mzAnou STUDY OF NO, Thesis for the -D9gfee"of.Ph-. D. A . Wm . . m, ‘ SM . . V M ‘ . . M W... M . A, N m m M TDIO v PHO WW. LIBRARY Michigan State University cfig'ir ABSTRACT PHOTOIONIZATION MASS SPECTROMETRY: A STUDY OF N02 AND NO BY Paul Charles Killgoar, Jr. The discrepancies among the reported values for the ionization potential of N02 led to a reinvestigation of this molecule by means of photoionization mass spectrometry. A careful study of the threshold region revealed considerable autoionization structure. The measured appearance potential of N02+ is 9.62 eV, which combined with other data gives 9.25.: I.P.(N02) :.9.62 eV. The results of this study sug— gest that the value is closer to the upper limit. Fragmentation of N02 into NO+ and 0 also was studied and the dependence of the dissociative lifetime of meta- stable N02+* on the quantum state of its bending vibration was investigated. The observed autoionization structure of the N02+ photo- ionization efficiency curve has been assigned to Rydberg series converging to higher ionization potentials. The photoionization efficiency curve of NO+ was also obtained. Considerable fine structure was observed in the Paul Charles Killgoar, Jr. threshold region (9.25—10.0 eV) which is attributed to vibrational autoionization. This fine structure has been assigned to Rydberg series converging to the excited vibra- tional levels of the ground NO+ ion (12+). The autoioniza- tion structure above threshold has been correlated with existing Rydberg series of the NO molecule. PHOTOIONIZATION MASS SPECTROMETRY: A STUDY OF N02 AND NO BY Paul Charles Killgoar, Jr. A THESIS Submitted to Michigan State University in partial fulfillment of the requirements of the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1972 ACKNOWLEDGMENTS I wish to express my sincere appreciation to Dr. George Leroi for his guidance and encouragement without which this work would never have been completed. I also wish to thank Dr. Joseph Berkowitz and Dr. William Chupka of Argonne National Laboratory for allowing me access to their instrument, their patience with my mis— takes and their sincere interest in my education. I would like to thank the members of the Molecular Spec- troscopy Group and Dr. Daniel O'Hare for their friendship and many stimulating discussions, scientific and other, in the course of this work. Financial support from Michigan State University, the Office of Naval Research and the Associated Midwest Universi- ties of Argonne "Thesis Parts" program is appreciated. Finally, I wish to thank my wife for her patience and understanding during these past three years. ii TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . . . . The Instrument . . . . . . . . . . . . . . . The Ionization of N02 . . . . . . . . . . . . The Fragmentation of N02 . . . . . . . . . . Ion nSO O O O O O O O O O O O O O O O O A new Rydberg series converging to 18.86 eV: 2b2 _>' npU o o o o o o o o o o o o o o o o A possible Rydberg series converging to 14 .07 eV: 1.62 ->’ inr o o o o O o o o o o o A Rydberg series converging to 18.86 eV: 2b2 _> ndC o o o o o o o o o o o o o o o 0 Comparison of experimentally observed Rydberg series with series I reported by Tanaka and Jursa: 2b2 —% an'—> 18.86 eV . . . . . . iv Page 24 33 42 45 45 55 65 65 73 76 77 78 79 LIST OF TABLES (Cont.) TABLE XIV. XVI. XVII. XVIII. XIX. XXI. XXII. XXIII. XXIV. Comparison of experimentally observed Rydberg series with series III reported by Tanaka and Jursa: 2b2 —% nso -> 18.86 eV. . . . . . . Comparison of experimentally observed Rydberg series with series IV reported by Tanaka and Jursa: 2b2 _>' ndé _> 18.86 eV 0 o o o o 0 Assignment of autoionization structure on the first vibrational step of NO . . . . . . . Assignment of autoionization structure on the second vibrational step of NO . . . . . . Assignment of autoionization structure on the third vibrational step of NO . . . . . . . A Rydberg series of NO: 1W -¢ nso (—w a3Z+ at 15 .65 eV) 0 O O O O O O O O O 0' O O O O A Rydberg series of NO: 50 —> npO (~+ b3TT' at 16.56 ev) . . . . . . . . . .V. . . . . A Rydberg series of NO: 1w -+ nso (—> w3A at 16.86 ev) . . . . . . . . . . ,,. . . . A Rydberg series of —+ nso (—> b'3Z- at 17.59 ev) . . . . . . . . . . . . . . . A Rydberg series of —> nso (—> W'A at 18 .07 eV) . O O Q . O O . . . . ' . . O . A Rydberg series of NO: 50-—> npO (—> A'TT- at 18.32 ev) . . . . . . . . . I ' Unassigned autoionization structure in NO . Page 81 81 91 93 93 99 99 100 101 102 103 105 j I» h FIGURE 10. 11. 12. 13. 14. LIST OF FIGURES The photoionization mass spectrometer . . . . Schematic drawing of the cold cell . . . . . The collision source . . . . . . . . . . . . Walsh diagram for XYZ molecule . . . . . . Threshold region of N02+ photoionization efficiency curve . . . . . . . . . . . . . . Schematic representation for the autoioniza- tion structure in the threshold region of N02 Photoionization efficiency curve for NO+ from N02 in the threshold region (1000 - 920 R). . Potential energy curves representing fragmenta— tion Of N02 0 o o o o o o o o o c o o o o o o Photoionization efficie cy curve for the meta- stable production of NO from N02 . . . . . . Voltage scan of metastable peak at m/e = 19.57 . + . . . Comparison of NO and N02+ photoronization efficiency curves . . . . . . . . . . . . . . Rdeerg series Of N02 0 O o o o o o o o o o o Photoionization efficiency curve of NO+ from N0 in the threshold region . . . . . . . . . The photoionization §fficiency curve of NO+ from N0 (1350 — 600 ) . . . . . . . . . . . Page 15 19 27 31 36 51 57 60 68 84 89 97 INTRODUCTION A photoionization mass spectrometer is simply a mass spectrometer with the conventional electron bombardment ionization source replaced by the radiation emanating from a vacuum ultraviolet monochromator. This modification in- creases the versatility and precision of the instrument. One of the main advantages is the energy resolution avail— able using a monochromator instead of the electron gun. In an ideal case the energy spread of an electron beam is approximately 0.1 eV, but with appropriate slits a mono- chromator can give an energy resolution of 0.01 eV. A second advantage which is much more useful is the energy dependence of the photoionization cross section at threshold. Wigner and others [1—3] have shown that the threshold law for direct ionization is En_1 a cross section, (1) where n is the number of electrons leaving the collision complex. It is obvious that for electron impact the cross section is linear with energy; that is, the cross section rises very slowly from the baseline. This makes it very difficult to determine exactly where the onset actually occurs. For photon impact there is no energy dependence. 2 That means at the threshold the cross section ideally rises to some maximum value, after which it should vary very slowly with increasing energy, until the next threshold. This makes it much easier to determine the onset for ioniza- tion (the appearance potential). The photoionization cross section is defined as the number of ions produced per photon absorbed Oi = O Ni/(IO ' I): (2) where oi is the photoionization cross section, 0 is the absorption cross section, Ni is the number of ions produced per second, I is the transmitted photon intensity and Io is the incident photon intensity. To measure this quantity with a single beam instrument is very difficult since I and Io cannot be measured simultaneously. To circumvent this, a quantity called the photoionization efficiency (P.I.E) is used instead. This is the number of ions pro— duced per photon transmitted P.I.E. = Ni/I . (3) It can be shown [4] that for absorptions of less than 10% of the total incident beam the ratio of the photoionization cross section to the photoionization efficiency is nearly unity. At the outset, this research project involved the de- sign and construction of a photoionization mass spectrometer. This instrument was to consist of a one-meter off-plane 3 Eagle mount vacuum ultraviolet monochromator coupled to a quadrupole mass Spectrometer, each of which had been ini- } tially designed and constructed at the Frick Chemical Laboratory at Princeton University. One major difficulty encountered with this instrument was the limited flexibility of the vacuum ultraviolet monochromator. In the off-plane } Eagle design the entrance and exit slits are located sym— ' metrically above and below the Rowland circle [5]. When the instrument was initially designed the distance between the centers of the entrance and exit slits was only one ._ -....-._—q—-L_.___- . - inch. This distance was increased to four inches by modi- fying the front plate of the monochromator. Even with this adjustment, however, there was insufficient room to attach the quadrupole mass spectrometer to the monochromator and still have enough room remaining to couple a light source to the entrance slit. It thus appeared that the experimental research which was to be the foundation of this dissertation would have to be significantly delayed while the photoionization instru— { ment at Michigan State was extensively redesigned and con- structed. In light of this tenuous future, a better instru— ment was sought, and the main portion of the experimental work described in this thesis was performed at the Argonne National Laboratory with the cooperation and guidance of Drs. Joseph Berkowitz and William Chupka of the Physics Division. 4 Initially the experimental portion of the research was to include a determination of the ionization potential of N02, for which a wide dispartiy existed in the reported values. It was hoped that a new investigation by various independent techniques would yield a consistent value. The methods to be used included ion-molecule reactions, frag- mentation by photon-impact, and direct ionization of the N02 molecule itself. The ion-molecule reaction chosen was that between NO+ and N02. Possible choices for the frag- mentation experiments were CH3N02 and nitric acid. The criterion used in choosing molecules for the fragmentation experiments was that the fragmentation into N02+ upon photon impact was likely to be the first process. This point was important because to determine an accurate ionization po— tential from a fragmentation process an accurate appearance potential for the fragment of interest is necessary. Thus if the fragment is not formed as the first process it will have to compete with the primary fragmentation process, and its observed appearance potential will be somewhat higher than the true value. To complete this study of the N02 molecule, the fragmentation into NO+ and O was investigated in detail, including the study of a metastable fragmenta- tion. A Rydberg analysis of the autoionization structure was also performed. During the performance of the various experiments a photoionization efficiency curve of NO+ was obtained. Care- ful examination of the threshold region revealed some 5 unusual structure. The molecule NO has been well studied; however, no interpretation has been put forward for the structure (indeed it has only been mentioned once in the open literature). This fine structure has been interpreted in terms of Rydberg series converging to excited vibrational levels of the ground ionic state. The autoionization structure above threshold has been correlated with existing Rydberg series of the NO molecule. EXPERIMENTAL The_;nstrument The photoionization efficiency data reported in this work were obtained using the one meter instrument at the Argonne National Laboratory [6,7]. This consisted of a one meter vacuum ultraviolet monochromator built by McPherson Instrument Co. (a modified Model 225) coupled to a 600 magnetic-sector mass spectrometer as illustrated in Figure 1. The light sources employed were the hydrogen many-lined source generated by a continuous d.d. discharge and the continua of argon and helium, which were produced by a pulsed d.c. discharge. These lamps provided a range of photon energies from 9.0 to 20.7 eV (1300 R - 600 A). In order to utilize the higher energy photons (above 11 ev) the lamp was run windowless. To do this three stages of differential pumping were employed, in which the third stage was a pump on the monochromator chamber. With the argon source a typical Operating pressure within the lamp was 150 torr and in the monochromator the pressure was main— tained at or below 10-5 torr by the differential pumping. The grating used in these experiments had 1200 lines/mm and a dispersion of 8.3 R/mm. Entrance and exit slits of .Houofionuommm mmmE GOHDMNHCOAODOQQ 038 .H musmflm .H onsmflm m0hdw¢om200202 > D 33:04) quoaamo(og)2 (ou)2(og)2(ou)2 (”u)‘ (vg)4 (wu)° '1 l 1 1 l ] 3 1 ' N02(134°) (1a1)2 (2a1)2 (1b1)2 (1a2)2 (3b2)2 ’ (1b2)2 (2b2)2 (3a1)2 (4a1)1 It is the presence of the extra electron in the 4a1 orbital which contributes most to the change in geometry. Walsh [27] has discussed the influence of the different or— bitals on the geometry of XY2 molecules, and as may be seen from the appropriate Walsh diagram (Figure 4), it is the 4a1 orbital which is most stabilized by the departure from linearity. The first ionization of N02 results from the removal of this 4a1 electron, thus leaving a linear ion isoelectronic with C02. In order to accurately determine the adiabatic ioniza— tion potential of a molecule by means of photon interaction techniques it must be possible to populate the ground vibra— tional level (all vibrational quantum numbers equal zero) of the ion. However, in the case of N02 the difference in geometry between the neutral molecule and the ion is so drastic that it is virtually impossible to populate the v' = 0 level of the bending vibration of the NO2+ ion 26 Figure 4. Walsh diagram for an XY2 molecule. 27 4b2 in, 30‘ \I“ “‘\‘\“"“‘:1ts A ; , In“ Ibllrlr. .. -1. NEE; 2c:I \‘i 209 1b2 10; lo, 106 90 — — 1é0° Figure 4. 28 directly. There is, however, an indirect method by which one still may accurately determine the adiabatic ionization potential, and that is to populate the lower vibrational levels of the ion by means of autoionization. This process relies upon the existence of another electronic state, usually a Rydberg level of the neutral molecule which con- verges to a higher ionization potential, overlapping the ionization continuum of interest. Essentially what is in— volved is the interaction of a discrete state with the ionization continuum of some electronic state. If a transi- tion is made to the discrete state a radiationless transi— tion (autoionization) to the continuum can occur. The Franck-Condon factors coupling the discrete to the con- tinuous states also play an important role in the transi- tion probability [31]. The total transition probability to the ionic state from the ground state then depends on two Franck—Condon factors instead of just one. There are three different types of autoionization, the two important forms being electronic and vibrational and a less important type being rotational [32]. The most commonly observed autoionization structure is attributed to electronic auto— ionization. Rydberg states converging to excited states of the molecular ion involve an excited ion core and a Rydberg electron. When the Rydberg electron is in the vicinity of the core electronic autoionization may occur resulting in the ejection of an electron and a decrease in the electronic energy of the core. The perturbation which accounts for the 29 mixing of the Rydberg state with the ionization continuum is a configuration interaction [33,34]. The second form of autoionization may occur when the discrete state is a vibrationally excited Rydberg state. That is, the electronic state of the ion core of the Rydberg level is the same as the electronic state of the ion to which the neutral mole- cule autoionizes; however, the core has some vibrational excitation. This excitation is then transferred to the Rydberg electron and the state makes the transition into the ionization continuum. The perturbation which is respon- sible for the mixing of states in this case results in the breakdown of the Born-Oppenheimer approximation [31,35]. Autoionization structure is usually exhibited as a peak or series of peaks superimposed upon the ionization continuum. However, this need not be the case. Fano [33] has discussed line shapes due to autoionization and has shown that it is possible to have a valley or "window reso- nance" in the ionization continuum. The threshold region of the photoionization efficiency curve of N02+ is shown in Figure 5. As can be seen there is a significant amount of structure, all of which must be attributed to autoionization of Rydberg levels. Two vacuum ultraviolet absorption measurements in this region Show the existence of one member (n = 3) of a Rydberg series converging to 12.86 eV. Nakayama, 33.31. [11] and Tanaka and Jursa [15] have Shown that this state has vibra- . . -1 tional structure With an average spaC1ng of 640 cm . The ., . v-1 ,7 . [infilREMahfuéirhNIU-un .. ix... inuirltkaPPr 30 .w>nso AUGOHUSMMO COHDMNHGOHODOLQ NOZ mo QOHmwM UHozmeLE .m ousmflm + 31 .m onsmflm Q: :623 m>§> no.2 . new. _ .- .nlnmw - . new. . new. . new. . 3+9 . mum. x7 _ >. 22. T) N ell p.50 000 lln V n O n... 13 FullilriIiLi _ _ h L D r _ TI. 780 ~00 I. L L [Iii l_ 1: ll|11||L T! .533 .l r _ _ _ L - L _J (suun ‘QNYS‘IH 32 vibrational spacing is assigned to the v2 bending vibra- tion of N02* (electronically excited N02). This member of the series has been observed to autoionize in the photoioni— zation efficiency curve and is identified as T & J n = 3 in Figure 5. The observed vibrational spacing in the pres- ent case is 603 i 50 cm-l, which is within the experimental error of the measurements. Table II compares the peaks ob- served by Tanaka and Jursa with those of this experiment. It may be noted that the transition to v' = 0 reported to be at 1293 R‘by Tanaka and Jursa was not Observed in these experiments. To scan this region more carefully a special experiment was performed utilizing 10 minute count— ing times per point. Again no peak at 1293 A could be de- tected. Since it was Observed Optically it is not immedi- ately obvious why this transition is not Observed in the photoionization efficiency curve. There are two different rationalizations for this phenomenon. The first interpretation is the obvious conclusion that the V' = 0 level of this Rydberg state is below the adiabatic ionization potential of the N02 molecule. If this is the case, the v' = 0 level does not overlap the ionization continuum in any way and therefore no structure will be observed. This is not, however, the only interpre— tation. Even if the v' = 0 level is above the adiabatic ionization potential of the molecule, the Franck-Condon overlap of this level with the ionization continuum may be so small that the autoionization transition rate is too 33 Table II. Comparison of Optically observed peaks to observed autoionization peaks. V' Tanak: and Intensitya This Work Jursa (A) X 0 1292.8 6 -___ 1 1282.8 9 1282.30 2 1272.2 8 1272.80 3 1261.6 4 1262.50 4 1251.3 1 1253.55 5 1243.43 6 1234.30 See reference 15. 34 small to be Observed. There is a strong likelihood that this is really what is happening in the present case. The Rydberg series in question has been assigned by Edquist, gt El: [22] as one converging to the first excited ionic state at 12.86 ev. This excited state has been attributed to the removal of a 3b2 electron from N02. Upon examination of the Walsh diagram (Figure 4) it can be seen that the 3b2 electron stabilizes the linear config- uration of the molecule. Removal of this electron should result in the excited electronic state of the N02+ ion being more bent than the ground state of the molecule because of the increased influence of the 4a1 electron. Rydberg levels are expected to exhibit the same vibra- tional structure as the state to which they converge, this being especially true of the higher members of a series [32]. It is therefore probable that the n = 3 member of the Rydberg series converging to 12.86 eV is at least as bent as the ground state of the molecule. This point is borne out by the fact that intensity measurements by Tanaka and Jursa (see Table II) show the v' = 0,1,2 and 3 levels to be about equal in intensity, thus indicating that the Franck- Condon factors are favorable for the optical transitions. For this reason, the vibrational ground state of this Rydberg level is unlikely to overlap the ionic ground state very much more than the molecular ground state does. These points are illustrated schematically in Figure 6. 35 Figure 6. Schematic representation for the autoioniza— tion structure in the threshold region of N02. (:6 i .15.: - >0¢m2m \, -_--._——_...— O ANGLE - 04%“ Figure 6. 37 If a transition is made from the molecular ground state to a Rydberg level with similar geometry, such as that denoted by I in Figure 6, the molecule may make the transition, denoted by II in Figure 6, into the ground ionic state. If this occurs a peak will be Observed in the photoionization efficiency curve at the energy corre— sponding to transition I. If on the other hand a transi- tion is made from the molecular ground state to the Rydberg level denoted by III, Ufiie is no way that this level can make a transition into the ionic state, except possibly by a tunneling process. If tunneling does occur the cross section for the process is so small it cannot be detected in these experiments. It seems likely that this is the best explanation for the absence of the peak at 1293 8. However, even with the above arguments for the unfavorable Franck-Condom factors for the autoionization transition one cannot exclude the possibility that the 1293 A level is lower than the adia- batic ionization potential. In Figure 5 two other vibrational progressions with the same average spacing can be seen in the 1300 - 1230 8 region. The assignment of these progressions is not im— mediately clear. It may be that these progressions are converging to the same excited state Of the N02+ ion as the progression identified by Tanaka and Jursa [15]. One argu- ment against this possibility is the complete absence of any similar structure in the next member of the Rydberg 38 series; that is, only one progression is seen in this region. Another possible explanation for the existence of these progressions is that they can be due to vibrational autoionization of Rydberg series converging to vibrationally excited states of the ground electronic state Of the ion. Although this is a good possibility, it remains unprovable at present. The possibility that these peaks are due to vibrational hot band structure can almost certainly be ruled out, because the structure did not diminish when the gas was cooled in some experiments. At the temperature of these experiments only ~ 2% of the molecules are expected to be in vibrationally excited states, and with the low cross sections Observed this will never be an important contribution. As far as evaluation of the ionization poten— tial is concerned, however, these progressions are of little relevance. The lowest observable ion current due to photoioniza— tion occurs at 9.62 i 0.01 eV (1288.8 3). From the spectra in Figure 5, one predicts that the next peak at longer wavelength should occur at 1293 A, and this is not seen. Therefore, from this investigation it appears that the ionization potential of N02 cannot be any greater than 9.62 i 0.01 eV. This value is lower than any of the pre- vious investigations made using photoionization techniques, and lower than all other reported values except one, that being 8.8 eV from the photoelectron investigations by Natalis and co—workers [21,36]. The higher sensitivity of 39 the instrument employed in this work accounts for the dis- crepancy between this result and the other photoionization results; however, it is difficult to explain why the photo— electron result of Natalis, EE.§£° [21,36] is so much lower, and indeed, so much lower than the value Obtained by other photoelectron experiments. In Natalis' photoelectron work resonance lines of Ar (1048 - 1067 A) were used as the excitation source. In order for the photoelectron result to yield a value lower than the photoionization measurements, the exciting line must coincide with an autoionizing level which can populate the v‘ = 0 level of the ionic ground state. Upon examination of the photoionization efficiency curve in the region of the 1067 A exciting line very strong autoionization structure is Observed. This autoionization structure has been attributed to the n = 4 member of the Rydberg series converging to 12.86 eV, and as stated earlier the molecular geometry in this Rydberg series is expected to be more bent than the molecular ground state, and for this reason it is unlikely that the autoionizing level could populate the v' = 0 level of the ion. The most damaging evidence against the value of 8.8 eV for the ionization potential of N02 is the work reported by Fehsenfeld, Ferguson and Mosesman [37]. This experiment involved the charge- transfer reaction between N02+ and NO. A buffer gas was employed to make certain the N02+ was not in an excited state when it reacted with the NO. The N02+ initially was produced by charge transfer from Ar+. It was observed that 40 the charge transfer was exothermic and thus it must be con- cluded that the minimum value for the ionization potential of N02 is greater than 9.25 i 0.02 eV, the ionization poten— tial of NO. Several other experimental attempts were made to ob— tain a consistent value for the ionization potential of N02 gig different approaches. One of these was from the frag— mentation of CH3N02. Kandell [17] reported an electron im- pact value for the appearance potential of N02+ from CH3N02 of 9.91 eV, and it looked like a promising molecule to study. Investigation of the fragmentation showed that the first process was not dissociation into CH3 and N02+ as hoped, but that these products were formed by the third or possibly fourth fragmentation process. With this in mind, the value of 10.0 ev for the ionization potential of N02+ Obtained from the present CH3N02 photoionization experiment is within the range of interest, but still too high to be confirmatory. Another experiment was performed in which the endo— ergic charge—transfer between NO+ and N02 was attempted. The rationale for the experiment was based on the observa— tion by Fehsenfeld, gt _1. [37] of the exothermic charge— transfer and the hope that the reverse charge transfer could be Observed. The experiment proved futile as there was no significant amount of N02+ produced by the reaction. There- fore, although the threshold for the reverse reaction is low 41 the cross section is apparently too small to be measured with our present arrangement. One further experiment that was performed was the photodissociation of HNOS. Good photoionization efficiency curves were obtained for this process, and it appears that fragmentation to the OH radical and N02+ is the first process. Using readily available thermodynamic data [38] for the heats of formation of the fragments of HN03, which are listed in Table III, the heat of reaction for the process: HNO3 ———> N02 + H0 (4) can be calculated. From equation (5), = AHO NO + AHO HO — AH° HNO 5 AHRXN £298( 2) f298( ) f293( 3) ( ) and using the values from Table III, the heat of reaction is calculated to be 12.087 eV. Using this value, the ap- pearance potential, and equation (6) I.P.(N02) = A.P.(N02+) — AHRXN (6) the value for the ionization potential of N02 is calculated to be 9.94 eV. Again, this value is higher than the photo— ionization result. A possible explanation for this could be the existence of a potential barrier to the fragmentation, in which case more energy is needed to dissociate the mole- cule, and thus the fragments leave with excess kinetic or internal energy. Another possibility exists which requires 42 Table III. Heats of formation for HNO3, C2H5ON02 and their fragments. Ang AHO Fragment 293 £298 Ref (Kcal/mole) (ev) HN03 -32.28 -1.40 38 C2H5ON02 -36.82 -1.60 40 HO 9.31 0.404 38 C2H5O - 8.5 —0.369 40 N02 7.93 0.344 38 43 further investigation; it is possible that the heat of form- ation Of one of the fragments or the parent itself is in error. It is highly unlikely that the heats of formation of N02 and OH are poorly established, thus indicating that the heat of formation of HNO3 is incorrect. Evidence which supports the observed direct ionization potential of N02 from this study is obtained from the fragmentation pattern of C2H5ON02 reported by Victor Fong [39]. From the reaction hv + C2H50N02 ——+ C2H50 + N02+ (7) Fong reported the appearance potential for N02+ to be 11.26 i 0.01 eV. Again as in the case of HN03, using readily available thermodynamic data [38,40] which are presented in Table III the heat of reaction, AHRXN’ to form the neutral products can be calculated using equation (8): = AHO c H + AHO NO - AHO c H ONO . 8 AHRXN £293( 2 50) £298( 2) £298( 2 5 2) ( ) The AHRXN Obtained is 1.56 ev. Employing the formula given in equation (6), the ionization potential of N02 is found to be 9.7 i 0.1 ev. This process is not the first fragmentation and as a result the value calculated is prob— ably high; the true ionization potential should be somewhat lower. Thus, in conclusion it appears that the ionization potential of N02 is greater than 9.25 i 0.02 eV and less 44 than or equal to 9.62 i 0.01 eV. There are indications that the value is not very far below the upper limit. The Dissociation of N02 The dissociative ionization of N02 into NO+ and an O atom in a mass spectrometer has been reported by Dibeler, ‘§£_§l, [23], Weissler, gt_al. [25], Kiser and Hisatsune [19] and Collin and Lossing [16]. As with the ionization poten- tial of N02, there is disagreement in the reported appearance potentials of NO+. However, as may be seen in Table IV, there is a smaller range in the reported values. The appearance potential of NO+ from N02 can be calcu- lated for the reaction written in equation (9): 2 +1+ 3 .— N02( A1) + hv ———> NO ( z ) + o( P) + e (9) by using standard thermodynamic values given in Table V, the ionization potential of 9.25 eV for NO [41] and equation (10): (NO) + AHO A.P.(NO+) = I.P.(NO) + AHO f0 f0 (0) - AH%O(NOZ).(10) From the calculation the appearance potential of NO+ is found to be 12.37 eV, based on the 00K values. Dibeler and coaworkers [23], as well as Kiser and Hisatsune [19], seem to have found the threshold, within the quoted experi— mental errors Of the techniques employed, where it is pre- dicted to be. 45 Table IV. Reported threshold values for formation of NO+ from N02. . Threshold Reference Technique (ev) Electron Impact 10.1 i 0.2 16 Electron Impact 12.48 i 0.43 19 Photoionization 12.34 23 Photoionization 11.3 i 0.04 25 Table V. Heats of formation of N02 and its dissociation products.a AHO AHO AHO AHO Fragment f0 f0 f298 f298 (Kcal/mole) (ev) (Kcal/mole) (ev) N02 (2A1) 8.59 0.3725 7.91 0.3430 NO (12+) 21.46 0.9306 ’ 21.58 0.9358 0 (392) 58.989 2.5581 59.559 2.5828 aFrom reference 42. 46 As stated in the portion of the experimental section concerned with the fragmentation of N02 (Chapter II), NO impurity in the sample was the single most Significant problem in these experiments. Even with the provisions outlined in Chapter II to remove this NO impurity, approxi- mately 5% NO impurity remained. The background ion contri- bution to the efficiency curve from the impurity was sub- tracted from the raw data by the computer. Figure 7 shows the photoionization efficiency curve for the NO+ fragment from N02 in the threshold region (1100 — 920 A). There are two significant aspects to this curve; the first is the possible existence of a very small step 12.37 eV, the thermodynamic threshold, and the second a much stronger series of steps starting at approximately 12.90 eV. This result is in very good agreement with that reported by Dibeler and co-workers [23]. They reported a small peak at 12.34 eV, but not a continuous dissociation, and also observed the strong rise at 13.01 eV. Examination of the photoionization efficiency curve of NO+ from N0 in this region shows the existence of a very strong autoioniza- tion peak at 12.34 eV, and it is probable that what Dibeler and OUIIS believe to be NO+ from N02 is in fact NO+ from NO impurity. As can be seen in Figure 7 the step observed at 12.37 eV is very small and it may well be the result of NO impurity not sufficiently corrected for in the data. Al— though this step could be due to dissociation of the N02 molecule, there is no conclusive evidence to make a decision 47 .Am one -oooHv aoemon OHOLOOHLD mnu CH NOZ Eoum +02 How O>Hdo xocOHOHmmm COHDONHQOHODOLO .h Onsmflm coo. 48 .>. und— >o era, _ .b musmam $502533 owe owe _ >. omdp _ 62% 2050335530 > _ _ O — ON— mNp . _ o _ N .9 map vamp P _ owe oflo '3'l'd 49 on this point.. A possible explanation for this step will be discussed in more detail subsequently. The interesting region of this efficiency curve is from 12.96 to 13.33 eV. The sudden rise in this region can be correlated with an excited ionic state observed by Brundle, 23 a1. [12] and Edquist, gt_ 1. [22] in their photoelectron Spectra of N02, which is located at 12.86 eV and has a vibrational progres- sion associated with it. The vibrational structure Observed in this photoelectron band is shown in Figure 7 and has been assigned as a progression in v2, the bending mode of N02+*. It can be seen in Figure 7 that the v5 = 0 and v; = 1 vibrational levels Show only slight evidence of dis— sociation while V; 2.2 are strongly dissociated. It is clear that this dissociation can occur only via some form of predissociation mechanism. In order to discuss this predissociation there are some important points which should be made. These are represented schematically in Figure 8. The first is that the ionic ground state of N02+ is a singlet state. Utilizing the WigneréWitmer correlation rules [32] and using the isoelec- tronic molecule C02 as an example, this lowest singlet state does not correlate with ground state dissociation products because of the violation of the spin conservation rule. However, the first excited ionic state at 12.86 eV, the pre- dissociating state, is a triplet and can correlate with the ground state dissociation products. A second aspect of this dissociation is that the ground state dissociation limit 50 Figure 8. Potential energy curves representing fragmentation of N02. 51 NOTE) +000) 14— / + I “02‘352’ \X 13- ‘I’ Nofi‘fn 0(3P) 12- ‘ ‘ ¥ H- N02 10- 3 NOZCAO >- 9-— 0 , E 2 / 2 Lu 0— NOQ ON"'O Figure 8. 52 (12.37 eV) is lower than the zero point energy of the potential curve for the first excited ionic state at 12.86 ev. The next dissociation limit corresponds to the production of excited oxygen atoms and will be located 1.967 ev above the ground state dissociation limit [43]. In order for the state at 12.86 eV to be dissociated it must go to ground state products at 12.37 ev and thus there must be some form of potential barrier to this process. Again, if one uses the isoelectronic case Of C02 as a guide, it is expected that the ground state dissociation products can be corre— lated with a repulsive curve (reference 30, page 431). If this state in N02+ is not linear as it is in C02, the re— pulsive curve originating from it will be split due to the symmetry change giving two repulsive curves. For the dis— sociation to occur it must go through the antisymmetric stretch. Since there are at least six bound vibrational levels in the bending mode, there must be at least zero point energy in each of the other normal modes. The potential barrier which must exist to some measure is probably due to the interaction of one of the repulsive surfaces correlating with ground state products with the surface of the excited ionic state at 12.86 ev. If a re— pulsive state and the excited ionic state are Of the same symmetry species, as indeed must be the case for one of the repulsive surfaces arising from ground state products, the interaction of the two electronic states will be so strong as to result in an "avoided crossing". This "avoided 53 crossing" of course is a result of the non-crossing rule [30], and is represented schematically in Figure 8. The dissociation coordinate represented in the figure is the ON-O bond length. The ground state molecule has a bond length of 1.193 A [30]. Because the first ionic State of N02+ is isoelectronic with ground state C02 it will probably have a comparable bond length, which for C02 is 1.16 R (or perhaps somewhat shorter due to the increased nuclear charge). The first excited ionic state corresponds to the promotion of a ”g electron in the C02 molecule to the Wu orbital. This will result in the molecule becoming more bent, but also since the Wu orbital is a formally anti- bonding orbital it will weaken the molecule and the bond strength. Because of this it is probable that the equi— librium bond length of the first excited ionic state is no shorter than in the ground state of the neutral molecule. In Figure 8 only one vibration in v3, the antisymmetric stretch, is shown for the 3B2 state of N02+; however, this is merely for purposes of the drawing, since there is no way of telling just how many bound vibrations there are in this degree of freedom. Looking at the ionization efficiency curve in Figure 7 it is evident that the v; = 0 and 1 vibrational levels dissociate very little, whereas significant amounts of NO+ are formed from the higher vibrational states. It thus ap— pears that the effectiveness of the predissociation in some measure depends upon the ability of the bending vibration 54 to couple with the antisymmetric stretch. If this is the case, the lifetime of the excited ion will be a function of the vibrational level from which the dissociation initiates. If a comparison is made of peak intensities for each vibrational level observed in the fragmentation spectrum to the peak intensities of these same vibrational levels ob- served in the photoelectron band, it indeed becomes clear that the probability for dissociation varies as a function of the bending coordinate Of the 3B2 ion. This is shown in Table VI as the ratio of NO+ Observed to that of the N02+ from which it originates. This ratio reaches an approxi— mate plateau at v5 = 3 and above implying that all of the N02+ formed in these vibrational states dissociates within the time it is in the ion source, or that all these vibra- tional levels have the same half—life for dissociation. This point can be checked experimentally by Searching for the production of NO+ by the unimolecular dissociation of metastable N02+ in the mass spectrometer. If an ion decom- poses after it has been accelerated from the source chamber, but before it enters the magnetic field, it appears generally at a non-integral mass position given by the formula mmeta = (mf)2/mp, and it is said to be metastable. (mf and mp are the masses of the fragment and parent ions respective— ly.) Newton and Sciamanna have reported the observation by electron impact [44,45] of a metastable dissociation of N02+ yielding NO+ fragments. They observed two separate contri- butions to the metastable peaks, one of low kinetic energy 55 release and one of high kinetic energy release. It there— fore appeared that a search for NO+ from metastable N02+ produced by photoionization might prove fruitful. Table VI. Ratio of NO+ peak height from fragmentation t N02 peak height from photoelectron spectrum. v' NO+/N02+ 0 0.33 r 0.3 1 0.24 r 0.2 2 1.29 r 0.1 3 2.38 r 0.1 4 2.15 r 0.2 5 1.93 i 0.3 a From reference 22. Shown in Figure 9 is the photoionization efficiency curve for the metastable production of NO+ from NOZ+; also shown are the positions of the vibrational levels Observed in the 12.86 eV photoelectron band. As can be seen, the v; z 0 level shows only very slight evidence of dissociating in the time it takes the N02+ parent ion to move from the ion source through the field-free region. The vé = 1 level shows a substantial amount of metastable production, as does v; = 2. For the higher vibrations, v; = 3 and greater, there does not appear to be any contribution whatever to the 56 .noz gone +oz no COHDUSOOHQ magnummuoe OLD How O>HOO mosOHOHmmm COHDMNflcOHODOLm .m Onsmflm .m mnsmflm :3 £96.05; 00.0— 000 omo omo owe 0mm ONO 0mg 99“ ‘0 o 9 O «O 0.9 ’ o 0 3 o o o o o o \ o 00 o o \ o I o o \ o 00 I I 0.00 to C O .9 99 O ‘00 80 \‘IO‘ .09 000 O .900 00, 88 ’0 o h . .9, "90 .‘k. .‘ ‘ O. .9 I... Q ‘ . O .9 . O 0‘ fl 0. .9. O O ‘ O O Q 9 O 0‘ 8V 3 O o 09 9 9 o 9t O 57 .0 o _ e. .. . .. . Oz _ _ _ . '3'l'd 58 metastable curve, which shows that all the parent (or most of it) dissociates in the time it takes to reach the field- free region. An examination of the behavior of the mass peak as a function of the ion accelerating voltage was made to determine the kinetic energy release to the fragments and thus identify which of the two metastables reported by Newton and Sciamanna was being observed. Shown in Figure 10 is a voltage scan of the metastable peak at m/e = 19.55. The exciting wavelength for the measurement was 950 A (13.04 eV). To calculate the kinetic energy release to the fragments, equation (11) taken from the paper of Newton and Sciamanna [44], was employed. 4m: m2 T 1/2 d = ) , (11) m0 m1 eVA m0 is the mass Of the parent, ml the mass Of the Observed fragment, m2 the mass of the other fragment, T is the kinetic energy release, VA is the ion accelerating voltage and d is the half-width of the voltage scan at 70% full height in mass units. Making this calculation yields a kinetic energy release of 0.59 eV to the fragments, which is in excellent agreement with the value of 0.51 eV reported by Newton and Sciamanna for their low kinetic energy com— ponent [44]. The relatively steep sides of the curve in Figure 10 indicate that there is relatively little spread in the kinetic energy release. Further, by producing the -)(- parent N02+ with a photon beam Of 13.04 eV it is 59 .hm.mH u O\E pm Xmom OHQMDOMDOE mo cmom Omwuao> .oH musmflm 60 .oH musmflm 7:03 ._<_th...0._ 02_h<¢mauuu< 20. oomV oo—V ooov Iooop rooou rooon 61- elm SanOD NOI £9 61 selectively made in vibrational levels of V; :.2. With the thermodynamic threshold at 12.37 eV the parent ion N02+ has excess energy of 0.67 eV when in v; = 2, but on the average about 0.59 eV of this exess energy is given up to the fragments as kinetic energy. The lowest vibrational spacing in NO+ is 0.28 eV so it is clear that very little if any NO+ is formed with excess vibrational energy. From information about the production of NO+ fragments within the ion source and the metastable production of NO+ originating from the same electronic state of N02+*, it is possible to calculate the half-lives of the various vibra- tional levels for this dissociation. To make this calcula- tion one needs to know the residence time of the N02+ parent ions in each region of the mass spectrometer. These times can be calculated directly from the knowledge of the dimen— sions of each region, the strengths of any electric fields present, and simple classic mechanics. First, it is easiest to calculate the half—life of the, v; = 3 level. Because no detectable amount of NO+ origin— ating from the vibration is seen in the metastable curve one may make the assumption that at least 90% of the N02+ formed in v; = 3 dissociates in a time shorter than 16 usec, the time it takes the ions to reach the field-free region in this experiment. (Repeller voltage = 0.8 VJ Using this conservative estimate and assuming the decay is first order, the half—life is calculated to be i 4.8 usec. To calculate the half—life of the v; = 2 level the information in Table VI 62 must be utilized. As can be seen, the ratio of the peak heights is diminished 55% from the value expected if the half-lives of the v; = 3 and v; = 2 states were the same and all other effects were equal. By calculating how much N02+ in the v; = 3 level had dissociated in the time the ions are in the source chamber, and assuming 55% less dis- sociation for v; = 2, a rough estimate of how much N02+ in the v; = 2 level has dissociated in the same amount of time can be made. Again assuming first-order decay, the half-life of v; = 2 is approximately 11 usec. The calculation of the half-life of v' = 1 and v' = 0 is somewhat more involved. The information needed to do this calculation must come from the metastable photoioniza- tion curve. All the intensity in this curve comes from the dissociation of N02+* while the electronically excited ion is in the field-free region. The intensity (area) of each peak can be expressed mathematically assuming first order decay kinetics. ta is the time it takes the N02+* to reach the field-free region from the source and tb is the time it takes to pass from the source through the field-free region. The concentration of N02+* that decays while in the field-free region can be calculated. The amount of N02+* in some particular vibration ‘v; = X present at time ta is AXa and the amount present after passing through the field— free region is AXb' The difference between these two num- 9(- bers is the amount of N02+ which has decomposed, and therefore 63 the amount of NO+ that should be observed in the metastable photoionization efficiency curve. The equations necessary to perform this calculation are: -k t x a A... = Ame , <12a> AXb = AXO e_kxtJO , and (12b) Ikxta -kxtb Zara-Airwxae -e >' (He) where AxO is the concentration of N02+* in the vibrational state v; = X at t = 0 and kX is the rate constant for dissociation. To continue the calculation, measurements of the ratio of the observed peak heights in the metastable curve relative to v; = 2 can be made. This ratio is called Rm* and is given by equation (13): P.H. v' = 2 R * = ____£_g____l.l X : 0,1. (13) m 1>.H.(v2 = x) Rm* may be expressed mathematically using equation (12C): A _A A "kzta — e-k2 tb R * _ 2a 2b _ 20 m — - _ —k t -k ’ AXa AXb AXo(e x a _ e xtb) x = 0,1.(14) In this equation k2 is known from previous arguments, and therefore equation (14) reduces to: 64 A20 0.134 R* = . (15) m _ Akxta -kth AXO(e _ e ) In equation (15) there are three unknowns: A20, AXO, and kx' The ratio AZO/AXO may be found utilizing the photo- elctron results of Brundle [12] by taking the reported peak heights. It should be noted that the effects of autoioni— zation are absent in the photoelectron spectra, and that the photoelectron cross-section is energy dependent. Thus, the ratio of the photoelectron peak heights is not an exact measure of the relative N02+* concentrations in the photo- ionization experiment. This ratio of the Observed Franck- Condon factors is Called R and is given by equation F.C. (13) except Rm* is replaced by R Substituting this F.C.. value into equation (15) and rearranging yields: -k t -k R 0.134 R 'X- m There is only one unknown in this equation, kx' which may be determined graphically. The measured values of RF.C. and Rm* for v; = 0 and 1 are given in Table VII. These same calculation can be repeated starting with the assumption that 100% of the N02+* formed in v; = 3 dissoci- ates in a time shorter than 10.15 usec, the time it takes the ions to reach the field-free region. The results Of both calculations are presented in Table VIII; they indicate a striking dependence of the half-life of the N02+* ion on the number of bending vibrational quanta excited. 65 Table VII. Experimental data used to solve equation (16). I * a V2 Rm RF.C. 0 18.7 3.2 1 3.1 1.3 aSee reference 12. Table VIII. Summary of results from the calculation of the lifetimes for dissociation of the vibrational levels of the 3B2 state of N02+. Upper Limit Lower Limit v' 11/2 (u sec) Tl/z (usec) 0 154.6 147.4 1 57.8 53.3 2 16.2 13.9 .i 3 .2 4.8 .2 2.4 The small rise in the photoionization efficiency curve (Figure 7) at 12.37 eV requires some explanation. It is probably due to NO impurity; however, it might be accounted for in two other ways. First there may be a Rydberg level in the vicinity which Simultaneously autoionizes and pre- dissociates into one of the repulsive curves originating from the ground state dissociation products. One problem 66 with this interpretation is that after the onset there is almost continuous fragmentation. If this originates from a Rydberg level it should be discrete unless there is an extremely dense pOpulation of predissociating Rydberg levels. The second possibility is that an optical transition is made to the ground ionic state, and in the time the N02+ ion remains in the ionization chamber it makes a forbidden transition into the dissociation continuum. This would re- quire a spin change, but this may happen since the sample resides in the source ~ 11 usec, which may be long enough for the multiplicity change. One argument against this process is that the Franck—Condom factors for the Optical transition to the ionic curve are most probably so small that the process would not be observed. In Figure 11 the NO+ photoionization efficiency curve is compared with the N02+ parent curve. It may be noted that the two curves show identical structure, which is due to the fact that the Rydberg levels which appear as auto— ionization structure in the parent curve may also competi— tively predissociate, therefore appearing in the fragment curve as well. This is illustrated by equation (17): + _ N02 + e //z \i + _ NO + O + e N02 ‘1' hV In conclusion, if the NO+ observed at the thermo- chemical limit is not an artifact, it is probably due to 67 .mm>uso hocmfloflwmo QOHDONHCOHODOSQ +aoz paw +02 mo QOmHHmmEOU .HH onsmflh ocu— ow: . HH meDmanH 23 1523 m>§> 23— Ge» 00% p p 5 F 1.. GR +02 . 3/ E» '3'l'd ’ n \ . u . I. m, .W..u....— 6‘ u .h... .n. a. . . .H A. / .s :1 m.....l.. K). - 69 some form of predissociation. There is, however, no way to say with certainty that this onset is real. The striking structure Observed at and above 12.86 eV has been explained as due to predissociation of the first excited ionic state (3B2) and the lifetimes of the various bending vibrational levels of this state have been calculated. However, the details of the dissociation mechanism(s) involved are still uncertain. RYDBERG ANALYSIS OF N02 Introduction A series of lines corresponding to transitions from the ground state of a molecule to more and more highly excited states of a molecule form a Rydberg series, in analogy with atoms. In 1885 Balmer identified a series of lines in the spectrum Of atomic hydrogen due to promotion of the ls elec— tron to higher and higher atomic orbitals. He found an equation to describe this series which was generalized by Rydberg and this series is now called a Rydberg series. In the more highly excited states of a molecule the higher the orbital energy the more the orbital looks like an atomic orbital. Thus transitions to these orbitals behave like atomic spectra and form Rydberg series. Rydberg series in molecules may be expressed mathematically by the Rydberg formula: vn=v— ——B——, (18) (n-é)2 where Vn is the observed frequency Of the Rydberg transi- tion, v is the frequency of the convergence limit (this corresponds to complete removal of the electron), R is the Rydberg constant, n is the quantum number of the Rydberg level, and 0 is the quantum defect which is a 70 71 correction term which accounts for effects such as Rydberg orbital penetration into the molecular core. In molecules, Rydberg orbitals have designations analogous to atomic or- bitals np, ns, etc. which come from the united atom ap- proximation followed by the symmetry of molecular orbital formed, a1, 0, b2, etc. (For a more detailed discussion of Rydberg series see references 30,32,46.) The Rydberg analysis of N02 has proven to be a very difficult problem. Tanaka and Jursa [15], Nakayoma, Kita— mura and Watanabe [11], and Price and Simpson [14] have investigated the vacuum ultraviolet absorption spectrum of this molecule and assigned several Rydberg transitions. All had similar problems in their analyses. The first was the diffusness of the spectra, due in part to the transitions themselves, due to the continuous background absorption, and due to the effect of the sample gas on the photographic plate. Second, and more significant, these investigators did not know in advance the values of the ionization poten- tials of N02. Brundle [12] and Edquist, et al. [22] have since established the positions and accompanying vibrational structure of the ionization potentials of N02 up to 21 eV by photoelectron spectroscopy. The vibrational structure near the ionization limits is extremely useful when trying to assign Rydberg transitions, since each member of a series should exhibit the same type of structure as the limit to which it converges. Utilizing the information from their photoelectron spectra, the previously observed absorption 72 spectra, and the photoionization efficiency curve reported by Dibeler, gt_al. [23], Edquist and coaworkers suggested assignments for many of the Rydberg transitions in N02. The photoionization efficiency curve of N02+ obtained in this investigation exhibits substantial autoionization structure. Using the assignments of Edquist, et al. [22] as a guideline, most of this structure can be identified. Assignment of the Autoionization Structure The first excited ionic state at 12.86 eV corresponds to the removal of an electron from the 3b2 orbital. Tanaka and Jursa [15] and Nakayama, 35 al. [11] have observed a series converging to 12.86 ev. Edquist, g£_al, [22] have assigned this as the 3b2 ——e nso transition, based on the quantum defect having a value of approximately 1.00. While only two members of this Rydberg series have been seen pre— viously, four members of the series are observed in the photoionization spectrum. Table IX lists the positions of each member of the series and the positions and quantum de- fects of the vibrational levels associated with each member. This vibrational structure has been assigned as a progression in v2, the bending vibration. The v; = 0 vibrational level is not Observed for the n = 3 and n = 4 members of the Rydberg series in autoionization. A possible reason for this behavior is that the v; = 0 vibrational levels of these states do not overlap the ionization continuum suffici- ently to autoionize (see discussion on the ionization 73 Table IX. A Rydberg series converging to 12.86 eV: 3b2 -—> nso n V' Ohs§§ved T33:::(gnd 5 3 0 1293.3 1 1282.30 1283.2 0.96 2 1272.80 1273.2 0.96 3 1262.50 1263.4 0.96 4 1253.55 1253.8 0.97 5 1243.43 0.97 6 1234.30 0.97 4 0 1092.8 1 1.085.46 1085.4 1.01 2 1077.66 1078.1 1.01 3 1070.46 1070.9 1.01 4 1063.46 1063.8 1.01 5 1056.26 1057.0 1.01 5 0 1033.15 1.02 1 1026.40 1.02 2 1019.40 1.02 6 1006.15 0.99 1 999.90 0.99 aFrom Reference [15]. 74 potential of N02). When going to higher members of the series this problem is no longer a factor and the v; 0 level Can autoionize. There is much more structure in the region of the convergence limit, probably due to the higher members of the Rydberg series piling up. Because of this effect many lines are closely spaced and probably perturb one another making it impossible to assign all these lines. Edquist, 33 al. [22] have assigned the strong band at 1142.6 8 in the photoionization curve as due to a 1a2‘—> npo Rydberg series converging to 13.60 eV. This bandlis strong and exhibits little evidence of vibrational structure implying, based on the photoelectron band shapes of Edquist, et al. [22], that this band may be converging to either 14.07 eV or 18.86 eV, if it is due to a Rydberg transition. If this band is a member of a Rydberg series converging to 14.07 eV it would have a quantum defect of 0.96, indicating a nso type transition. It is probable that the ionic state at 14.07 eV is bent; since this state results from the promotion of a lag electron the transition is for— bidden because of the dipole selection rule. A second argu- ment against this interpretation is that no other member of the series can be identified. If, on the other hand this band is a member of a Rydberg series converging to 18.86 eV it will have a quantum defect of 0.70, in line with a npo type transition. The state at 18.86 eV is also probably bent, however, the npO transition is dipole allowed. Utilizing this quantum defect two more members of this series 75 may be assigned. Table X presents a summary of the positions of each member of the series and the quantum defects of each member. Four other series have previously been reported con- verging to 18.86 eV by Tanaks and Jursa [15], but this pos- sible series was not so designated. The band at 1142.6 8 was assigned as the v; = 0 member of a progression Tanaka and Jursa [15] had placed in a Rydberg series con— verging in the vicinity of 12.0 eV. The first and third members of that series have since been assigned as the first two members (n = 3,4) of a Rydberg series converging to 12.86 eV [12]. The remaining part of this progression re— ported by Tanaka and Jursa [15] has been assigned by Edquist, .EE.§£° [22] as the 3b2 ~> nd0 transition converging to 12.86 ev. Examination of the photoionization efficiency curve shows the presence of three members of this progres— sion. However, there is no evidence for any higher members of the Rydberg series converging to 12.86 eV reported by Edquist, _£ 91, [22]. The second member (n = 4) of this Rydberg series is expected to appear at approximately the same place the second member of the nso series converging to 12.86 eV; perhaps it is obscured by the intense lines due to the transition. Thus it remains unclear whether the assignment of this progression as a Rydberg series con— verging to 12.86 eV by Edquist, g; l. [22] is correct and no conclusion may be made on the basis of this data. 76 Table X. A new Rydberg series converging to 18.86 eV: 2b2‘—> npo. , Observed 0 n V (A) (calc) 2 0 1142.60 0.72 3 0 760.40 0.70 4 0 702.60 0.66 In the vicinity Of 1020 A three bands tend to stand out more prominently than the rest. Most of the structure in this region has been attributed to the piling up of the Rydberg series converging to 12.86 eV (968 8). However, because of the anomalous intensity of these bands it might be assumed that they are converging to a higher ionization potential. From the photoelectron Spectrum Of Edquist, 35 a1, [22] it is possible that these bands are converging to 14.07 eV. If this is the case, the series would have a quantum defect of approximately 0.40 in line with a 1a2 —> npv type transition which is optically allowed. An extrapo- lation to the next member of the series predicts a band at approximately 950 X where such a band is located. The 950 A band also exhibits a similar intensity distribution in its vibrational structure as the preceding member of the series. NO other member of this series can be found. Therefore the assignment must be considered tentative. Table XI presents the information about this series. 77 Table XI. A possible Rydberg series converging to 14.07 eV: 1a2 —e npw. nu v1 v2 v3 Obi§§ved 5 3 0 0 0 1022.40 0.35 0 1 0 1015.65 0.36 1 0 0 1012.40 0.35 0 2 0 1009.15 0.38 4 0 0 0 954.90 0.47 0 1 0 950.65 0.51 1 0 0 946.65 0.47 In the region of 1089 A there is a very smooth and pronounced dip in the photoionization efficiency curve. This dip is not an artifact produced by the light source and is quite reproducible. There is some evidence of a second dip in this region, but it occurs between two strong autoionization lines. It is possible these dips are a manifestation of the phenomenon described by Fano [33] as a window resonance. If this is a window resonance there might be other such "resonances" in the photoionization efficiency curve which would form a Rydberg series. Several attempts were made to fit this band to a Rydberg series. If 18.86 eV is taken as the convergence limit then the band does seem to fit into some type of Rydberg series. A second 78 member may be seen in the vicintiy Of 758 A. This band also appears to have a second component with it. A third member of the series is expected to be at 700 A. A very small dip is present in this region; but its assignment as a member of the series is open to question. Table XII presents the information relative to this sereis, which has an average quantum defect of 0.60 in line with a ndo type transition. Table XII. A Rydberg series converging to 18.86 eV: 2b2 —> ndo. . Observed 6 n V (X) (calc) 2 0 1089.65 0.65 1 1081.15 0.65 3 0 752.60 0.62 1 746.80 0.60 4 () 698 .60 0 .50 The majority of the remaining structure in the photo- ionization efficiency curve may be assigned as Rydberg series converging to 18.86 eV which have been identified by Tanaka and Jursa [15]. Table XIII presents a comparison of series I reported by Tanaka and Jursa [15] with the auto— ionization structure of this work. All seven members of the 79 Table XIII. Comparison of experimentally observed Rydberg serie with series I reported by Tanaka and Jursa : 2b2'—> npw —> 18.86 ev. , Observed Tanaka and n V (R) Jursaa 5 2 0 897.10 897.02 0.36 3 0 '721.46 721.20 0.16 4 0 691.06 691.01 0.16 5 0 678.26 678.31 0.16, 6 0 671.26 671.59 0.18 7 0 666.86 667.39 0.04 8 0 665.26 664.96 0.14 a From reference 15. 80 series reported by Tanaka and Jursa[15] are found in the present study; however, the n = 6,7, and 8 members have an anomalous intensity behavior. The n = 6 band has an in— tensity significantly greater than any other member of the series and it seems likely that this particular band is a member of some other series converging to a higher ioniza— tion potential. This line is included in Table XIII for comparison, although its validity in the series is not established. The n = 2 member of this series was reported by Edquist, g3 31. [22] to be due to NO impurity lines in the spectrum of Tanaka and Jursa [15]. Since the line is observed in the photoionization efficiency curve and a mass spectrometer is used as the ion detector, the possibility that this line is due to NO impurity is precluded. Series II reported by Tanaka and Jursa [15] is not ob— served in the photoionization efficiency curve at all. They reported this series to be weak in absorption, and if this is coupled with an unfavorable autoionization cross section its absence would be understandable. The remaining two series, III and IV, reported by Tanaka and Jursa [15] have been combined into one by Edquist, gt al. [22]. This was prompted by the belief that one member of series IV was again due to NO impurity; however, the line in question (783.86 A) has been observed in the photoionization effi- ciency curve and therefore cannot be due to NO+. Thus the two series have been kept intact. Table IV presents a com- parison of Series III of Tanaka and Jursa [15] with the 81 Table XIV. Comparison of experimentally observed Rydberg series with Series III reported by Tanaka and Jursa : 2b2 —> ns0 -0 18.86 ev. n Obs§rved Tanaka a d 6 (‘) Jursa( ) (calc) 3 ' 810.66 1.05 805.86 806.06 1.03 4 717.00 717.57 1.06 715.86 715.67 1.03 5 689.66 689.80 1.07 688.66 688.56 1.02 a From reference 15. Table XV. Comparison of experimentally observed Rydberg serieg with Series IV reported by Tanaka and Jursa : 2b2 -> ndé —> 18.86 eV. Observed Tanaka and 0 n (R) Jursa (A) (calc) 2 792.86 —0.05 783.86 784.93 -0.14 3 712.86 712.25 -0.04 710.46 710.43 -0.10 4 687.66 687.52 -0.04 686.66 686.01 -0.10 ea From reference 15. 82 experimental data. Each member of this series was reported to be a doublet by Tanaka and Jursa [15]; however, they did not see the second component. This component has been ob— served in this photoionization study and is included in the table. Table XV presents a similar comparison between series IV Of Tanaka and Jursa [15] and the experimentally observed lines. Again each member of the series was de- scribed as a doublet and the second component of n = 3 has been added to the table. There are several other lines in this region (650 A), but they do not fit well into any of these series. Tanaka and Jursa [15] have suggested that because so many series are converging on this limit the lines may be perturbing each other, making assignment virtu- ally impossible. The line at 671.26 A, which as previously stated ap- pears tO be too intense to be a member of series I converg- ing to 18.86 eV, may belong to a Rydberg series converging to a higher ionization potential. From the photoelectron spectrum the only remaining possibility is the limit at 21.26 eV. Making this assumption yields a quantum defect of 0.79 for n = 3 converging to this limit; however, no other line can be associated with this one. Possibly it is converging to a still higher ionization potential. Figure 12 presents the photoionization efficiency curve of N02 with a summary of the various Rydberg series assignments . 83 .moz mo mOHHOm mnmnpwm .NH Onsmflm NA ensues 84 2.: 1.55353 00.n— - - OOSN— . l + co...— . . . . 60.0— . _ _ _ _ 8.. . r .L t h s p l m _ p :1! ill l s: n incl _ _: 9 00a O N . . _ em.-- . _ - , _ -_: i. _ _ _ i 2 >33. 6 >636. >683: 33.2 mflw__ [II_ av: v. n .>1AXW¢— « _ _ in: = L +«02 _ [W To) J H L _ _ C i L a] Jim 6. _ q C u L THE PHOTOIONIZATION OF NO Introduction It has been pointed out by Wigner and others [1—3] that the threshold law for direct ionization by photon impact should be approximately a step function. That is, the cross section for ionization at the threshold should be finite and vary only slowly with energy in the region above this ionization limit until a new limit is encountered. In order to observe this behavior, however, certain other factors must be favorable. In all polyatomic and diatomic molecules there exist Rydberg series converging to the first and each subsequent ionization limit. Those excited neutral states which lie above the first ionization poten- tial of the molecule may autoionize, predissociate or fluoresce. The rate of fluorescence emission is usually much smaller than that of the first two processes and is usually negligible. If these states autoionize they will cause peak-like structure to be superimposed on the step and therefore it may become difficult to distinguish where the next step occurs. In order to observe step—like be— ‘havior these excited neutral states must be depopulated by some mechanism which does not yield the charge species of 85 86 interest. The mechanism most able to do this is predis- sociation Of the excited neutral states into neutral frag— ments or charged fragments. The restriction in this case is that the predissociation rate be much faster than the competing autoionization process. The classic example used to demonstrate step function behavior at threshold has been the NO molecule. Watanabe, Marmo and Inn [47] reported one of the earliest photoioni- zation studies, which revealed three steps in the threshold region Of the NO+ ion corresponding to vibrational levels of the ion. Since that time various other investigators have reported the photoionization cross section curve for this molecule [48-51]. All of these results are in essen- tial agreement that the molecule exhibits step—like behavior over the first four vibrational levels of the ion. One of these investigators [49] used the fact that the cross sec- tion was a step function as a standard to calibrate the photon detector. This step—like behavior has been assumed to exist because the neutral excited states are completely predissociated. As the preliminary step for a charge-transfer experi— ment a cursory photoionization study of NO was undertaken using 300 micron slits and the hydrogen lamp. Examination of the photoionization efficiency curve in the threshold region revealed some unusual structure on the steps, which obscured the step somewhat. In an attempt to determine if 87 this structure was real or an artifact of the hydrogen lamp a careful study of NO was undertaken. The Threshold Region (1350 - 1220 2) Shown in Figure 13 is the photoionization efficiency curve of NO in the threshold region, obtained utilizing 100 micron slits. As is evident, there is significant structure superimposed upon the Step. Since these data were taken using the smooth Argon continuum as a light source rather than the many-lined psuedo-continuum of the hydrogen lamp, the structure must be considered real. Watanabe [48] and Miescher [52] have investigated the absorption Spectrum of NO in this region (1400 - 1300 A). Miescher [52] used a 10-meter Spectrograph for his study and has performed a very detailed analysis of the spectrum. He identified several Rydberg series converging to excited vibrational levels, v' > 0, of the ground state X'Z+ of the NO+ ion. This structure may be correlated with the autoionization struc- ture in the photoionization efficiency curve. Because these Rydberg states are converging to vibrationally excited states of the first ionic state it is evident that this autoionization structure is due to vibrational autoioniza- tion. AS can be seen in Figure 13 most of these autoioniza- tion lines are weak and are fairly well predissociated, but a small probability (of the order of a few percent) of com— peting autoionization does exist. 88 .coamou paonmmszu 020 CH 02 Eonm +02 mo O>Hso mocmHOHmmm COHDMNHCOAODOEm .mH musmflm 89 3... >O mud 2.3 8.2 mon— .mH ousmflm . A VIhOZm4m><>> www.— < mofl 93.. 8a.. '3'l'd 90 An attempt has been made to assign the structure in the photoionization cross section curve using the assign- ments of Miescher [52]. Since the instrument used in this experiment does not have the resolution available to Miescher, several possibilities exist for some assignments. Presented in Table XVI is the assignment of the lines Observed on the first vibrational step of the photoioniza- tion efficiency curve. A comparison is made between the Observed lines and tnose reported by Miescher [52]. In some cases two assignments were possible; in these instances the quantum defect was used to make a decision between the two. Both possibilities are presented in the Table for comparison, with an asterisk used to denote the more likely choice. The most important aSpect of these assignments is the convergence limits of each line. As can be seen from Table XVI, all but one line can be assigned as Rydberg states con— verging to v' 2.2 of the ground ionic state. Therefore, all of these states converging to v' Z_2 must make a vibrational energy change of at least two quanta in order to autoionize. Berry [35] has proposed a propensity rule for autoionization by vibration in which he states the probe ability of autoionization should be greatest for a vibration- al energy change Av = 1, with decreasing probability as Av increases. The autoionization structure observed on the first step does not disprove or confirm this propensity rule. In this case if these states are to autoionize at all they 91 Table XVI. Assignment of autoionization+s§ructure on the first vibrational step of NO . Observed Mie§cherb Assi nmentc 0 0b (3) ( ) g (calc) 1337.17 1340.21(h) 6so -e v' = 2 +1.19 +1.208 1336.17 1335.92 5d0-—> v' = 2 +0.16 +0.078 1334.17 1333.81 5f —> v' = 2 +0.12 +0.020 5So-—> v' = 3 +1.00 +1.201 1331°13 "' 4d0 -> v' = 3* +0.00 -0.033 5d _>- V' : 2 —0.07 —..._. 1328°42 “‘ 4d —e v' = 3* -0.05 -0.033 1326.42 —-- 6pw —> v' = 2 +0.84 +0.76 9pw -> v' = 1* +0.75 240.76 1324-42 “’ 6pc —e v' = 2 +0.77 +0.69 6pc —> v' = 2* +0.68 +0.69 1322'42 “‘ 5pW-—> v' = 3 +0.82 +0.76 1317.67 --- 5pw —> v' = 3 +0.74 +0.76 1314.42 1212.80 7so-—> v' = 2 +1.38 +1.217 1311.67 --- 5po-—> v' = 3 +0.62 +0.69 1309.92 1310.07(h) 6f —> v' = 2 +0.13 +0.029 aThe vibrational quantum number is that of the limit to which the series converges and also that of the ion core of the Rydberg state. The ion resulting from autoionization will of course have a lower vibrational quantum number. bSee reference 52. CWhen more than one assignment is possible an asterisk is used to indicate the author's choice of the two. 92 must do so by changing vibrational energy by two quanta. Also since there is only one line converging to v' = 1 and it is also very weak it is impossible to estimate if this state has a higher probability of autoionizing than any other state. The intensity of the observed autoioniza- tion structure is not only a function of the autoionization rate, but of the rates of the competing processes. Pre— dissociation will have a rate which is dependent (often strongly) upon the vibrational state of the Rydberg level. Thus strong autoionization structure might imply that the predissociation is weak rather than the autoionization is strong. The absence of any other v' = 1 lines may be due to a much stronger predissociation of this vibrational level. Examination of Miescher's spectrum [52] shows that he does not observe in absorption any v' = 1 levels above the first ionization limit. It is possible these levels are so predissociated they blend into the continuum and thus one would not expect to see any contribution from these in the photoionization eXperiment. Table XVII presents the assignments of the observed autoionization structure superimposed on the second vibra— tional step of the photoionization efficiency curve. There are three strong lines with decreasing intensity and they all have the same quantum defect and appear to belong to the same Rydberg series. In this case these states are con- verging to the v' 2 vibrational level of the NO ionic ground state. It is interesting to note that these lines 93 Table XVII. Assignment of autoionization structure on the second vibrational step of NO+. Obs§§ved Assignment (cglc) 0a 1297.42 8pv —+ v' = 2 0.77 0.76 1290.42 5f —> v' = 3 0.03 0.029 1288.92 9pw —> v' = 2 0.75 0.76 1285.42 ? 1283.17 10pF —+ v' = 2 0.78 0.76 1274.17 6d —> v' = 3 —0.06 -0.03 1269.92 7pc —> v' = 3 0.68 0.69 1267.67 ? a . . Average value for observed series from Miescher, reference 52. Table XVIII. Assignment of autoionization structure on the third vibrational step of NO+. obsggved Assignment (cglc) 0a 1260.42 8pF —> v' = 3 0.77 0.76 1258.42 8pc —> v' = 3 0.62 0.69 1252.67 9pv —> v' = 3 0.75 0.76 1250.92 9pc -> v' = 3 0.54 0.69 1247.42 10pw —> v' = 3 0.78 0.76 a . . Average value for observed series, from Miescher, reference 52. 94 are much more intense than any of the others which are all converging to the v' = 3 vibrational level of the ion. Again, for these v' = 3 levels to autoionize a vibrational energy change of two quanta is necessary. This may be evidence supporting the propensity rule of Berry [35]; how— ever, it is by no means definitive. Two lines remain un- assigned as they did not fit any Rydberg series. It is quite likely that these states are being perturbed by other Rydberg states. Mischer [52] has pointed out that there is some perturbation between states with different 3 com- ponents but the same m2 quantum number. The next vibrational step, v' - 2 has very little structure superimposed on it. This is probably a result of the fact that the Rydberg levels in this region must be of a very high principal quantum number and the transition probability to these states is relatively small, because of n—3 dependence and a small Franck-Condom factor for the v' = 3 level. Table XVIII presents the data on the five observed autoionization lines on this step. All of these Rydberg states are converging to v' = 3 as expected, since only four vibrational states were observed in the photoelectron spectrum of Turner and May [53]. It is interesting to note that if one looks in the literature only one paper may be found which mentions fine structure on the vibrational steps of NO+ [54]. Cantone, Emma and Grasso [54] did an electron impact study of NO and a first derivative plot of their data does indeed show 95 some strong structure. This structure was attributed to autoionization by Cantone, gt al. [54] but they did not speculate on the type of autoionization. A careful examina- tion of the photoionization efficiency curves of Watanabe [48], Hurzeler, et al. [49] and Reese and Rosenstock [51] shows that the structure reported here is also observable in all these curves. None of these authors, however, make any mention of this structure. The 1220 — 600 2 Region Figure 14 is the photoionization efficiency curve for N0+. The remaining structure between 1220 R and 600 R has been discussed in detail by various authors [41,55,56] in terms of Rydberg series converging to higher ionization potentials. Edquist, 35 al. [56] have investigated the photoelectron spectrum of NO and utilizing this information have taken all the available data on the NO spectrum and compiled it into a group of Rydberg series. Except for one difficulty most of the structure in the photoionization curve can be correlated with this compilation. Edquist, _£‘_l. [56] do not distinguish between lines reported by two different authors. For example a line reported at 800.4 2 by Tanaka [55] is assigned as the n = 4, v' = 1 member of a Rydberg series converging to 16.86 eV while a line at 800.4 2 reported by Huber [41] is assigned as n = 3, v' = 0 of a Rydberg series converging to 18.32 ev. It is obvious that this line can be only one or the other (unless 96 AM com I ommfiv oz Eonm +02 mo m>nso mommaoflmmm coaumuflcoflouonm one .vH ousmfim .vH musmflm 97 2.: x5233; 8a.. . B: . on? . J g))§5lilfll¢llr( . r I .51.)... ...\.\(/./\.) \/ / . .2. x) . .1/ , . . m . on... _ 3n . on... 9x.- ) . >1 «221/232. ../J_..).. . / \/./ / .5. 22.22/12. 2. 22 :2 2.2/2... . 2. 2 . . .2. 522? i 2. . _ . 2. .. 2 2 . .2. ..2._. 2.222.222.2/ 98 superimposed). Therefore when such a controversy exists only one will be chosen. Table XIX presents a comparison of the observed auto- ionization structure with that reported as belonging to an nso series converging to 15.65 ev. Edquist, gt gt. [56] did not report any lines for the n = 3 member of this series. Information concerning this member of the series was obtained from the photoionization curve, the v' = 0 and three other vibrational components being observed. Table XX through Table XXII present a comparison be- tween the observed photoionization structure and the Rydberg series of Edquist, gt__t. [56] which converge to 16.56 eV, 16.86 eV and 17.59 eV reSpectively. Table XXIII presents a comparison between the Rydberg series assigned by Edquist, gt gt. [56] converging to 18.07 eV and the structure in the photoionization efficiency curve. Most of the reported structure of this series has been assigned by this author to other series and it seems possible that many of the lines in this series are not re— solved well enough from other lines in the spectrum. Those lines which may be unambiguously assigned to this series are listed in the Table. Table XXIV presents the Rydberg series converging to 18.32 eV. All previously reported members of this series are observed and the series is extended to n = 7 with the addition of the line observed at 689.92 X. This accounts for all the structure between 1150 X and 600 X, 99 Table XIX. A Rydber series of NO: lvi—> nso(—> a3z+ at 15.65 eV . a n v' x 0 obs xrep (calc) 3 0 1013.68 '1.01 1 993.80 -3 995.8 0.98 2 983.30 21983.9 0.99 3 973.60 21972.4 1.00 4 0 877.80 878.20 1.02 5 0 839.30 838.9 1.07 aSee reference 56. Table XX. A Rydber series of NO: 50-—> np; (—> b3TT'at 16.56 eV . a n v k h 5 obs rep (calc) 3pv 0 897.30 897.3 0.77 1 892.05 882.6 0.82 3p0 0 885.55 885.4 0.70 1 --— 872.0 4pW 0 811.05 810.9 0.63 1 --- 799.2 4pc 0 808.55 809.5 0.68 1 --- 797.2 5p0 0 783.80 783.7 0.71 V 1 771.30 772.8 0.59 aSee reference 56. 100 Table XXI. A Rydberg series of NO: 1v‘—> nso (—> w3A at 16.86 ev). n V xobs xrepa (cglc) 3 0 915.55 916.5 0.98 1 907.55 907.4 0.99 2 --- 897.5 3 889.50 889.8 1.00 4 880.55 881.8 1.01 4 0 808.55 809.5 1.02 1 --- 800.4 2 --- 792.3 3 —-- 784.4 4 777.30 776.5 1.03 5 770.05 769.7 1.03 5 0 --- 775.2 1 765.00 767.2 0.90 2 760.55 760.3 1.04 3 752.80 753.3 1.05 aSee reference 56. Table XXII. A Rydberg series of NO: 101 at 17.59 eV). 17 —> nso (—> b'3z- n v 7\obs xrep (cglc) 3 0 868.55 868.3 0.98 1 860.05 859.9 0.98 2 850.35 850.0 0.98 3 842.55 841.7 0.98 4 833.42 833.6 0.98 5 826.30 825.7 0.98 6 -—— 818.1 4 3 751.30 751.0 1.04 4 744.80 744.6 1.04 5 738.20 738.3 1.05 6 732.05 732.2 1.04 7 --- 726.2 8 --- 719.2 5 3 ——- 721.3 4 714.80 715.1 1.03 5 709.30 709.2 1.06 6 703.05 703.4 1.03 7 --- 697.6 aSee reference 56 Table XXIII. A Rydberg series of NO: 1w —> nso (—> w'A at 18.07 eV). n V 7\obs xrep (cglc) 3 1 -—- 830.0 2 821.30 821.0 0.96 3 812.80 812.7 0.96 4 804.55 804.9 0.96 5 796.30 795.8 0.95 6 787.30 787.1 0.93 aSee reference 56. Table XXIV. at 18.32 eV). 103 A Rydberg series of NO: 50 —+ np: (—> A'TT' n v x 0 obs rep (calc) 3pv 0 800.80 800.40 0.81 1 790.30 789.8 0.81 3p0 0 793.05 793.0 0.75 1 783.20 —-— 0.75 4pw 729.30 728.9 0.79 1 720.30 720.3 0.78 4pc 726.55 726.7 0.72 1 718.05 717.8 0.70 5p 706.05 706.0 0.77 1 697.80 697.6 0.74 6p 0 695.55 695.4 0.78 1 687.55 --— 0.73 7p 0 698.92 --— 0.77 aSee reference 56. 104 Between 1220 R and 1150 8 there are eighteen lines. Reese and Rosenstock [51] reported fifteen lines in this region and placed eight of them into two progressions. However, two of their fifteen lines are not observed in this study. The reason for this is not clear since both experiments should have yielded the same results. Possibly some arti— fact was introduced into the experiment of Reese and Rosen— stock [51]. Eight of the peaks in this region may still be placed into two progressions as follows: (a) 1077.93 2, 1064.93 2, 1053.55 2, and 1041.80 3; (b) 1056.66 2, 1044.80 2, 1033.68 2, and 1024.93 2. The remainder of the structure is not accounted for and may belong to the auto- ionization of non-Rydberg type transitions. The eighteen lines observed are listed in Table XXV. In conclusion the autoionization structure in the thres- hold region has been assigned as vibrational autoionization, and the structure in the remaining portion of the photoioni- zation efficiency curve has been correlated with the pro— posed Rydberg series of Edquist, gt_ 1. [56]. 105 Table XXV. Unassigned autoionization structure in NO. xobs 1128.56 1110.31 1086.56 1082.18 1077.93 1064.93 1056.68 1053.55 1044.80 1041.80 1037.18 1033.68 1024.93 1020.68 1017.18 1001.80 REFERENCES 10. 11. 12. 13. 14. 15. REFERENCES E. Wigner, Phys. Rev. 73} 1002 (1948). G. H. Wannier, Phys. Rev. 92/ 817 (1953). S. Geltman, Phys. Rev. 102, 171 (1956). A. J. C. Nicholson, J. Chem. Phys. 32/ 954 (1963). J; A. R. Samson, Techniques of‘Vacuum‘Ultraytolgt Spectroscopy, John Wiley and Sons: New York, 1967 . Jigggfikowitz and w. A. Chupka, J. Chem. Phys. 45, 1287 W. A.)Chupka and J. Berkowitz, J. Chem. Phys. 47, 4320 1967 . N” R. Allison, J. Burns and A. J. Tuzzolino, J. Opt. Soc. Am, 54/ 747 (1964). G. V. Marr, Photoionization Processes in Gases, Academic Press: New York, 1967. W. A.)Chupka and J. Berkowitz, J. Chem. Phys. 47, 2921 1967 . 'VV T. Nakayama, M. Y. Kitamura, and K. Watanabe, J. Chem. Phys. 32, 1180 (1959). C. R. Brundle, Chem. Phys. Letters 5, 410 (1970); C. R. Brundle, D. Neumann, W. C. Price, D. Evans, A. W. Potts, and D. G. Streets, J. Chem. Phys. 53, 705 (1970). Later measurements indicate N02 introduced pure but breaks down (by reaction with hot mercury in the pumps, etc.) to give NO which is not pumped rapidly, W. A. Chupka, Private Communication. W. C. Price and D. M. Simpson, Trans. Far. Soc. 31, 106 (1941). Y. Tanaka and A. S. Jursa, J. Chem. Phys. 36/ 2493 (1962). 106 107 16. J. Collin and F. P. Lossing, J. Chem. Phys. 28, 900 (1958). 17. R. J. Kandell, J. Chem. Phys. 23, 84 (1955). 18. S. Tsuda and W". H. Hamill, Adv. Mass Spectry. 3, 249 1966 19. R. W. Kiser and I. C. Hisatsune, J. Phys. Chem. 65, 1444 (1961). 20. E. C. G. Stuecklberg and H. D. Smyth, Phys. Rev. 36, 478 (1930). 21. P. Natalis and J. E. Collin, Chem. Phys. Letters 2, 79 (1968). 22. O. Edquist, E. Lindholm, L. E. Selin, L. Asbrink, C. E. Kuyatt, S. R. Mielczarek, J. A. Simpson and I. Fischer— Hjalmers, Physica Scripta l, 172 (1970). 23. V. H. Dibeler, J. A. walker and S. K. Liston, J. Res. Natl. Bur. Std. 71a, 371 (1967) 24. D. C. Frost, D. Mak and C. A. McDowell, Can. J. Chem. 42, 1064 (1962). 25. G. L. Weissler, J. A. R. Samson, M. Ogawa and G. R. Cook, J. opt. Soc. Am. 49, 338 (1959). 26. R. s. Mulliken, J. Chem. Phys. 3, 720 (1935). 27. A. D. Walsh, J. Chem. Soc., 2266 (1953). 28. L. Harris and G. W. King, J. Chem. Phys. 8, 775 (1940). 29. S. Claesson, J. Donohue and V. Schomaker, J. Chem. Phys. 12, 207 (1948). 30. G. Herzberg, Electronic Spectra and Electronic Struc— ture of Molecules, Van Nostrandi 'New Jersey, 1967. 31. J. N. Bardsley, Chem. Phys. Letters 2, 329 (1968). 32. G. Herzberg, Spectra of Diatomic Molecules, Van Nostrand: New Jersey, 1950. 33. U. Fano, Phys. Rev. 124, 1866 (1961). 34. U. Fano and J. W. Cooper, Rev. Mod. Phys. 40, 441 (1968). 35. R. S. Berry, J. Chem. Phyg. 45, 1228 (1966). 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 108 P. Natalis, J. Delwiche and J. E. Collin, Chem. Phys. Letters 3, 139 (1971). F. C. Fehsenfeld, E. E. Ferguson and M. Mosesman, Chem;_Phys.4Lettgrs 3, 73 (1969). J. L. Franklin, J. G. Dillard, H. M. Rosenstock, J. T. Herron, K. Draxl and F. H. Field, Ionization Potentials, Appearance Potentials, and Heats of Formation of Gaseous Positive Ions, (NSRDS-NBS 32, 1969)} Yicto§ Fong, Masters Thesis, Michigan State University 1970 . P. Gray and A. Williams, Chem. Rev. 32, 239 (1959). K. P. Huber, Helv. Phys. Acta 33, 929 (1961). JANAF Thermochemical Tables (Dow Chemical Company, Clearinghouse for Federal Scientific and Technical Information, Springfield, Va., 1967). C. E. Moore, Ionization Potentials and Ionization Limits Derived frgmythe Analysis of thical Spectra, (NSRDS-NBS 33, 19701 A. S. Newton and A. F. Sciamanna, J. Chem. Phys. $3, 4327 (1966). A. S. Newton and A. F. Sciamanna, J. Chem. Phys. 32, 327 (1970). A. B. F. Duncan, Rydberg Series in Atoms and Molecules, Academic Press: New York, 1971. K. Watanabe, F. F. Marmo and E. C. Y. Inn, Phys. Rev. 33, 1155 (1953). K. Watanabe, J. Chem. Phys. 33, 1564 (1954). H. Hurzeler, M. G. Inghram and J. D. Morrison, J. Chem. Phys. 33, 76 (1958). A. J. C. Nicholson, J. Chem. Phys. 33, 1171 (1965). R. M. Reese and H. M. Rosenstock, J. Chem. Phys. $3, 2007 (1966). E. Miescher, J. Mol. Spectry. 32, 130 (1966). 53. 54. 55. 56. 109 D. W} Turner and D. P. May, J. Chem. Phys. 45, 471 (1966). B. Cantone, V. Emma and F. Grasso, in Advances in Mass Spectrometry, E. Kendrik, Editor, The Institute of Petroleum: London, 1968. Vol. IV, p. 599. Y. Tanaka, Sci. Papers. Inst. Phys. Chem. Res. (Tokyo) gig, 456 (1942). O. Edquist, E. Lindholm, L. E. Selin, H. Sjorgen and L. Asbrink, Arkiv Fysik 42, 439 (1970).