 4R1R2 +2Rle +2Rl 2R2. For Rl< 2R2, the result is MM 6‘) “if (5) Q4=[l€+ 2 - Rl A singularity occurs when R1 = 2R2. In this case, the end of the effective collision volume is determined by the collision of the product ions with the wall of the analyzer. It is of interest to see what the equations (4) and (5) predict for the length of the effective collision volume for the reactions studied in the present work. These results are tabulated in Table 2. The reaction N2+———-:>»N+ + N is an example of the singular case, and the calculated 9“ is determined by collision of the product ions with the wall of the analyzer. In the instrument used here, as in Coggeshall's instrument (52), this occurs when the ions have traversed an arc of 17°, or 3.77 cm. TABLE 2 Theoretical Q“ for Several Reactions Reaction 94 (cm) 002++M—>c++o++o+m 0.522 CO++M—->C++O+M 1.23 M“ + M—>Ar+ + M*- 0.271 N2++M-—=‘>N++N+M 3.77 N20+ + M ——> No+ + N + M 0.782 79 ifiue approximate nature of the above analysis should be emphasized. We have estimated the maximum distance that a reactant ion could penetrate the analyzer. In all cases except R1 = 2R2, this maximum penetration is achieved by an ion which passes very near the slits, and the collision volume for an average ion would be much smaller. For in- ’Wah_ stance, if the collision volume is triangular, the average ; length i; would be only half of the value calculated above. i An analysis of the shape of the collision volume would be E very difficult and would probably require computer calcula- g jJ tion. k“! A second criticism of the analysis above is that the kinetic energy released in the dissociation was not con- sidered. A semi-empirical study of metastable peak shapes by Newton (78) has shown that this parameter must be in- cluded in such an analysis. It would be expected to be of even greater importance in collision—induced reactions, since kinetic energy may be transferred to the target molecule. Until the trajectories of ions undergoing reaction in a spectrometer are better understood, it is felt that the method of calibration employed in the present work is the best approximation. m-xlmb“ 10. 11. 12. 13. 14. 15. 16. IX. REFERENCES F. W. Aston, Proc. Cambridge Phil. Soc., 19, 317 (1920). F. W. Aston, Mass Spectra and Isotopes, 2nd ed. New York: Longmans, Green, and Co., 1941. K. T. Bainbridge and E. B. Jordan, Phys. Rev., 51, 595 (1937). J. Mattauch and H. Lichtblau, Phys. Zeit., 49, 16 (1939). H. D. Smyth, Phys. Rev., g2, 452 (1925). H. D. Smyth, Rev. Mod. Phys., 3, 347 (1931). A. J. Dempster, Phys. Rev., 2;, 64 (1938). J. A. Hipple and E. U. Condon, Phys. Rev., 68, 54 ' (1945). J. A. Hipple, R. E. Fox, and E. U. Condon, Phys ég. 347 (1946). . Rev., J. A. Hipple, Phys. Rev., 71, 594 (1947). O. Eisenhut, R. Conrad, Zeit. Electrochem., 6, 654 (1930). W. Schutze, Wiss. Veroffentl. Siemens-Werken., 16, 89 (1937). H . Ewald and A. Henglein, Z. Naturforsch,.§a, 463 (1951). A. Henglein, Z. Naturforsch., 13, 165 (1952). (L F. Wells and C. E. Melton, Rev. Sci. Inst., 28, 1065 (1957)- %.IL Melton and G. F. Wells, J. Chem. Phys., 21, 1132 1957). 80 17. 18. 19. 20. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 81 C. E. Melton, M. M. Bretscher, and R. Baldock, J. Chem. Phys., 26, 1302 (1957). C. E. Melton and H. M. Rosenstock, J. Chem. Phys., 26, 568 (1957). H. M. Rosenstock and C. E. Melton, J. Chem. Phys., 36. 314 (1957). C. E. Melton and G. A. Ropp, J. Chem. Phys., 29, 400 (1958). Path L. Kerwin, Rev. Sci.,Inst., 21, 96 (1950). 3 J. D. Carette and L. Kerwin, Can. J. Phys., 39, 1300 L (1961). i W. McGowan and L. Kerwin, Can. J. Phys., 38, 642 7 (1960). . 5 £— ,2 W. McGowan and L. Kerwin, Can. J. Phys., 41, 316 ~22 (1963). N. M. Tunitskii, P. M. Smirnova, and M. V. Tikhomirov, Doklady Akad. Nauk. SSSR, 101, 1083 (1955). A. P. Irsa and L. Friedman, J. Chem. Phys., 34, 330 (1960). M. V. Tikhomirov, V. M. Kolotyrkin, and N. M. Tunitskii, Doklady Akad. Nauk. SSSR, 101, 903 (1955). V. M. Kolotyrkin, M. V. Tikhomirov, N. M. Tunitskii, Doklady Akad. Nauk. SSSR, 92, 1193 (1953). G. M. Begun and L. Landau, J. Chem. Phys., 3:, 547 (1961). G. M. Begun and L. Landau, J. Chem. Phys., 36, 1083 (1961). L. Friedman and A. P. Irsa, J. Chem. Phys., 36, 1082 (1961). W. A. Chupka and K. M. A. Refaey, Argonne National Laboratory Informal Report PHY-1966A, May 1966. E. Lindholm, Proc. Phys. Soc., 665, 1068 (1953). E. Lindholm, Arkiv Fys., 8, 433 (1954). E. Lindholm and H. von Koch, Arkiv Fys., 12, 123 (1961). j36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 82 1%. V2. Fedorenko, Zhur. Tekh. Fiz., 24, 769 (1954). F. M. Rourke, J. C. Sheffield, W. D. Davis, and F. A. White, J. Chem. Phys., _3_1, 193 (1959). J.-(3. Abbe and J.-P. Adloff, C. R. Acad. Sc. Paris., 258, 3003 (1964) J.-C. Abbe and J.-P. Adloff, Physics Letters, 1, 28 (1964). .J.-C. Abbe and J.-P. Adloff, Bull. Soc. Chim. France, 1964, 1212. 14. In ram and R. Gomer, Z. Naturforsch., 102, 863 (1955 . “- J. B. Hasted, Proc. Conf. Ioniz. Phen. Gases Munich. 1961. p. 11. J. B. Hasted, J. T. Scott, and A. Y. J. Chong, Proc. Conf. Ioniz. Phen. Gases Uppsala., 1959. p. 34. H. B. Gilbody, J. B. Hasted, J. V. Ireland, A. R. Lee, E. W. Thomas, and A. S. Whiteman, Proc. Roy. Soc. (London), 274a, 40 (1963). M. Islam, J. B. Hasted, H. B. Gilbody, and J. V. Ireland, Proc. Phys. Soc., 12, 1118 (1962). J. B. Hasted and R. A. Smith, Proc. Roy. Soc. (LondonL 2353, 354 (1956)- J. B. Hasted in Bates, ed., Atomic and Molecular Pro- cesses, New York: Academic PFess, 1965. W} W. Hunt, R. E. Huffman, J. Saari, G. Wassel, J. F. Beta“ E. H. Pauve, W. Wyess, and R. A. Flugge, Air Fbmce Cambridge Research Laboratories Research Pro- Reanw AFCRL—63-708 (1963). W.)L Hunt, R. E. Huffman, K. E. McGee, Rev. Sci. Inst., 3:, 82 (1964). R.IL Ferguson, K. E. McCulloh, and H. M. Rosenstock, J.