‘H—fli LIBRARY Michigan Stan: . University at, A ' rm This is to certify that the thesis entitled A PHOTOIONIZATION MASS SPECTROMETRIC STUDY OF ACETONITRILE AND ACETONITRILE-d3 presented by Gary William Ray has been accepted towards fulfillment of the requirements for _P.h+D..__degree in Chemical Physics 7,37 1; (Jam Major professor Date February 20, 1978 0-7639 A PHOTOIONIZATION MASS SPECTROMETRIC STUDY OF ACETONITRILE AND ACETONITRILE-d3 By Gary William Bay A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry Chemical Physics Program 1978 ABSTRACT A PHOTOIONIZATION MASS SPECTROMETRIC STUDY OF ACETONITRILE AND ACETONITRILE-d3 By Gary William Ray Due to its importance as an interstellar molecule and the paucity of vacuum UV spectral data on nitriles, aceto- nitrile has been studied by means of photoionization mass spectrometry (PIMS). Its deuterated analog, D3CCN, has been included in the study to aid in the data analysis. The photoionization efficiency (P.I.E.) curves from threshold 2*, 3+ and D3CCN+ resulting from the photoionization of acetonitrile and to 600 X of H3CCN+, H2CCN+, HCCN+, CH CH acetonitrile-d3 have been determined for the first time. Their appearance potentials have been determined and also 2+ and CD3+. The appearance po- tentials of the deuterated ions as well as that of CH3+ have not been reported heretofore. An autoionizing Rydberg those of D2CCN+, DCCN+, CD series has been observed in the P.I.E. curves of H3CCN+ and D3CCN+ and assigned to an nso series converging to the AZA excited state of the parent ion. All of the fragment 1 ions studied have appearance potentials that lie between the thresholds of the A2A1 and §2E state of the parent ions. Gary William Ray At lower energies they appear to be formed by the predis- sociation of autoionizing Rydberg states. However, after the energy of the exciting light has reached the threshold of the E2 E state of the parent ion the fragment ion P.I.E. curves rise sharply, indicating that most of their intensity is the result of the predissociation of this state. Fin- ally, the appearance potentials reported here have been used to derive several thermodynamic parameters. This dissertation is dedicated to my parents, who got me started, and to my wife, who helped me finish the task. May my efforts be worthy of theirs. ii ACKNOWLEDGMENTS I Would like to thank my advisor, Professor George E. Leroi, for his friendship, patience, guidance and en- couragement during the course of my studies. His efforts will long be remembered. I would also like to thank my second reader, Professor Paul M. Parker, for his many helpful suggestions during the writing of this disserta- tion. I am especially grateful to Dr. Edward Darland, who designed the lion's share of the MSU PIMS apparatus, and to Mr. David Rider, whose help was so vital during the final stages of my experiments. I would also like to thank Dr. Darland for allowing me to use his drawing of the ap- paratus in Figure II-l. My gratitude also goes to the many members of the Mol- ecular Spectroscopy Group for their friendship over the years. Special thanks go to Mrs. Naomi Hack for her interest— ing discussions and willingness to do a "quick typing job" at a moment's notice. The financial support of the Office of Naval Research and the NSF is gratefully acknowledged. iii Chapter LIST OF LIST OF I. IIA. IIB. IIC. IIIA. IIIB. IIIC. REFEREN TABLE OF TABLES FIGURES. INTRODUCTION. THE INSTRUMENT. SAMPLES CONTENTS EXPERIMENTAL PROCEDURE. THE MASS SPECTRUM . THE PHOTOIONIZATION EF CURVES. . . . H2CCN+ and D2CCN+ HCCN+ and DCCN+ CH + and CD + . . . . 2 2 + + CH3 and CD3 SUMMARY CES iv FICIENCY Page vi ll 23 2H 27 39 5A 70 76 85 93 99 Table III-l III-2 III-3 III—u III-5 III-6 III-7 LIST OF TABLES Page Acetonitrile and acetonitrile-d3 photoionization mass spectra. Ionizing energy = 21.21 eV; sample pressure = 0.20 mtorr; repeller voltage = +10 V. . . . . . . . . . . . 28 Appearance potentials calculated for the primary fragment ions ob— served by Franklin, et al. in the CH3CN mass spectrum. . . . . . . . . . 31 Acetonitrile—d3 mass spectrum at a source pressure of 0.8 mtorr and a repeller voltage of +0.5 V . . . . . 37 A comparison of the values for the first ionization potential of CH CN 3 and CD3CN as reported in this and other works. . . . . . . . . . . . . . “7 Analysis of the autoionizing Rydberg series in CH3CN and CD3CN. . . . . . . 50 Appearance potentials of the frag— ment ions of HBCCN+ and D3CCN+ cor- rected to 0 °K . . . . . . . . . . . . 97 Derived thermodynamic quantities . . . 98 Figure II-l II—2 III-l III—2 III-3 III-u III-5 III-6 III-7 III-8 III—9 LIST OF FIGURES Location of components within the vacuum chambers. Schematic diagram of the ion source Threshold region P.I.E. curve Threshold region P.I.E. curve The H3 threshold to 600 The D3 threshold to 600 Threshold region P.I.E. curve Threshold region P.I.E. curve The H2CCN+ P.I.E. threshold to 600 Threshold region P.I.E. curve Threshold region P.I.E. curve CCN+ P.I.E. CCN+ P.I.E. of the of the curve curve of the of the curve of the of the vi H3CCN+ D3CCN+ from from £12ch+ D2CCN+ from HCCN+ DCCN+ Page l3 16 U3 “5 56 58 60 62 68 72 7M Figure III-10 III-ll III-12 III-l3 III-1U III-15 III-l6 The HCCN+ P.I.E. curve from threshold to 600 A . . . . 4. Threshold region of the CH2 P.I.E. curve + Threshold region of the CD2 P.I.E. curve The CH2+ P.I.E. curve from threshold to 600 K . . Threshold region of the CH3+ P.I.E. curve . . . . . . . . . + 3 P.I.E. curve . . . . . . . . . The CH3+ P.I.E. curve from the Threshold region of the CD threshold to 600 A . . vii Page 78 80 82 87 89 92 95 CHAPTER I INTRODUCTION The presence of several molecular species in the galactic environment (e.g., interstellar nebulae and stel- 1 As detec- lar atmospheres) has been known for some time. tion techniques have improved, more molecules of increasing complexity have been discovered. Recent theoretical and laboratory work has interpreted a newly discovered set of radio frequency spectral lines as belonging to a mole- cule with an eleven carbon backbone. These discoveries have naturally led to many investigations into the nature of interstellar chemistry.2 The energy source for inter- stellar chemical reactions is still not well understood. At the average temperature of interstellar nebulae (about 100°K) most neutral—neutral reactions are inhibited as their potential energy surface barrier heights exceed the available thermal energy. Moreover, photodissociation and photoionization must be ruled out as the available ultraviolet light from nearby stars is heavily attenuated 8 particles/cm3) by the material of the dense (about 10 nebulae in which the more complex molecules have been dis- covered.3 Recently, a new quantitative theory of inter— stellar chemistry has been developed by Herbst and Klemper- er.)4 They suggest that the driving force of interstellar chemistry is ionization by cosmic rays. A substantial l cosmic ray flux is provided by stars nearby to or imbedded in the interstellar nebulae. Significantly, this flux is not attenuated by the nebular material. Since the initial step in this model is ionization by cosmic rays, one would expect interstellar chemistry to be predominantly ion- molecule reactions with the rare exception of a neutral— neutral reaction of low activation energy. Herbst and Klemperer have shown that their model can account for most of the molecules so far discovered in interstellar space. In order for the theory to receive a quantitative test, the thermodynamic parameters (e.g., heats of forma- tion, bond dissociation energies, electron and proton affinities) associated with each ion and molecule in a given ion-molecule reaction must be known. The technique of photoionization mass spectrometry (PIMS) is one of the best and most widely used methods employed to determine such parameters. It is not my intention here to provide more than a brief review of PIMS, since there are several fine, though somewhat dated, reviews available.5’6 However, I wish to provide enough basic information to allow the reader to follow this dissertation without undue effort. As is well known, the mass spectrum of any molecule is comprised of the molecular (parent) ion and several frag- ment ions — provided that the energy of the ionizing radiation is sufficiently large. In order to derive the aforementioned thermodynamic quantities, it is necessary to measure the energy thresholds for the formation of the parent ion and each of its fragment ions. This requires a mass spectrometer that is capable of accurately measur- ing the energy dependence of the cross section for the formation of each of the ions. This has often been ac- .complished with an instrument equipped with an electron bombardment source whose voltage can be accurately varied. A photoionization mass spectrometer performs the same task with improved energy resolution with a continuum source of vacuum ultraviolet radiation and a vacuum monochromator. The photoionization cross section, 0 is defined as the is number of ions produced per absorbed photon or 01 E oNi/(IO - I), (l) where o is the absorption cross section, Ni is the number of ions produced per second, I0 is the incident photon intensity and I is the transmitted photon intensity.7 Actually, a PIMS instrument is incapable of measuring 01 since vacuum monochromators are only single beam instru- ments. In addition, the difference between IO and I would be very difficult to measure anyway as the absorption cross sections in the vacuum ultraviolet region are very small (10"18 cm2 or less). Consequently, a quantity called the photoionization efficiency is measured for each ion as a function of wavelength or energy. The photoionization efficiency (P.I.E.) is defined as the number of ions formed per transmitted photon or P.I.E. s Ni/I. (2) It has been shown that for sufficiently low pressures (where I/IO is at least 90%) the ratio of the P.I.E. to 8 ai is nearly unity. A PIMS instrument is somewhat less useful for routine structural, qualitative and quantita- tive analysis than its electron bombardment counterpart due to the relatively low intensity (fl? most sources of vacuum ultraviolet continua. However, the use of photoionization gives this technique two distinct advantages over conven— tional instruments in the type of study of interest here. First, as mentioned earlier, thermodynamic data on a molecule and its photofragments (both charged and neutral) may be derived if the energies at which a molecule's frag- ment ions appear are known. The derivation of such quanti— ties will be discussed in more detail later. Obviously, the more accurately one determines these thresholds, the more accurate will be the resulting thermodynamic quanti— ties. Under favorable conditions, the energy resolution in a PIMS experiment may be as high as 0.001 eV due to the wavelength resolution available from vacuum mono- chromators. This is two orders of magnitude better than the best results attainable with an electron bombardment mass spectrometer, since the resolution of electron bom- bardment sources is always limited by the kinetic energy distribution of the electrons leaving the heated filament. A second advantage of PIMS is that the thresholds for direct photoionization and ionic fragmentation are much easier to observe than in the case of electron bombard- ment. It has been shown that the cross section for direct ionization is related to the energy of the ionizing radia- tion in the following way: Oi a En—l (3) where n is the number of electrons leaving the collision complex.9 The cross section for ionization by electron impact thus displays a linear, slowly rising dependence on the electron energy, because two electrons leave the collision complex. The onset of ionization is often dif- ficult to observe, because the cross section is nearly zero at threshold. However, the cross section for photo- ionization is constant and has a finite value at threshold (i.e., it is a step function). This makes the direct photoionization threshold relatively easy to observe. The thresholds for higher energy processes (transitions to excited vibrational and electronic states of the ion) are then superimposed upon the constant direct ionization cross section (P.I.E.). However, other factors may inter- vene that complicate this obviously ideal case. Since photoionization is a vertical process, the relative in- tensities of the observed transitions are controlled by their Franck-Condom factors. In those favorable cases where the ionic ground state geometry does not differ sig— nificantly from that of the neutral molecule, one of the vertical transitions will also be the adiabatic transi— tion. The adiabatic transition is between the ground electronic, vibrational and rotational state of the neutral molecule and the ground electronic, vibrational and rota- tional state of the ion. However, in some cases the Franck- Condon factor for this transition may be so small that it is observed as an exceedingly weak step or is totally ab- sent from the P.I.E.10 A further complication may occur when neutral molecules in excited rotational or vibrational states are photo- ionized. The result is usually weak structure (hot bands) below the energy of the true vertical transition from the molecular ground state. Rotational hot bands are usually diffuse and structureless as the rotational levels are too closely spaced to be resolved by most PIMS instruments. Vibrational hot bands often appear as fairly well defined steps. Either type of hot band masks the position of the actual threshold. Cooling the sample will, in most cases, remove such structure. As mentioned previously, the thresholds for transitions to excited states of the ion will be superimposed on the P.I.E. curve for ionization to the ionic ground state at energies above the first ionization potential. However, these thresholds are often obscured by additional struc- ture that is also superimposed on the P.I.E. curve. In addition to its valence electronic levels, every molecule displays many sets of Rydberg levels at higher energies.11 As in the case of hydrogen, some of the Rydberg levels occur in series that converge to the first ionization potential of the molecule. However, in the cases of molecules and many—electron atoms, other series converge to excited states of the ion as well. Necessarily, some of these Rydberg states occur at energies above that of the first ionization potential and are imbedded in the first ionization continuum, and the corresponding wave- functions are actually mixed Rydberg and continuum wave- functions.12 Thus, electrons excited to these neutral molecule states may "cross over" to the continuum and leave the molecule. When this occurs, "autoionization" structure will appear on the P.I.E. curve. This structure is not observed in those cases where predissociation of the Rydberg states (leading to neutral fragments) depopu- lates them faster than the autoionization rate. Fragment ions are formed when the energy of the ioniz- ing photons exceeds the first ionization potential suf- ficiently to cause dissociation of the parent ion.13 If we photoionize a molecule AB, where A and B are atoms or groups of atoms, two fragment ion forming processes may be obServed, dissociative ionization + AB + hv = A + B + e' (A) and ion pair formation + _ AB+hv=A +13. (5) Generally, fragment ions are formed by the ionization and excitation of a molecule to an excited ionic state or to an autoionizing Rydberg state followed by predissocia- tion.lu In some cases, the state of the parent molecular ion reached by photoionization may have a repulsive po- tential surface. Either situation precludes the observa- tion of step-function like onsets for fragment ion forma— tion. Hot rotational bands may also serve to mask the threshold further. Kinetic factors may be at work here as well. Most ionic fragmentation processes have a minimum "activation energy" that must be exceeded. This is termed the "kinetic shift". The presence of the kinetic shift will cause the measured fragmentation threshold to be higher than the true thermodynamic value. These factors often prevent precise determination of ionic fragmentation thresholds. Consequently, the values reported are termed "appearance potentials". If the kinetic shift is not large, the appearance potentials can be reported with known limits of accuracy. If the kinetic shift is large, the accuracy of a reported value may be questionable and will at best be an upper bound to the actual value. Once the ionization potential and fragment ion ap— pearance potentials of a molecule have been determined, the values may be incorporated in thermodynamic cycle calcula- tions (along with other data) to derive thermodynamic quan- tities of interest. As an example, let us consider the dissociative ionization of our general molecule AB: AB + hv = A+ + B + e‘. (6) The appearance potential of A+ (AP(A+)) corresponds to the heat of reaction for this endoergic process. When the ap- pearance potential is corrected to its 0°K value, we may write + + AP(A ) = AHf(A ) + AHf(B) — AHF(AB), (7) since AHf = AGf at that temperature. If any two of the heats of formation are known, the third can be determined. In addition, since AHf(A+) = AHf(A) + IP(A), (8) 10 where IP(A) is the ionization potential of the fragment A, Equation (7) becomes AP(A+) = AHf(A) + IP(A) + AHf(B) - AHf(AB) (9) or AP(A+) = IP(A) + DO(A-B) (10) where DO(A-B) is the energy of the A-B bond. Other quan- tities may be derived depending upon the data obtained and the thermodynamic data available in the literature. Acetonitrile was chosen as the subject for this inves- tigation for several reasons. It has been discovered in interstellar nebulae. It provides an interesting problem in mass spectrometry (as will be seen in Chapter III). In addition, very few data exist in the literature on nitriles despite the fact that they are an important class of compounds.15 CHAPTER II EXPERIMENTAL A. The Instrument The Michigan State University PIMS apparatus and its associated data acquisition equipment have already been described in great detail.16 A general description of the instrument is sufficient for the purposes of this dissertation. During this discussion the reader should refer to the diagram of the MSU PIMS instrument in Figure II-l. Vacuum ultraviolet radiation is produced by a Hinter- regger-type discharge lamp (LP). For the experiments des- cribed here, the well known Hopfield continuum of He and pseudo-continuum of H2 were employed. They were excited by high voltage pulsed D.C. and high voltage D.C. dis- charges, respectively. The Hopfield continuum provides usable light intensity in the region from about 600 A to 1000 A. The H2 pseudo-continuum provides usable intensity from about 930 A to 1800 A. For the purposes of obtaining mass spectra, a D.C. discharge through low pressure He (about 1 torr) was used to produce the intense HeI atomic resonance line (58A.32 A or 21.218 eV). As the lamp must be operated in a windowless configuration, two stages of 11 BA EN EX GR IS IT LP P1 P2 P3 P14 P5 PT QP Q8 12 Baffle Entrance Slit Exit Slit Grating Ion Source Ion Transducer Lamp First Differential Pumping Port Second Differential Pumping Port Monochromator Pumping Port Sample Chamber Pumping Port Quadrupole Chamber Pumping Port Photon Transducer Quadrupole Mass Filter Quadrupole Support Figure II-l. Location of components within the vacuum chambers. 13 P4 75 P5 MK“ LEE ° / r” x" A \ i/ r I, ’1’, rr P] km r *5 I, O 0 IS 'J In II_‘QP 1 E] ®OOOpso mo msficmm poaaoqom m opmam pmaaoaom u mm .oopzom 20H map mo Empwmfic oameonom .NIHH mpswfim l6 mum J “I mH GIN mm ll ll It. (1 mam 17 insulated from the rest of the monochromator. The ion source is fitted with entrance and exit slits that are wide enough to allow the uninhibited passage of the light beam. For future experiments, the exit slit should be made narrower to allow for higher sample pressures while maintaining a low background pressure in the vacuum cham- ber. Of course, photoelectrons would be produced by the light striking the interior of the ion source. However, these can be trapped by placing a small metal plate, elec— trically insulated from the rest of the ion source, below the entrance slit. A small positive potential on the plate will attract the unwanted photoelectrons. Ions formed in the ion source exit through a circular aperture transverse to the light beam. The diameter of the aperture may be varied and is covered with a fine gold coated mesh (90% transmissivity) to insure a uniform poten- tial within the source itself. The ions are pushed out of the ion source by a positive potential placed on the repeller. In most PIMS instruments the repeller is a simple flat plate that is electrically insulated from the rest of the ion source. One PIMS research group has designed a focused repeller, which consists of a metal plate bent into a modified U-shape.l7 However, this did not provide effective focusing of the ions as the repeller voltage could be varied by $0.5 volts without any change in the measured ion intensity.18 In view of the many analogies 18 between charged particle and physical optics, our repeller has been fashioned in the shape of a spherical mirror. The repeller is a 311 stainless steel plate with a spheri- cal depression machined into one face. The focal point of a spherical mirror is on a normal to the mirror's surface at a distance of R/2, where R is the mirror's radius of curvature. Consequently, the radius of curvature of our repeller is twice the distance from the repeller to the first element of the mass filter lens system. This seems to provide a closer approximation to a truly focused system as a variation of i0.l volts or less will cause a signi- ficant change in the measured ion intensity. Sample gases are introduced by means of a stainless steel tube entering behind the repeller. An identical tube leads out of the ion source, transversely to the inlet tube, to a capacitance manometer that monitors the sample pressure. After passing through the ion source, the transmitted photons are detected by the photon transducer (PT). For these experiments, the transducer is a sodium salicylate phosphor - RCA 8850 photomultiplier system. The vacuum ultraviolet radiation is absorbed by a thin layer of sodium salicylate deposited on a quartz disc. The excited phos- phor in turn fluoresces in a wide band whose peak intensity occurs at approximately A200 3. The fluorescence is con- ducted down a Lucite light pipe that Joins the quartz 19 disc to an RCA 8850 photomultiplier in a Pacific Photo- metrics thermoelectrically cooled housing. The advantage of such a system is that sodium salicylate has a nearly constant quantum yield from around 200 A to 1200 3.19 Since the photomultiplier tube is always presented with light of the same wavelength, there results a detection system whose sensitivity as a function of wavelength is constant. This is in marked contrast to the other detectors that are often used, such as bare photomultipliers, electron multipliers and simple photocathodes whose sensitivity varies considerably with wavelength. Accordingly, they must be calibrated with the sodium salicylate phosphor- photomultiplier detection system. However, the latter system has its shortcomings as well. It is sensitive to the near ultraviolet and visible light that is produced by our lamp, which is scattered, undiffracted, off the grating. This produces a large background photon signal. Moreover, photomultipliers are much noisier than electron multipliers. Perhaps the two greatest shortcomings are that the sodium salicylate is often chemically degraded by a significant number of compounds and it is slowly aged by contamination by diffusion pump oil vapors. The scattered light effects may be corrected for during post-experiment data analysis. Photomultiplier noise is considerably reduced when the multiplier is cooled. More— over, the relatively weak ion signals require long counting 20 times that improve the photon signal-to-noise ratio even further. When reactive gases are in use, electron multi- pliers and bare metal photocathodes must be used. Fre- quent changing of the sodium salicylate phosphor will pre- vent aging effects from altering the data. Ion focusing and mass analysis are performed by an Extranuclear quadrupole (QP) mass filter and its associated electrostatic lenses (LE). The ion transducer (IT) is a Johnston Laboratories MM-l focused mesh electron multi- plier. The vacuum chamber that houses the ion source, ion optics, quadrupole mass filter and photon and ion trans- ducers is pumped by two six inch oil diffusion pumps, PM and P5. All of the oil diffusion pumps on the apparatus are trapped by freon refrigerated baffles in order to re- duce backstreaming of the pump oil into the vacuum chamber. The Roots-type blower, mechanical backing pumps and re- frigeration units are housed in a small room in one corner of the laboratory in order to minimize the amount of noise and heat in the main room. All of the pump exhaust ports are connected to a fume hood to keep pump oil vapors and toxic gases out of the laboratory. Sample gases are fed into the ion source by means of a variable leak valve that is incorporated into a metal vacuum manifold. One half of the manifold, for chemically inert gases, is constructed from copper tubing and brass 21 valves while the remainder is constructed entirely from stainless steel in order to handle more corrosive species. The substitution of brass and copper for stainless steel wherever possible has resulted in a considerable savings on the construction cost of the manifold. The MSU PIMS instrument is interfaced to a Digital Equipment Corporation PDP-8/M minicomputer. Signals from the photon and ion transducers are amplified by a Keithley A27 current amplifier and a Keithley A17 high speed pico- ammeter, respectively. The output signals of the ampli- fiers are fed to voltage-to-frequency converters and the converter signals are interfaced into the computer. Once the initial instrumental operating parameters are set, the entire experiment is automated with the computer in control of all aspects of data acquisition. The computer changes the wavelength setting by an amount chosen by the user throughElstepping motor attached to the monochromator drive screw. At the beginning and end of each experiment, mea- surements are taken to estimate the scattered light contri- bution to the photon and ion signals. Throughout the experiment, measurements of the dark currents of the detectors are taken at prespecified intervals. In addi- tion, reference measurements at a prespecified wavelength are taken throughout the experiment to monitor drifts in lamp intensity, sample pressure and the detectors. In order for the operator to Judge the quality of the data 22 during the actual experiment, a P.I.E. curve approximately corrected for stray light and detector dark current drift is generated in real time. A unique aspect of the data acquisition capabilities of the MSU instrument is the ability to achieve a constant signal to noise ratio throughout an experiment. This is accomplished by a variable counting time option in the operating program of the computer. The user specifies the minimum number of ion counts for the computer to ac— cumulate and minimum and maximum counting time limits. At each data point the computer counts for the minimum specified time and then computes the ion count rate. If the minimum specified number of counts has been accumulated, the P.I.E. (ions per sec/photons per sec) will be computed, logged and plotted and the computer will step the monochro- mator to the wavelength of the next datum. If the mini- mum count has not been accumulated, but the computed count rate indicates that it will be reached within the maximum allowed counting time, the computer will continue to count for the time required to reach the minimum count. If neither of the above is achieved, the computer will plot the P.I.E. for that datum and move to the next wavelength. Such a procedure provides the advantage of minimizing the duration of each experimental run by counting only as long as necessary to achieve the desired signal-to—noise ratio. This is important as PIMS experiments are typically long 23 (as much as 60+ hours) due to the often very low ion count rate (sometimes as low as I count per sec or less). B. Samples Acetonitrile (99% purity) and acetonitrile-d3 (99 atom % purity) were obtained from the Aldrich Chemical Company. Although these purities are quite sufficient for most PIMS studies, further purification was undertaken in order to remove dissolved air from the samples. The sample containers were spherical pyrex vessels of approximately 0.5 liters in volume. Samples were intro- duced through a 0.5 inch O.D. glass-to-metal transition tube. A 0.25 inch O.D. glass-to-metal transition tube leads to the sample inlet manifold by way of a polyethylene tube. The tube is connected to the manifold and vessel by means of Cajon "Ultra-torr" unions. Similarly, the 0.5 inch O.D. sample introduction tube is sealed off by means of an appropriate union with one end blocked off by a stain- less steel rod. Before the samples are placed in a vessel, it is pumped down to approximately 6 x 10"7 torr by the sample-mass spectrometer chamber diffusion pumps for one hour. The samples are then subjected to three freeze-pump- thaw purification cycles at dry ice temperature (-78.5°C) until they could be pumped down to 6 x 10'7 torr or less while frozen. 2A 0. Experimental Procedure As mentioned previously, mass spectra of acetonitrile and acetonitrile—d3 were taken using the HeI 58A A line. The intensity of this line is comparable to that of the Hop- field continuum when viewed at central image. Since the HeI line is positioned near the high energy end of the Hop- field continuum (600 A), one is able to measure the rela- tive intensities of all of the ions created from a given parent compound at energies between 21 eV and threshold. Mass spectra were taken at several pressures on the order of one mtorr (micron) to determine the pressure dependence of the ion intensities (see Chapter III-A). The ion op- tics were adjusted in order to maximize the measured ion intensities (this matter is treated in more detail in Chap- ter III-A). The quadrupole mass filter was adjusted to provide mass resolution sufficient to cause the ion cur- rent to go to zero between adjacent peaks in the mass spectrum. The ion transducer (Johnston MM-l electron multiplier) voltage was kept at 3.0 kV. Higher voltages increased the intensity of the ion signals with an equi- valent increase in the noise and background signals. For this reason, they were not used. The mass spectra were measured by hand-tuning the mass control on the mass filter control unit while observing the ion intensities indicated on the picoammeter voltmeter. The very weak signals (10-12 11 to 10- amps) of some of the ions were extremely noisy. 25 Since these signals were not integrated, the reliability of their absolute values is questionable - probably no better than i25%. The photoionization efficiency curves were measured using the He Hopfield continuum for the 600 A to 950 A region, while the H2 pseudo-continuum was used from 930 A to 1030 A where the intensity of the Hopfield continuum becomes comparable to that of the scattered light. The sample pressures were kept around 0.6 mtorr. The voltage on the photon transducer (RCA 8850) was adjusted to maxi— mize the multiplier current with a minimum of noise (typi— cally 1.2 kV). All other instrumental operating parameters were identical with those used in obtaining the mass spectra. It is difficult to accurately determine the average signal—to-noise ratio for these experiments as the output frequency of the voltage-to-frequency converters is propor- tional to the current amplifier (picoammeter) output vol— tage - not to the actual ion count rate. A high speed pulse counter was available for some of the preliminary experiments. However, a malfunction caused it to be un- usable for the bulk of this work. A comparison between a picoammeter and the pulse counter indicated that an ion current of 3 x 10"11 amps corresponded approximately to a count rate of 100 ions/sec. Using this count rate as a standard, integration times were set (see Chapter II-A) 26 to achieve a signal—to-noise ratio of approximately 100:1 provided that all of the observed noise was random. How- ever, it was not possible to achieve such a high signal- to—noise ratio with the lower intensity ions without re— sorting to unreasonable integration times (10“ sec/datum at a count rate of l ion/sec). In such cases, the entire experiment was carried out at the longest possible integra- tion time available in the current operating program which is #05 seconds. CHAPTER III RESULTS AND DISCUSSION A. The Mass Spectrum The mass spectrum of acetonitrile is surprisingly rich for a molecule comprised of only six atoms. Franklin, Wada, Natalis and Heirl reported the mass spectrum along with a study of the ion-molecule reactions between the aceto- nitrile molecular ion and the neutral molecule.2O They reported thirteen primary ion mass peaks and three ion- molecule reaction products for H CON and eleven primary 3 ion mass peaks and four ion-molecule reaction products for D3CCN. Unfortunately, the energy of their ioniz— ing electron beam is not clearly stated. If their earlier works are any indication, it was probably in excess of 30 eV. Table III-l. presents mass spectra for H3CCN and D3CCN taken on the MSU PIMS apparatus. The ionizing radia- tion in this case was provided by the HeI resonance line (58A.33U 3, 21.218 eV). The ion source pressure was 0.2 mtorr and the repeller voltage was set at +10 V with respect to the ion source potential in order to minimize the time spent in the ion source by the primary ions. With the exception of the ions m/e = 2A and m/e = 25, we have observed the same ions with similar relative intensities 27 28 +2mom0 H0.0 +2m0=m NH.0 me -- ..... +2m0mm 0.H H: m m m . +2 00 0H.0 +2 0 m 00 0 0: -- ..... +zmom 0H.0 0m +2m0 s00.0 +200 Hm0.0 0m +m0 0H.0 +m0 0H.0 mm mamo. 2000 0H0.0 -- --- 0m + m+m m m m +200. 0 0. z mm.0 20 m. z =m.0 00 + + +m N + -- ..... + m 0.+zom 0H0.0 em 000. 20 000.0 ammo. 20 000.0 mm + m + + + + 0o Hmo.o ---- ----- 0H +000 HH.0 -- ----- 0H ..-.. ----- +000 mmee 3 +00 000.0 +mm0 000.0 0H -- ----- +mo mHo.o ma +0 oomph +0 momma ma mHSEpom Azzuo\ev +20 mHsEpom Aazuo\ev +zom o\E ofiofiamod ofiofiaood o>HpmHom mpfimcoch o>HpmHom mpfimcman pofloaoh masons om.o u oLSwmonQ oHQEmm m>m Hmém .mppoodw mmme :oHpmNHcoHouosq mclofifispHCOpoom cum mafiupfiGOpoo< > oa+ n mmmpflo> mwmmzo wcHNHCOH .HIHHH manna 29 +2m000 000.0 -- --- 0m +Zm ma moms» III- III-- mm fim m0 m00.0 +2m000 m0.0 am -- --- :2m0m 0H0.0 mm +2N0=0 000.0 -- --- 0: +zN0m0 0.H -- --- a: «asapom Aazuo\ev +zomoo on masspom Aaauo\sv +zommo on o\E ofioaaeod oflofiamom seapoflom spemcopsH o>HpmHom mpfimcopCH .Umscfipcoo \ .HIHHH mHan 30 as Franklin and coworkers. Differences are bound to exist due to the use of different ionizing energies, source pres- sures and mass spectrometers. Of course, the question arises as to why the m/e = 2A and m/e = 25 peaks were not observed in the present study. This question may be ans- wered by comparing the appearance potentials of these ions to the energy of our radiation source. Table III-2 lists the m/e peaks observed by Franklin and coworkers along with their possible formulae. With the choices of the neutral fragments listed in the third column, zero- degree Kelvin appearance potentials have been calculated for each ion from the latest thermodynamic data in the literature.21 Appearance potentials were not calculated for the deuterated ions as their values should be very similar to those of the protonated species. For the m/e = 2“ ion, assumed to be C2+, four appearance potentials are possible. If the neutral fragments are N + H + H, 2 NH2 + H or NH + H2, the calculated appearance potentials are 26.8 eV, 23.6 eV, and 23.1 eV, respectively. These are all clearly above the energy of the HeI resonance line. The other calculated value is 19.3 eV, assuming that the neutral fragment is NH3. Thus it would appear that the absence of C2+ (m/e = 2“) from the photoionization mass spectrum is due to its relatively high appearance poten- tial. A similar situation exists for m/e = 25 (C2H+). If the neutral fragments are NH + H, then its absence 31 Table III-2. Appearance potentials calculated for the primary fragment ions observed by Franklin, et al. in the CH3CN mass spectrum. Assumed Assumed Neutral Appearance m/e Formula Fragments Potential 12 0+ HCN+H2 19.02 eV 13 CH+ HCN+H 19.u ev 1A CH2+ HCN lu.9o eV 15 CH3+ CN 1h.96 eV 2A 02* N+H2+H 26.8 eV 02+ NH2+H 23.6 eV c2+ NH+H2 23.1 eV c2+ NH3 19.3 eV 25 02H+ NH+H 22.0 eV 02H+ N+H2 21.22 eV C2H+ NH2 18.1 eV 26 CN+ CH3 19.1 ev 26 02H2+ NH 16.2 ev 27 HCN+ CH2 18.1 ev 27 C2H3+ N 15.6 eV 28 H20N+ CH 17.9 ev 38 02N+ H2+H 19.1 eV 39 HC2N+ H2 15.1 eV no H202N+ H lu.01 eV 32 from the photoionization mass spectrum would be explained by the high calculated appearance potential of 22.0 eV. An appearance potential of 21.22 eV is calculated if the assumed neutral fragments are N + H2. This energy falls right on the HeI resonance line. Since m/e = 25 is a frag- ment ion, the cross section for its formation should be very small at threshold. As a result, the intensity of the signal might be too weak to be observable when the HeI lamp is used as a source of ionizing radiation. Were the true value closer to the third and lowest calculated appearance potential, the cross section would be large enough at 21.2 eV to make the ion observable. Thus, it is concluded that the 02H+ (m/e = 25) appearance potential is at least 21.22 eV. Several fragment ion peaks are really comprised of two different ions. In the electron bombardment mass spectrum of CH3CN, Franklin and coworkers concluded that m/e = 26 is comprised of approximately 60% CN+ and “0% 02H2+ by means of a high resolution time-of—flight mass spectrometer. Similarly, the m/e = 27 peak was found to be approximately 85% HCN+ and 15% C2H3+. A glance at Table III-2 shows that the appearance potentials of the ions in each of these pairs differ widely. Such a situa- tion might allow the determination of the P.I.E. curve of the ion with the lowest appearance potential up to the threshold for the formation of the second member of the 33 composite mass peak. However, the total intensity of each of these composite mass peaks is so low that the MSU PIMS apparatus is unable to measure the P.I.E.'s or appearance potentials of any of the ions comprising them. The other multiple ion mass peak in the CH3CN mass spectrum is m/e = 28. The two possible ions are N2+ (ionization potential = 15.58 eV) and H2CN+. That m/e = 28 contains a high N + component is suggested by the 2 low m/e = 28 intensity in the CH3CN mass spectrum of Frank— lin and coworkers and the strong m/e = 32 peak in the photoionization mass spectrum. Efforts were made to remove air from the samples, but they were not completely successful. The calculated appearance potential of H2CN+ is greater than that of N +, so its threshold might show 2 up as a weak (at best) step amidst the well-known strong, sharp autoionization structure of the N2+ P.I.E. The decision to study acetonitrile—d3 in addition to the protonated compound was based on the hope that deuterium substitution would help distinguish the ions in the CH3CN °multip1e ion mass peaks (m/e = 26,27). Unfortunately, the situation in the CD3CN mass spectrum is even more compli- cated than its protonated counterpart. Three possible multiple ion mass peaks may be present: m/e = 26, m/e = 28 and m/e = 30. As in CH3CN, the peak at m/e = 28 is + comprised of two low intensity ions mixed with N2 The + peak at m/e 30 is probably comprised of D20N+ and C2D3 , 3“ both of which have signals that are too weak to measure 26 deserves special com- meaningfully. The peak at m/e ment.' Although two ions are possible at this mass, CN+ and C D+, only the CN+ fragment is likely to be present 2 at the ionizing energy used here as the appearance poten- tial of 02D+ is close to that of C2H+. This would explain the very low relative intensity of m/e = 26 (CN+) in CD3CN in comparison so that of m/e = 26 (CN+ and C +) in CH3CN. 2H2 Of the remaining primary ions in both photoionization mass spectra, only five in each case were of sufficient intensity to enable meaningful P.I.E. curves to be ob- tained. The CH3+ and CD3+ fragment ions represented the upper sensitivity limit of the MSU PIMS instrument. The mass spectra of both compounds display peaks that result from reactions between the parent (molecular) ion and the neutral molecule. The pressure and time depen- dence of these reactions were also studied by Franklin and coworkers. As expected for secondary reactions, they found a squared dependence of the product ion signals on ion source pressure (sample pressure) and the residence time of the ions in the source. These ion-molecule reac- tions are not of interest to this study. However, the reaction H(D)3CCN+ + H(D)3CCN + H(D)uCCN+ + H(D)2CCN (1) 35 proved to be particularly troublesome. The presence of an ion at m/e = A2 in CH3CN and m/e = A6 in CD3CN made it necesSary to employ relatively high mass resolution even at ionizing energies below the appearance potential of the first fragment ion. Franklin and coworkers reported that the appearance potential of the CHuCN+ ion is the same as the first ionization potential of the neutral molecule, which indicated that the reaction is at least thermoneutral if not exoergic. This was confirmed by Haney and Frank— lin, who estimated the AH° of reaction (1) to be -25 kcal/ 22 A more serious problem caused by mol in a later study. reaction (1) is that it diminishes the intensity of the parent ion signals at higher source pressures and longer ion source residence times (lower repeller voltages). This is dramatically demonstrated in Table III-3. Whereas the mass spectra in Table III-l were measured under condi- tions of low source pressure and high repeller voltage, the CD3CN mass spectrum in Table III-3 was measured at a repeller voltage of +0.5 V relative to the ion source po- tential and a source pressure of 0.8 mtorr. Note that the relative intensities of the m/e = 58, m/e = A6 and m/e = A2 peaks are A, 36 and 5 times greater than that of the parent ion (m/e = AA), respectively. In addition, a second complication may result. Since the neutral product of reaction (1) is CHZCN, the CH CN+ signal may be en- 2 hanced by direct ionization of this fragment. Obviously, 36 a good solution to this problem would be to measure the P.I.E.'s of the molecular and fragment ions under the conditions of Table 111-1. However, under those condi- tions the intensity of the parent ion signal at the peak of the Hopfield continuum (815 A, 15.2 eV) is only 3 x 10"11 amps or about 100 counts/sec. Since the intensities of many of the fragment ion signals are less than 10% of the parent ion signal, a compromise had to be reached. Consequently, the experimental results presented in the next section were obtained at an intermediate average source pressure of approximately 0.5 mtorr and a potential of +10 V on the repeller with respect to the ion source potential. Under these conditions, the intensity of m/e = A1 (AA) relative to the intensity of m/e = A2 (A6) is approximately 3:1 at an ionizing energy of 15.2 eV. This minimized the effects of reaction (1) while allowing the achievement of better signal-to-noise ratios in more reasonable (though still long) times. One might expect that the complete mass spectrum of acetonitrile would display negative ions in addition to the parent ions just discussed, due to the presence of a -CN group with its pseudohalogen characteristics. In their earlier incomplete PIMS study of acetonitrile, Dibeler and Liston searched unsuccessfully for negative ions resulting from ion—pair formation processes.23 At the time of this writing the MSU PIMS apparatus is not yet 37 Table III—3. Acetonitrile—d3 mass spectrum at a source pressure of 0.8 mtorr and a repeller vol- tage of +0.5 V. m/e Intensity Relative to CD3CN+ (m/e=AA) l2 trace 1A 0.01 16 0.15 18 0.A 26 —-- 28 0.13 30 0.05 32 0.075 38 0.010 A0 0.050 A2 5.0 AA 1 A6 36.3 5A 0.035 56 0.25 58 A.13 38 set up for negative ion detection. However, attempts were made to detect the positive fragment from the reaction H(D)3CCN + hv + H(D)3c+ + CN‘. (2) The CH3+ appearance potential in reaction (2) may be cal- culated in the following way: + A.P.(CH3 ) = DO(H3C—CN) + I.P.(CH3) — E.A.(CN) (3) 01" +) 2A A.P.(CH = 5.26 eV + 9.8A2 ev25 - 3.82 ev26 (Aa) 3 A.P.(CH +) = 11.28 eV or 1099 A, (Ab) 3 where DO(H3C-CN) is the bond energy of the acetonitrile carbon-carbon bond, I.P.(CH3) is the first ionization potential of CH3, and E.A.(CN) is the electron affinity of the cyanogen radical. Searches were made for CH3+ + and CD 3 lines in the Hopfield continuum and the H2 pseudo con- in this wavelength region using atomic impurity tinuum as sources of radiation. The results, like those of Dibeler and Liston, were negative. However, as the light sources in this wavelength region are comparatively weak, absence of ion-pair formation is by no means con- clusively demonstrated. 39 B. The Photoionization Efficiency Curves Acetonitrile was the subject of two early photoioniza- tion (PI) studie527’28 in addition to the PIMS investigation mentioned earlier.23 Because these two early works lacked mass analysis, they were limited to the determination of the first ionization potential. In their complete PIMS study of acetonitrile, Dibeler and Liston23 reported only the first ionization potential and the appearance poten- tials of the three most intense fragment ions - H2CCN+, HCCN+, and CH2+. Moreover, they did not present photo- ionization efficiency curves for any of these ions. Their data were reported in the second part of a paper the primary objective of which was the determination of the heat of formation of the cyanogen radical (ON) by means of a PIMS study of HCN. Their inability to measure the appearance potentials of the CH3+ and CN+ fragments, which would have provided a second independent measurement of the cyanogen radical's heat of formation, probably caused them to lose interest in further work on acetonitrile. More- over, the first 70 A of the parent ion P.I.E. curve lie in a wavelength region where both the He Hopfield continuum and the H pseudo-continuum are of low intensity (950 A — 2 1020 A). This makes studies of the CH CN+ threshold region 3 very difficult. The goal of the present investigation was to provide A0 as complete a PIMS study of acetonitrile as possible, con- sidering the nature of the mass spectrum. It is unfortu- nate that the appearance potential and P.I.E. curve of the CN+ fragment could not be observed. However, it was possible to measure them for the CH3+ fragment ion, mark- ing the first time that they have been reported. A knowl- edge of the CH + appearance potential allows the calculation 3 of the carbon-carbon bond energy, DO(H3C-CN), of acetoni- trile. This is an important quantity to know when trying to unravel the energetics of the interstellar chemistry of acetonitrile. In addition the work was extended to include the deuterated compound. Preliminary experiments revealed that the P.I.E. curves of the deuterated ions were within experimental error iden- tical to those of their protonated counterparts. For this reason, only the threshold regions of their P.I.E. curves will be presented here, except for the deuterated molecular ion CD3CN+, where the complete curve is shown. All ap- pearance potentials were measured with 100 micron entrance and exit slits on the monochromator. This arrangement provided a photon bandwidth of 0.83 A. In addition, the P.I.E. curves of CH3CN+, CD30N+ and CH2CN+ from threshold to 600 A (20.7 eV) were obtained under these resolution conditions. The remaining ions were found to be of low intensity and to possess essentially featureless P.I.E. curves. Consequently, their overall P.I.E. curves were A1 obtained using 300 micron entrance and exit slits (2.5 A photon bandwidth). When 100 micron slits were in use, measurements were taken every 0.25 A. In the case of 300 micron slits, data were collected every 0.5 A. As men- tioned in Chapter II-B, the P.I.E. curves have been cor- rected for light source, sample pressure and detector drift, stray light detector drift and stray light. In addition, the overall P.I.E. curves were corrected for the wavelength dependence of the photon transducer sensitivity by means of the sodium salicylate - photomultiplier detec- tion system. With the exception of H2CCN+, the threshold regions of all of the fragment ion P.I.E. curves were mea- sured with an electron multiplier as the photon transducer. This was done in order to take advantage of the electron multiplier's lower levels of noise and dark current in these regions of weak ion intensity. The lack of cor- rection for the wavelength dependence of the electron multiplier sensitivity is not critical over the narrow threshold ranges (20 A to A0 A) and does not change the positions of the thresholds themselves. Figures III-l and III—2 depict the thresholds and first 80 A of the P.I.E. curves of CH3CN+ and CD30N+, respec- tively. In addition, the thresholds have been replotted on an expanded vertical scale in order to facilitate mea— surement of the first ionization potentials. Using the A2 Figure III-1. Threshold region of the H CCN+ P.I.E. curve. 3 (ARBITRARY UNITS) P.I.E. A3 0.940 NAVELENGTH (ANGSTROMS) X10 1.020 AA Figure III-2. Threshold region of the D CCN+ P.I.E. curve. 3 (ARBITRARY UNITS) 'P.I.E. 0.940 A5 WAVELENGTH )(10 1.020 (ANGSTROMS) A6 method of Guyon and Berkowitz32 , we may assume that the 0°K photoionization threshold is a step function and obtain the ionization potentials by extrapolating the linear portion of the P.I.E. curves to the horizontal axis and subtract- ing 0.A A to compensate for the half-width of the photon beam. The values so obtained are compared to the current literature values in Table III-A. The value of the aceto- nitrile first ionization potential reported here agrees well with previous determinations. A few remarks on some of the other studies are in order. The low resolution photoelectron spectroscopic study of Frost, Herring, Mc- Dowell and Stenhouse30 seems to be too low, especially since the value reported is a vertical ionization potential rather than the adiabatic value reported here. The value reported by Watanabe, Nakayama and Mottl27 is probably high, because the ion current was measured between two parallel plates across which a low voltage was applied. The sensitivity of such an arrangement is undoubtedly much lower than that of an electron multiplier. Finally, the pr601810ncd‘the value reported by Nicholson28 is probably unfounded since the width of the photon beam was A.A A in contrast to 0.83 A in the present case. The acetonitrile-d3 ionization potential reported by Lake and Thompson in a photoelectron spectroscopic study is the only value currently in the literature.29 It is probably higher than the present result because it A7 Table III-A. A comparison of the values for the first ionization potential of CH3CN and CD3CN as reported in this and other works. Molecule Reported Value Method Reference CH3CN 12.202i0.005 eV PIMS This work l2.l9i0.01 eV PIMS 23 12.20510.00A eV PI 28 12.2210.01 PI 27 12.20i0.01 PE 29 12.18 PE 30 12.21 PE 31 CD3CN 12.21u10.005 eV PIMS This work 12.2310.01 eV PE 29 A8 is a vertical ionization potential. It is interesting that acetonitrile-d3 was also studied in reference 30, but no result was reported. The very small difference in first ionization potentials between the protonated and deuterated compounds is expected. All of the photoelectron (PE) studies have revealed that the most weakly bound electron of acetonitrile is in the CN group n-orbital (3e).33’3l4 Three vibrational modes of the ion are excited upon removal of this electron. They have been assigned to the CN stretch- ing mode (v2), the symmetric CH(D)3 deformation mode (93) and the C-0 stretching mode (VA) with wavenumbers 2010 cm-1, 1 1A30 cm- and 810 cm'1 in H3CCN+, respectively.31 The same 1 1 and 920 om‘l, modes in the neutral are 2267 cm” , 1385 cm- respectively.35 The mode that appears in the PE spectrum with the most intensity is 02. In comparison, v3 is only very weakly excited. These results suggest that the n— orbital is C-N and C-0 bonding and weakly C-H antibonding. Thus, the substitution of deuterium for hydrogen should have little effect upon the first ionization potential. The difference is probably due to the small differences in the zero-point energies of the neutrals and ions. The region immediately above threshold (about 1017 A) to approximately 990 A (Figures III-1,2) is difficult to interpret. The small peaks correspond to lines in the H2 pseudo-continuum. They appear because the intensity of the scattered light is nearly equal to the intensity of A9 the incident beam. Apparently, the computer program cur- rently used for data treatment is unable to make the proper corrections in regions of such low light intensity. The overall shape of the P.I.E. curves in this region is sug- gestive of a vibronic threshold; however the onset is not clear. This is probably due to the fact that three vibra— tional modes are excited in this region (see the previous paragraph). If the CH3CN+ P.I.E. curve were simply an integral photoelectron spectrum, vibronic thresholds would appear at 1006 A (9“), 1000 A (03), 995 A (02) and 993 A (293). Weak autoionization structure superimposed on the P.I.E. curves may be masking these thresholds. The three peaks in the region from 983 A to 96A A, with a fourth suggested at 959 A, undoubtedly are part of a previously unobserved autoionizing Rydberg series that converges to an excited state of the ion. Their positions and the analysis proposed here are presented in Table III-5. Attempts were made to assign this series as converging to one of the vibrationally excited states of the ground state ion (72E).33’3u However, the quantum defects calculated from the position of each peak were all completely different (and sometimes absurd as well!). The peaks are not vibronic components of a single member of an autoionizing Rydberg series, because their separation diminishes with increasing energy. Furthermore, the positions of these peaks are un- altered upon deuteration with the exception of the third 50 Table III-5. Analysis of the autoionizing Rydberg series in CH CN and CD CN. 3 3 Assumed Peak Position Quantum Molecule Limit (observed) Defect (6) n CH3CN 13.1A ev29 983.0 A;101,729 cm"1 .92 6 -1 105’981 cm 971.0 A;102,987 om'l .95 7 96A.0 A;103,73A our1 .01 8 959.0 A;10A,275 om’l .98 9 Mean 6=0.97 CD3CN 13.1A eV 983.0 A;101,729 om’l .92 6 105,981 cm“1 1 971.0 A;102,987 om' .95 7 963.5 A;103,788 cm"1 .93 8 959.0 A;10A,275 em"1 .98 9 Mean 6=0.95 51 component. In H3CCN+, this peak lies between vibronic thres- holds at 968 A (Av3) and 962 A (Av2), whereas the A02 thres- hold (CD3 symmetric deformation) lies at a much longer wave- length in the deuterated compound. Perhaps the position of this component of the H3CCN Rydberg series is slightly distorted by the structure underlying it. The only reason- able choice for the series limit seems to be the threshold for the first electronic excited state of the molecular ion (A2Al).33’3u PE studies have placed the origin of this excited state at 13.1A eV (9A3.6 A) for both the protonated and deuterated ions.29 This corresponds to approximately the midpoint of the sharp step observed above the Rydberg series. The PE results and CNDO calculations by Frost, et al. suggest that the electron ionized is from the nitrogen "lone-pair" orbital.30 Although the intensities of the peaks decrease with energy, they are on a "baseline" that is increasing with energy. This is probably the result of the series being superimposed over several vibronic thres- holds of the ionic ground state. These thresholds have all been identified in the PE spectra (best seen in Reference 31). In addition to the autoionization structure, the broad structureless region in the P.I.E. curve between 959 A and 9A3 A probably results from further vibronic thresholds, masked by higher unresolvable members of the autoionizing Rydberg series. These observations, plus the magnitudes of the calculated quantum defects, lead to the assignment 52 of these peaks to part of an nso autoionizing Rydberg series, where n = 6-9. That is, the series is formed by the excitation of an electron from the nitrogen "lone-pair" orbital of the neutral molecule into an empty s-orbital with subsequent autoionization. It is difficult to test the validity of this assign- ment since the vacuum ultraviolet spectroscopic data on acetonitrile are poor and far from complete. The most recent vacuum UV absorption study is thirty-one years old and terminates at 1060 A.36 Three Rydberg peaks were ob- served and assigned an incorrect convergence limit of 11.96 eV. No hint of lower members of the series identified in this work was observed. Recently, an electron impact energy loss spectrum of acetonitrile was reported.37 The energy resolution was only 100 meV, but three Rydberg series con- verging to the ground state of the molecular ion were ob- served. Calculation of the energies of the n = 3, A and 5 components of the series identified here, suggests that the electron impact spectrum should display peaks at 9.93 eV, 11.69 eV and 12.31 eV, respectively. There are peaks in all three of these regions, but the energy resolution is insufficient to distinguish these features from the other series suggested by the authors as contributors to the broad bands which are observed.37 The most interesting aSpect of the electron impact spectrum is a broad, weak 2 A "band" that cuts-off at 13.11: eV, the threshold of the A’ l 53 state of the parent ion. The lower energy limit of the band is 12.55 eV. It thus encompasses the region of the series observed in this work. The resolution of the electron im- pact study was too low to resolve this feature, but it is interesting to note that a shoulder and peak on it cor— respond roughly in energy to the n = 6 and 7 members of the proposed nso series. The only disquieting aspect of the assignment proposed here is that the n = 5 component, which should fall within the rangecxfthe P.I.E. curve at 1006 A, is not observed. Provided that the proposed assignment is correct, a pos- sible explanation is that the n = 5 component has been com- pletely depopulated by predissociation into neutral frag- ments. The broadness and low intensity of the observed peaks suggest that the entire series (within the limits of the P.I.E. curve) is subject to rapid predissociation. Indeed, if this were not the case, the threshold of the AaAl state would be unobservable, because the series would 2 converge smoothly to the A|A limit without displaying 1 a discontinuity at that point.5a It seems reasonable to suggest that the n = 5 member of the series is crossed by a repulsive potential surface of the neutral molecule. In such a case, it would be completely depopulated while the higher energy members of the series would be pre- dissociated to a lesser extent due to their greater sepa- ration from the repulsive surface. 5A The H CCN+ and D CCN+ P.I.E. curves from threshold to 3 3 600 A are plotted in Figures III—3 and III-A, respectively. As with the threshold regions, they are completely identi- cal. Other than the steep step at 9A3.6 A (13.1A eV), no additional thresholds are observed. It is worth noting that from 600 A to 9A5 A the P.I.E.'s are remarkably similar to the HCN+ P.I.E. curve reported by Dibeler and Liston along with their acetonitrile data.23 The overall decrease in the P.I.E.'s at shorter wavelengths is probably due to the formation of fragment ions. Indeed, all of the higher energy bands (above the 13.1A eV band) in the PE spectrum are highly predissociated. Very weak autoionization struc- ture is present in the region from 820 A to 850 A. This structure may represent one or more Rydberg series whose limit is the second electronic excited state (BQE)33’3u of the parent ion. All of the PE spectroscopic studies have placed this threshold at 15.13 ev or 819.5 A. The weakness of this structure is another indication of the dom- inance of predissociation in this energy region. gchN+ and D CCN+ 2 The threshold regions of the H2CCN+ and D CCN+ P.I.E. 2 curves are pictured in Figures III-5. and III-6, respec- tively. In common with most fragment ion P.I.E.'s, they are approximately linear sloping curves with thermal tail- ing at threshold. Guyon and Berkowitz32 have shown that 55 .< oo 0 m 00 waocmopnp Eon0 o>L30 .m.H.m +200 m m 0:9 .m-HHH otsmfim 56 m- 0; Amzomhmozgso .m.H.0 +200 m Q 0:9 .e-HHH barman 58 m- 0; Amzomhmoz950 .m.H.m +zOomm one .NIHHH madman 68 - 0; Aazomewozhso .m.H.m + m mo 0:9 .ma-HHH oesmam 87 - 0; Amzomhmozmhdo .m.H.m + m mo 0:9 .0H-HHH chewaa 95 m- 0; Amzomhmoz0 m0.0000.m A20-0m20 o0 Aflos\H002 000.0000 >0 H.00mm.HH - A+m200 Hmma +020 Aaoexaaox m.0amm.m00 >0 00.0000.= u A0200 W20 Afloa\H002 m.0as.mmm0 >0 00.0002.0H - “+0200 mea +020 Afloaxamox 0.0Amsm0 >0 20.00ma.0fi - A+20020 W20 +2002 Aaos\H002 Ham.msv >0 H.000H.m u A20020-20 o0 Aaoexaeox 002.200 >0 H.0000.H - “200020 omma “Hosxfieox ma.0000.0000 >0 000.00000.NH - A+200020 omma +20002 zpfipcmsa OHEmcmcoEpoce GOH .mmfipfiusmsc afiemsmcoEpmSp 00>Hpma .NIHHH mHan REFERENCES 5a. 5b. 5c. 10. ll. l2. 13. REFERENCES Turner, Barry E., Scientific American, 228, 50 (1973). Gordon, M. A. and Snyder, L. E., Molecules in the Galactic Environment, Wiley-Interscience, New York, 1973. Solomon, Philip M., Physics Today, 26(3), 32 (1973). Herbst, Eric and Klemperer, William, Physics Today 39(6), 32 (1976). Chupka, W. A., "Ion-Molecule Reactions by Photoioniza- tion Techniques", in Ion-Molecule Reactions, Vol. I, J. L. Franklin, Ed., PIEnum Press, New—York, pp. 33-76 (1972). Chupka, W. A., "Photoionization and Fragmentation of Polyatomic Molecules", in Chemical Spectroscgpy and Photochemistry ig_the Vacuum Ultraviolet, Sandorfy, 0., Ausloos, P. J. and Robin, M. B., D. Reidel Pub- lishing 00., pp. A3l-A63 (197A). Chupka, W. A., "Effects of Internal Energy Effects on Ion-Molecule Reactions", in Interactign Between Ions agd MQlecules, Ausloos, P. J., Ed., Plenum Press, New York, pp. 2A9-262 (1975). Reid, N. W., Int. J. Mass Spectrom. Ion Phys. 6, l (1971). Samson, J. A. R., Sci. Tech. Aerosp. Rep. 13(12), (1976). Nicholson, A. J. 0., J. Chem. Phys. 39, 95A (1963). Wigner, E., Phys. Rev. 13, 1002 (19A8). Chupka, W. A., Berkowitz, J. and Rafaey, K. M. A., J. Chem. Phys. 50(5), 1938 (1969). Duncan, A. B. F., Rydberg Series in Atoms and Molecules, Academic Press, New York, 1 71. Fano, U. and Cooper, J. W., Rev. Mod. Phys. Ag, AAl (1968). Chupka, W. A., J. Chem. Phys. 10(1), 191 (1959). 99 1A. 15. 16. 17. 18. 19. 20. 21. 22. 23. 2A. 25. 26. 27. 28. 29. 30. 31. 100 Chupka, w. A., J. Chem. Phys. 51(5), 1936 (1971). Robin, M. B., Higher Excited States of Polyatomic Molefi les, Vol. II, Academic Press, New York, p. 292 197 ). Darland, E. J., Doctoral Dissertation, Michigan State University, 1977. Buttrill, S. E., Jr., J. Chem. Phys. 61(1), 619 (197A). Williamson, A. D., Doctoral Dissertation, California Institute of Technology, 1975. Samson, J. A. R., Techniques pf Vacuum Ultraviolet Spegtrgsggpy, John Wiley and Sons, New York, 1967. Franklin, J. L., Wada, Y., Natalis, P. and Heirl, P. M., J. Phys. Chem. 10, 2353 (1966). Energetics g: Gaseous Ions, in J. Phys. Chem. Ref. Data 6(1), Rosenstock, H. M., Draxl, K., Steiner, B. W., and Herron, J. T., Editors, 1977. Haney, Max and Franklin, J. L., J. Phys. Chem. 13(12), A328 (1969). Dibeler, V. H. and Liston, S. K., J. Chem. Phys. 58, A755 (1968). Okabe, H. and Dibeler, V. H., J. Chem. Phys. 59, 2A30 1973). Herzgerg, G. and Shoosmith, J., Can. J. Phys. 35, 523 (195 ). Binding Energies lg Atomic Negative Ions, in J. Phys. Chem. Ref. Data 3(3) Hotop, H. and Lineberger, W. 0., Editors, 539 (1976). Watanabe, K., Nakayama, T. and Mottl, J., J. Quant. Spectry. Radiative Transfer g, 369 (1962). Nicholson, A. J. C., J. Chem. Phys. A3, 1171 (1905). Lake, R. F. and Thompson, H., Proc. Roy. Soc. Lond. A. 3110 187 (1970). Frost, D. C., Herring, F. G., McDowell, C. A. and Stenhouse, I. A., Phys. Letters A, 533 (1970). Turner, D. W., Baker, C., Baker, A. D. and Brundle, C. R., Molecular Photoelectron Spectroscopy, Wiley- Interscience, LOnHEn, 1970. 32. 33. 3A. 35. 36. 37. 38. 39. A0. A1. A2. A3. AA. A5. A6. 101 Guyon, P. M. and Berkowitz, J., J. Chem. Phys. 99 (A), 181A (1971). Yee, D. S. C. and Brion, C. E., J. Electron Spec. and Rel. Phenomena Q, 313 (1976). Baybritt, P., Guest, M. F. and Hillier, I. H., Mol. Phys. 25(5), 1025 (1973). Shimanouchi, T., Tables g£_Molecu1ar Viggational Fre- quencies, Consolidated Volume I, NSRDS-NBS 99, 8A 1972). Cutler, J. A., J. Chem. Phys. 6(2), 136 (19A8). Stradling, R. S. and Loudon, A. C., JCFTBS 13(5), 623 (1977). Dibeler, B. H., Walker, J. A., McCulloh, K. E. and Rosenstock, H. M., Int. J. Mass Spectrom. Ion Phys. 1, 209 (1971). Wagman, D. D., Evans, W. H., Parker, V. B., Halow, I., Bailey, S. M. and Schumm, R. H., NBS Tech. Note 270-3 (U68. Government Printing Office, Washington, D.C., l9 ). Pottie, R. F. and Lossing, F. P., J. Am. Chem. Soc., 93. A737 (1961). Haney, M. A. and Franklin, J. L., J. Chem. Phys. 99 (9), A093 (1968). Guyon, P. M., Chupka, W. A. and Berkowitz, J., J. Chem. Phys. §9(A), lAl9 (1976). Chupka, W. A., J. Chem. Phys. fifi, 2337 (1968). Herzberg, 0., Can. J. Phys. 99, 1511 (1961). Berkowitz, J., Chupka, W. A. and Walter, T. A., J. Chem. Phys. 99(A), 1A97 (1969). McCulloh, K. E. and Dibeler, v. H., J. Chem. Phys. 53(11). nuns (1976). -.