UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

High resolution photoelectron spectroscopy of some polyatomic molecules Katrib, Ali 1972

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1972_A1 K38.pdf [ 7.8MB ]
JSON: 831-1.0060133.json
JSON-LD: 831-1.0060133-ld.json
RDF/XML (Pretty): 831-1.0060133-rdf.xml
RDF/JSON: 831-1.0060133-rdf.json
Turtle: 831-1.0060133-turtle.txt
N-Triples: 831-1.0060133-rdf-ntriples.txt
Original Record: 831-1.0060133-source.json
Full Text

Full Text

\\22Q HIGH RESOLUTION PHOTOELECTRON SPECTROSCOPY OF SOME POLYATOMIC MOLECULES BY A. KATRIB B. SC, University of Damascus (1965). C. E.S., University of Strasbourg, (1967). Doctorate 3rd cycle (Nuclear Chemistry), University Strasbourg, (1969). A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in the Department of CHEMISTRY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March 1972 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia , I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and s tudy . I f u r t h e r agree t h a t pe rmiss ion fo r e x t e n s i v e copying o f t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department o r by h i s r e p r e s e n t a t i v e s . I t i s understood that copying o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l ga in s h a l l not be a l lowed wi thout my w r i t t e n p e r m i s s i o n . Department of The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8, Canada Date - i i -ABSTRACT Ay The construction of/^high resolution photoelectron spectrometer (15 mV) has enabled detailed study of molecular ionization potentials. A b r i e f account of hi s t o r i c a l and experimental developments is described. Spectra of some "divalent" sulphur-containing compounds have been recorded, and the effect of different alkyl groups on the sulphur "non-bonding" orbital i s discussed. 3d orbita l participation in some thionyl and sulphury1 halides (S0C12, S0 2C1 2, SOF2 and S O ^ ) i s considered in the interpretation of the spectra. The inductive and resonance effects i n some dibromoethylenes i s discussed, and the assignment of bromine "lone .pair" orbitals i s made. Evidence of (P d) TT bonding in some halisalanes SiH^X (X = F, Cl, Br) and SiH^ (X = Cl) i s discussed, and compared with analogous carbon containing compounds. CNDO/2 and INDO semi-empirical calculations are used i n order to assist the assignment of the PE spectra. - i i i -TABLE OF CONTENTS Page CHAPTER I. INTRODUCTION 1 CHAPTER II. THEORETICAL CONSIDERATIONS 5 .2.1.1 Ionization Potentials 5 2.1.2 Interpretation of PE Spectra 7 2.1.3 Auto-ionization 12 2.1.4 Jahn-Teller Splitting .. 1 2 2.1.5 Spin-Orbit Interaction 13 2.2 Experimental ................ I 6 2.2.1 Angular Distribution of Photoelectrons .... 16 2.2.2 Factors Influencing Resolution in Photo-electron Spectroscopy 18 2.2.3 The Photoelectron Spectrometer 20 Hel (584 A) Source .......... . 20 Electrostatic Energy Analyzer .... 22 Magnetic Shielding, Helmholtz Coils 2 5 The Sample Inlet System, Calibration 2 5 of the Spectra 2 8 Vacuum System 2 9 The Detector 29 CHAPTER III. PHOTOELECTRON SPECTROSCOPY OF SOME DIVALENT SULPHUR COMPOUNDS 3 2 3.1 Introduction 3 2 3.2 Sulphur 3p "Non-Bonding" Electrons 33 - iv -Page 3.3 Individual Molecules 3 6 3.3.1 Hydrogen and Deuterium Sulphide 36 3.3.2 Methyl Mercaptan (CH^H) 5 5 3.3.3 a-Toluenthiol and Benzenthiol 5 9 3.3.4 Dimethyl Sulphide 6 2 3.3.5 Ethylene Sulphide 6 5 3.4 Photoelectron Spectra of Some Disulphide Compounds 69 3.4.1 Introduction 69 3.4.2 S 2C1 2 ..... ........ 70 3.4.3 (CH 3) 2S 2 7 3 3.4.4 (CH 3CH 2) 2S 2 7 4 3.4.5 (CF 3) 2S 2 7 5 CHAPTER IV. PHOTOELECTRON SPECTRA OF SOME THIONYL AND SULPHURYL HALIDES 7 6 4.1 Introduction 7^ 79, 4.2 Individual Molecules 4.2.1 Sulphur Dioxide 7 9 . 4.2.2 Sulphuryl Fluoride ( S O ^ ) 8 2 4.2.3 Sulphuryl Chloride (S0 2C1 2) 9 4 4.2.4 Thionyl Fluoride (SOF2> 1 0 5 4.2.5 Thionyl Chloride (S0C12) m 4.2.6 Bonding in Sulphuryl and Thionyl Halides... 114 CHAPTER V. PHOTOELECTRON SPECTRA OF SOME BROMOETHYLENES ... 118 5.1 Introduction 118 - v -Page 5.2 Individual Molecules 118 5.2.1 Vinyl Bromide 118 5.2.2 2-Bromopropene 122 5.2.3 Cis-l,2-dibromoethylene 124 5.2.4 Trans-l,2-dibromoethylene 126 CHAPTER VI. PHOTOELECTRON SPECTRA OF SOME HALOSILANES 130 6.1 Introduction 130 6.2 Experimental 132 6.3 Individual Molecules 133 6.3.1 SiH. and SiD. ... 133 4 4 6.3.2 H 3SiF ... 135 6.3.3 H 2SiF 2 138 6.3.4 SiH 3Cl 141 6.3.5 S i H 2 C l 2 144 6.3.6 SiHCl 3 146 6.3.7 SiH 3Br 149 6.4 Conclusions 151 APPENDIX 1.1 The Roothaan Method 157 1.2 The Zero-Differential Overlap Approximation (ZDO). 159 1.3 Complete Neglect of Differential Overlap (CNDO)... 160 1.4 The CNDO/2 Parameterization 164 1.5 Intermediate Neglect of Differential Overlap (INDO) 166 1.6 Koopmans' Theorem 167 1.7 The Use of CNDO/2 and INDO i n Photoelectron Spectro-scopy 172 BIBLIOGRAPHY 174 - v i -LIST OF FIGURES Figure Page 1 The energy levels - ionization process 6 2 The Franck-Condon principle in photoelectron spectroscopy 9 3 Light source and 180° hemispherical analyzer unit... 23 4 180° Hemispherical analyzer design and operating parameters 26 + 2 2 5 Spin-orbit components of Ar ( P-jy2> ^ 3 / 2 ^ ^ 6 Scheme of the PE spectrometer 30 7 Plate of the photoelectron spectrometer 31 8 PE spectra of sulphur lone pair orbital in some "divalent" sulphur compounds 34 + + 9 PE spectrum of the f i r s t band of H^ S and D2S 37 10 PE spectrum of the f i r s t band of I^S using Ar o (1048/1067 A) resonance lines 40 2 + 11 The A1 band PE spectrum of H^ S 41 12 Plot of vibrational interval against binding energy 2 + for the band of H^ S . The E arid TT levels are indicated by 0 and A respectively, and the estimated • errors are shown 43 2 13 The A1 band PE spectrum of D2S 46 14 Plot of vibrational interval against binding energy for the 2A1 band of D£S 47 15 2 B 2 band PE spectrum of D2S 51 2 16 B 2 band PE spectrum of H^ S 53 - v i i -Figure Page. 17 PE spectrum of CH3SH .. ... 56 18 PE spectrum of C ^ S H ... 61 19 PE spectrum of (CH^S 63 20 PE spectrum of CH. - CH 66 21 PE spectra of some disulphide compounds (S^X^, X = Cl, CH3, C 2H 5, CF 3) ..... 71 22 PE spectrum of S0 2 80 23 PE spectrum of S O ^ '. . 84 24 Band (a) PE spectrum of S0 2F 2 86 25 Band (b) PE spectrum of S0 2F 2 88 26 Band (c) PE spectrum of S0 2F 2 90 27 Band (e) PE spectrum of S O ^ 93 28 PE spectrum of S0 2C1 2 95 29 Band (a) PE spectrum of S0 2C1 2 97 30 Band (b) PE spectrum of S0 2C1 2 99 31 Band (c) PE spectrum of S0 2C1 2 101 32 Band (e) PE spectrum of S0 2C1 2 103 33 PE spectrum of S0F 2 1° 6 34 Band (a) PE spectrum of S0F 2 ... 108 35 Band (c) PE spectrum of S0F 2 ... 109 36 PE spectrum of S0C12 112 37 The PE spectrum of CH2=CHBr 119 38 The PE spectrum of 2-CH3~CBr= C N 2 1 2 3 39 The PE spectrum of cis-l,2-CHBr=CHBr 125 - v i i i -Figure Page 40 The PE spectrum of trans-1,2-CHBr=CHBr 127 41 PE spectra of SiH^ and SiD^ : 134 42 PE spectra of fluorosilanes - 140 43 PE spectra of chlorosilanes 142 44 PE spectrum of SiH^Br 150 45 IP correlation diagram of some difluorocompounds ... 152 46 IP correlation diagram of some dichlorocompounds ... 153 - ix -LIST OF TABLES Table : Page 1 Ionization Potentials of the Sulphur "Lone Pair" i n some RSR* Compbunds 35 2 Ionization Potentials for I^S and Tt^S and Vibrational Frequencies of Ions Produced 38 + 2 3 Ionization Data for R^S ( A^) Photoelectron Band ... 45 + 2 4 Ionization Data for the D^S ( A^) Photoelectron Band 50 2 5 The B 2 Band of D,,S and H2S - The Most Prominant Progression 49 6 Vertical Ionization Potentials of Methyl Mercaptan (eV) 57 7 Vertical Ionization Potentials of Dimethyl Sulphoxide (eV) 64 8 Vertical Ionization Potentials of Ethylene Oxide and Sulphide (eV) 67 9 Vertical Ionization Potentials of Some Disulphide Compounds (eV) 72 10 Ionization Data for S O ^ 83 11 Vibrational Structure on PE Spectrum of S0 2C1 2 96 12 Ionization Data for S0F 2 105 13 Ionization Data for S0C12 113 14 Properties of the SO Bond 115 15 Some Average Energies (in eV) for Orbitals Mainly Localized on the Oxygen Atoms 115 16 Vertical Ionization Potentials of Bromoethylenes (eV) 120 17 Vertical IP's of SiH^F and CH3F (eV) 137 18 Ionization Potentials for SiH F 0 and CH F (eV) ...... 139 - x -Table Page 19 Vertical IP's of SiH 3Cl and CH3C1 (eV) 143 20 Vertical IP's of S i H 2 C l 2 and CH2C12 (eV) 145 21 Vertical IP's of SiHCl 3 and CHC13 (eV) 147 r - xi -ACKNOWLEDGEMENTS I would like to take this opportunity to express my deepest appreciation and gratitude to Professor D.C. Frost for his invaluable help, encouragement and guidance throughout the course of this work. I also wish to thank Professor CA. McDowell for his support and continuous interest i n this work. Thanks are also due to Dr. F.G. Herring, Dr. A.J. Merer, Dr. R.A.N. McLean, Dr. D. Chadwick, Dr. A.B. Cornford and A. Bain for their helpful discussions and assistance. I wish to gratefully acknowledge the s k i l l of the staffs of the Mechanical, Electronic and Illustrations workshops, and especially J. Wyngaarden for assistance i n construction and maintenance of the photoelectron spectrometer. I express my sincere gratitude to my wife for her encouragement during this work. - 1 -CHAPTER I INTRODUCTION Molecular photoelectron spectroscopy (PES) is a relatively new technique for determining ionization potentials (IP's). The 1-3 f i r s t experiments were carried out independently by Turner's and Terenin's^ groups in the late 1950's. Monochromatic ultraviolet radiation usually of energy below 60 eV is used to photoeject electrons from-a target of free atoms or molecules. Kinetic energy analysis of the ejected photoelectrons produces a photoelectron spectrum, which yields the IP's of the species under study. The kinetic energy E g of the ejected electrons is given by the Einstein relation E = h v - ( I + E + + E+., + E + J (1-1) e e n vxb rot I i s the adiabatic IP for the molecular orbital n e + E is the translational energy imparted to the ion + + E ., and E ^ are the vibrational and rotational energy of the ion vxb rot b } with respect to the neutral molecule. The translational energy of the ion is usually negligible, because conservation of momentum requires that the energy be almost entirely imparted to the electron. In the case of the hydrogen molecule, the mass difference i s approximately 4000 to 1 and the electron i s - 2 -therefore produced with about 4000 times the kinetic energy of the ion. This ratio i s larger for heavier molecules. In the case of an atom, there is no vibrational or rotational energy involved, and the IP as given by the relation (1-1) corresponds purely to an electronic transition to excited states. In high resolu-tion photoelectron spectroscopy i t is quite common to observe vibrational structure in most molecular spectra. Quite recently i t has become possible to resolve rotational structure on the PE spectra of H* 6 and H 20 +. 7 As mentioned above, two groups originated molecular photoelectron 4 spectroscopy; Vilesov et a l . used a vacuum UV monochromator as the source of ionizing radiation, separated by an LiF window from an ionization chamber which contained a pair of cylindrical grids for electron energy analysis. vThese workers were restricted to an ionizing energy cf less than 11.7 eV. Al-Joboury and Turner"'" employed a windowless system, with He (I) o resonance line (584 A = 21.2168 eV) as the source of incident radiation. 9 ° Price has used the He.(II) resonance line (304 A = 40.8 eV) by employing much higher current density and reducing the helium pressure in the light source. Other workers have also used the He(II) resonance line to obtain higher IP's of molecules. 8 ^ Frost et a l . ^ used o He 537 A resonance line at 23.09 eV, which i s 2% of the intensity of o the He(I) 584 A line to obtain the PE spectrum of neon. Argon and krypton have been used in the light source to lead to microwave discharge 11-13 radiation of energies less than 17 eV. Collin and co-workers have used Ar and Ne resonance lines to obtain the "PE spectra of NO - 3 -and H_S. Price was able to obtain a very high resolution oxygen spectrum by using the Ne excitation lines. There are a considerable number of review articles which deal with PE spectroscopy 15-26 up to 1970. A related technique involves the use of X-rays as an ionizing source. The K lines of Al (1486.6 eV) and Mg (1253.6 eV) are usually ct used as X-ray sources, which enables the study of core electron binding energies. This i s called ESCA or "Electron Spectroscopy for Chemical Analysis", by K. Siegbahn et a l . 27-28 This technique also has been described by Hercules. 29 The resolving power of an ESCA spectrometer is usually ~1 eV for the best working conditions. A new commercial ESCA 200 (AEI) spectrometer is able to obtain a line width of 0.90 eV (Ag 3d peak) by using an X-ray monochromator. It i s usually not possible to observe IP's separated by less than 1 eV in energy, spin-orbit couplings, Jahn-Teller s p l i t t i n g , vibrational structure, etc. In some cases the ESCA technique i s useful to clear up discrepancies regarding IP's i n the region of 21 eV (due to the presence of hydrogen Lyman a and 6 i n the light source) and this w i l l be discussed later. There are several different methods which have been used to obtain IP's of atoms and molecules. The most important are: optical spectro-scopy (most accurate); electron impact (most generally applicable) and photoionization. These have been discussed i n different review articles; we refer here to the most recent ones which give information about the importance and applicability of these methods. - 4 -The main advantage of molecular photoelectron spectroscopy (PES) compared to other electron spectroscopic techniques is i t s capacity to yield information about different electronic states of the positive ions. This technique is therefore very useful to quantum chemists who * may use the IP s to test molecular energy level calculations. PES i s not a threshold technique, consequently the spectra are not generally complicated by processes such as autoionization (from bound excited state(s)). The use of Koopmans' theorem and the semi-empirical calculations in PES are discussed in the appendix. - 5 -CHAPTER II THEORETICAL CONSIDERATIONS 2.1.1 Ionization Potentials It i s important to distinguish between adiabatic and ve r t i c a l ionization potentials. The former i s the difference i n energy between the ground vibrational levels of the molecule and the resulting ion, and generally corresponds to the f i r s t peak or the threshold of the PE spectral band. The latter ionization potential corresponds to the most probable ionization transition from the ground state of the molecule. The ionization potentials given by electron impact are often close to the "ve r t i c a l " values, however, those given by optical spectroscopic or photoionization techniques are usually adiabatic. The f i r s t ionization potential (IP) may be interpreted as involving the removal of an electron from the most weakly bound orbital (Figure 1), which i s represented by the f i r s t band on the PE spectrum, the second ionization potential (H^) corresponds to removal of an electron .from the second most weakly bound molecular orbital and leaves the ion in its. f i r s t excited s t a t e , a n d so on. Molecular-Configuration Possible ionic Configuration .Figure 1 Energy of Photoelectron hi/- IP hv-IF| - 7 -2.1.2 Interpretation of PE Spectra The interpretation of the band shapes in PE spectra is based largely on the application of the Franck-Condon principle, which states that an electronic transition takes place so rapidly that a vibrating molecule does not change i t s internuclear geometry appreciably during tie transition. It i s assumed generally that a l l the transitions occur from the ground vibrational state, as can be seen in Figure 2 for diatomic molecules. Any transitions occurring from excited vibrational states w i l l result in "hot bands", which may lead to misinterpretation of the spectra. The transition probability from the vibrational ground state (v" = 0) to the various ionic vibrational states (v' = 0,1,2...) is given by the Born-Oppenheimer equation where G e, e„ i s the electronic transition moment, which varies slowly with internuclear distance r, and J^>^tip^„dr i s the vibrational overlap integral between the i n i t i a l and f i n a l states. Since the electronic transitions occur so rapidly compared to vibrational motion by the molecule, the maximum probability for transition corresponds to the central v e r t i c a l line in Figure 2. The shaded area i s called the Franck-Condon region. An effective Franck-Condon region for transitions from the i n i t i a l state can be defined in terms of the maximum and minimum internuclear distance from which observable transitions can occur. Transitions w i l l be observed only to those upper states which (2 -D - 8 -have vibrational levels of high probability within this internuclear separation range. The width of the Franck-Condon region i s effectively o quite small (0.1-0.2 A) for molecules with medium to large bond strength. The adiabatic IP refers to the transition between the ground molecular state v" = 0 and the corresponding ionic state v 1 = 0. The ver t i c a l IP corresponds to the most probable transition between the ground state of the molecule v" = 0 and the corresponding ionic state. This can be seen as the most intense peak in a band with vibrational structure, otherwise the ver t i c a l IP i s measured from the maximum of a structure-less band. It i s sometimes not possible to obtain adiabatic IP's from the spectra of some molecules when the internuclear distance of the ion 30 r t i s much larger than the corresponding molecule r „, i.e. CH, , 31 NH^  and ND^ . Figure 2 shows the most common transitions between the ground state of a molecule and some different ionic states, and the resulting band shapes expected i n the PE spectrum. If the ionized electron i s from a non-bonding MO the potential surface of the ion A w i l l be almost exactly above that of the molecule M. The corresponding band shape w i l l consist of a strong peak (0 *• 0 transition) possibly accompanied by one or two small peaks of much lesser intensity. In this case the adiabatic IP i s the same as the vertical value. If the ionized electron has antibonding character, the inter-nuclear distance r g , w i l l decrease with respect to the ground state 'Xi of the molecule, and the potential surface B w i l l be shifted to smaller r £ . If a bonding electron i s removed on ionization, then the inter-nuclear distance of the ion r , w i l l be larger than r „. The bonding THE F R A N C K - C O N D O N PRINCIPLE IN PHOTOELECTRON SPECTROSCOPY INTERNUCLEAR DISTANCE, r — Figure 2 - 10 -being weaker, the potential energy surface of the ion C w i l l be displaced to greater r^. The f i r s t vibrational peak in the band corresponds to the adiabatic IP, and the most intense one, which corresponds to the v' = 4 «- v" = 0 transition, corresponds to the ve r t i c a l IP. The vibrational peaks in the spectrum ion are approximately equally spaced, however, this spacing decreases constantly i n case D due to the anharmonicity of the upper state causing the excited levels to converge. This is followed by a loss of structure i n the spectrum which indicates a dissociation process. Ionization to a purely repulsive surface il w i l l result i n a smooth structureless band. A predissociation process may occur when the upper state curve i s crossed by an unstable state of short lifetime F*. This exemplified by broadening of the PE peaks followed by a less of structure on the corresponding band. The energies of vibrational levels of diatomic molecules are given by 1 1 2 Ev = (v* + 2")hu) ~ (y* + f) hxco + s m a l l e r terms (2-2) where v* = vibrational quantum number in the ion h = Planck's constant X = anharmonicity constant • i fir ui = y r j ^ " > t n e vibrational frequency K = force constant of the vibration M = the reduced mass - 11 -K i s a measure of the curvature of the potential well at the equilibrium internuclear distance, and indirectly i s a measure of the bond dissociation energy. If a non-bonding electron i s removed upon ionization, the bond length remains essentially unchanged,consequently there w i l l be no change in K ( r e») a n a w> and the vibrational spacing w i l l remain unchanged. If a bonding electron is removed, K and w w i l l decrease with respect to the neutral molecule, since the bonding i s now weaker. The inter-nuclear distance r £ , w i l l increase. The magnitude of the vibrational spacing decreases to an extent depending on the nature of the bonding, and can be reduced as much as 50% for a very strong bonding orbital. If the ionized electron i s from an antibonding orbital, K and w w i l l increase, so Ihe vibrational spacing of the resulting ion w i l l be equal or higher than i s the case in the neutral molecule, and r g , w i l l decrease. The above discussion can be extended to polyatomic molecules where multi-dimensional potential surfaces are involved, and changes in bond angles and bond length upon ionization w i l l possibly result i f several vibrational modes may be excited simultaneously. Analysis of these spectral frequencies i s not always straightforward, and can sometimes lead to misinterpretation of the spectra. 32 Turner obtained an empirical linear relationship between the vibrational spacing in the neutral molecule and the resulting ion, in the case where r g changes l i t t l e upon ionization. Since the difference between the vertical (I T T) and adiabatic (I.) IP's (IP -IP.) i s a V A V A function of the fractional change in internuclear distance. He found that: - 12 -(2-3) this relationship can be expressed as follows by considering Mecke's relation wa r AE = IP V - IP A = 1 . 2 < £ - 1) (2-4) for a large number of small molecules. Where w" and w' are the vibrational frequencies of the neutral molecule and the resultant ion. 2.1.3 Autoionization If a molecule is excited with radiation of wavelength which coincides with an autoionizable state, then there i s a certain probability that the molecule w i l l ionize. This process has been 33 described by Price as one in which the excited inner electron gives up i t s energy to a more loosely bound electron, and ejects i t by a kind of internal electron bombardment. He observed this effect in oxygen, nitrogen and NO by using a neon-resonance line source (16.67 and 16.85 eV). Similar effects have been observed for nitrogen and carbon 34 disulphide by Collin et a l . The autoionization process in diatomics has been discussed in several art i c l e s . ' 2.1.4 Jahn-Teller Splitting 37 The Jahn-Teller theorem states that a non-linear molecule in a degenerate electronic state i s unstable toward nuclear distortions which lower the molecular symmetry and thereby remove the electronic - 13 degeneracy. For a t r i p l y degenerate electronic state, we may obtain three non-degenerate or one non-degenerate and one doubly-degenerate state depending on the type of displacement. For one doubly-degenerate electronic state, we obtain for the displaced position of the nuclei two non-degenerate electronic states of slightly different energy. The magnitude of the Jahn-Teller interaction in a given degenerate electronic state i s d i f f i c u l t to predict. It. depends on the way in which the various f i l l e d molecular orbitals affect the vibrational 38 motion. Liehr has shown that a molecule in a degenerate electronic state w i l l reduce in symmetry only as far as the nearest point group which w i l l remove i t s degeneracy. The Jahn-Teller s p l i t t i n g i n the PE spectrum of methane has been 30 39 discussed by Siegbahn et a l . and Dixon, and has been observed by 15 9 Turner et a l . and Price et a l . The t r i p l y degenerate ( l t 2 ) MO in the ground state of this molecule is s p l i t into three non-degenerate electronic states in the corresponding molecule ion. Jahn-Teller 9 s p l i t t i n g has been observed i n the PE spectra of methyl halides, 40 41 42 43 benzene, cyclopropane, SiH^, * and SiD^ (see Chapter VI). The corresponding effect i n degenerate linear molecules i s called the Renner-Teller effect. 2.1.5 Spin-Orbit Interaction In molecules (C^ v or greater) of sufficient symmetry, the electronic spin angular momentum may couple with the electronic orbital angular momentum. This spin-orbit interaction w i l l lead to two separate states o f different energy. Removal of one electron from a closed shell rare - 14 -gas atom produces a doublet spin state and the odd electron has 2 angular momentum I - 1. The spin-orbit interaction leads to a n <* 2 j?l/2 states i n the ion. The magnitude of the spin-orbit interaction depends on the nature of the atoms in a given molecule, and has been 9 44 observed in the PE spectra of several molecules. ' In order to calculate the extent of this interaction, the factors influencing the spin-orbit interaction are now given. An electron i n i t s motion experiences time-varying electric fields due to i t s motion relative to the other charged particles in a molecule. Through Maxwell's equations these variations produce magnetic fields which interact with the spin magnetic moment. The energy of interaction of a magnetic f i e l d B and a magnetic moment u i s given as: • » - - < " > - - " * ? < £ > " V S . (2-5) c r where y = -gBS (2-6) from the relation (2-5), the spin-orbit interaction i s given as: H < j n - " ^ 9 E S < — K T ^ ) * S (2-7) SO „ 2 e n r d r en e zmc en en where £ = r x mv en en e i s the spin angular momentum g is a constant arising from r e l a t i v i t y en 8 (Bohr magneton) i s the unit of magnetic moment - r — 2m - 15 -is the orbital angular momentum of electron e with respect to nucleus n r i s the position vector of the electron with respect to the nucleus en r v & i s the velocity of the electron with respect to the nucleus V i s the potential at the electron arising from the nucleus en- ° This equation is used only after the following assumptions: (a) only nuclei have been considered i n the elect r i c fields E (b~) the nuclei have been assumed at rest (c) the potential V , i s spherical with respect to r e n> a n ^ (d) for several electrons, the energy i s additive. In considering further approximations: (1) the potentials V" e n as Coulombic potentials for each nucleus, i.e., eZ V = (2-8) en 4TTE r o en where Z i s the effective charge on the nucleus, and Aire i s the e i i o permittivity of free space. E q is dielectric constant, then . . dV -eZ 1 en eff ( 2 _ 9 ) r dr . 3 en en 4ire r o en (2) assuming hydrogen-like atomic wavefunctions, the average value of becomes ^ —x Becomes —x—~ z r a n 4(4 -4 ) (£+1) en o z a i s the Bohr radius o n and 4 are quantum numbers - 16 -If Z i s taken as Z^^* the spin-orbit interaction is proprotional 4 45 to ( z e£f)• This reflects the importance of the interaction in heavier atoms; in other words, because the outer electron i s at a large distance from the nucleus, the electrostatic interaction with the inner core of electrons around the nucleus is weak. The spin-orbit splittings i n the rare gas atoms are 0.178, 0.665 and + + + 1.306 eV for Ar , Kr and Xe respectively. This increase with respect 46 47 49 to the atomic number can be seen also i n halogen atoms and molecules ' as F = 404 cm"1, Cl = 881 cm - 1, Br .= 3685 cm"1 and I = 7603 cm"1, CH3C1 = 630 cm"1, CH^Br = 2560 and CH3I = 5050 cm"1. In the case of polyatomic molecules, the spin-orbit interaction i s dominated by contributions from heavy atoms, also the spin-orbit splittings w i l l be reduced when a.heavy atom is bonded to light atoms. 2.2 Experimental 2.2.1 Angular Distribution of Photoelectrons The angular distribution of electrons ejected from atoms and molecules by photon impact has been the subject of several theoretical^^ 53—58 and experimental studies. For an atom or a randomly oriented molecule, such as in the gas phase, an unpolarized photon beam gives rise to an angular distribution 1(6), given by the general form 1(9) = ^ [1 - | (3cosV 1)] (2-10) o" t o t represents the total cross section f o r ionization 0 the angle between the direction of the photoelectron and the incident photon beam - 17 -£ i s a parameter dependent on the energy of the photoelectron and the nature of the molecular orbital. It i s commonly used to characterize the angular distribution. The dependence of angular distribution on the nature of the MO 59 can be used to assign the symmetry of this MO. Manson et a l . have studied the angular distribution of photoelectrons from ethylene. They observe that the angular distribution of the electrons from the f i r s t band, representing the (C=C)iv MO, is different from the other two 2 + (a MO's). This difference has been observed also for the £ and g 2 58 TT states of N- and CO by Carlson and Jonas, u 2 J It i s expected on theoretical grounds that the ratio of the vibrational peaks within an electronic band w i l l be angle-dependent, since Franck-Condon factors generally are angle-independent. Niehaus 56 2 2 et a l . have measured the ratio y(6) of the two Ar ^ij2 a n c* ^3/2 o spin-orbit components by using Ne 744 and 736 A resonance lines and the relationship , I ( e)( 2P, / ?,744 A) R(0) = Z 5 - = C"1 y ( 6 ) (2-11) I ( 9 ) ( P3/2' 7 3 6 A ) o o where C is the intensity ratio of the 736 A to the 744 A radiation. y(6) was' found to be angle independent and equal to 0.54. The angular distribution for a vibrational peak in a given electronic band i s angle dependent. In the case of the Kr spin-orbit components the maximum intensity was found at 0 = 36.5° by using Ne 0 58 736 and 744 A resonance lines. It i s also found that the angular distribution at high pressure i s pressure-dependent, therefore, the measurement should be taken under conditions where the pressure effects - 18 -are negligible. This can be done by keeping the pressure low in the col l i s i o n chamber during the period of measurement. The angular distribution experiments can be carried either by 56 58 rotating the detector or the co l l i s i o n chamber. This technique w i l l probably prove useful for identifying the symmetry for a given MO, and further theoretical and experimental developments are anticipated. 2.2.2 Factors Influencing the Resolution in PE Spectroscopy One of the advantages of PES over electron impact techniques i s i t s capacity to obtain high resolution spectra which enable one to observe the vibrational and in some cases rotational'' structure of molecule ..ions., .also to observe .splitting in the PE bands due to other effects, i.e. spin-orbit coupling, Jahn-Teller splitting...etc. Unfortunately there are some limitations on the resolution i n PES. The resolution i s usually expressed by the f u l l width measured at + 2 half the maximum height (fwhm) of the Ar 3^/2 c o m P o n e n t ' Most of the factors influencing the resolving power of the instrument have been d i s c u s s e d ^ a n d they are as follows: 1. Instrumental Factors (i) the energy analyzer, where the theoretical resolution AE of a 180° spherical analyzer i s given by the relation*'"' - 19 -E the energy of the photoelectrons u) width of the entrance or exit s l i t R the radius described by the electrons (mean radius of the o J analyzer) ( i i ) Field gradients, and fluctuation of the magnetic f i e l d near the analyzer. Magnetic shielding in the form of Helmholz coils or u-metal shielding is necessary to reduce these fields to a few milligauss. ( i i i ) Electrode surface potentials. It i s very important to minimize any f i e l d distortion due to local variations i n the potentials of the electrodes, particularly when measuring low energy electrons. In order to obtain constant potentials over a l l exposed electrodes; they can be gold plated or treated with Aquadag (colloidal suspension of carbon). (iv) The radiation source (a) Natural broadening (lifetime of the resonant state) which _ c o o i s of the order of 3.3 x 10 A for He 584 A. —6 ° (b) Pressure (resonance) broadening which i s 7.5 x 10 A/Torr. The microwave discharge i s operated at the lowest possible pressure, o which results in a half width of ^ 1 mV for He 584 A. (v) The nature of molecules studied The lifetime of the ionic states produced in the c o l l i s i o n chamber w i l l be affected by: (a) Radiative transitions to lower ionic levels; these lead to a broadening of ^ 10 7 eV. (b) Non-radiative transitions to other ionic states (dissociation) w i l l impose lifetimes of =10 sec and a broadening of 'v.lO * eV.^ For predissociation states, the lifetime changes from state to state. In the case of large polyatomic molecules, where many degrees of freedom - 20 -are involved, many vibrational modes are expected to overlap with each other. (c) Doppler broadening, which results from the motion of the molecules and electrons with respect to the stationary detector. This can be expressed by the relation.^ AE = 2m V V (2-13) e e m where m is the mass of the electron, V ' and V are the average e ' e m velocities of the electron and the molecule respectively. For a given PE band therefore, the resolution i s dependent on the energy of the incident radiation. —3 61 AT room temperature for m = 100 and E = 20 eV, AE = 2.5 x 10~ eV. (d) The pressure of the sample in the co l l i s i o n chamber. This should be kept as low as possible. The best resolution that has been obtained i n PE spectroscopy i s between 4 to 7 mV,^  however, most of the workers i n this f i e l d now commonly employ;- a resolution of the order of 15-20 mV. 2.2.3 The Photoelectron Spectrometer o The Hel (584 A) Source Potential ionizing sources for electron spectroscopic studies include electrons, photons, X-rays, and metastable atoms. The advantage of using resonance radiation i s mainly because i t is practically monoenergetic. The most commonly used photon source o i s the helium I resonance line at 584.33 A (21.2168 eV) which i s very - 21 -strong line emission, and may be generated by a microwave or o condensed discharge in helium. Some 537 A radiation (23.09 eV) i s o present (^  2% of the helium 584 A resonance l i n e ) , which results in a low intensity 'ghost' spectrum shifted by 1.88 eV to the higher kinetic energy side of the 584 A spectrum. This can be used to obtain spectra of electrons of binding energy 21-23 eV, e.g. Ne with an IP at 21.565 eV."^ The presence of impurities i n the helium w i l l of course cause unwanted radiation; for example hydrogen Lyman a (1215 o A = 10.20 eV), and Lyman g (12.09 eV) appear to be quite common. The presence of these impurity lines may lead to false interpretations of the spectra. 9 8 61 ° Price et a l . and Brundle et a l . ' have used the 304 A (40.8 eV) o line of. the "He"(IT) to obtain IP's between 30 and 40 eV. The 304 A line is obtained by employing a high current density and a low helium pressure (^  5 microns). Other sources of energetic photons are discharges in rieon and argon. Argon has two resonance lines at 1067 ° 11 and 1048 A (11.62 and 11.84 eV). Collin et a l . have used these emissions to obtain the spectra of ^ S and ^ Se. The main reason for the use of the Ne and Ar resonance lines i s to obtain spectra of better resolution and to observe variations i n spectra with incident photon energy. The main disadvantage of these lines i s of course their doublet nature. In the research reported here, the discharge takes place i n a 62 quartz tube within a resonant cavity, and the power Is supplied by a 100 watt 2450 mC/S Raytheon Microtherm generator. The power supply i s regulated by an A.C. voltage regulator which improves the s t a b i l i t y - 22 -of the discharge in the resonant cavity. The light source appears to be a yellowish peach colour when i t i s operating correctly. When a l i t t l e a i r i s introduced the colour changes to blue-violet. Tank helium (Canadian Liquid Air Co.) i s used without any special treatment. —6 The helium pressure in the instrument is 1 x 10 torr, and the helium flow i s controlled by an Edwards high vacuum OSIC stainless steel needle valve, through an axial quartz discharge tube 25.2 cm and diameter of 3 cm. A brass sheathing into which f i t s both the quartz discharge tube and collimating capillary (10 cm long and 0.5 mm diameter) i s constrained to a linear configuration by the solid brass housing, and each is sealed by 009 Viton 0-rings. The space between the quartz discharge tube and the collimating capillary i s diff e r e n t i a l l y pumped to prevent helium from entering the c o l l i s i o n chamber and the analyzer. These paths are aligned carefully to obtain the maximum transmitted o flux of He I 584 A radiation (Figure 3). Forced a i r i s used to cool the discharge tube. Electrostatic Energy Analyzer 4-5 The type of energy analyzer originally used by Vielsov et a l . 1-3 and Turner was based on retarding potential grids. Ionization takes place along the axis of two cylindrical grids, and a retarding f i e l d isapplied between them. IP's w i l l be distinguished i n the spectrum by steps or peaks. This type of analyzer has rather poor resolution, the main reason being the ejected electrons do not a l l move at right angles to the cylinder axis. This problem can be avoided by using a spherical grid arrangement where electrons are ejected nearly - 23 -A - 180° Hemispherical analyzer B - Vacuum Enclosure C - Channeltron M u l t i p l i e r D - C o l l i s i o n Chamber E - Lens F - SamDle Gas i n l e t G - C o l l i m a t i n g C a p i l l a r y H - 0-Ring Seal I - Boror N i t r i d e C o n s t r u c t i o n J - He p>;mp o f f K - Quartz Discharge Tube L - Microwave Cavit y M - A i r Cooling N - Tuning Stub 0 - Microthern J u n c t i o n P - He i n l e t Q - Needle Valve o :ar. Figure 3 - 24 -at right angles to the retarding f i e l d . A best resolution obtained in 6 3 these laboratories for this type of analyzer was 30 mV. This analyzer does have the advantage of almost total collection of electrons. The main problem of this kind of analyzer appears to be f i e l d penetration through the grids, and accumulation of space charge on the grid surfaces. A 127° electrostatic velocity selector has been 64 described by Turner which has a resolving power of 15 mV for the 2 2 Ar ( P^^2» P i / 2 ^ Pea^-S* The mean radius of this electrostatic selector was 10 cm, which led to a maximum solid angle of 0.0128 Sr compared with 0.63 Sr for a f u l l y u t i l i z e d hemispherical condenser of the same radius. Usually only electrons emitted at right angles to the photon beam may be collected. In the present work a 180° hemispherical electrostatic energy analyzer has been constructed. The capability of the hemispherical analyzer to accept a relatively large solid angle permits transmission of a relatively high electron flux. This type of analyzer i s similar 31 to that described previously by Branton et a l . The analyzer electrodes, a hemishere and cup, machined from solid brass, have ra d i i 8.75 and 11.25 cm respectively, with a mean electron trajectory of 10 cm. The electrode surfaces exposed to the electron trajectory are treated with Aquadag in order to minimize any f i e l d distortion. The two electrodes are fixed on the upper cover of the vacuum enclosure by nylon screws and boron nitride washers for e l e c t r i c a l insulation. The target chamber was constructed from brass i n the form of a cylinder 1.5 cm long and 2.2 cm internal diameter. Photoelectrons pass through an exit hole (diameter 0.85 mm) at 90° to the photon beam. - 25 -This hole i s d r i l l e d in the center of the co l l i s i o n chamber. The lens element (Figure 3) with circular aperture (diameter 1.85 mm) is e l e c t r i c a l l y isolated from the target chamber by a teflon washer. This lens i s used in focusing the photoelectrons emerging from the 65 target chamber, and defining their launch angle of incidence a (Figure 4) upon the entrance s l i t to the analyzer. A high voltage (^  60 volts) applied on the lens causes strong focusing and optimum collection of electrons while increasing the total effective launch angle of electrons at the analyzer surface. A lower voltage (less than 10 volts) results in reduced intensity and possible non-focussing at the analyzer entrance with a subsequent decrease in resolution. The alignment of the holes in the target chamber and the lens i s carried out by fixing the positions of the c o l l i s i o n chamber and the lens i n a teflon housing inside the vacuum enclosure. The target chamber potential i s varied to scan the PE spectrum, with the potential across the two hemispheres and therefore the resolution constant. The practical resolution obtained on the Ar doublet i s of the order of 15 mV (Figure 5) with maximum intensity 5000 c/s and an Argon pressure of 4 x 10 ^ torr. Magnetic Shielding, Helmholtz Coils The efficiency of an electrostatic energy analyzer i s dependent upon several factors, and possibly one of the most important are the magnetic fields in the region of the electron trajectory. These undoubtly affect the transmission and the resolving power of a PE spectrometer. The magnetic fields originate from various sources in and out of the laboratory, and of course the earth's magnetic f i e l d . target chamber X 2 : r j / i ? 0 = = - ( - ^/^o)+2 (A£/£ 0)-2a 2, 180° HEMISPHERICAL ANALYSER UNIT FIGURE 4 15759 eV L _ i i . 1570 15-90 Figure 5 - 28 -Shielding of some sort is needed to cancel these fi e l d s . Three pairs of Helmholtz coils mounted at right angles were used in our case. The current through each pair i s controlled by a 20 volt, 5.7 amp. lambda power supply. Different combinations of current through the pairs of Helmnoltz coils are possible in order to focus the 27 electron beam on the exit aperature. Siegbahn et a l . have used a servo system in conjunction with coils to counteract magnetic f i e l d changes and fluctuations occurring during the course of a run. The use of u-metal shielding may replace the Helmholtz coils. The Sample Inlet System, Calibration of the Spectra In the case of gases, the sample is introduced through a Veeco valve to a small reservoir (300 ml) and is passed to the c o l l i s i o n chamber via a Granville-Phillips 203 variable leak valve. Liquid samples are degassed by using liquid nitrogen traps. A 4 l i t e r glass reservoir i s used to store argon for calibration purposes. It is. of course very important to ensure the purity of the sample under investigation. Common gases may be identified by their PE spectra, however, in the case of unstable compounds they may interact with the sample line surface and decompose. We have observed that several of the halocompounds have remained inside the sample line to interact with subsequent samples and lead to various impurity peaks i n the spectra. Where possible, purity of the samples studied here has been verified by instrumental apalysis (i.e., mass spectrometry, IR, NMR, etc.). - 29 - Vacuum System The vacuum system consists of: 1. the sample system roughing pump, 2. the light source differential pump, which removes the helium. The main chamber is evacuated by an o i l diffusion pump (C.E.C. type MCF 60) coupled with a liquid nitrogen trap, which i s backed by a rotary pump (Welch Duo Seal, No. 1405H). The pressure i n the vacuum chamber was measured with a Veeco ionization gauge RG75K. The pressure measured in the system i s of the order 2 x 10 ^  torr, and the working pressure when the sample is introduced is ^  2 x 10 "*torr. This i s dependent on the nature of the sample under investigation. The vacuum system can be seen in Figure 6. The Detector The detector chamber consists of a circular lens with a central hole (diameter 0.85 mm) at earth potential. The photoelectrons are accelerated into a Mullard channeltron multiplierj which has a gain 4 5 of 10 to 10 at 3.2 KV. The spectrum i s scanned by varying the potential applied between the c o l l i s i o n chamber and the analyzer. The scanning potential is obtained by amplifying a 4-volts ramp originating from the multichannel analyzer. This ramp also serves as internal trigger for the multiple scan. The whole system i s similar to 31 that described previously by Branton et a l . , except that there i s no dif f e r e n t i a l pump connected to the detector chamber (Figure 6,7). 24 22- 21 - 20 ; 23 1. 180° Hemispherical analyzer 2. Vacuum enclosure 3. Collision Chamber 4. Vacuum U.V. Light Source 5. He Cylinder 6. Microwave discharge Power supply 7. Rotary fore pump 8. Channeltron Multiplier 9. Head Amplifier 10. Argon reservoir 11. Diffusion pump 12. Rotary Fore pump 13. Rotary Fore pump 14. Scanning potential For Collision Chamber 15. Controls for Electron lens System 16. Controls for Energy Analyzer 17. Ion Gauge 18. Acceleration Voltage 19..' Pulse Amplifier 20. Discriminator 21. Fabriteck 1000 Chan. Analyzer 22. Magnetic tape 23. X-Y Plotter 24. Ramp to Energy Analyzer Figure 6 - 32 -CHAPTER III PHOTOELECTRON SPECTROSCOPY OF SOME DIVALENT SULPHUR COMPOUNDS 3.1 Introduction In mercaptans, the sulphur atom is generally thought to participate in bonding via a bonds involving mainly two of the three 3p orbitals on the sulphur atom leaving the 3s orbital and the remaining 3p orbital to accommodate the four non-bonding electrons. The involvement of the 76 sulphur 3d orbitals in the bonding seems to be negligible. The formally 3s non-bonding pair is much more tightly held by the sulphur atom than the 3p "non-bonding" pair. The replacement of hydrogen atoms in E^S by electron donating groups leads to a destabilization of these electron pairs, which i s reflected in the IP's of these molecules. Photoelectron spectroscopic studies have been carried out previously on divalent sulphur compounds. In particular the PE spectrum of hydrogen sulphide is the subject of research by Delwich et a l . 1 1 and Turner et a l . 1 ^ Both groups observed a regular progression of five -1 2 peaks (mean spacing 'v 850 cm ) in the second band of H^ S ( A^) followed by a loss of structure at 13.4 eV. As the appearance + 71 potential of S /E^S is at 13.4 eV, they attributed this loss of Structure to the predissociation - 33 -+ 2 H2SVA2) H-^E + ) + S 2 g + The PE spectrum of H2S presented in this . work show additional 2 structure on the band after 13.4 eV. 3.2 Sulphur 3p "Non-Bonding" Electrons It is usually quite easy to recognize the band corresponding to the sulphur non-bonding orbital "lone pair" in the PE spectra of divalent sulphur compounds. It can be seen as a relatively sharp peak compared to most other bands in the spectrum which are much broader. This is due to the negligible change in the Franck-Condon envelope between the neutral species and the ion. In a l l compounds studied in this work, the sulphur "lone pair" corresponds to the f i r s t IP in the PE spectrum, except in C^H^CH2SH where the alkyl substituent has an IP to lower energy. Figure 8 shows the band containing the sulphur "lone pair" in the PE spectra of a series of compounds RSR'. We see that the strong peak in H2S which corresponds to ionization of a (mainly) 3p non-bonding orbital of the sulphur atom becomes less sharp in the mercaptans, and is often accompanied by vibrational structure. In the mercaptans (RSH) we observe a constant decrease in the IP of the sulphur "lone pair" as the number of carbon atoms in the alkyl group R increases. If we consider A.,^ , as the difference between the IP of H„S ("lone pair") at 10.47 eV and RSR', we see that A , increases as the number K K of carbon atoms in R increases when R' = H (Table I). The only exception appears to be C,H CH.SH, which has a higher IP than C,H SH. The reason - 34 -10.47 —i 1 8.5 9 eV ( C H 3 ) 3 C S C 2 H 5 8.5 eV C L h L S H 6 5 Figure 8 - 35 -Table 1. Ionization Potentials of the Sulphur "Lone Pair" in Some RSR' Compounds. R R' I P (eV> ARR' ( e V ) = I P ( H 2 S ) CH2—CH2 -IP (RSR*) H H 10.47 0 CH3 H 9.42 1.05 C 2H 5 H 9.26 1.21 C 3H ? H 9.1953 .1.28 n-C.Hn H 9.14a 1.33 4 9 tert-(CH 3) 3C H 9.053 1.42 C,HC H_ H 9.25 1.22 C,HC .H .8.28 2.19 o J CH3 CH3 .8.65 1.82 CH3 C 2H 5 8.55 1.92 C 2H 5 C 2H 5 8.43a 2.04 CH2=CH-CH2 CH=CH-CH2 8.54 1.93 (CH 3) 3C C 2H 5 8.29 2.18 C 3H ? C 3H ? 8.33 2.17 9.05 1.42 CH -CH -CH-3 \J  2 8.84 1.63 See reference 66. - 36 -seems to be that the C(H_)-S-H plane in C,R,_CH~SR is perpendicular to L 6 5 2 the ring as shown by CNDO/2 calculations [see 3.3.3]; consequently there is no interaction between the sulphur non-bonding orbital and the benzene ring (as predicted by CNDO/2 calculations). In the case of thiophenol the calculations show that the f i r s t IP arises from a molecular orbital which is formed mainly from the sulphur 3p atomic IT orbital and IT MO of benzene. The decrease in IP of the sulphur "lone pair" for different substituent alkyl groups can be attributed to an ' inductive effect, and to a lesser extent hyperconjugation. The difference in A^- i between the n-butyl and tert-butyl mercaptans (a, 0.09 K K eV) arises because of the inductive effect, the substituent methyl groups being closer to the S atom in the latter species. The decrease in IP observed for mercaptans as R becomes larger i s progressive when both R and Rf are alkyl groups, but not cumulative, i.e., i f R=R' we find that A„„ < 2kr.„ (see Table I ) , to the f i r s t approximation we K K K t i can say that k i s a direct measurement of the inductive effect K K of alkyl groups. It would be interesting to correlate these values with the chemical shift studied by ESCA on the core electron binding energies. 3.3 Individual Molecules 3.3.1 Hydrogen and Deuterium Sulphide 2 The B 1 band The f i r s t band of H^ S consists of two peaks of relative intensities 96/4 (Figure 9) . The strongest one corresponds to an adiabatic IP of '10.47 eV (Table II), and the separation between the two peaks i s - 37 -Table 2- Ionization Potentials for H_S and D„S and Vibrational Frequencies of Ions Produced Ionic State Adiabatic 0 IP (eV) V Vertical IP (eV) b V l -1 (cm ) b V 2 - i (cm ) Adiabatic IP (eV) D2S Ver t i c a l 0 b (cm ) b (cm ) 10.47 10.47 2540 1250 10.47 10.47 1830 950 12.75 13.44 1120 12.75 13.44 923 14.60 15.45 2450 1450 14.65 15.53 1800 1060 22 neutral 2614.6 1182.7 1891.6 934 A l l values + 0.02 eV b *~X 2 A l l values + 50 cm except and v 2 for + 100 c by extrapolation d Ref. 67 for H2S, Ref. 68 for D2S. + 0.5 eV, from Ref. 28 2540 cm ^. This corresponds to v-^ > the symmetric stretching mode. -1 This frequency i s % 1830 cm in D^S, and these values are in good 11 15 ' agreement with previous results. ' The ionized orbital can be assigned as an almost completely non-bonding sulphur "lone pair". + 2 A symmetry on the (0,0,0) (0,0,0) peak of H2S ( led Natalis and Delwich to suspect the presence of a small peak at an interval of ^ 1000 cm ^ which would correspond to v2> t n e bending mode, which is 1182.7 cm ^ in the neutral. We observe a partially resolved peak at 10.62 eV (interval 1240 cm "*"). This spectrum was obtained using the o argon "1048/1067 A discharge (Figure 10). The 1k1 band 6 9 As expected for the removal of a bonding electron, the band i s f a i r l y broad, extending over~1.7 eV. The vibrational structure below 13.32 eV consists of six f a i r l y well defined peaks (Figure 11). These peaks seem to be asymmetric, and the vibrational spacing decreases constantly from 1120 to 800 cm"1 for R^ S and 940 to 600 cm"1 for D2S (Figure 13). The vibration i s v2> the bending mode. However, in H2S, after a complex section including two broad peaks between 13.32 eV and the sharp peak at 13.54 eV, there begins a new series of about twelve peaks with spacing between 490 and 600 cm \ (A close examination of Turner et al.'s spectrum"^ indicates some similar peaks though not as well resolved.). At f i r s t this apparent continuation of structure after a dissociation limit was disturbing as i t was believed that a 4 predissociation to the A 2 state was involved because of the complex nature of the peaks just at 13.4 eV the value for the appearance potential - 40 -Figure 10 - 41 -- 42 -The most probable explanation of this structure comes from considering the vibrational-rotational levels involved, especially as Ihe bending vibration i s excited and there w i l l be angle change on ionization. An elegant study of the levels produced on straightening a bent triatomic molecule has been performed by Dixon,^ using a harmonic potential function perturbed by an exponential barrier 2 a exp (-gq ), and thereby the irregular vibrational structure of the electronic spectrum of PH^ has been explained. It was shown that, for suitable values of a and 3 , the vibrational spacing between levels with the rotational angular momentum quantum number (K) equal to zero have a minimum near the potential energy maximum corresponding to the linear configuration. For any particular value of a and 3 and for different- K'-s, the minimum becomes less pronounced for higher K and 2 + moves slightly to lower energy. As in the A^ band of H^ S we are 2 dealing with the upper component of a IT electronic state (A = + 1), where the Renner-Teller interaction has s p l i t the degenerate state to give two minima away from the linear configuration, the situation i s complex. In this case, the E levels (K = 0) arise from K = £ = 1 curve obtained by Dixon (Ref. 70, Figure 3), and K = 1 levels from a combination of £ = 0 and £ = 2 and so on. With the selection rule AK = + 1 for a transition of the type A-^ -A^  in H^ S below the potential barrier to linearity, the PE spectrum should not be too complex as the K rotational levels are almost coincident, (Ref. 70, Figure 4). However, just above the barrier the pattern w i l l change, and Dixon has shown that the odd-numbered levels how l i e together almost exactly half-way between the even-numbered levels. The pattern of the PE spectrum must 1 2 0 0 E o ^ 1000 i 8 0 0 12.6 13.2 . 13.6 14.0 B INDING E N E R G Y Figure 12 14.4 eV - 44 -now correspond to ionization to alternately even and odd-numbered K 2 rotational levels of the state. This must produce a resultant spacing in PE peaks of about half the value obtained below the barrier. 2 Referring now to the A^ band in the PE spectrum of H^ S (Figure 11, and Table 3) we see that below the barrier, between 12.75 and 13.37 eV, the K levels are very closely spaced so that ionization to the ivbrational levels = 0-5, produces single peaks. Their symmetry could possibly be associated with unresolved K rotational structure degraded to higher energies, indicating an increase in bond angle. Above about 13.37 eV, however, the spacing i s suddenly halved and this must be due to ionization to the odd- and even-numbered K levels separately. The E levels, i.e. those not affected by Renner-Teller interaction, may be easily identified by extension of the simple progression above the barrier. The remaining strong peaks are assigned as transitions to the odd-numbered K levels; principally II levels (Table 3). Of course transitions to other odd and even K levels w i l l also be involved but these should overlap the transitions to 2 and n levels at energies much in excess of the barrier. However, between 13.32 and 13.50 eV there i s no doubt that the broadening of the peaks can be partly explained by transitions of these levels. Figure 12 shows the variation of the vibrational spacings for the 2 A^ band of with binding energy. Clearly the curves are in excellent agreement with Dixon's theory because the intervals in the Z * 7 Recently Asbrink and Rabalais were able to resolve the rotational structure on the zA n band of water. - 45 -+ 2 Table 3. Ionization Data for the H2S ( Photoelectr on Band 2 Levels IT Levels eV (cm ) eV (cm 1) 0 12.752 1088 1 12.887 955 2 13.005 934 3 13.121 852 4 13.226 804 13.252 5 13.326 890 13.372 968 6 13.436 911 13.487 928 7 13.549 999 13.611 1000 8 13.673 1040 13.741 1048 9 13.802 1070 13.881 1129 10 13.935 1105 14.019 1113 11 14.072 1120 14.158 1121 12 14.212 1130 14.295 1105 13 14.352 14.435 1129 CD ID jq ro - 47 -- 48 -progressions have a minimum exactly where the transitions to the odd-numbered K levels are f i r s t seen, i.e. at the top of the barrier to linearity. If we take 13.35 eV to correspond to the maximum in the + 2 potential curve for H^ S ( A^), then the height of the barrier i s 4800 + 200 cm This is somewhat less than for the f i r s t excited state of PR^. Obviously, the shapes of the Renner-Teller component states must differ a l i t t l e between PH^ and E^S+ because the vibrational spacing below the barrier for H^S* i s larger than for PIL^  and the percentage decrease in vibrational spacing on reaching the potential maximum is larger in H^ s"*" than in PH^. For the same pattern is observed below the barrier but although we do observe bare signs of structure after i t i s not well enough resolved to be assigned with confidence (Figure 13). If the situation were exactly analogous to H^ S we should expect a separation of = 300 cm \ which we should just be able to resolve. However, i f the separation + + of odd and even levels i s not quite as distinct in D2S as in , the structure may not be observed. The loss of structure in D2S could also be partly due to predissociation. The curve of vibrational spacing vs. binding enerby is shown in Figure 14, Table 4. 2 The B 2 band The vibrational structure observed on the third band of H^S* seems 15 to be similar to that previously obtained. The peaks are very asymmetric with the possibility of complex series of progressions. The problem is further complicated by the overlap of the t a i l end of - 49 -Table 5. 2 The B 2 Band of D2S and H2S - The Most Prominant Progression. D0S 2 2 Energy eV (cm 1) eV (cm ) Level (000) 14.65 1800* Unobserved (100) 14.88 1770* 14.90 2300* (200) 15.11 1745+ 15.18 4* 2170 3 15.32 *f* 15.45 2016+ 1677T 4 15.53 1599+ 15.78 1936+ 5 15.74 15.95 1855f 1464T 6 15.92 1387+ 16.17 1770* 7 16.09 1290+ 16.39 1690* 8 16.25 1250* 16.66 1600* 9 16.40 1210* 16.80 1530* 10 16.55 1180* 16.99 11 16.70 1140* 12 16.84 1100* 13 16.98 The error in these measurements is + 50 cm * _ l The error in these these measurements is + 150 cm - 50 -Table 4. Ionization Data for the D0S (A..) Photoelectron Band. IP (eV) (cm 1) 0 12.755 943 1 12.872 790 2 12.970 742 3 13.052 661 4 13.144 629 5 13.222 604 6 13.297 588 7 13.370 the second band and the start of the third band. However, as can be 2 seen i n Figure 15, the B^ band of D^S is less complex than that of ^ S . The principal progression, the symmetric stretching frequency [(0,0,0),(1,0,0),(2,0,0),...etc.], i s obvious despite the uncertainty on the f i r s t two peaks and the breadth of the peaks above (7,0,0) level. Table 5 shows that this series tends to converge rapidly towards the dissociation limit, since the vibrational spacing decreases markedly. The uncertainty i n the energy of the levels [except for (2,0,0)-(7,0,0)] is quite high as a l l the other maxima except the peaks respresenting these six levels are quite broad, due to the presence of another vibrational progression. This i s clearly seen on the low energy - 51 -- 52 -side of the (5,0,0) peak, where another maxima i s apparent. The separation between the (4,0,0) level and the maximum belonging to the other series on the low energy side of the (5,0,0) peak is 1060 + 50 cm By extending the second progression on either side of this strong peak, we obviously have another series in displaced to higher energy from the f i r s t series by^1060 cm. \ This almost certainly corresponds to v 2, the bending frequency, which i s 943 cm ^ i n the ground state of the molecule. This significant increase in the bending vibration i s indicative of the antibonding interaction between the two deuterium atoms in the D2S orbital involved, and this further confirms that ionization has taken place from the b 2 orbital. The -1 + decrease in the stretching frequency from D2S (1891.6 cm ) to D2S , 2 -1 B 2 (1800 cm ) and the convergence of the vibrational progression shows the S-D bonding nature of the b 2 orbital. + For H^ S the maxima for the principal progression i n the symmetric stretching frequency are not so easy to pick out as the peaks are much broader and more asymmetric. The (0,0,0) peak is not observed because of the overlap with the second band, but the rest of the levels show the same rapid convergence as in D 2S +. The low energy freatures are more distinct as would be expected since there is a bigger petcentage difference between the stretching and bending frequency for H2S than for D2S (Table 2). Also i n the case of H2S there are features on the high energy side which are as distinct, at least on the f i r s t two peaks. The series on the low energy we again assign as the (1,0,1),(1,1,0),(2,1,0)...progression and in fact the shoulder at - 53 -Figure 16 - 54 -14.76 eV, the (1,0,1) level, i s the f i r s t distinguishable feature in this band. The shoulder (2,1,0) on the third distinct peak (3,0,0) i s the best observed. The separation between this and the (2,0,0) level is = 1450 + 100 cm \ and indicates that the value of for the + 2 -1 I^S ion is greater than that of H^ S (1182.7 cm ). The series on the high, energy of the main peaks seems to best correspond to a progression on the symmetric stretch displaced by two quanta of V2» i.e., starting at the high energy side of (1,0,0) at 14.96 eV we have (0,2,0),(1,2,0),(2,2,0), etc. It i s also possible that a f u l l series on the bend i s present as i n addition to (0,1,0) and (0,2,0),(0,3,0) could overlap with (1,1,0),(0,4,0) and (1,2,0) etc. One alternative that we believe i s highly improbable i s that one of the series i s a progression in v^, the antisymmetric stretch instead of x>2- The reduction in frequency from the neutral species to the ion is considered to be too great. Table 5 l i s t s the peaks in the principal ' progression (1,0,0),(2,0,0) etc. and Figure 17 indicates our assignment of the three series. In both H^S* and D2S+ i t i s d i f f i c u l t to pick out the subsidiary progressions after the f i r s t few members because of the large anharmonicity on the stretching vibration causing greater overlap of the progressions towards the higher members. Also in both species but particularly noticeable in D 2 S there i s a loss of intensity and broadening at the =16.2 eV peak. Obviously some change has taken place to cause the vibrational states to be less stable. + 2 Turner explains the broadening of a l l the peaks in H^ S as being due to the ins t a b i l i t y of the levels with respect to predissociation + 4 to SH via the crossing A" state (see Ref.71afor potential energy curves). - 55 -The above discussion, especially of D2S^, indicates that the presence of more than one progression causes a considerable amount of the 2 + broadening. A dissociation process of the state of H^ S into + 71a H 2 and S as can be seen from Fiquet-Fayard and Guyonlo potential energy curves at 16.2 eV is not excluded. 3.3.2 Methyl Mercaptan (CH^SK) The photoelectron spectrum of methyl mercaptan shows the presence of four bands below 17 eV (Figure 17). The f i r s t IP at 9.42 eV (Figure 8) consists of a strong peak accompanied by two vibrational peaks with frequencies of 680 + 40 and 1250 + 80 cm respectively. The f i r s t obviously corresponds to the C-S stretching vibration, which i s expected to be excited i f there is any interaction between the sulphur p orbital and a carbon p orbital. As the C-S stretching frequency i n the ground state of the molecule i s 704 cm 6 7 i t is d i f f i c u l t to say whether the highest orbital i s bonding or antibonding with respect to the carbon-sulphur bond, but whichever i t is the interaction i s f a i r l y small. CNDO/2 calculations predict that this orbital i s mainly a sulphur non-bonding orbital with samll participation of carbon 2p orbitals in the form of a C-S TT MO (Table 6). The 1250 cm ^ vibration almost certainly corresponds to the symmetrical deformation of the CH^ group (reduced from 1335 cm in the neutral species). As can be seen, the sulphur lone pair i s destabilized compared to that of I^S by interaction with the methyl group. - 57 -Table 6. Vertical Ionization Potentials of Methylmercaptan (eV) Experimental CNDO/2 Electronic Orbital Type density 9.42 8.4 0.93 S(3P) n(S3P) a " 12.0 10.8 0.13 S(35) a(S-C) a' 0.48 S(3P) 0.16 C(2P) 0.15 H(1S) 13.9 12.5 0.40 S(3P) a(S-H) a' 0.29 C(2P) 0.27 H(1S) 15.0 17.1 0.48 C(2P) cr(CHj a" 0.44 H(1S) 15.5 17.3 0.45 C(2P) a(CH ) a 1 0.40 H(1S) =20.0 21.4 0.64 S(3S) n(S,3S) a' 0.10 C(2P) 0.20 H(1S) - 58 -The second and third IP's are at 12.01 and 13.9 eV, neither band having any resolvable vibrational structure (the lack of structure in the second band i s probably due to a dissociation to CH-jS+ at 11.2 + 73 0.5 eV ). The second IP i s assigned to the C-S bonding orbital by analogy to dimethyl sulphide where the two S-C bonding orbitals are at 11.2 and 12.6 eV, and ethylene sulphide at 11.32 and 11.72 eV respectively. The third IP is assigned to the S-H bonding orbital. The CNDO^ calculations show an almost equal amount of sulphur and hydrogen atomic orbitals in this MO. As i n the alkyl halides the higher energy IP's below 21 eV can be assigned as from mainly CH^ bonding orbitals. Around 15 eV in the alkyl halides there i s a band with a Jahn-Teller contour assigned as 9 49 ionization from the e(CH.j) bonding orbital. ' In methanthiol there i s a broad band of intensity twice that of the 12.0 and 13.9 eV bands with two maxima at 15.0 and 15.5 eV which i s very similar to a band 9 49 in the PE spectra of the methyl halides. ' It could almost be considered that the CH^ group i s only observing the SH group as a single atom, and in fact CNDO/2 calculations indicate this as the two low energy bonding CH^ IP's are predicted to be almost degenerate at 17.11 and 17.27 eV. In ethanthiol, the f i r s t IP at 9.26 eV shows a strong peak similar to that in methanthiol accompanied by a resolved vibrational peak at 12.76 cm ^  (Figure 8). This corresponds to the CH^ symmetrical -1 74 deformation (1385 cm in the neutral ). The MO from which the electron i s removed can be assigned as a mainly sulphur non-bonding 3p orbital. - 59 -3.3.3 g-Toluenthiol and Benzenthiol As seen in the case of methyl mercaptan (also cf. dimethyl sulphide 3.3.4) the TT interaction of a sulphur lone pair with adjacent methyl groups i s relatively small. However, we shall now see that the opposite is the case when sulphur i s attached to a phenyl group. This can best be seen by comparing the PE spectra of benzenthiol (C,HCSH) and toluenthiol (C,HCCH„SH). In benzene the lowest IP, which o 5 o 5 2. i s due to ionization from the ir e, orbital i s at 9.25 eV [see Ref. 15, CH. 10] and in monosubstituted benzene ( C 2 v o r i 0 w e r symmetry) this i s s p l i t into two orbitals, one having a n interaction with the substituent (b^ in C^) and one almost non-interacting (a^ i n . The separation between these two levels i s greatest with substituents halogen .lone .pairs and least for a substituent with no lone pairs [see Ref. 15, CH. 11]. Toluene is of this latter type and the PE band exhibits only a slight broadening and a shift to lower energy compared to benzene [see Ref. 15, CH. 11]. Replacement of one of the hydrogen atoms of the CH^ group by a SH group produces no remarkable changes in the PE spectrum except for the superposition of a new sharp peak (at 9.25 eV), (see Figure 8) with at least one other vibrational component (spacing =1220 cm "*"). The f i r s t v e r t i c a l IP i s 8.85 eV (cf. 8.81 eV in toluene). It can be seen in this case that the sulphur lone pair has an almost negligible effect on the orbitals of the benzene ring. The inverse is also true, as besides the sharp band PhC^SH has i t s sulphur lone pair IP at 9.25 eV compared with that of CH3CH2SH at 9.26 eV. The vibrational spacing of 1220 + 50 cm"1 probably corresponds to CH2 wagging (at = 1400 cm 1 in the neutral) as - 60 -there w i l l be some. interaction between the sulphur lone pair and the CH^ orbitals. In contrast to the case of a-toluenthiol, the replacement of one H atom in benzene by a SH group has an enormous effect. Below 11 eV instead of the one IP, there are three at vertical IP's; 8.28, 9.38, 10.65 eV respectively (see Figure 18). The 9.38 eV band obviously corresponds to the non-interacting benzene ring orbital. The separation between the other two IP's (2.37 eV) shows the extent of interaction between the sulphur lone pair and the Tr(b^) orbital of the benzene ring. Iodobenzene also shows a large interaction but the separation between the two b^ IP's i s only 1.78 eV and the lowest IP is 8.67 eV [see Ref. 15, CH. 11]. The 8.28 eV band in the PE spectrum of C ^ S H is slightly narrower than the 10.-65 eV band and because the 0 -> o transition i s the vertical one, the corresponding orbital i s believed to have more sulphur lone pair character. The vibrational structure on the band consists of a progression of eight peaks each separated by =s 400 cm \ which probably corresponds to the substituent-sensitive -1 75 ring breathing mode at 412 cm in the neutral thiophenol. CNDO/2 calculations carried out on a-toluenthiol assuming the CO^-S-H plane to be perpendicular to the ring predict the second IP to be associated with the removal of an electron from the sulphur lone pair orbital. This lone pair i s almost completely non-interacting as regards the benzene ring. Slight deivations i n this configuration are predicted to cause l i t t l e change i n the IP. However, i f the CO^-S-H plane coincides with the plane of the benzene ring and becomes the origin of the lowest IP comparison of the calculations and PE. - 61 -- 62 -spectra indicates that the former configuration i s preferable. In the case of thiophenol the calculations show that the f i r s t IP arises from an o r b i t a l which i s formed mainly from the sulphur lone pair with a considerable amount of benzene ring character, which i s i n agreement with the interpretation of the PE spectrum suggested above. 3.3.4 Dimethyl Sulphide The PE spectrum of dimethyl sulphide (Figure 19) shows the presence of four main bands below the energy of 21 eV. The f i r s t IP at 8.65 eV consists of a r e l a t i v e l y strong peak accompanied by two v i b r a t i o n a l peaks with frequency of 750 + 60 cm ^  (Figure 8). CNDO/2 calculations predict s l i g h t antibonding character between the sulphur and the adjacent carbon atoms for the highest occupied MO, but the MO can be considered mainly as a sulphur lone pair o r b i t a l . The v i b r a t i o n a l spacing of 750 cm ^  i s compared to C-S -1 stretching mode which i s at 685 cm i n the ground state of the molecule [see Ref. 67, p. 652]. I t seems that the large s h i f t of the f i r s t IP to lower energy compared to U^S i s due to a large inductive effect as discussed previously. The second and t h i r d IP's at 11.2 (a^) and 12.6 0^) eV are assigned to S-C a bonding, and do not show any vi b r a t i o n a l structure. These are followed by a band which extends over a range of 1.5 eV and shows the presence of four components, a l l assigned to (CH^) bonding. The CNDO/2 calculations indicate a, much larger than observed range for these CH^ o r b i t a l s and probably over-estimate the interaction between o r b i t a l s on the two CH^ groups. The experimental and calculated IP!s are given i n Table 7. - 63 -- 64 -Table 7. Vertical Ionization Potentials of Dimethyl Sulphide (eV). , a Experimental Experimental CNDO/2 Electronic Density « t • i b Orbital type 8.68 8.65 7.7 0.88 S(3P) n(S3P) b l 10.96 11.2 10.4 0.13 0.46 0.28 S(3S) S(3P) C(2P) o(S-C) a l 12.16 12.6 11.5 0.32 0.48 0.18 S(3P) C(2P) H(1S) a(S-C) b2 13.68 14.2 15.1 0.10 0.50 0.38 S(3P) C(2P) HQS) a(CH3) b2 14.8 16.1 0.50 0.48 S(3P) H(1S) a(CH3) a2 15.4 17.5 0.11 0.48 0.40 S(3P) C(2P) H(1S) 3' h '1 15.7 18.4 0.52 0.42 C(2P) HQS) 0(CH3) a l 19.7 20.4 0.61 0.23 0.12 S(3S) C(2P) HQS) ^n(S,3S) a± See Reference 72. Assuming C 0 symmetry. - 65 -3.3.5 Ethylene Sulphide The PE spectrum of ethylene sulphide consists of six bands below 18 eV (Figure 20). The f i r s t band (adiabatic IP 9.05 eV) shows four vibrational peaks with mean spacing of 1090 cm \ This band is similar to the f i r s t band of ethylene oxide 7 7 where the vibronic structure of this band parallels reasonably well that observed on the optical Rydberg absorptions. 7 8> 7 9 The combination of optical spectroscopy and PES indicate that the three optical bands observed for ethylene oxide terminate at 58380, 63610, and 96680 cm and corresponded to 3s, 3p and 3d Rydberg orbitals, l i k e l y centered on the oxygen atom. CNDO/2 calculations on ethylene sulphide indicate that the highest occupied b^(n) orbital i s 90% localized on the sulphur atom (Table 8) suggested that a Rydberg series for this molecule could be observed in the UV electronic spectrum. The vacuum UV spectrum of ethylene 80 sulphide has been obtained by Basco and Morse, and the expected Rydberg series are observed. The optical spectrum consists of at least three series, the f i r s t , with an origin of 47310 cm ^ is rather diffuse, but four vibrational peaks can be observed with a spacing of 1100 cm The second and third series are sharp with origins of 51880 and 57510 cm ^ and vibrational spacings.of 1130 and 1100 cm respectively. The observed origins of these Rydberg series and the measured f i r s t IP of ethylene sulphide (72995 cm ^) can be used to estimate n* (= n-6, where n i s a principal quantum number and <5 the so-called quantum defect) 81 from the relationship Ethylene Oxide Ethylene Sulphide CNDO/2 INDO CNDO/2 obs. IP's Calc. Electronic Symm. Symm. Calc. Electronic Obs. IP's Calc. Electronic Symm. Density Density Density 10.57 11.26 0.78 0.06 0.16 0(2p) C(2p) H(ls) b l b l 10.18 0.80 0.04 0.16 0(2p) C(2p) H(ls) 9.05 7.69 0.91 0.01 0.01 S(3p) C(2p) H(ls) . b l 11.7 12 0.04 0.20 0.72 0(2s) 0(2p) C(2p) a l a l 10.95 0.07 0.27 0.63 0(2s) 0(2p) C(2p) 11.32 9.69 0.51 0.05 0.36 0.08 S(3p) S(3dyz) C(2p) H(ls) b2 13.7 13.25 0.39 0.60 C(2p) H(ls) a2 b2 13 0.62 0.14 0(2p) C(2p) 11.72 11.12 0.27 0.55 S(3p) C(2p) a l ~14.2 13.46 0.62 0.28 0(2p) C(2p) b2 a2 13.19 0.37 0.63 C(2p) H(ls) 13.59 13.2 0.02 0.39 0.59 S(3dyy) C(2p) H(ls) a2 16.6 17.82 0.08 0.29 0.45 0.18 0(2s) 0(2p) C(2p) H(ls) a l a l 16.7 0.11 0.24 0.47 0.18 0(2s) 0(2p) C(2p) H(ls) 15.33 16.06 0.18 0.55 0.15 S(3s) C(2p) H(ls) 3 1 . 17.4 23 0.20 0.50 0.30 0(2p) C(2p) H(ls) b l b l 22.7 0.17 0.50 0.32 C(2p) C(2p) H(ls) 16.69 20.98 0.45 0.34 0.16 S(3s) C(2p) H(ls) a l See Reference 77. - 68 -v = IP - — ^ - 9 -(n*) Z where v i s the frequency of the origin (cm and R is the Rydberg constant. The values of n* obtained here are 2.07, 2.28 and 2.68 for the 47310, 51880, and 57510 cm 1 bands respectively. In comparison with ethylene oxide^ we assign the f i r s t two series to 4s and 4p Rydberg orbitals and the third to a 3d Rydberg orbital, then the values for the quantum defects are <5, = 1.93, 6, = 1.72 and 6„, = 4s 4p 3d 67 0.32, the value of 6 = 0.32,corresponds to a d type Rydberg orbital, which confirms our assignment. A pseudo-orbital energy may be defined on the basis of Koopman's 82 theorem, i.e. the pseudo orbital energy is equal to the negative of the IP. The pseudo orbital energy for the 3p orbital in a neutral sulphur atom is -10.4 eV which may be compared to the value of -9.1 eV for the b^ (3p) orbital in ethylene sulphide. The pseudo-orbital energies of the Rydberg 4s, 4p and 3d states in enutral sulphur atom are -3.9, -2.5, and -1.9 eV respectively. On the basis of the assignment of the Rydberg series in ethylene sulphide the pseudo orbital energies correspondingly are -3.2, -2.6, and -1.9 eV. This correspondence lends support to the assignment of the Rydberg series i n ethylene sulphide. The vibrational frequency observed on both the f i r s t band of the PE spectrum and the Rydberg series no doubt corresponds to the CI^ —1 83 bending mode, (1107 cm in the neutral). The f i r s t vibrational peak in this f i r s t band appears to be asymmetric with a s p l i t t i n g estimated - 69 -to be 85 cm "S this may be due to rotational structure (cf. the PE spectrum of water'') . It has been found that CNDO/2 LCAO-SCF calculations gave a reasonable description of the PE spectrum of ethylene oxide (Table 7), therefore i t was decided to employ these calculations as an aid to assigning the other bands in the PE spectrum of ethylene sulphide. 84 The geometry reported by Cunningham et a l . was used. The calculations predict a rj* character between the sulphur atom and the adjacent carbon atoms, this being responsible for the decrease i n the sulphur "non-bonding" 3p orbital IP (9.05 eV) compared to the corresponding one in hydrogen sulphide (10.47 eV); this has been explained as due to an inductive effect. The presence of the carbon-carbon bond reduces the number of IP's below 20 eV relative to dimethyl sulphide (3.3.4). The second and third IP's are assigned to (S-C) a bonding compared to those observed i n dimethyl sulphide. The other IP's are reported in Table 8. CNDO/2 and INDO calculations on ethylene oxide are in good agreement with the experimental values and are reported in Table 8. 3.4 Photoelectron Spectra of Some Disulphide Compounds 3.4.1 Introduction In previous studies, the effect of one or two alkyl groups on the energy of the sulphur 3p non-bonding orbital has been discussed. It was observed in a series of compounds containing one divalent sulphur atom that the IP of the sulphur "lone pair" orbital decreases as the number of carbon atoms in one alkyl group increases. When two groups are bonded to the sulphur atom the effect although larger than the - 70 -effect of only one i n the molecule is not cumulative. This decrease in IP i s due mainly to an inductive effect. The case of disulphide compounds, there are two sulphur 3p non-bonding orbitals involved, therefore the affect of halogen and alkyl groups on these sulphur "lone pair" orbital IP's must be considered. The other IP's w i l l be reported, especially those of the (S-S) MO in different compounds. 3.4.2 S 2C1 2 The PE spectrum of ^ C ^ (Figure 21) does not show any vibrational structure. The CNDO/2 calculations have been carried by choosing different dihedral angles 0°, 45°, 82°, 90° and 180° of S 2C1 2 planes to assist"the assignment of the experimental IP's. The molecular energy by CNDO/2 is minimized for the dihedral angle of 90°, which is very close 85 to the experimental value 82.5°. It i s observed that the parameteriza-tion of the CNDO/2 method dealing with molecules of the second row elements are not sufficiently accurate (i.e. see Ch. IV). The bond lengths and angles used i n these calculations are those given by u . _ 85 Hxrota. The f i r s t band indicates two IP's at 9.77 and 9.91 eV. These are assigned (by considering the CNDO/2 calculations) to the sulphur 3p non-bonding orbitals. For dihedral angle of 90°, these atomic orbitals are orthogonal to each other with equal electronic density. There appears to be small chlorine participation in the form of (S-Cl) a bonding for one sulphur atom and a* for the other. The slight bonding and antibonding - 71 -- i i i 1 1 1 , , , r 11 13 15 17 19 eV 21 Figure 21 Table 9. Vertical IP's in (eV) ci2s2 (CH3) 2S2 ( C2 H5 )2 S2 (CF 3) 2S2 CNDO/2 Calc. Obs. Orb. Obs. Orb. Obs. Orb. Obs. Orb. 8.68 9.77 S n(3p) 8.7 S n(3p) 8.65 s n(3p) 10.74 S (3p)n 8.73 9.91 S n(3p) 9.94 S n(3p) 8.94 S n(3p) 11.08 S n(3p) 9.88 11.08 (S-S)P a(Cl) (S-C1)P 11.25 (S-C) 10.95 (S-C) CT 13.25 C-S 11.38 11.88 CI n(3p) 12.35 (S-C) 11.67 (S-C) CT 13.81 (C-S) 11.89 12.2 CI n(3p) 13.4 (s-s) P A [11.67-17]' a(C 2H 5) 14.47 (S-S) 12.05 12.54 CI n(3p) [13.4-17] a(CH3) 13.89 13.87 (S-S) P CT [15.43-20] jc(CF 3) 14.04 15.85 (S-Cl) P IF (2 P) 14.14 16.57 (S-Cl) P characters' have different effects on the sulphur "lone pair" orbital IP's, that is why we observe two non-bonding orbitals instead of degenerate ones. The third IP at 11.08 eV has some (S-S)P and (S-Cl) TT character, a with relatively high electronic density on the chlorine atoms. This is assigned with the other three IP's at 11.88, 12.2 and 12.54 eV as each has a chlorine "lone pair" orbital. These are at about the 87 sameenergy as those of similar type observed in dichloroethylenes 9 and methylene chloride. The sharpness of the;  peaks is also indicative of the non-bonding character of these MO's. The seventh IP at 13.87 eV is assigned to a (S-S) P MO with some 0 participation of chlorines i n the form of (S-Cl) TT bonding. The other two IP's at "16.33 and 17.33 eV relate to (S-Cl) P MO's. a 3.4.3 (CH 3) 2S 2 The PE spectrum of dimethyl disulphide (Figure 21) does not show any vibrational structure. The f i r s t band seems to infer two distinct IP's at 8.7 and 8.94 eV. These can be assigned to the sulphur "lone pair" orbitals in a similar way as some bands i n the PE spectrum of 87 S 2C1 2- Yamabe et a l . predicted the sulphur lone pairs in this molecule to be orthogonal (with a slight difference in energy) by using a semi-empirical SCF method at 10.35 and 10.37 eV. It i s surprising to find that the sulphur non-bonding orbitals i n (CH 3) 2S 2 have about the same binding energy as similar orbitals in dimethyl sulphide. The explanation for this may be either the CH^ S group has the same inductive effect as the CH„ group, or the electronic 74 -density on each sulphur non-bonding orbital in (012)2^2 ^ s n a x f °f that i n (CH^^S as shown by CNDO/2 calculations. Therefore the magnitude of the interaction between the non-bonding orbital i n sulphur atom and the methyl group is much larger, which results in a large decrease of the IP's. The reason of observing two non-degenerate sulphur 3p non-bonding orbitals can be explained in the same way as in S 2 C I 2 . The two IP's at 11.25 and 12.35 eV are assigned to the (S-C) a bonding MO's. These are about the same energy as the corresponding ones in dimethyl sulphide (see Table 7) (at 11.2 and 12.6 eV) and in ethylene sulphide (at 11.32 and 11.72 eV). The (S-S) a bonding MO i s at 13.4 eV compared to that i n C1 2S 2 at 13.87 eV. The inductive effect of the chlorine atoms is responsible for the stabilization of this bond. There i s a broad band which extends over the region 13-17 eV and is thought to contain many IP's. It is d i f f i c u l t i f not impossible to recognize each IP individually. By comparison with dimethyl sulphide and other methyl containing molecules, we assign this band to a(CE^) MO's. The diffuse band 18-19.5 eV may be assigned to the S(3s) orbital. 3.4.4 (CH 3CH 2) 2S 2 The sulphur non-bonding orbitals in this molecule can be seen in the f i r s t band of the PE spectrum (Figure 21) at 8.65 and 8.94 eV. The decrease in IP with respect to the corresponding orbitals i n dimethyl disulphide i s due to the larger inductive effect of CH^CI^ groups. This effect i s also reflected on the (S-C)a MO's at 10.95 and 11.67 eV. It i s very difficult to recognize the IP of the (S-S) a MO due to i t s - 75 -overlap with the IP's from CH^CH^ groups. These can be seen i n two broad bands between 12 and 17 eV. The S(3s) orbital IP i s ^ 20 eV. 3.4.5 (CF 3 ) 2 S 2 The inductive effect of fluorine atoms in CF^ groups i s responsible for the stabilization of the sulphur 3p non-bonding orbitals at 10.74 and 11.08 eV. These can be recognized in the f i r s t band of the PE spectrum of (CF.j) 2S 2 (Figure 21). This i s also reflected in the (S-C) o MO's at 13.25 and 13.81 eV, meanwhile the third IP in the second band at 14.47 eV is assigned to the (S-S) a MO. The increase in energy of this MO compared to the corresponding one in (CR 3) 2S 2 and C1 2S 2 is attributed to the inductive effect. It i s very d i f f i c u l t to assign the IP's in the region between 15 and 18 eV where many MO's are involved, especially those of the fluorine "lone pair" orbitals. It i s interesting to observe that the band maximum at 17.33 eV has almost the same energy as the fluorine 9 lone pair orbitals at 17.25 eV in CF^H, and 17.2 eV in CH3CF0. T h e r e f o r e a s s i g n this band to fluorine "lone pair" orbitals. Obviously this band contains evidence for more than one IP, consequently more than one fluorine "lone pair" orbital i s involved. A l l the observed and calculated IP's of these molecules are reported in Table 9. - 76 -CHAPTER IV PHOTOELECTRON SPECTRA OF SOME THIONYL AND SULPHURYL HALIDES 4.1 Introduction There has been considerable interest 82,88-91 in the bonding in inorganic sulphur compounds, particularly where the sulphur atom has a formal valence state higher than two, with a tendency to form multiple bonds with other atoms. Some of the physical properties of frequencies, ' have been correlated with changes in the TT bond orders. As there are a large number available, most of these studies have been carried out on compounds with sulphur-oxygen bonds. Photoelectron spectroscopy can yield information on multiple \ bonding involving second row atoms (e.g. the halosilanes (see Chapter 94 95 VI)) but besides that of S0 2 and SO no photoelectron spectra have been published of sulphur-oxygen species. In these species considerable changes in molecular structure (bond lengths and angles) and properties (particularly S-0 stretching vibrations) are effected by changing the other atom(s) bonded to the sulphur, and i t i s of interest to see i f the molecular energy levels obtained from the photoelectron spectra can predict or be at least consistent with these effects. Sulphuryl halides (S0 2X 2) and thionyl halides (S0X 2), in particular are typical examples which show these effects to a large degree. In SO F„, which the type of compound particularly bond lengths 88-89 and vibrational 92,93 2-is isoelectronic with SO^ , we should expect that i f the TT bonding between sulphur and a l l other four atoms were equal the molecule would be tetrahedral. However, although the S-F bond length is ° 96 ° 97 98 reduced only slightly from 1.56 A in SF5OF to 1.53 A in > 0 the S-0 bond length is reduced considerably from 1.64 A in SF..OF o to 1.405 A in ^^2^2 * Obviously there must be a consierable increase 99 in the bonding interaction (presumably mainly n bonding ) in the S-0 bond compared to the S-F bond in the latter. This is also evident i n thebond angles. The greater TT electron density in the S-0 bonds causes greater mutual repulsion than with the S-F bonds, and hence 0S0 = 124° and FSF = 96°. A satisfactory bond order-bond length relationship i s obtained for S-0 bonds in tetra co-ordinate sulphur compounds by considering that TT bonding involves 2p orbitals on the oxygen atoms 88 and only the two d orbitals on the sulphur with the best overlap. The shorter SO bond length and wider 0S0 angle for SO2F2 compared to 0 1 on 0 i ni S0 2C1 2 (1.43 A and 120° respectively ) and S0 2 (1.432 A and 119° ) is caused by the greater contraction of the sulphur d orbitals and hence greater TT overlap with the oxygen produced by the more electronegative fluorine atoms. However, i t also i s possible in this case that the chlorine atoms are more involved in TT bonding than the fluorine atoms and this seem to be borne out by the large Cl-S-Cl angle of 111°. The PE spectra should give some indication of which picture is more accurate here. 94 102 Both the PE spectrum of and a recent good SCF calculation on S0 2 indicate that there is a large stabilisation of the oxygen pu orbitals involved in extensive TT bonding (a^,b^) compared to the others (a^tby)' Hopefully these levels in 50^2 would give evidence for the even stronger bonding in this molecule. To aid the assignment of the photoelectron spectrum of SC^]?^, the spectra of SC^C^, SC^ and SOC^ were also recorded. The assignment of the mainly halogen lone pair IP's can be aided by comparison with 103 the photoelectron spectra of a number of the dihalo compounds, including the thionyl halides. The bonding in the thionyl halides i s complicated by the sulphur lone pair, but desite this the X-S-0 and X-S-X angles are f a i r l y close to those for the analogous sulphuryl compounds - in F-SO, F-S-0 = 107 F-S-F = 93° and in C12S0, Cl-S-0 = 106° and Cl-S-Cl = 114°. A l l 1he bond lengths are just significantly longer - in ^ 2^^ r(^-F) = o o o 1.585 A r(S-0) = 1.412 A and in CI SO r(S-Cl) = 2.07^ r(S-O) = 1.45 A. As with the sulfuryl compounds, the trend in the thionyl compounds have been explained on the basis of a simple LCA0 M0 treatment."^ 1 The lowering of the molecular symmetry in the thionyl compounds cause complications in interpretation of the PE spectra, but again some of the trends in the bonding should be able to be rationalized. Experimental The S 0 2 F 2 sample was purchased from the Mtheson Co. Ltd., SOF2 from the Pierce Chemical Company, and the SO2CI2 and SOC^ from K&K Laboratories. The purity of the samples was demonstrated by infrared and mass spectrometric analysis. It was very d i f f i c u l t to obtain the spectra of SO2CI2 and SOC^ because these compounds seem to decompose as they pass through the metallic sample inlet tube in the spectrometer to give S0„. To avoid this problem we used a glass inlet system, - 79 -however despite this precaution there s t i l l appeared to be some decomposition inside the c o l l i s i o n chamber togive SC^ after a few minutes 4.2 Individual Molecules 4.2.1 Sulphur Dioxide To aid the assignment particularly of the sulphuryl species the PE spectrum of SC^ (Fig. 22) is f i r s t discussed. The three bands in the PE spectrum of SC^ have been assigned as follows: (a) 12.5 eV band - from 6a^ orbital sulphur lone pair. (b) 13-14 eV band - two IP's from 4b^ and la^ orbitals, in plane and out of plane non-bonding oxygen pu combinations ve r t i c a l IP's. (c) 16-17.5 eV band - at least two IP's from 5a^ and l b ^ orbitals S-0 u bonding combinations, which can be considered as the bonding combinations of oxygen lone pair orbitals. A recent good MO calculation on SO^ confirms this general 102 pattern and the separation between the three groups, although i t also puts the next highest b^ orbital i n the third group, which was 94 in fact suggested as a possibility in the original work. For comparison with the sulfuryl compounds one is particularly interested in the order of the components of the second and third bands. The calculation estimates the separation of the two components of the second band to be 0.6 eV (experimental 0.3 eV) apart with the orbital at highest orbital energy. As in most species of C„ symmetry - 81 -with two atoms with lone pairs the b^ orbital i s at higher energy 103 than the a^ orbital and as the interaction with sulphur d 99 orbitals would cause greater stabilization of the a^ orbital, this order is surprising. Since no orbital populations were given, i t i s impossible to give a f u l l explanation. However, a close examination of the PE band between 13 and 14 eV seems to indicate agreement with the order obtained in the calculations. The PE spectrum and the analysis of the band is almost identical to 94 that of Eland and Danby. It is noted that the f i r s t component with adiabatic IP 13.01 + 0.05 eV is narrower than the second and has a v e r t i c a l IP of 13.24 eV. The second component has a ve r t i c a l IP of 13.47 eV but the adiabatic i s not obvious. Eland and Danby prefer to place this at about the same energy as the ve r t i c a l IP of the f i r s t component (^13.25 eV), but i t is thought that i t may be at least one and possibly up to three vibrational intervals (v^ ) below this. Vibrational analysis of the f i r s t component indicate excitation of both the stretching (1100 cm 1) and bending vibrations (^400 cm "*") . The second component shows only the bending vibration ( — 480 cm ^) . Taking into consideration the oxygen-oxygen antibonding nature of both orbitals involved, i t would be very d i f f i c u l t to distinguish between them, except in the case of d orbital participation by the sulphur atom which is much larger in a^ than b^, therefore the a^ orbital w i l l be stabilized and w i l l be considered as the second component. This PE band is expected to be broad. The third band in the spectrum i s also complex, with irregularities in the vibrational spacings and peak shapes. However, two progressions -1 -1 94 of 800 cm and 950 cm can be distinguished. This appears to be - 82 -due to at least two states of the ion with adiabatic IP's at 15.99 and 16.33 eV respectively. We agree with the assignment given by Eland 94 and Danby that the ionized states involve b^ and a^ type orbitals on thebasis that the possibility of interaction between sulphur 3d 2 z- 82 and 3d 2_ 2 ( a q) symmetry is much larger than the other 3d orbitals. . x y x This will stabilize the a^ orbital. Also comparison with the S f X ^ species indicates that the a^ orbital i s li k e l y to produce a sharper band and consequently the IP of 16.33 eV is assigned a^ and the 15.99 eV IP is assigned b^. 4.2.2 Sulphuryl Fluoride (SO,,F2) By comparison with SC^ and difluoro compounds e.g. (C^, ^2^2 and SiH^F2 [see. Ch. V]) one should expect at least eight IP's i n the regionbelow 21 eV, corresponding to two "lone pair", four fluorine "lone pair" orbitals, and two (S-0) TT MO's. These "lone pair" orbitals involve in most cases differing electronic participation from other atoms in the molecule, therefore, we w i l l not observe a very sharp band as we saw previously in the case of divalent sulphur containing compounds. Considering the symmetry elements for the SO2 moiety to be analogous to those in the SO2 molecule, we expect two low IP's from mainly non-bonding orbitals (a2,b2) around 13-14 eV as i n S0 2 (since there w i l l be l i t t l e interaction i n these orbitals with the rest of the system) and two higher IP's from sulphur-oxygen IT bonding orbitals (a^,b^). Similarly by comparison with F2O , OU^F^ a n d SiH 2F 2 the two non-bonding fluorine "lone pair" IP's (a 2,b 2) are expected together at about 16 eV and there w i l l also be two higher - 83 -energy fluorine "lone pair" IP's (a^,b^) which w i l l also have a considerable amount of a bonding character. Turning now to the PE spectrum of $62^2 (figure 23) i t is noted that there are six distinct bands at (a) 13-14 eV (b) 15-16 eV (c) 16.7-17 eV (d) 18-18.5 eV (e) 19.7-20.1 eV (f) 20.80 eV. The f i r s t two bands have an integrated intensity about twice that of each of the next two and the band contours and vibrational structure indicate the presence of two components in each. Bands (a) and (b) are then assigned as the non-bonding oxygen and fluorine lone pairs respectively in agreement with the above argument. We now discuss each of the bands separately. The vertical IP's and observed vibrational spacings are l i s t e d i n Table 10. Table 10. Ionisation Data for S O ^ . Bandc Vertical IP (in eV) Vibrational Frequencies 0 (in cm-i) Ionic c State (a) 13.55 - - (430) -13.61 - 510 340 (b) 15.13 - - - 330 15.35 (1050) 1050 - -(c) 16.68 1130 840 520 -(d) 18.36 - - - -(e) 19.82 - 800 506 -19.88 (f) ~21 1269 848 544 384 See text; b + 40 cm 1 at c best; Ref. 98(b). B2 neutral molecule 0 - 84 -- 85 -Band (a) (Figure 24) This band has two components. The f i r s t is broad (vertical IP * 13.55 eV) with a barely visible progression 430 cm , i.e. the OSO bending vibration (544 cm in SO2F2) . The second component is narrower (vertical IP 13.61 eV) and overlaps the center of the f i r s t , with clearly visible but complex vibrational structure. The -1 -1 principal progression (340 cm ) is on (which i s 384 cm i n S0 2F 2) and there i s another on displaced by 500 cm \ one quantum of Vy The (000) peak position i s not absolutely certain as i t may be contained in the other component. It is also possible that there i s a f u l l progression on as there are some other weak shoulders on some of the peaks. Excitation of and V 2 is also not out of the question as the vibrational spacings could be.multiples of and v^. The assignment of which of the components i s a2 (or b2) of the mainly oxygen non-bonding orbitals is not straightforward. In SO2 the second component was broader with a progression on the band, but in this case the f i r s t component seems to be broader. There i s also l i t t l e similarity between the vibrational structure on the f i r s t component in SO2 and the second component in SO2F2. From qualitative overlap considerations and comparison with other similar lone pairs the order should be t>2 > a2, more especially i f there i s considerable d orbital interaction which w i l l be much larger for the a^ orbital. The loss of structure to the high energy side of the second component may be due to a dissociation process, or continuation of the f i r s t one. Figure 24 - 87 -Band (b) (Figure 25) This band i s clearly made up of two components each with one main progression. The f i r s t , with adiabatic IP 14.89 eV, has a regular progression of at least eleven peaks with separation 330 + 50 cm \ the eighth peak coinciding with the f i r s t peak of the second component. The vibration which i s excited is obviously (384 cm 1 in the ground state of the molecule), and using Turner's relationship W IP(V) - IP(A) = 1.2 ( neutral _ ^ w. ion assuming W. = 320 cm 1 the vertical IP is at 15.13 eV. The second b ion component appears to commence at 15.23 eV and this i s considered as the adiabatic IP (vertical IP at 15.35 eV) followed by 7 distinct vibrational peaks. The vibrational spacing of 1050 + 50 cm 1 appears to correspond to v^, the symmetric SO stretch, at 1269 cm 1 in the neutral molecule or to v 2 > the symmetric S-F stretch (848 cm 1 i n the neutral). There are also weak peaks between those of the main progression, which correspond to the addition of one quantum of (546 cm"1 in S O ^ ) . The progression on i n the f i r s t component of this band agrees with our prediction of ionization from a mainly non-bonding fluorine lone pair orbital. The slight reduction in frequency i n the ion i s possibly due to a reduction in the bond angle to improve d orbital overlap. In such cases the a 2 orbital w i l l be bonding with the 3d of the atom of the same symmetry, and this would explain the reduction in the vibrational frequency. Assignment of the structure on the second - 88 -1050 r 330 cm' i i i i i i i i i 1 1 r 1 5 1 5 . 5 Figure 25 - 89 -component to either or i s rather d i f f i c u l t . F i r s t l y i t would be strange for an orbital considered as a mainly fluorine lone pair to involve excitation of although of course this is possible i f there is a considerable amount of electron density on the oxygen atoms. However, this can be explained only in the case of d orbital participation in the form of the 3d^ z (b^) orbital, which forms an antibonding MO with the fluorine atoms. These are antibonding with respect to each other. The increase in vibrational frequency i s relevant with the antibonding character between the sulphur and the fluorine atoms. There i s an alternative, however, i f one considers the out-of-plane b^ MO which i s bonding between the fluorine atoms and has some antibonding character with the sulphur 3d^z of the same symmetry. This i s li k e l y to be the case, therefore our assignment i s and b^ for the f i r s t and the second components respectively. Band (c) This band with vertical IP 16.68 eV (Figure 26) appears to involve only one ionization process, with excitation involving three vibrational states. The principal progression from the (0,0,0,0) level of the three quanta with spacing 1130 cm \ i.e. the SO stretching vibra-tion, i s followed by two others involving the same vibration but displaced to higher energy by one quantum of (520 cm ^) and v 2 (840 cm ^ ) respectively. Only i s significantly decreased from the value i n the neutral species (see Table 10). The sharpness of the bond indicates removal of a relatively non-bonding electron. The major excitation of the SO stretching vibration and the secondary excitation of the 0S0 - 90 -1130 cm i -530-r- 840 1 6 . 5 1 6 . 7 5 Figure 26 1 7 ev - 91 -bending vibration and the bands in this region in S C ^ suggest that a sulphur-oxygen IT electron i s involved. This i s further confirmed by the spectrum of S C ^ C ^ which has a band with very similar vibrational structure in the same region. Comparing with SC^, there are two possible assignments for this ionisation from the a^ or b^ SO TT orbital. For various reasons we prefer the former. F i r s t l y , there are no a^ orbitals at higher energy than this one but there are many lower whereas the:mainly fluorine lone pair b^ orbital i s just over one eV higher in energy. Secondly the b^ orbital is expected to be more strongly ir bonding, thus at higher energy and the band produced on ionization from this orbital would not be expected to be as v e r t i c a l . The structure on this band certainly does not correspond to the f i r s t component of the band (c) of S0 9, but there are similarities with the second component. The assignment of this second component, in SO^ to ionization from the a^ orbital, i s in agreement with the calculated order. In both SO2 and SO2F2 this orbital w i l l contribute less to SOTT bonding on overlap grounds than the b^ SO TT orbital. It i s probably at lower energy than the b^ orbital in SO2 (but not in SO2F2) because of i t s involvement in the ir bonding system and the presence of a higher energy a^ orbital (S lone pair). Band (d) This band has a vertical IP at 18.36 eV and adiabatic IP apparently at 17.89 eV. Because of i t s narrow width and poorly resolved progression on v^, the 0S0 bend, (- 500 cm 1 ) , we believe that ionization i s most li k e l y to be from a mainly SO TT orbital, i.e. the TT bonding b^ - 92 -orbital. This orbital i s then about 2 eV lower in energy than in SC^ and must be stabilized by the effect of the higher energy "F lone pair" orbital, by the greater inductive effect and better overlap. This orbital probably contributes considerably to producing a stronger SO bond in S0 2F 2 than in S0 2 as discussed in the Introduction. Band (e) The f i r s t four peaks in this band appear to be well-defined therefore the peaks become broader and very asymmetric. This is probably due to the i n s t a b i l i t y of the vibrational states but there is also the possibility that a second ionization process i s involved. This latter possibility i s preferred as we expect to observe the two in plane (b2,a^) non-bonding fluorine orbitals. The f i r s t ionization band of adiabatic IP 19.7 eV and v e r t i c a l IP 19.82 eV (Figure 27) contains one progression with spacing of 850-800 cm and obviously corresponds to v 2 the S-F stretching vibration (848 cm in neutral S0 2F 2). Considering the vibrational structure and bonding picture, ionization must be from an orbital mainly localized on the fluorine without a considerable amount of bonding character. It is postulated that this the mainly b 2 fluorine lone pair orbital. The other ve r t i c a l IP 19.88 eV (a^ fluorine lone pair) has a progression with the same vibrational spacing as the f i r s t component displaced by 500 cm ^, corresponding to the 0S0 bending vibration (544 cm ^ in the neutral). The large increase in the IP of this fluorine lone pair with respect to the corresponding one in SiH9F„ and CH_F9 is due primarily to 3d orbital participation. - 94 -Band (f) This band whose ver t i c a l IP is in the energy region of ^21 eV, is p a r t i a l l y resolved i n our spectra, because of the cut off (Figure 23). It i s considered to be (S-F) a MO. He(II) source w i l l be useful to obtain the complete spectrum of this band. The orbi t a l configuration of SO2F2 can be written as follows: b2(nO) > a2(nO) > a 2(nF) > b^nF) > a^SOir) > b^SO-rr) b 2(nF) > a^nF) > b 2(S-F) 4.2.3 Sulphuryl Chloride (S0 2C1 2) We expect to observe at least 10 IP's below 21 eV which are: four chlorine "lone pair" orbitals (b2»a2,b^,a^), two oxygen "lone pair" orbitals (a 2 >b 2) , two (S-0) TT ( a ^ b ^ MO's, and two (S-Cl) a (b^a.^) MO's. However i t appears by comparison with other compounds (see Table 9) , S i ^ C ^ (see Table 20) , S0 2 and S0 2F 2 that the f i r s t six orbitals would be f a i r l y close together below 15 eV. Eight distinct bands are observed in the PE spectrum (Figure 28, Table 11) as follows: (a) 12-12.75 eV (b) 13-13.5 eV (c) 13.65-14 eV (d) ~13.8-14.75 eV (e) 16.9-17.2 eV (f) 17.3-17.9 eV (g) 17.9-18.5 eV (h) 20.12 eV. Although only four bands are observed below 15 eV, i t i s obvious from the intensities and band shapes that (a) and (b) have at least two components each. It is also possible that (g) has also two components. - 95 -- 96 -Table 11. Vibrational Structure on Photoelectron Spectrum of s o 2 c i 2 . Band3 Vertical Vibrational Frequencies (in cm Ionic IP (eV) State ^ 2^ v3 v4 (a) (12.41 400 - \ (12.41 \ (b) 13.2 500 r \ (c) 13.74 640 \ (d) (14.14 980 400 V (14.28 \ (e) 16.93 1170 380 600 200 \ (f) 17.61 \ (g) (18.21 \ (18.21 \ Neutral molecule^ 1182 405 560 218 a See text; Ref. 93. Band (a) This band (Figure 29) has a barely resolved progression with separation 400 cm ^ on the asymmetric low IP side, probably corresponding to the S-Cl stretch at 408 cm"1 in S0 2C1 2- There are at least two IP's in this band considering the intensity ratio and band shapes. However, i t i s d i f f i c u l t to resolve them thus we place a l l the ver t i c a l IP's at 12.41 eV (the maximum of the band). The adiabatic value of the f i r s t component i s at 12.05 eV (the onset). - 97 -- 98 -Assignment of the two possible components requires consideration of a l l of the "lonepair" ionization processes possible i n this energy region. In the unperturbed state the oxygen "lone pairs" (e.g. in SC^), are at comparable energy to chlorine lone pairs in ^ other dichloro compounds (e.g., 12.8 eV i n Cl^O , 13.2 eV in C^CO). However, the oxygen "lone pair" orbitals in SG^ and SC^^ a r e '*'n t b e e n e r § y region of 13-14 eV, where the ligands have no effect on this energy. Therefore we expect to observe the oxygen "lone pair" a2,b2 orbitals at 13-14 eV. In this case we assign these two IP's to the out-of-plane chlorine "lone pair" b^ and a2 orbitals. These IP's are considerably higher than the corresponding ones in S i ^ C ^ and C^C^, and are explained in terms of interaction with the sulphur atom via 3d orbitals. The f i r s t component is b.ro.ader and can be assigned to the chlorine "lone pair" b^ orbital. The second IP is an a2 chlorine "lone pair" orbital. Band (b) This very odd-shaped band (Figure 30) with adiabatic IP ~13 eV and vertical IP 13.2 eV has a poorly resolved progression of broad peaks separated by 500 + 60 cm This corresponds to the 0S0 bend which is at 560 cm 1 in the ground state of the molecule. This IP can be assigned to ftie oxygen "lone pair" b^ orbital which is bonding with the sulphur 3dyz of the same symmetry, and bonding also between the oxygen atoms. This bonding character is more li k e l y to excite the 0S0 bending mode in SO2CI2. This M0 has the same IP as the corresponding one i n S0„ and S0„F,,. - 66 -Band (c) This band with ve r t i c a l IP 13.74 eV (Figure 31) has two intense peaks separated by 640 cm ^ and certainly at least one other peak at higher energy and shoulder at lower energy. The overlap with band (d) adds a considerable complication but for various reasons this band must be assigned to ionization from the oxygen "lone pair" orbital. F i r s t i t has been recognized previously that the a^ non-bonding orbital are sharper than the others, also the more oxygen-oxygen anti-bonding a^ orbital explains the increase in the vibrational frequency of the 0S0 bending mode. The shift to high energy with respect to the b 2 orbital i s expected due to the possibility of interaction between the oxygen "lone pair" a^ orbital and the chlorine "lone pair" with the same symmetry. Band (d) This band is very broad and almost certainly, from i t s relative intensity, has at least two components. The low energy side of the band gives some indication of progressions on both the S-0 and S-Cl stretching vibrations (980 and 400 cm respectively). This band can be assigned to the chlorine "lone pair" b^ and a^ in-plane orbitals. These two IP's are higher in energy than the corresponding ones i n 9 methylene chloride and dichlorosilane in a similar way to the (a2,b^) orbitals, but with larger magnitude. This i s explained by a larger interaction with the sulphur atom via 3d orbitals. The vibrational frequency of 400 cm ^ indicates an antibonding MO i s involved upon ionization; this i s so in the case of the chlorine "lone pair" b 2 orbital, which i s antibonding between the chlorine-chlorine atoms, and forms an - 101 -- 102 -antibonding MO with the sulphur 3d orbital of the same symmetry. Therefore the second component can be assigned to the chlorine "lone pair" a^ orbital, which also may interact with the sulphur 3d^2 ^2 ( a i ^ orbital. Band (e) This band (vertical and adiabatic IP 16.93 eV) has the best defined structure (Figure 32) and this i s similar to that observed in the corresponding band in S0 2F 2 (kand > Figure 26). It i s possible to analyze the band in the same way as i n S0 2F 2 with the addition of a fourth progression, displaced by one quantum of the Cl-S-Cl bending vibration. However, in addition to the series (n,0,0,0), (l,n,0,0), (1,0,n,0) and (1,0,0,n) up to n = '2 in a l l cases and n = 3 for (n,0,0,0), there appears to be some other weak peaks indicating that other vibrational states are populated. However, one cannot separate out a l l the peaks for these series as w i l l be almost half and twice v^. Either method of assignment is possible, or more l i k e l y , a combination of both as the band looks very similar to that in S0 2F 2 (16.68 eV). There is very l i t t l e change i n the vibrational frequencies of to in the ion compared to the neutral species. In contrast, in S0 2F 2 there i s a significant decrease in (from 1269 cm 1 to 1130 cm 1 in the ion) and this maybean indication of stronger S-0 Tr-bonding in this molecule. Another indication of this i s obtained indirectly by considering the greater excitation of the halogen-involved vibration for S0 2C1 2. This i s shown by excitation of the Cl-S-Cl bend whereas excitation of the F-S-F bend is lacking in the corresponding band in 17 17-25 17^5 eV Figure 32 - 104 -SO2F2 and also by the increased population of levels involving the S-Cl stretch. It follows that i f there is more electron density in the S-Cl bonds there must be less in the S-0 bonds. Despite this, i t fe interesting to note that this band is at higher energy in SO2CI2 than in SO2F2, and this may be due to larger 3d^2 (a^) orbital participation. However, this seems to lend further support for the assignment of the a^ n S-0 orbital. The shape of this band in both molecules' PE spectra makes i t clear that the orbital is very much the lesser bonding of the two SO TT orbitals. Bands (f) and (g) These bands exhibit no vibrational structure, and are assigned to the three MO's b^(S-O) TT and b2 and a^ (S-Cl) 0 . The f i r s t maximum at 17.61 eV could correspond to the b^ (S-0) TT which is at the same order of energy as i t is in SO2F2. The other two bands overlap, and only one maximum can be distinguished, at 18.12 eV. These two orbitals are more stabilized than the corresponding ones in S2CI2 which are at 15.85 and 16.5 7 eV respectively. This increase in energy can be rationalized by involving 3d^ (a^) orbital participation, and the interaction with the orbital (S-0) TT of the same symmetry (a^). The orbital configuration of SO2CI2 can be written as follows: b^nCl) > a 2(nCl) > b 2(n0) > a2(nO) > b 2(nCl) > a^nCl)- > a 1(S0 Tr) > b1(S0Tr) > b 2(S-Cl) > a^S-Cl). - 105 -4.2.4 Thionyl Fluoride ( S O F ? ) By comparison with S®2^2 a n d ^2* a t -*-east s e v e n IP's are expected for S O F 2 below 21.2 eV, corresponding roughly to ionization from the sulphur lone pair, oxygen lone pair, four fluorine lone pairs and the sulphur-oxygen ir orbital. Unfortunately i t was not easy to obtain a good PE spectrum of SOF2 (Figure 33), but five ionization processes are clearly v i s i b l e below 17 e,V (Table 12) and in the broad band extending from 18 to 20.5 eV there i s l i k e l y to be more than one component. Table 12 Ionization Data for S0F 2 Band 3 Vertical IP (eV) Vibrational V l V2 Frequencies V3 V4 Ionic 3 State (a) 12.60 - - (440) 440 2A< (b) 14.14 (1000) - - - V (c) /14.8 I14.85 - 800 800 - 350 V (d) 16.6 J17.0 l l 7 . 0 - - -V neutral moleculeC 1308 806 530 326 a See text; b + 40 cm"1; C Ref. 93. - 107 -Band (a) This band with an apparanet adiabatic IP of 12.25 eV and ve r t i c a l IP of 12.60 eV (Figure 34), has clear vibrational structure on the low energy side. The peaks are quite irregular and there i s the possibility of a double progression, especially as there i s clearly an extra peak just below the vertical IP. The spacing of the eight principal maxima varies from 440 to 380 cm \ with a decrease through the progression except between the seventh and eighth member (440 cm "*") . This vibrational frequency could correspond to two frequencies in the neutral species, i.e. the bending of the SO bond relative to the F-S-F plane -1 -1 (326 cm ) or the F-S-F bending vibrations (526 cm ). It is possible to infer from the width of the peak that progressions involving both frequencies are present. The loss of structure just above the vertical transition could be due to predissociation, or a geometry change producing a greater number of energy levels. The ionization process producing this band must involve an a orbital localized mainly on the sulphur atom (i.e. the sulphur lone pair). The corresponding v e r t i c a l IP in SO^ to which S0F 2 can be compared as the resultant of the two SF bonds is at 120° to the SO bond is 12.50 eV. The CNDO/2 calculations which give f a i r l y good agreement for the f i r s t few IP's also agree with this assignment. This orbital w i l l almost certainly be antibonding with respect to the oxygen and fluorine atoms. It should have almost negligible interaction with the next higher a' orbital which has mostly TT (S-0) character. This should almost certainly mean that the S-0 wagging vibration is excited. This requires an increase from 326 cm ^ in the neutral species to 440 cm ^ i n the ion which would only be possible i f there was a strong antibonding interaction between sulphur and oxygen. -.108 -1 1 1 1 2 . 2 5 1 2 . 5 0 1 2 . 7 5 Figure 34 —I 1 3 . 0 0 B> ev - 109 -Although the ionic frequency i s closer to the F-S-F bending vibration of the neutral (526 cm ^ ) , there seems no mechanism for exciting this vibration. Band (b) This band, with v e r t i c a l IP 14.14 eV, seems to have a progression of peaks with separation of 1000 cm As the ionization i s from the non-bonding oxygen a" orbital, excitation of the S-0 stretching frequency (1308 cm 1 in SOF^) is probably involved. This orbital in SOF2 is stabilized by comparison with the corresponding oxygen lone pair orbital in SO2 (a2 and b2). The fluorine atoms probably partly produce this stabilization by decreasing the size of the d orbital enabling better overlap with the oxygen lone pair. Also the lower symmetry of SOF2 allows more SO IT character i n the oxygen lone pair orbital, as well as the orbital contraction effect mentioned above. Band (c) This band shows vibrational structure. The peaks are quite sharp but the structure is quite complex (Figure 35). However, two progressions with vibrational spacing of 790 cm 1 are obvious with an additional quantum of frequency 310 cm The former corresponds to the S-F stretching vibration (806 cm 1 in SOF2) and the latter to the 0-S wagging vibration (326 cm 1 in SOF2). It i s possible that there are two ionization processes involved in this band which correspond to the fluorine "lone pair" orbitals. The f i r s t component has an adiabatic IP of 14.56 eV and a v e r t i c a l IP of 14.8 eV. The second one has a possible vertical IP 14.85 eV which is very close to the f i r s t one, and corresponds also - 110 --1100 cmL Figure 35 - I l l -to the fluorine "lone pair" orbital. It i s d i f f i c u l t to obtain the adiabatic IP, since i t may overlap with the f i r s t progression. Band Cd) This band (15.5 eV to 17.5 eV) has at least three structureless components. The f i r s t is extremely broad and i t s vertical IP i s d i f f i c u l t to determine but is probably about 16.6 eV. By comparison with S0 2 > S0 2F 2 and SC>2C12 we find that this IP i s of the same energy as the (S-0) TT bonding MO. Despite the difference in band shape, which in this case indicates a stronger bonding character, we assign this component to the (S-0) ir a' MO. This is followed by the two remaining fluorine "lone pair" bands, and only one maximum can be distinguished, which corresponds to a vertical IP of 17.0 eV. 4.2.5 Thionyl Chloride S0C12 By comparison with the other species nine IP's could be observed for the molecule below 21.2 eV, corresponding roughly to the following orbitals; sulphur lone pair, four chlorine long pairs, oxygen lone pair, two sulphur-chlorine a and a sulphur-oxygen TT. A S i t was extremely d i f f i c u l t to obtain a spectrum of S0C12 free of S0 2 (see Experimental), no structure was observed on any of the bands. Therefore we expect that the band between 13-14 eV contains some S0 2 > similarly for the band between 16-17 eV which seems to be purely S0 2. Assignment of the IP's below 17 eV was possible by considering the spectrum of S0 2C1 2« The f i r s t band whose vertical IP is 11.3 eV (Figure 36, Table 13) i s assigned to the a' sulphur "lone pair" orbital. Also the f i r s t IP of - 113 -dimethylsulfoxide and thionyl fluoride (at 9.3 and 12.6 eV respectively) are assigned as sulphur lone pairs. Table 13. S0C12 3. SL Band Vertical Ionic State IP (eV) (a) GO (b) (c) Cd) " See text. With this compound i t is li k e l y that the mainly oxygen lone pair (a") orbital w i l l be at higher energy than the two chlorine a" lone pairs, as SC"2 has i t s second IP higher than that of most chlorine lone pair orbitals. This i s the case in SC^C^ as has been discussed previously. Therefore we assign the two IP's at 11.9 and 12.21 eV to chlorine "lone pair" a" orbitals. The IP at 12.55 eV is assinged to the oxygen "lone pair" a" orbital, followed by the remaining two chlorine "lone pair" a' orbitals, which have the ver t i c a l IP 13.4 eV for the f i r s t component. The maximum for the second component is not apparent i n the band and i t overlaps with the f i r s t one. We expect to observe bands for the (S-0) TT and the two (S-Cl) a orbitals at the energy of { 11.3 11.9 12.21 12.55 13.4 13.4 2A< 2A" 2A" 2A' 2A' 2A' - 114 -16-18 eV, but the presence of the band in this region (see spectrum) is very similar to that in SO^, therefore i t is impossible to decide which one is which. 4.2.6 Bonding in Sulphuryl and Thionyl Halides We now compare the values of the IP's for these compounds with some of the other physical properties to see i f there is agreement with the current bonding picture. By considering the assignment in the previous discussion, we made the oversimplification i n considering the orbitals to be localized, but despite this we believe the general argument to be valid. F i r s t l y , looking at the sulphyr-halogen bonds, we have already observed that these are significantly shorter in the sulphuryl compounds compared to the thionyl compounds. This is in agreement with the stabilization of the fluorine lone pair orbitals in ^O^F^ compared to SOF2 (by increased sulphur 3d orbital participation). The average of the mainly fluorine lone pair (out-of-plane) IP's i s 15.26 eV for SO2F2 and 14.85 eV for SC^. The same trend has been observed for sulfuryl and thionyl chlorides, where the IP is 12.8 eV for SC^C^ and 12.05 eV for SOC^. In the introduction an argument based on bonding changes was used to explain the properties of SO bonds in these compounds. Many 99 of these properties have been compared for a large series of sulphur-oxygen compounds and the relevant data for the five compounds is given in Table 15. These numbers a l l indicate that the SO bond order decreases in the order S0 2F 2 > S0F 2 > S0 2Cl 2-> S0 2 > SOC^. - 115 -93 Table 14. Properties of the SO Bond ° a - 1 5 Compound r g 0 (A) v g 0 (cm ) k a 10 S0 2F 2 1.405 + 0.003 1385 12.05 S0 2C1 2 1.43+0.02 1298 10.75 S0 2 1.432+0.002 1259 9.96 S0F 2 1.412+0.001 1308 10.95 S0C12 1.45 + 0.02 1229 9.49 Average SO stretching frequency. b SO force constant. Table 15. Some Average Ionization Energies (in eV) for Orbitals Mainly Localized on the Oxygen Atoms I II III S0 2F 2 15.55 _ 17.51 13.58 s o 2 c i 2 15.37 17.27 13.47 so 2 14.99 16.60 13.36 S0F 2 15.57 17.00 14.14 soci 2 14.7 16.00 13.40 I = average of the mainly oxygen ir type orbitals: four for the SO species and two for the SO species. II = average of the SO TT orbitals. III = average of the SO lone pair orbitals. - 116 -We now examine the IP's for the mainly SO orbitals (on the above assignments) to see i f there i s any correlation. Table 15 l i s t s some pertinent average IP's for (I) a l l the mainly oxygen TT type orbitals both lone pair and SO TT, (II) the SO ir orbitals and (III) the mainly oxygen lone pair type orbitals. The energy of the b^ TT SO orbital i s also given for the di-oxygen species. Although none of the averages show exactly the above trend, the overall average of the energy levels i s f a i r l y close to i t , i.e. SO„F„ ^ S 0 F O > S 0 o C l o > S 0 O > S 0 C 1 - . The biggest influence in this Z Z ^ Z Z Z z z order i s the (SO) levels. For SO^ and the halo species the range of the average TT SO levels i s 1.5 eV whereas the range of the oxygen lone pair levels is only 0.8 eV. It seems reasonable that a TT bonding orbital would be more influenced by effective change i n the vi r t u a l l y non-bonding orbitals on the oxygen atoms. However, most bonding pictures would indicate that the and oxygen lone pair orbitals have some interaction with the sulphur d orbitals and that this should increase with the charge on the sulphur atoms. From SO^ to S O 2 F 2 the lone pairs (a^yb^) are stabilized on average by only 0 . 2 2 eV whereas the TT orbitals (a^»b^) are stabilized on an average by 0 . 9 eV. This is due mainly to larger 3d orbital participation in the form of 3d 9 and 3d o o (a.) orbital, also the interaction with orbitals of the z^ x^-y^ 1 ' same symmetry would cause considerable changes in the IP's. This i s especially noticeable in the b^ TT (SO) which moves from 16.6 eV i n S O 2 to 18.36 eV in S 0 2 F 2 (see Table 1 0 ) . The anamolously large stabilization of the a" oxygen lone pair orbital in S O F 2 could be partly due to better d overlap due to the less rigorous symmetry requirements compared to the C„ species. - 117 -If suffices to state that the trends in bonding indicated i n Table 14 are in general agreement with those given in Table 15. It i s concluded that there is an increase in ir bond strength through the series S0 2< S0 2C1 2 < SOF2 < S0 2F 2- It is interesting to note that i n the sulphur IV state one oxygen atom has an intermediate effect between two chlorines and two fluorines. - 118 -CHAPTER V PHOTOELECTRON SPECTRA OF SOME BROMOETHYLENES 5.1 Introduction The electronic structure of ethylene 1^ and i t s derivatives has been considered with special interest.^' 1^"' In the case of haloethylenes, one can use photoelectron spectroscopy (PES) to study the inductive and resonance effects. Evidence for the latter can be seen in direct interactions between thehalogen "lone pair" orbital, whose axis i s perpendicular to the plane of the molecule, and the C=C TT system. In view of the recent work on the PE spectra of chloro^' 86 and fluoroethylenes, i t is of interest to examine the "TT" interaction -• in some bromoethylenes. 5.2 Individual Molecules 5.2.1 Vinyl Bromide The.PE spectrum of vinyl bromide (Figure 38) shows at least seven bands below 21.22 eV. The low resolution PE spectrum of this molecule reported by Turner et a l . [see Ref. 15, p. 246] does not reveal any vibrational structure. The low symmetry of vinyl bromide (C ) gives s rise to only two irreducible representations, a' and a". The IP's corresponding to these MO's are reported in Table 16. - 119 -1 - 120 -Table 16. Vertical Ionization Potentials of Bromoethylenes (eV) Vinyl bromide 9.80 (1340,740, 290) 10.90 (1200,460) 12.28 (650) 12.94 15.02 16.21 19.20 cis-1,2 dibromoethylene 9.44 (1280,650) 10.74 11.24 11.56 12.85 13.27 14.80 16.48 trans-1,2 dibromoethylene 9.44 (1310) 11.05 11.13 (560) 11.6 (680) 12.96 ( 950) 13.24 15.92 19.14 2-bromopropene 9.58 (1200,290) 10.63 11.62 12.40 13.53 15.15 15.84 Values are + 0.02 eV; vibrational frequencies (cm ), where resolved, are given in parentheses and are + 40 cm \ The f i r s t band (Figure 37) shows the presence of three vibrational progressions of spacings 290, 740 and 1340 cm \ The adiabatic IP i s 9.8 eV, and the band i s assigned to the (C=C) IT MO, with a large interaction of bromine 4p lone pair in the form of a (C-Br) TT MO 108 in similar way to that found in vinyl chloride. It is interesting to note the 0.7 eV decrease in IP with respect to the (C=C) TT IP in ethylene at 10.51 eV. The inductive effect of the bromine atom would tend to promote the inverse. The resonance effect i s therefore dominant in vinyl bromide. Vinyl chloride and fluoride have an IP of 10.18 and 10.31 eV respectively. However, in these two compounds the inductive effect of the chlorine and fluorine atoms is larger than that in the bromide, but s t i l l less in magnitude than the resonance effects. 1 - 121 -The 1340 cm ^  vibration can be assigned to the (C=C) TT stretching frequency, which is 1595 cm in the ground state of the neutral molecule."^ 9 The 740 cm ^  vibration is assigned to the (C-Br) stretching mode (598 cm ^  in the neutral). This reflects the anti-bonding character of the MO involved. The 290 cm vibration corresponds to the 6 (C-CBr) bending mode (345 cm ^  in the neutral).'''"'"^ The second band is relatively sharp and indicates an adiabatic IP at 10.9 eV, and is accompanied by two progressions of 1200 and 460 cm ^ . This IP can be assigned to the bromine non-bonding 4p orbital by comparison with vinyl chloride. The 1200 cm vibration in the ion may correspond to the (C-H) bending mode (1260 cm in the ground state 110 of the molecule ). This may be compared with the vibrational frequency of 1240 cm ^  which accompanies the spin-orbit components of 9 the bromine non-bonding orbital at 10,8 eV in CH^Br. This frequency is assigned to he symmetrical CH^ deformation vibration (1305 cm ^  in the neutral). The 460 cm ^  vibration corresponds to the (C-Br) stretching mode, reduced from 598 cm i n the neutral which indicates slight C-Br bonding in the ionized orbitals. 86 Lake and Thompson observed that the two chlorine "lone pair" orbitals in vinyl chloride almost degenerate at 11.72 and 11.87 eV. Klasson and Manne,"*"^ 7 and Turner et.al. [see Ref. 15, p. 246] do not agree with this interpretation on the ground that the halogen "lone pair" orbitals cannot be degenerate, since one i s interacting strongly with the (C=C) IT system; w i l l therefore lose i t s non-bonding character. Also the spectrum of Turner et a l . was significantly different - 122 -i n this energy region. Klasson and Manne attribute Lake and Thompsons' 11.87 eV peak to vibrational structure. In the case of vinyl bromide, the bromine 4p non-bonding orbitals are assigned at 10.90 and 12.28 eV. Therefore we expect that the separation between the two non-bonding chlorine electrons to be at least the same, this w i l l be discussed elsewhere. The third IP at 12.28 eV indicates the presence of some vibrational peaks of ~650 cm This can be assigned to the (C-Br) stretching mode (598 cm ^ i n the neutral). This IP i s assigned to the other 4p bromine non-bonding orbitals with (a") symmetry. The broaden-ing of the band indicates the bonding nature of the MO involved. The other IP's are reported in Table 16. 5.2.2 2-Bromopropene The PE spectrum of 2-bromopropene shows the presence of at least eight bands below 21.22 eV (Figure 38); the f i r s t two exhibit vibrational structure of 290 and 1200 cm ^ respectively. The f i r s t band indicates a vertical IP at 9.58 eV which corresponds to ionization from (C=C) ir MO. This IP considerably less than the f i r s t IP of ethylene or even vinyl bromide because of the CH^ group inductive effect. This w i l l add to the resonance effects present i n vinyl bromide and result in a large decrease of the (C=C)ir IP. [This M0 has a counterpart in vinyl bromide, and can be described as a direct interaction between one of the bromine out-of-plane "lone pair" 4p orbitals and the (C=C) ir M0]. The 1200 cm ^ vibration is assigned to the C=C stretch (1636 cm in the neutral"*"^9). Figure 38 - 124 -The second band in the PE spectrum i s rather sharp and has an adiabatic IP at 10.51 eV (Figure 38(c)). This i s assigned to the bromine "lone pair" 4p orbital, with some carbon participation. This i s reflected in the 4p decrease of the C-Br stretching frequency -1 -1 (545 cm i n the neutral) of ~ 350 cm . The other bromine "lone pair" 4p orbital (a") interacts with the (C=C) bond and has some (C-Br) t bonding character, and i t has a ver t i c a l IP at 12.4 eV, compared to the corresponding one at 12.9 eV in vinyl bromide. The decrease of 0.5 eV i s due to the presence of methyl group i n 2-bromopropene. The broad band between 14 and 17 eV has counterparts in methyl 9 bromide, methanol and methyl mercaptan, and can be assigned to the o (CH3) MO. The other IP's are reported in Table 16. 5.2.3 Cls-1,2-dibromoethylene Here there w i l l be four orbitals derived from 4p "lone pair" atomic orbitals of the bromine atoms. One has the correct symmetry to conjugate with the C=C TT orbitals, consequently this i s expected to lead to a higher IP than the other three, also i t naturally loses some of its non-bonding character, to give a broader spectral band shape compared to the other three. Cis-1,2-dibromoethylene has symmetry, and the PE spectrum shows the presence of eight distinct bands below 21 eV (Figure 39). The f i r s t band has adiabatic and vertical IP's at 9.32 and 9.44 eV respectively. The band shows two progressions, at 650 and 1280 cm 1 and corresponds to the (C=C) IT (b^) ionization. The vertical IP is in good agreement with that reported by Bralsford et a l . and is lower than the corresponding one in vinyl bromide and 2-bromopropene (Table 16). It i s reasonable to - 126 -suppose that the magnitude of the interaction between the (C=C) TT bond and the bromine atom i s much larger than they are in vinyl bromide and 2-bromopropene. The frequency observed at 650 cm 1 is assigned to the (C-Br) stretch (580 cm 1 in the neutral). This indicates the antibonding character of the MO involved. The 1280 cm 1 vibration is assigned to the (C=C) stretch which i s reduced from 1584 cm - 1 in the ground state of the molecule. 1 1^ The second band (vertical IP 10.74 eV) i s shown in Figure 36 and originates from ionization of a MO. The band shows a 450 cm 1 vibration which can be assigned only to the C-Br stretching mode (580 cm 1 in the neutral) and so we conclude there is slight interaction between the bromine "lone pair" orbital and the carbon atoms. The two IP's at .11.24 and 11.56 eV are assigned to the bromine "lone pair" orbitals a^ and a^ (TT) respectively. The 350 cm 1 vibration on the latter band reflects the bonding character of the (C-Br) MO. The fourth bromine (b^) "lone pair" which interacts with the ^ (C=C) bond has vertical IP at 14.80 eV. The other IP's are reported in Table 1 6 5.2.4 Trans-1,2-Dibromoethylene The PE spectrum (Figure 40) of trans-1,2-dibromoethylene (symmetry ^2h^ s n o w s t n e presence of at least six bands below 21 eV and the f i r s t three show vibrational structure. The f i r s t band (Figure 40b) is very similar to the corresponding one in cis-l,2-bromoethylene and is assigned to the TT (C=C) MO, and has adiabatic and vertical IP's of 9.30 and 9.44 eV respectively. The magnitude of interaction between the (C=C) TT band and the bromine non-- 127 -- 128 -bonding electrons is almost the same as in _cis-l,2-dibromoethylene, which probably accounts for the (C=C) TT IP's being the same. Since the ionized MO has symmetry a^ we expect that the bromine "lone pair" w i l l have a higher IP than the other three "non-bonding" orbitals. The vibrational frequencies observed in the f i r s t band are 290 and 1310 cm These are assigned to the (C-Br) bending and the C=C stretching modes respectively (218 and 1578 cm ^ i n the neutral"*""^) . The increase in the vibrational frequency of the bending mode of the * molecule ion reflects the IT character of the (C-Br) MO. Since the bromine atoms are relatively far from each other with respect to cis-1,2-dibromoethylene, there can be no direct interaction between them, and we expect to find the non-bonding 4p orbitals almost degenerate. The second band shows the presence of vibrational structure (Figure 40c), and the f i r s t strong peak seems to be asymmetric, also the vibrational spacing between v' = 0 <- v" = 0 and v' = 1 •«- v" = 0 i s 650 cm which is larger than the regular spacing of the other vibrational peaks at 560 cm \ This led us to suspect two IP's to be associated and their adiabatic values are 11.05 eV (a ) and 11.13 eV (b ) 8 u respectively. The 560 cm ^ vibration which follows the f i r s t peak relates to the adiabatic IP of 11.13 eV and is due to the (C-Br) symmetrical stretch (748 cm ^ in the neutral). The third band is also assigned as bromine "non-bonding" 4p orbital and has an adiabatic IP of 11.6 eV (b ) with vibrational structure of 8 680 cm This also corresponds to the C-Br symmetrical stretch. The slight lowering of frequency on ionization reflects the almost entirely non-bonding character of the M0. The other bromine "lone pair" which - 129 -interacts with the (C=C) IT bond and has the same symmetry (a^) has ver t i c a l IP of 13.24 eV. The other IP's are reported in Table 16. - 130 -CHAPTER VI PHOTOELECTRON SPECTRA OF SOME HALOSILANES 1 1 6* 1 1 7 6.1 Introduction Many of the properties of compounds containing s i l i c o n bonded to elements with TT electrons, either lone pair or bonding, have been rationalized in terms of the interaction of the TT orbital with a symmetry related d orbital on si l i c o n to form a (p d) TT bond. A 112 recent review gives an excellent survey of the available evidence. In this work i t is referred particularly to the physical properties in a series of s i l i c o n containing compounds, which do not conform to what would be expected by simply comparing with the carbon analogues, taking into account the electronegativity difference. Among such properties are changes in shape and opening of bond angles, and bond energies which are greater than expected. The use of PE spectroscopy helps to reveal more of the nature of the IT or "lone pair" electrons on the atom adjacent to s i l i c o n . One would expect on electronegativity grounds that the IP's of these TT electrons would move to lower energies compared to the carbon analogues, and any deviation from this would indicate that these electrons were involved in bonding with the s i l i c o n atoms or groups including the sil i c o n atom. Examination of the other IP's of the molecule could aid I 131 -with the assignment of the type of interaction involved. If the electrons on the carbon analogue were vir t u a l l y non-bonding, ionization would produce a species with a very similar Franck-Condon envelope and hence a f a i r l y sharp band i n PE spectrum. Should the electrons on the si l i c o n species be involved in some bonding there would be a change in geometry on ionization and excitation to several vibrational levels could be expected, i.e. a broader band would be produced. As many of the properties of the halosilanes depend on the postulated (P -> d) TT bonding i t was decided to use PE spectroscopy to try to determine how much influence i t has. It was thought that a study of one whole series SiR^X (X = F, C l , Br, I) and at least one series-SiH X. (here X = Cl) would be sufficient (with comparison n 4-n with the previously studied carbon analogues) to give considerable information 811^2 was also studied. The interpretation of the PE spectra can be aided by using the results of semi-empirical calculations such as CNDO/2, which also predict the amount of d orbital involvement i n the various molecular orbitals. 113 Cradock and Ebsworth published a note on some of the molecules studied here, i.e. the SiH^X species. In this, the breadth of the lowest energy band in the PE spectra of X= C l , Br, I, assigned to ionization of the mainly "lone pair" electrons, and the shift to higher IP's compared to the carbon compounds, was cited as evidence for (P -> d) ir bonding. The possibility of P -> a (SiH^) bonding was not considered as there was no excitation of any SiH vibration. However, there are a number of cases in which expected vibrations do not appear in PE spectral bands (e.g. in the second IP of PH^ the P-H stretch is excited 1"' 114 whereas for CH^F of the same symmetry a CH^ deformation i s excited ). - 132 -6.2 Experimental SiH^Cl and S i l ^ C ^ were prepared according to the method of Ring et al."^""* by passing SiH^ (Matheson Co.) over heated silver chloride about ten times, and collecting the products at -126° (methyl cyclo-hexane slush). Pure SiH^Cl was obtained by passing the products through a -112° trap (n-butyl bromide slush) . 95% pure S i l ^ C ^ can be obtained by passing the residue in the -112° trap through a -95° trap (toluene slush), the residue in this case being SiHCl^. SiD^Cl and 810^012 were prepared as above from SiD^ and AgCl. 118 SiD^ was prepared by the action of LiAlD^ on SiCl^ in diglyme i n vacuo. The products are d i s t i l l e d through a -126° trap to remove SiCl ^ and solvent, and pure SiD^ i s collected at -196°. 119 SiH^F and 5111^2 were prepared by the reaction of SiH^Cl and SiH^C^ respectively with antimony trifluoride and purified by passing through a -126° trap (to remove chloro-silanes). These species, especially were kept in the gas phase as short a time as possible. SiH^Br and S i l ^ B ^ were prepared by the reaction of a three to 120 one excess of SiH, with BBr„ at 0°C for 6 hours. SiH-Br was 4 3 3 obtained by passing the products through a'-: - 95° trap. The residue from the -95° trap yielded SiH^Br^ by- passing through a -78° trap (acetone to solid CQ^) . Again these species were always kept at -196°. No spectrum of SiR^B^ was obtained as i t seemed to decompose to SiH^, HBr and SiH^Br in our spectrometer. * The s i l i c o n compounds mentioned in this work were prepared by N.P.C. Westwood, Department of Chemistry, University of Windsor, Canada - 133 -The purity of a l l samples was tested by comparing the NMR and IR 121-123 spectra with the literature values for the pure species. The samples were stored at -196° unt i l used. The bond lengths employed for CNDO/2 calculations were obtained 10/ i Q r 1 O A 10 7 for each .molecule as follows: H^SiF, H 2 S i F 2 > ^ S i C l , H 2 S i C l 2 > 128 and HSiClg. The bond angles were assumed to be tetrahedral 6.3 Individual Molecules 6.3.1 SiH. and SiD. — 4 4 The PE spectrum of silane (Figure 41) shows the presence of two bands below the energy of 21 eV. The f i r s t band is quite broad and extends over the range of 11.66-14 eV, where the three components at 12.46, 12.71 and 13 eV can be distinguished. In the low energy side of the band a vibrational progression of nine peaks with frequency of -1 42 <v 800 cm is similar to that observed by Pullen et a l . The ionized electron i s from a t 2 molecular orbital. This i s similar to the case 9 15 30 of methane. ' Recently Rabalais et a l . reported the high resolution PE spectrum of methane, where the f i r s t band has a similar pattern to that of silane but with more complicated vibrational progressions. The 2 + T„ electronic state of the CH, ion which results from the ionization 2 4 of a t 2 electron leads to three overlapping electronic bands owing to Jahn-Teller effects, and these electronic bands have lower symmetry than the parent molecule. The observed three IP's in the f i r s t band of silane can be explained in the same way. The decrease in energy from 30 methane (adiabatic IP = 12.51 eV ) to silane (adiabatic IP = 11.66 eV) can be explained on electronegativity grounds. The vibrational frequency - 134 -Figure 41 - 135 -of 850 cm in the molecule ion i s assigned to (950 cm ^ in the neutral [see Ref. 68, p. 167]). At 17.94 eV there is a very low intensity progression of three peaks of mean spacing of ^ 1800 cm which i s assigned to the Si(3s)a^ MO. The frequency corresponds to the stretching mode (2187 cm ^ i n 42 the neutral). This band has not been observed by Pullen et a l . 129 however, i t was reported recently in a note by Cradock. The corresponding IP in halosilanes SiH^X i s as follows SiH^F = 19.29 eV, SiH 3Cl = 18.23 eV and SiR^Br = 18.04 eV (see text). The shift to high eenergy can be attributed to the inductive effect of the halogen atoms. The PE spectrum of SiD^ (Figure 42) shows the same pattern as the SiH^ spectrum. The vibrational structure observed between 11.66 and 12.54 eV of spacing % 590 cm ^ i s expected in the deuterated silane. The maxima of the Jahn-Teller components are 12.55, 12.81 and 12.99 eV. In SiD^, the 3a^ band i s not observed due to low intensity of the vibrational peaks. 6.3.2 R 3SiF The PE spectrum of this molecule ( C 3 v symmetry) (Figure 42) has three bands of relative intensities 2:3:1 with resolved vibrational structure in the last one and on the high energy side of the second band. This latter structure and the intensity ratios indicate two ionization processes in the second band. The f i r s t band has two components, but as this is assigned to the e (SiH 3) orbital, i t i s expected to have a Jahn-Teller contour. The ve r t i c a l IP's for the four processes are 12.8, = 16.13, <\» 16.38 and 19.29 eV. - 136 -CNDO/2 calculations including 3d orbitals on HgSiF predict four IP's below 21 eV; at 12.3, 15.75, 17.27 and 18.42 eV, corresponding to ionization from 3e C aSiH^), 7a^ (aF-SiH^), 2e (nF), and 6a^ (Si 3s) orbitals respectively. Considering the approximations involved the agreement appears reasonable. The f i r s t band, corresponding to ionization from the 3e orbital, has maxima at 12.58 and 13.02 eV, i.e. a Jahn-Teller s p l i t t i n g of 0.44 eV which i s somewhat less than that in CH^F (0.5 eV). The CNDO/2 calculations indicate that there i s some Si-F TT character in this mainly Si-H bonding orbital. This effect i s even 114 more obvious in CR^F, and this can be seen by comparing the f i r s t IP's of CR"3F and SiH^F with those of CH^ and SiH^. As expected there i s quite a large drop in IP from CH^ (13.97 eV) to SiH^ (~12.5 eV). However, the f i r s t IP in CH3F (13.04 eV) is very close to that of SiH 3F (12.80 eV). The unexpectedly large reduction i n the f i r s t IP of CH3F * is due to the much larger TT character compared to the SiH 3F. This in turn, could be caused by better overlap i n the orbitals i n CH3F and also by the poss i b i l i t y of F -> Si., interaction i n SiH 0F which i s ' pir dir 3 not available in CH3F. As mentioned above, the second band extending from 15.5-16.8 eV has two components with ve r t i c a l IP's at 16.13 and 16.58 eV, and i t would appear that these correspond to the 7a^ and 2e orbitals predicted by CNDO/2 calculations to be at 15.75 and 17.27 eV respectively. However _ an .intensity analysis of the band indicates relative intensities 16.13 and 16.38 eV of ratio — 1:1. Normally i t has been observed that when two IP's are close together a doubly degenerate one w i l l have a - 137 -Table 17. Vertical IP's of SiH^F and CH3F (eV) 9 Experimental Calc. Electronic Orbital Type CH„F CNDO/2 Density (eV) 12.08(2) 12.3 0.25 Si(3p) 0.11 Si(3d) a(Si-H) 13.04 0.64 H(ls) T r * ( S i - F ) > e 16.13 15.75 0.16 Si(3s) 0.14 Si(3p) , . ) 0.27 H(ls) a ( - b l H ; a l U7.06 { 3 ) 0.40 F(2p) 16.58 17.27 0.88 F(2p) n(F) e 19.29 (1) 18.42 0.29 Si(3s) 23.40 0.19 H(ls) 0.49 F(2p) band about twice the area of a non-degenerate one, but there is no real indication of this in the present spectrum. It would have been hoped that the structure on the second component of the band would confirm the order predicted by the CNDO/2 calculations. The spacing of the five peaks of the second component are 1470, 1420, 1400 and 1340 cm \ which is most l i k e l y to correspond to a progression in the symmetric SiH^ stretching vibration, considerably reduced from 2206 cm ^  i n the ground 121 state of the molecule SiH^F. It would have been thought more li k e l y for this mode to be excited by ionization from the a^ (mainly a SiH^) orbital, but i f as the calculation suggests the higher energy component is the e (nF), then there is a considerable bonding interaction - 138 -between the fluorine "lone pair" orbital and the SiH^ bonding orbital. In fact i t suggests that the 3e and 2e levels can be considered as antibonding and bonding levels respectively produced by the interaction of the fluorine 2p and a(SiH„) orbitals. Although the assignment of the structure on the basis of ionization from an a^ orbital would be preferred i t i s possible to reconcile either assignment the order given by the calculation i s followed. The 2e orbital i s almost a l l fluorine lone pair with an almost equal small amount (5%) of SiPand d character, but from the breadth of the PE band this seems to be a considerable underestimation. In CH^F the band appears to be somewhat wider and i s at a higher energy (^  17 eV) which i s in agreement with the suggestion that there i s a greater interaction between the F orbitals and the CH. moiety i n & pir 3 CH^F than i n SiH^F. The fact that the F lone pair electrons are at a higher orbi t a l energy i n SiH^F than i n CH^F probably also means that there i s no large stabilization of these electrons by the Si d TT orbital (as also predicted by CNDO/2 calculations). The 7a^ orbital i s mainly Si-H and Si-F bonding. The third band in the spectrum (6a^) consists of a progression of eight vibrational peaks from 19.10-19.81 eV with mean spacing of ^ 800 cm This vibrational frequency i s assigned to the H^Si wagging mode (990 cm 1 in the ground state of the molecule). 6.3.3 SiH2F_2 The PE spectrum of SiH.2F2 below 21 eV has five distinct broad bands in the range (a) 12.3-13.4 eV; (b) 14.6-15.54 eV; (c) 15.54-16.67 eV; - 139 -(d) 16.92-18.77 eV; and (e) 19.83-20.85 eV.with no apparent v i b r a t i o n a l structure. This pattern i s exactly analogous to the spectrum of CH^F^, except that i n general the bands have moved to lower energy i n S i H 2 F 2 . CNDO/2 ca l c u l a t i o n s p r e d i c t eight IP's below 21 eV. In f a c t bands (c) and (d) above appear to contain two and three components r e s p e c t i v e l y and eight v e r t i c a l IP's can be distinguished, these are l i s t e d i n Table 18 along with calculated values. Due to the loss of structure on these bands, the c a l c u l a t i o n s are used to a s s i s t the assignment of the MO's. The comparison with CR^F^ and SiH^F i s also used i n t h i s order. Table 18. I o n i z a t i o n P o t e n t i a l s for S i H 2 F 2 and o V e r t i c a l Calculated O r b i t a l Type C H o F ? IP (eV) IP (eV) Symmetry (eV) 12.85 (3) 12.4 b l a(Si-H) 13.29 15.20 (4) 14.79 (a 1) a l a(Si-H) 15.25 16.07 (8) 16.03 (b 2) b2 n(F) 15.40 16.37 17.16 (b 2) a2 n(F) 15.58 17.6 17.18 (a 2) b l n(F) ,ir(Si-F) 18.97 17.93 (12) 17.58 a l n(F ) , T r(Si-F) 18.97 18.30 18.59 (a 1) b l a(Si-F) 18.97 20.19 (3) 18.67 (a x) a l a(Si-F) 23.9 I 1 I I I 1 -I L_ L _ 12 14 16 18 2 0 Figure 42 - 141 -The f i r s t two bands are assigned to b^ and a^ a (SiH^) orbitals on the basis that (a) these orbitals are derived from the t^ orbital of SiH^ at 'v 12.60 eV (the lowest energy t^ orbital of SiF^ is at 16.46 eV), (b) the CNDO/2 calculations give f a i r l y good agreement with the observed IP's; the calculated energies being too low may be due to overestimation of the interaction with lower orbitals, or since CNDO/2 calculations give the f i r s t IP's of the second row elements 1 eV low with respect to the experimental value, (c) the order b^ > a^ i s also the order of the highest energy orbitals in 01^2-The followed four IP's with symmetries b2> a2» b^ and a^ are assigned to fluorine "non-bonding" orbitals. This assignment is more lik e l y to be the case, than that predicted by CNDO/2 (two b2 orbitals n(F), a(Si-F) with small difference in energy). The four IP's are shown to be more stabilized than the corresponding ones in CR2F2* This can be explained in terms of (P d) TT bonding. The b2 and a^ orbitals are assigned to o(Si-F) bonding (Table 18). 6.3.4 SiH 3Cl The PE spectrum of SiH^Cl shows the presence of three bands below 21 eV (Figure 43, Table 19) which are i n the ranges of energy 11.3-12.3, 13.1-14.7 and 17.9-18.9 eV with intensity ratios 6:9:2. One expects to observe four IP's below 21 eV corresponding to ionization from e[n(Cl)], a^aCSi-Cl) ], e[a(SiH 3)], and a ^ c K S i l ^ ) ] orbitals. The observed IP's on the spectrum are at 11.65, 13.51, 13.99 and 18.23 eV, - 142 -t2 l I I _i i I i _ _ i I l i 1 2 1 4 1 6 1 8 2 0 Figure 43 - 143 -Table 19. Vertical IP's of SiH 3Cl and CHgCr (eV) Experimental Calc. Electronic Orbital Type CH_C1 CNDO/2 Density 11.65 (6) 11.42 0.15 0.14 0.46 0.25 Si(3p) Si(3d) H(ls) Cl(3p) 13.51 (9) 13.08 0.14 0.12 0.63 Si(3p) H(ls) Cl(3p) 13.99 13.66 0.18 0.13 0.62 Si(3p) H(ls) Cl(3p) 18.23 17.16 0 0.43 0.33 0.15 Si(3s) H(ls) Cl(3p) n ( G 1 ) } e 11 3 a*(Si-H) > 6 i l , J a(Si-Cl) a± 14.02 a(Si-H) . 15.4 i r ( S i - C l ) ' 6 a(SiH 3Cl) a 21.5 compared to 11.42, 13.08, 13.66 and 17.16 eV obtained by CNDO/2 calculations. The f i r s t band has a barely resolved progression with a separation of 500 cm ^  which corresponds to the Si-Cl stretching vibration at 545 cm ^  i n the ground state of the molecule. This would be expected to be excited i f the Cl lone pair electrons were involved in bonding to the s i l i c o n . There are other indications of this bonding character. F i r s t l y , the band is much broader i n SiR^Cl (band width = e^) than in CH^Cl = 0.2 eV) and HC1 = 0 , 1 e V ^ * A l s ° t h e b a n d has moved to higher IP than in CR^Cl, which i s the opposite from what one would expect on electronegativity grounds. The slight decrease i n the vibrational frequency in the molecule ion also indicates the bonding - 144 -character of this orbital. The Jahn-Teller s p l i t t i n g in the e band is not obvious as in SiH^F and SiH^Br. The a1 band at 18.23 eV with adiabatic IP 18.13 eV has signs of structure. Two vibrational components separated by 1800 cm ^ are clearly observed but in this region the rising background and very poor signal/noise ratio prevent the assignment of further peaks. The PE spectrum of SiD^Cl showed this band much more clearly, and six components of separation 1320 cm ^ were observed, Figure 43. The vibration which i s excited i s obviously at 2201 cm ^ i n SiH^Cl and -1 121 1581 cm i n SiD^Cl. The reduction in stretching frequency in the ion i s as expected for the removal of an electron from a bonding orbital. The other bands in SiD^Cl spectrum are very similar to those i n SiH^Cl. 6.3.5 SiH 2Cl_ 2 The PE spectrum of SiH^Cl,, shows the presence of five bands with intensity ratio 1:1:2:3.5:1 (Figure 43 and Table 20). Obviously there are two processes involved in the third band which has a shoulder at 12.76 eV. Eight IP's are expected below 21 eV and as.a result of this and the intensity pattern three IP's must be present i n the fourth band. Intensity analyses revealed these components. Due to the loss of structure in a l l these bands, comparison with CH2C12, SiH 2F 2 and CNDO/2 calculations on S i H 2 C l 2 must be used to aid assignment of the IP's. If we consider that the Cl lone pair orbitals are not involved in bonding one would expect four low IP's from b 2, a2» b^ and a^ - 145 -Table 20. Vertical IP's of Si H 2 C l 2 and CH2 C l 2 (eV) Experimental Calc. CNDO/2 Electronic Density Orbital Type CH2C12 11.70 (2) 10.94(b1) 0.16 0.46 0.26 Si(3d) H(ls) Cl(3p) n(Cl) b 2 11.40 12.09 (2) 11.3(b2) 0.93 Cl(3p) n(Cl) b, a*(Si-H) 11.40 12.53 (4) 12.11(3^ 0.34 0.44 H(ls) Cl(3p) n(Cl) a 2 12.22 12.76 12.58(a2) 0.88 Cl(3p) n(Cl) a x 12.22 14.20 13.32(b2) 0.14 0.78 Si(3p) Cl(3p) o(Si-Cl) b 2 15.30 14.45 (7) 13.95(b1) 0.24 0.14 0.60 Si(3p) H(ls) Cl(3p) a(Si-Cl) a 1 15.94 14.60 14.62(3^ 0.12 0.68 Si(3p) Cl(3p) a(Si-H) b 1 16.77 18.32(1.5) 16.56(a;L) 0.34 0.32 0.28 Si(3s) H(ls) Cl(3p) a(H 2SiCl 2) a± 20.30 orbitals in increasing order of overlap. CH2C12 was assigned 9 according to these arguments, but reassigned with a 2 and b^ inter-changed i n the basis of the empirical calculations.^16,117 CND0/2 calculations of these IP's do not agree with the experimental values(Table 20 ) and predict a different order of symmetries. The assignment preferred for the non-bonding Cl electrons i n SiH„Cl9 places - 146 -the almost non-bonding a 2 orbital as the third band as i t is the sharpest, consistent with a small change in geometry on ionization. The other orbitals have the same order as given by purely overlap of lone pair considerations. However, the fact that the orbital in S i ^ C ^ comes at higher IP (12.53 eV) compared to the corresponding one in CI^C^ (12.22 eV) i s interesting. This is the only valence orbital which is completely non-bonding in CR^C^j ^ u t f t c a n be stabilized i n S i ^ C ^ by gaining some d^ character. The 18.32 eV band is as expected an a^ orbital being mainly Si (3s) but having ( S i - ^ ) and (SiCip bonding character. The assignment of the other three IP's around 14-15 eV is given in Table 20. 6.3.6 ,SiHCl 3 The PE spectrum of SiHCl^ (Figure 43) has five distinct bands with maxima at 11.94, 12.41, 13.07, 14.75 and 18.14 eV with relative intensities 1.2:3.4:2.1:3.4:1. There is a distinct shoulder at 14.98 eV on the fourth band which is broad enough and of intensity to imply the presence of two components. The f i r s t band (11.94 eV) i s assigned to an orbital which i s a pure Cl lone pair orbital. The sharpness of the band confirms this assignment. The shift of the a 2 band from 11.48 eV i n CHC13 to 11.94 eV i n SLHC13 indicates the effect of central atom size on this completely non-bonding orbital. The stabilization of this orbital in the SiHCl 3 shows that we cannot just consider shifts in otherwise non-bonding orbitals as an indication of (P -> d) TT bonding. - 147 -Table 21. Vertical IP's of SiHCl- and CHC1„ (eV) Experimental Calc. Electronic Orbital Type CHC1 CNDO/2 Density 11.94 10.43 0.99 Cl(3p) n(Cl) a 2 11.84 12.41 11.13 0.33 0.16 0.36 Cl(3p) Si(3d) H(ls) n(Cl) ax c*(Si-H) 11.91 12.41 11.49 0.93 Cl(3p) n(Cl) e 12.01 13.07 12.92 0.80 Cl(3p) n(Cl) e 12.85 14.75 14.64 0.80 0.10 0.10 Cl(3p) Si(3p) Si(3d) CT(Si-Cl) e 15.99 14.98 14.71 0.60 0.21 0.10 Cl(3p) Si(3p) H(ls) a(Si-H) a 1 16.96 .18,14 16.11 0.36 0.25 0.12 Cl(3p) Si(3s) H(ls) a(SiHCl 3) a 1 19.8 The second band i s assigned to a^ and e Cl non-bonding orbitals which seem to be accidentally degenerate at 12.41 eV. The other e Cl non-bonding orbitals i s at 13.07 eV. The order of these non-bonding 9 orbitals in SiHCl^ i s identical to CHCl^. CNDO/2 calculations also confirm the order (a 2,a 1 >e,e) (see Table 21). The calculated IP's are lower than the experimental values by 1 eV for the f i r s t two IP's. This has also been observed in SiH^Cl, S i H 2 C l 2 and divalent sulphur compounds discussed previously. The a^ orbital i s calculated to have a considerable amount of Si-H and Si-Cl bonding character as in the case for the lowest a^ orbital in SiH^Cl and SiH„Cl„. - 148 -The maximum at 14.75 eV i s assigned as mainly the e a(Si-Cl) and the shoulder at 14.98 to the a^ o(Si-Cl) on the basis of intensity and the relation to the a(Si-Cl) orbital of SiCl^ at 15.13 eV. The a^ (SiHCl^) band at 18.14 eV has three peaks separated by 2000 cm "*". This progression i s assigned to the v(a^) Si-H stretch, which has a -1 121 value of 2236 cm in the neutral molecule. 9 Comparison with the chloroform PE spectrum shows that the orbitals are i n the same order for SiHCl^ and CHCl^. The f i r s t four bands which correspond to the Cl non-bonding orbitals are at higher IP's in SiHCl^ than the corresponding values i n CHCl^. This can be explained in terms of (P -> d) TT bonding. Figure 44 shows that the lowest energy IP's i n the chlorosilanes between 12 and 13.5 eV are mostly derived * from the "lone pair electrons of SiCl, and that the next set of 4 bands (13.5-15 eV) are mainly due to the a(Si-Cl) MO. The spectra of the chlorosilanes are again very similar to the corresponding series of carbon compounds. One interesting point of comparison between the carbon series and s i l i c o n series is that i n the former the energy separation between the highest "lone pair" IP and the lowest of the next set of "a-bonding" IP's is a constant 3 eV through thewiole series wheras for the latter the separation is only 1.5-2 eV. This i s caused by both the greater stabilization of the Cl "lone-pair orbitals by bonding to the s i l i c o n atom, reducing the orbital energy of the lowest IP's, and also by a weaker a bonding system round the The PE spectrum of SiCl^ (Figure 4 3 ) s i m i l a r to the previously reported PE spectrum by Green et a l . - 149 -s i l i c o n atom increasing the orbital energy of the next lowest IP's. 6.3.7 SiH 3Br The PE spectrum of SiH^Br (Figure 44) shows the presence of four IP's at 11.00, 12.96, 13.63 and 18.04 eV, which correspond to ionization from e (Br), a^SiH^Br), e (SiH^) , and a^ (SiH^Br) orbitals respectively. The first band of half width 0.5 eV has a spin-orbit s p l i t t i n g of 0.20 eV, which is more reasonable than the value of 0.14 eV reported 113 49 previously, as HBr and CH^Br has splittings of ^ 0.30 eV. The half-width of the band i s considerably greater than that i n CH^Br and the IP is 0.30 eV higher. As discussed previously, a l l these observations are indicative that the p lone pairs on the bromine T r atom in SiH^Br have considerably more bonding character than those i n CH^Br. This bonding almost certainly involves the d^ orbitals on the s i l i c o n atom. There were some signs of barely resolved structure on the low energy spin-orbit components of the f i r s t band. The vibrational spacing of ^ 320 cm can be assigned to the Si-Br stretch, which -1 121 i s at 430 cm i n the neutral molecule. The second band with maxima at 12.96 eV (a^) symmetry has three resolvable peaks with a separation of 400 cm \ corresponding to the Si-Br stretch. This is the f i r s t time structure has been observed on the a^ band of an MH^ X system. Calculations on SiH^Cl indicate that this orbital i s nevertheless expected to be mainly Si-X a bonding with only a small amount of Si-H^ a bonding, although this i s less so for SiH^F. In agreement with this, a steady decrease i n this a^ IP i s - 150 -- 151 -observed through the series: SiH F (16.13), S i H ^ l (13.51) and S i H ^ r (12.96 eV) indicating a weakening of the o bonding system, especially the Si-X bond through the series. The third IP which i s assigned to e (SiH^), shows two possible maxima at-13.43 and 13.83 eV, which correspond to a Jahn-Teller s p l i t t i n g of 0.40 eV. As can be seen i n Figure 4 4 this orbital i s derived from the orbital in SiH^, and ionization from this or b i t a l produces a Jahn-Teller sp l i t t i n g of 0.5 eV. The sp l i t t i n g i s not observed in SiH^Cl as there i s an overlap with the a^ hand, but the splitting in SiH^F i s 0.46 eV. These splittings are somewhat less than those observed i n the analogous carbon containing compounds (0.5-0.7 eV) as would be expected for species with lower vibrational energies. The correlation diagrams for the halogen, non-bonding orbitals 103 in some dihalocompounds are shown in Figure 45 and 46. 6.4 Conclusions The PE spectra of halosilanes indicate that compared to the halo-methanes, the predominantly halogen "lone pair" orbitals have more bonding character. Among the evidence i s the shift to higher energy of the "lone pair" "IP's along with the reduction i n a l l other IP's for the halosilanes, and SiH^Cl and SiH^Br have broader bands than the corresponding CH^Cl and CH^Br. The stabilization of the halogen non-bonding electrons in the halosilanes i s due mainly to (P -> d) ir bonding. For the d i - , t r i - and tetrahalosilanes, the expected convergence of the "lone pair" orbitals due to reduced halogen-halogen (P -P ) TT TT interaction i s not observed. As a result i t i s postulated that some of - 152 -- 153 -- 154 -the "lone pair" orbitals are being stabilized more than others by bonding with the rest of the system. Other trends in the bonding in the halosilanes which can be deduced from the PE spectra, are: (1) . the strengthening of the Si-X a bond through the series SiH 3X (X = Br, Cl and F). (2) general weakening in the o bonding system compared to the halomethanes i s also evident from the lower IP's for the halosilanes and, as a result the physical properties of the halosilanes indicating stronger s i l i c o n halogen bonds must be caused by additional (P -> d) ir bonding. - 155 -CONCLUSIONS During the course of the work described in this thesis PE spectroscopy was used as a direct method of measuring IP's of many different molecules with high accuracy. Vibrational structure was observed in most of the molecular ions. In some "divalent" sulphur containing compounds, the sulphur non-bonding orbitals were easily identified as giving sharp peaks in the PE spectra. The effect of different alkyl groups on the energy of these sulphur non-bonding orbitals was measured. Considering hydrogen sulphide as a. standard molecule, i t was shown that the larger the alkyl group attached to the sulphur atom the smaller is the IP of thesulphur non-bonding orbital. This i s attributed to the inductive effect of the alkyl groups as electron donors, or in terms of MO theory, to the interaction between two orbitals of the same symmetry. CNDO/2 calculations predict a C-S ir* character between the sulphur p^ orbital and the adjacent carbon atoms. Some 3d orbital participation i n thionyl and sulphuryl halides has been observed to stabilize the MO's involved. The magnitude of this participation i s larger in sulphuryl halides. That i s , the average of the predominantly fluorine or chlorine lone pair (out-of-plane) IP's i s higher in SO2F2 and SO2CI2 than the corresponding ones in SOF2 and SOC^ respectively. In some halosilanes, i t was observed that the predominantly halogen "lone pair" orbitals have more bonding character, i.e. broader bands and higher IP's than the analogous orbitals in carbon containing compounds. The TT bonding between the halogen atoms and the s i l i c o n atom - 156 -is i n the form of (P •> d). CNDO/2 calculations on halosilanes predict this type of tending, but they f a i l to give the IP's i n the correct order. The calculated IP's in the case of sulphuryl and thionyl halides do not agree with the observed values. This i s due to the lack of parameterization of the CNDO/2 method when i t i s dealing with second-row elements. The CNDO/2 calculations can lead to misinterpretation of the PE spectra i f these effects are not taken into account. The vibrational structure in the f i r s t band of the PE spectrum of ethylene sulphide, which corresponds to a purely sulphur non-bonding orbital, led to the prediction of a Rydberg series in the UV spectrum of this molecule similar to that for ethylene oxide. Further PES studies on mercaptans -and other divalent sulphur compounds demonstrated the effect of the alkyl groups or halogen atoms on the predominantly sulphur non-bonding orbital. This coupled with more accurate calculations w i l l give more quantitative values about the magnitude of this interaction. The evidence of 3d orbital participation w i l l be demonstrated by studying more multivalent sulphur compounds. This i s also the case for halosilanes. The use of ESCA spectroscopy would be very useful to give a complete picture of the MO's in these molecules. - 157 -APPENDIX A 1. The Roothaan Method Molecular orbitals are usually expressed as a linear combination 131 of a set of basis functions in the form of atomic orbitals. The 132 133 resulting Hartree-Fock ' equations can then be solved by the methods of linear algebra. For a closed-shell molecule with a single determinant wave-function V «= 1^(1)0(1)^(2)6(2) .^n(2n)6(2n) (1) in which the molecular orbitals are orthonormal and are linear combinations of a set of atomic orbitals, <j> u n \\>. = E $ C . i ( l , n) (2) i = 1 Tu y i 131 Roothaan showed that the energy is minimum when the coefficients C _^  satisfy the secular equations n n E F C . = E S C .E. (u,i = 1 n) (3) v = 1 yv vx v = 1 pv v i i In these equations E^ is the orbital energy of the i t h molecular orbital. i s the overlap integral. Syv = K ( 1 ) * v ( 1 ) d T l ( 4 ) - 158 -where dr i s the one-electron volume element. If S i s zero, then <J> and <J> are orthogonal. The matrix elements of the Hartree-Fock Hamiltonian operator F ^ , can be expressed directly i n terms of atomic orbitals, and may be written as Fuv * F * v ( 1 ) d T l ( 5 ) This equation can be written as F =H + E P [(uvlXo) - -kuXlva)] (6) V V y V Xa Xa 2 H is the matrix element of the .one-.electron core Hamiltonian operator yv including the kinetic and potential energy in the electrostatic f i e l d of the core H y v = K a ) [ ~ h 2 ~ l \ K W d T l C 7 ) The core consists of the nuclei and any electrons not considered e x p l i c i t l y in the calculation. The population matrix i s defined by P = 2 E C .C . (8) yv ^ p i vx where the summation is over occupied molecular orbitals only. The quantities (yv|Xa) are two-electron interaction integrals over - 159 -atomic orbitals ( y v | A a ) = f A * Q H C D ZT- <f>*(2)(f> ( 2 ) d T , d T , ( 9 ) JJ y v 1 2 A O i. i. where r^ " i s the distance between the two electrons. This equation represents the repulsion between two electrons with specified charge distributions ii. and ib. are orthonormal i f ip.ib.dT = / CScp C .)(Ect> C . ) d x = EE C .C . /cb A dx = 6. . (Kronecker J 7 y % y i / v v j y v y x V J 7 V V x j d e ] _ t a ) = EE C* C . S 6 . . = 0 i 4 3 (10) = 1 i = j The Roothaan method has been used extensively in molecular orbital n 1 .. i 1 3 4 , 1 3 5 computatxons on small molecules. 2. The Zero-Differential Overlap Approximation (ZDO) The most d i f f i c u l t , and time consuming part of LCAO self-consistent molecular orbital calculations i s the evaluation and handling of the large number of electron interaction integrals (yv|Aa) defined i n equation ( 9 ) . Many of these electron repulsion integrals have values near zero, especially those involving d i f f e r e n t i a l overlap between different orbitals. The product (1)cj>^(1) i s called d i f f e r e n t i a l overlap of orbitals y and v. It i s integral over a l l space in the overlap integral S . yv - 160 -A useful approximation in developing semi-empirical molecular orbital theories i s the zero-differential overlap approximation (ZDO), in which differential overlap is assumed to be zero for different orbitals. Under the zero-differential overlap approximation (yvlxa) = (yyl'vv) <5 6, (11) ' 1 UV AO both the overlap integrals, and the electron interaction integrals representing their interaction with other charge distributions are neglected. 136 3. ..Complete Neglect of Differential Overlap (CNDO) 137 The CNDO method introduced by Pople, Santry, and Segal retains the main features of electron repulsion, only valence electrons are treated e x p l i c i t l y , the inner shells being treated as part of a rigi d core, so that they modify the nuclear potential i n the one-electron part of the Hamiltonian. The atomic orbital basis set i> is a valence y set (IS for hydrogen, 2S, 2P x > 2P^ and 2P^ for carbon, oxygen, etc.). The basic approximation is that the zero-differential overlap approximation i s used for a l l atomic orbitals pairs <))u())v» the Roothaan equations for the LCAO coefficients for a closed-shell molecule simplify to E F C . = E.C . (12) v yv v i l yx where the elements of the Hamiltonian matrix F are given yv 6 and 161 -F u u = - I P u y ( y v , m j ) + A E P ^ ( y y | U ) ( 1 3 ) A F = H - ^ P (yy|vv) (14) yv yv 2 yv 1 These equations are not invariant to orthogonal transformations among atomic orbitals on the same atom, so that further modifications must be made. To restore rotational invariance, the remaining electron interaction integrals (yylxx) are assigned a common value f for a l l AtS orbitals A and A on atoms A and B respectively. Thus y A a l l y on atom A (yy|xx) = y A B.'{ (15) a l l X on atom B is an average electrostatic repulsion between any electron on A and any electron on B. The next step i s to apply a related series of approximations to the matrix elements H of the core Hamiltonian operator yv E = j V 2 - E V R (16) 2 B B -Vg i s the interaction of any valence electron on atom A with the core of atom B. The diagonal matrix elements H are conveniently separated into yy one- and two-center contributions H = /A* (1) (- \ V 2 - V )d> (1) dx - E **(1)V RA (1) dx = U ~ E VAT* (17) - 162 -U is the one-center term, and is the diagonal matrix element of the yy uth orbital on atom A with respect to the kinetic energy and to the potential energy of the core atom A. The second term gives the electrostatic interaction of an electron in <t> on atom A with the cores of other atoms B. y If the off-diagonal core matrix elements are considered between different atomic orbitals d> and i> on the same atom A, then u v H = U - E <j>*(D V_*.(l) d T l yv yv B ? j A ry B v 1 = U - E (y|VR|v) (18) v v B^A U Is -the one-electron matrix element, and i s zero by symmetry. The yv next term represents the interaction of an overlap charge distribution with the potential of the core of the other atoms. In CNDO, the two-center term i s neglected in the case when (urv)» thus (y|VB|v) = 0 (19) If d> and <f> are on different atoms A and B, the differential u Tv overlap i s not neglected, since these elements take account of the basic bonding capacity of the overlap between the orbitals H = I f * yv J Yy (1) (- y V 2 - V -V )<j> (1) dx - E 6*(1)V4 (1) dx (20) 2 A B v l c ( f j A f B ) y C v 1 - 163 -The second term gives the interaction of the distribution d> d> with y v other cores. These integrals are neglected to be consistent with other approximations. The f i r s t term depends only on the local environment and i s a measure of the possible lowering of energy levels by an electron being in the electrostatic f i e l d of two atoms simultaneously. It i s referred to as a resonance integral (3 v ) • This i s assumed to be proportional to the overlap integral H '" = Pin s yv AB yv (21) yv yv yv the same for a l l atomic orbitals on two given atoms. F = U + ( P . . - 7 P ) Y A A + E [ - Q ^ Y A U + (VfA* - V A T L ) ] C 2 2 ) yy yy AA 2 yy 'AA B AB B'AB AB and 2 yv yfv (23) where Q i s the net charge on atom B. P. BB (24) Z i s the core charge of B. - 164 -The quantity "QgY^g represents the effect of the potential due to total charge on atom B, and ^ y^g-V^g represents the difference between the potentials due to the valence electrons and core of the neutral 138 atom B. This i s referred to as penetration integral. P i s the total valence shell electronic charge on atom A, defined by i P C25) AA ™ V The total energy i s the sum of the total electronic energy of the valence electrons and the repulsion energy between cores tot elec . „ A B AB A<B where E , = y E P (H + F ) (27) elec 2 yv yv uv pv 139 4. The CNDO/2 Parameterization A f u l l specification of a CNDO calculation requires values for the overlap integrals S , the core Hamiltonian elements V , V A T 1, the r uv yu AB electron repulsion integrals y^g a n d the bonding parameters fc^g* '^wo 139 140 versions of CNDO method have been proposed ' and referred to as CNDO/1 and CNDO/2. The CNDO/1 method can be used for atoms up to fluorine, and i t w i l l not be discussed in the following study. The CNDO/2 method i s rather more successful and has been more widely applied. It differs from - 165 -CNDO/1 in the way i t handles penetration integrals and the one-center atomic core integrals. The "penetration" effect in which electrons i n an orbital on one atom penetrate the shell of another leading to a net attraction is treated in CNDO/2 by neglecting the penetration integrals. Thus the electron-core potential integrals V^g are no longer evaluated separately but are related to the electron repulsion integrals by No theoretical j u s t i f i c a t i o n for this neglect of penetration can be given, but i t does appear to compensate errors of the opposite sign introduced by the neglect of overlap integrals. The second change in CNDO/2 concerns the way that the local matrix element U is estimated from atomic data, and given as I and A are the atomic ionization potential and electron a f f i n i t y y y respectively. The Fock matrix in the CNDO/2 method can be written V AB B 'AB - \ (I + A ) = U + (Z. - hy.. 2 p p pp A 2 'AA (28) F = - - k l + A") + [(P.. - Z.) - \ (P - l ) ] y A A + £ " yy 2 V p p AA A 2 pp 'AA B ( J A ) B B (29) and, - 166 -Fuv = *AB Spv " \ Pyv ^ AB ( 3 0 ) The CNDO/2 method is extended to.the second row elements by Santry and Segal. They consider three possible basis sets for a second row atom referred to as sp, spd and spd'. The sp set consists of 3s and 3p functions only and is similar to the calculation on first-row atoms. The spd includes five 3d atomic orbitals, while spd' has d functions more diffuse than in spd. In this case, the bonding parameters f3 ° are approximated by AB *AB = 1/2K( B;+ B;) (31) and Q o _ o 3s,3s 3p,3p B A " * C U 2 s , 2 s ( C ) + U 2 P , 2 p ( C ) ( } K i s constant and equal to 0.75 i f A or B is a second-row element. 141 5. Intermediate Neglect of Differential Overlap (INDO) The CNDO method does not consider the different interactions between two electrons with parallel or antiparallel spins, i f they are on the same atom. Tbis two-electron exchange integral i v | u v ) =JJl ] (u  = //4>y(1)^(2) ^ - 4>v(l)*v(2) dT± dx 2 (33) is neglected, and all~^nteractions between two electrons on atom A are replaced by irrespective of their spin. This results in the failure of CNDO calculations to give the separation of states arising - 167 -from the same electron configuration. 141 The INDO introduced by Pople, Beveridge and Dobosh retains the non atomic differential overlap, but only in one-center integral. The unrestricted F-matrix elements without approximations for the one-center integrals are then A • 1 • F =U + £ [P ( u p|xo) - y P. (yX|ya) + Z (P - Z ) Y yy yy ^ Xa 1 2 Xa 1 B(^A) (34) fi on atom A F a = U + E A [P, (yvlXa) - \ P, (yXlva)] y^v both on A yv yv , Xa 1 2 Xa 1 Xa (35) F ° , = T(6:+e:)S - \ P„ Y A R y on atom A, (36) yv 2 A B yv 2 yv'AB v on atom B In general the INDO and CNDO/2 methods are closely related, for the basic approximations are the same except for monoatomic terms. 6. The Energy Expression for a Closed-Chell Configuration In order to evaluate the energy expectation value < \p | H | ^ > where H is the total nonrelativistic Hamiltonian of an atom with 2 N electrons and given as H= H 1 + H 2 (37) - 168 -where H, = E H C ° r e 1(y) = - I V 2 - E Z A r ' l ,-r. 1 2 y A A yA (38) and H = H — (39) 2 r y<v yv core The quantity H i s the one-electron Hamiltonian corresponding to the motion of an electron in the f i e l d of the bare nuclei. The charge of nucleus A is Z,. r is the distance between two electrons. A yv ip i s the orbital wavefunction for a close-shell system and given as * - tr Z ( - l ) r P { ^ (1)0(1)^.(2)8(2)....* (2n)B(2n)} (40) P p where P i s a permutation of l,2....2n, (-1) is +1 or -1 for even or odd permutations. The energy expectation value can be written as <*|H|*> Hx|t|i> + <tp|H2|ir> (41) The one-electron term i s given as <*| HXU> = t ( 2 n - l ) l ] - : L x E /.../ P{^ 1(l)o(l)4' 1(2 ) e(2)...}H^ 0 r e x - r P { i p 1(l)a(l ) i P 1(2 ) e(2)...}dT 1 dT 2...dx 2 n (42) where the orbitals ii>. are assumed to be orthonormal. I - 169 -This yields to n <ip| H 1 iip> =2 E H (43) 1=1 which H.. is the expectation value of the one-electron core Hamiltonian corresponding to the molecular orbital H i i = ^ i ( 1 ) * * i ( 1 ) d T l ( 4 4 ) If electrons 1 and 2 are assigned to different spatial molecular orbitals I|K and ijj^  both may have a or g spin, and there w i l l be four contributions each equal to 1/2 J J i i = / / * i * C l ) * 1 * C 2 ) r * i ^ l H 1 C 2 ) dx, d x 2 (45) , yv J where J.. is the coulomb integral. If electrons 1 and 2 are assigned to the same molecular orbtial ty, they must have opposite spins and there are only two terms 1/2J^. Therefore the total contribution i s 2 E E J . . + I J . . • f I » X 1 ] . XX x j(^x) J X The exchange integral i s given as K i i = / ; V ( 1 ) V ( 2 ) T~ * i c i ) * i C 2 ) d T i d T2 C 4 6 ) J yv J - 170 -by considering that K. ^  = the total energy can be given as n n n E = 2 E H . . + E E (2J.. - K..) (47) 11 I J I J i i 3 The physical significance which can be given to the equation (47) i s 1) The one-electron integral K L ^ represents the energy of an electron in a molecular orbital in the f i e l d of the bare nuclei, and this multiplied by 2 since there are two electrons i n each MO. 2) The two-electron integral J^_. represents the energy due to electrostatic repulsion between a pair of electrons having charge 2 2 distributions (ip^(l)) and (IJJ^( I ) ) respectively. 3) The negative value of the exchange integral i s to reduce the energy of .in.terac.tipn,between electrons with parallel spins in different orbitals ib-. and i l ; . . The one-electron orbital energies is defined as n e. = H.. + E (2J.. - K..) (48) i n j 3 1 3 by using orbital energies, the total electronic energy can be written as n n n E = 2 E e. - E E (2J. . - K. .) (49) . x . . i i x i i i J n E = E (e. + H..) (50) . l i i l I n the case of removal of one electron k, therefore we are dealing with (2n-l) electron system, where the total energy E + i s given as - 171 -E + = E ( K ) (2n-l) = E (e. + H ) - e K (51) i and the energy of the electron e i s given by the relation E + - E = - e K (52) If an electron (m) is added to the 2n electron system, then the energy of the (2n+l) electron system is given as n E = Em(2n+1) = E (e. + H ) + e (53) l i i m i and the energy of the added electron i s given as E" - E = e ( m ) (54) By assuming that there i s no electronic reorganization of the remaining (2n-l) electrons upon the removal of one electron k, then e can be given a physical significance and referred to as the v e r t i c a l ionization potential, and electron a f f i n i t y for e^. This i s referred to as Koopmans' theorem. 144 Richards discussed carefully the use of Koopmans theorem and the error which may result from the approximations used i n the calculated IP's, this w i l l be reflected in the interpretation of the PE spectra. The approximations used i n eq. (52) can be resumed as: - 172 -1) It presumes that when the photoelectron is ejected, the orbitals and energies of a l l the remaining electrons are unchanged (frozen orbitals). In fact, the removal of an electron alters the potential acting on the remaining electrons, so that the..ion may attain a lower energy i f the orbitals are reorganized, leading to IP's lower than those given by equation (52). These are higher than the experimental values by almost 3 eV in some cases. 2) Relativistic effects are neglected. These are very important for inner electrons which have very large kinetic energies. R e l a t i v i s i t i c effects are very much smaller for valence shell electrons, and i t may be assumed that the r e l a t i v i s t i c energy i s the same i n the ion and the parent molecule. 3) The electron correlation energy i s the same for the molecule and the ion. In practice the correlation energy i s different i n the ion and the neutral molecule. This results i n calculated IP's lower than the experimental values. 7. The Use of CNDO/2, INDO and ab. i n i t i o i n Photoelectron Spectroscopy In most of the polyatomic molecules studied in this work, the CNDO/2 and INDO calculations are used to assist the interpretation of the spectra. The agreement between the calculated and experimental IP's i s generally good. Therefore one can depend on these calculations to assign the MO's of the molecules under investigation. In the case of the second-row elements, the calculated IP's by CNDO/2 are lower than the experimental IP's by one to two electron volts for the f i r s t three IP's, and higher by larger amounts for IP's above - 173 -the energy of 18 eV. The CNDO/2 calculations also overestimate the interaction between orbitals of the same symmetry,^7 this leads to misinterpretation of the PE spectra. However, the deduction of 4 eV from the calculated IP's in order to compare them with the experimental values i s large approximation, therefore much care should be paid in considering the calculated IP's for the assignment of the PE spectra. 145 Ab. i n i t i o calculations are used for small polyatomic molecules, because thecomputing time i s much longer for these calculations than for CNDO/2 and INDO methods. The calculated IP's agree f a i r l y well with the experimental values. In general, the PE spectra can be used to test the valid i t y of the calculations, and not vice versa. - 174 -BIBLIOGRAPHY 1. M.I.Al-Joboury and D.W. Turner, J. Chem. Phys., 37» 3 0 0 7 (1962). 2. M.I. Al-Joboury and D.W. Turner, J. Chem. Soc, 5141 (1963). 3. M.I. Al-Joboury and D.W. Turner, J. Chem. Soc, 4434 (1964). 4. F.I. Vilesov, B.L. Kurbatov and A.N. Terenin, Dokl. Akad. Nauk. SSSR, 138, 1329 (1961); Sov. phys. Dokl., 6, 490 (1961). 5. B.L. Kurbatov, F.I. Vilesov, and A.N. Terenin, Dokl. Akad. Nauk. SSSR, 140, 797 (1961); Sov. phys. Dokl., 6_, 883 (1962). 6. L. Asbrink, Chem. Phys. Letters, 7_, 549 (1970). O 7. L. Asbrink and J. Wayne Rabalais, Chem. Phys. Letters, 12, 182 (1971), 8. CR. Brundle, M.B. Robin and G.R. Jones, J. Chem. Phys. ,52, 3383 (1970). 9. A.W. "Potts, H.J. Lempka, D.G."Streets and W.C. Price, Phil. Trans, Roy. Soc. Lond. A. 268, 59 (1970). 10. G.R. Branton, D.C. Frost, T. Makita, CA. McDowell, and I.A. Seenhouse, J. Chem. Phys., 52, 802 (1970). 11. J. Delwiche, P. Natalis, and J.E. Collin , INter. J. Mass Spect. ion. Pbys., 5_, 443 (1970). 12. J.E. Collin and P. Natalis, INter. J. Mass Spect. ion. phys., 1, 483 (1968). 13. P. Natalis and J.E. Collin, Int. J. Mass Spect. ion phys., 2_, 221 (1969). 14. W.C. Price (private communication). 15. D.W. Turner, A.D. Baker, C. Baker, and CR. Brundle, High  Resolution Molecular Photoelectron Spectroscopy, John Wiley and Sons, Ltd., 1970. - 175 -16. C E . Brion "Recent Advances in Electron Spectroscopy" to be published i n MTP International Review of Science, 1971. 17. A.D. Baker and CR. Brundle, "Ionizing Photon Impact Phenomena" Academic Press, Inc., i n preparation. 18. CR. Brundle and M.B. Robin, "Photoelectron Spectroscopy", _Bell Telephone Laboratories, Inc., Murray H i l l , New Jersey, 07974. Determination Org. Struct. Phys. Methods, in press. 19. D.W. Turner, Advan. Phys. Org. Chem., 4_, 31 (1966). 20. D.W. Turner, Adv. Mass Spectrom., ^ , 755 (1968). 21. D.W. Turner, Phys. Methods Advan. Inorg. Chem., 74 (1968). 22. D.W. Turner, "Molecular Spectroscopy", p. Hepple, Ed., Institute of Petroleum, London, 1968, p. 209. 23. D.W. Turner, Chem. Brit., _4, 435 (1968). 24. R.S. Berry, Ann. Rev. Phys. Chem., 20.. 357 (1969). 25. D. Betteridge and A.D.Baker, Anal. Chem., 42, 43A (1970). 26. A.D. Baker, Accounts Chem. Res., J3, 17 (1970). 27. Electron Spectroscopy for Chemical Analysis, K. Siegbahn, C. Nordling, A. Fahlman, R. Nordberg, K. Hamrin, J. Hedman, G. Johansson, T. Bergmark, S. Karlsson, I. Lindgren, and B. Lindgren, Nova Acta Regia Socientatis Scientiorum Upplaliensis ser. IV, Vol. 20 (1967). 28. ESCA Applied to Free Molecules, K. Siegbahn, C. Nordling, G. Johansson, J. Hedman, P.F. Heden, K. Hamrin, U. Gelius, A.T. Bergmark, L.O. Werme, R. Manne, and Y. Baer, North Holland Publishing Co., Amsterdam, 1970. 29. D.M. Hercules, Anal. Chem., 42, 20 (1970). - 176 -30. J.W. Rabalais, T. Bergmark, L.O. Werme, L. Karlsson and K. Siegbahn, Physica Scripta. , J3, 13 (1971). 31. G.R. Branton, D.C. Frost, T. Makita, CA. McDowell and I.A. Stenhouse, P h i l . Trans. Roy. Soc. Lond. A., 268, 77 (1970). 32. D.W. Turner, Phil. Trans. Roy. Soc. Lond. A., 268, 7 (1970). ' 33. W.C Price, "Developments ih Photoelectron Spectroscopy", in Molecular Spectroscopy, London: Institute of Petroleum, 221 (1968). 34. J.E. Collin and P. Natalis, Int. J. Mass Spect. ion. Phys., 1_, 121 (1968). 35. A.J. Blake, J.L. Bahr, J.H. Carver and Vijay Kumar, Ph i l . Trans. Roy. Soc. Lond.„A., 268, 159 (1970). 36. A.L. Smith, Phil. Trans. Roy. Soc. Lond. A. 268, 169 (1970). 37. H.A. Jahn and E. Teller, Proc. Roy. Soc. 161 A, 220 (1937). 38. A.D. Liehr, J. Phys. Chem.,67, 389 (1963). 39. R.N. Dixon, Molec. Phys., 20, 113 (1971). 40. L. Asbrink, E. Lindholm and 0. Edqvist, Chem. Phys. Letters, _5_, 609 (1970). 41. C.G. Rowland, Chem. Phys. Letters, 9_, 169 (1971). 42. B.P. Pullen, Thomas A. Carlson, W.E. Moddeman, G.K. Schweitzer, W.E. Bull, and F.A. Grimm, J. Chem. Phys., _53, 768 (1970). 43. D.C. Frost, F.G. Herring, A. Katrib and CA. McDowell (to be published). 44. A.B. Cornford, D.C. Frost, CA. McDowell, J.L. Ragle and I.A. Stenhouse, J. Chem. Phys., _54, 2651 (1971). 45. J.N. Murrell, S.F.A. Kettle and J.M. Tedder, "Valence Theory", Chapter 9, John Wiley and Sons Ltd., London (1965). - 177 -46. F. Brogli and E. Heilbronner, Helv. Chim. Acta., 54, 1423 (1971). 47. W.C. Price, J. Chem. Phys., 4_, 539 (1936). 48. E. Heilbronner, V. Hornung and Else Kloster-Jensen, Helv., 53, 331 (1970). 49. J.L. Ragle, I.A. Stenhouse, D.C. Frost and CA. McDowell, J. Chem. Phys., 53, 178 (1970). 50. J.C. Tully, R. Stephan Berry and Bryan J. Dalton, Phys. Rev., 176, 95 (1968). 51. S.T. Manson and D.J. Kennedy, Chem. Phys. Letters, _7» 387 (1970). 52. F.H. Read, J. Phys. B (Proc. Phys. S o c ) , 1, 893 (1968). 53. D.A.Vroom, A.R. Comeaux and J. William McGowan, Chem. Phys. Letters, _3, 476 (1969). 54. H. Harrison, J. Chem. Phys., 52^ , 901 (1970). 55. J. Berkowitz and H. Enrhardt, Physics Letters, 21_, 531 (1966). 56. R. Morgenstern, A. Niehaus and M.W. Ruf, Chem. Phys. Letters., 4_, 635 (1970). 57. T.A. Carlson, Chem. Phys. Letters, 9_, 23 (1971). 58. T.A. Carlson and A.E. Jonas, J. Chem. Phys., _5_5, 4913 (1971). 59. D.C. Manson, A. Kuppermann and D.M. Mintz, "International Conference of Electron Spectroscopy", California (1971). 60. CR. Brundle and D.W. Turner, Proc. Roy. Soc. , 307, 26 (1968). 61. CR. Brundle and N.A. Kuebler, The Construction and Operation of a  High Resolution Photoelectron Spectrometer, B e l l Telephone Laboratories, Murray H i l l , New Jersey, Private communication. 62. D.C. Frost and CA. McDowell, USAF Report 1960, f i n a l report AF 19(604)-2275, Project 7635 (1960). - 178 -63. B. Brehm and F. Rudiger, Z. Naturforsch., 26, 523 (1971). 64. D.W. Turner, Proc. Roy. Soc. (London) A, j30_7, 15 (1968). 65. C E . Kuyatt and J. Arol Simpson, Rev. Sci. Instr., 38, 103 (1967). 66. K. Watanabe, T. Nakayama and J. Mottl, J. Quant. Spectro. Radiat. Transfer, 2, 369 (1962). 67. G. Herzberg "Electronic Spectra of Polyatomic Molecules", Van Nostrand, Princeton, 1967, p. 586. 68. G. Herzberg, "Infrared and Raman Spectra", Van Nostrand, Princeton 1966, p. 282. 69. D.C. Frost, A. Katrib, CA. McDowell and R.AtfN. McLean, Int. J. Mass Spectrom. ion. Phys., 7_, 485 (1971). 70. - R.N. Dixon, Trans. Faraday Soc., 60, 1363 (1964). 71. (a) F. Fiquet-Fayard and P.M. Guyon, Mol. Phys.,11, 17 (1966); (b) V.R. Dibeler and S.K. Liston, J. Chem. Phys. , 49_, 482 (1968); 72. W.R. Cullen, D.C. Frost and D.A. Vroom, Inorg. Chem., 8, 1803 (1969). 73. W.E.W. Ruska and J.L. Franklin, Int. J. Mass Sptrom. ion Phys., 3, 221 (1969). 74. N. Sheppard, J. Chem. Phys., JL7, 79 (1949). 75. D.W. Scott, J.P. McCullough, W.N. Hubbard, J.F. Messerly, I.A. Hossenlopp, F.R. Frow and G. Waddington, J. Amer. Chem. Soc.,78, 5463 (1956). 76. D.C. Frost, F.G. Herring, A. Katrib, CA. McDowell and R.A.N. McLean, J. Phys. Chem., (In press). 77. H. Basch, M.B. Robin, N.A. Kuebler, C. Baker and D.W. Turner, J. Chem. Phys., 51, 52 (1969). 78. T.K. Liu and A.B.F. Duncan, J. Chem. Phys.,17, 241 (1949). - 179 -79. A. Lowry and K. Watanabe, J. Chem. Phys., 28, 208 (1958). 80. N. Basco and R.D. Morse, Private communication (to be published). 81. A.B.F. Duncan, "Rydberg Series i n Atoms and Molecules", Academic Press, New York (1971). 82. D.W.J. Cruickshank and B.C. Webster i n "Inorganic Sulphur Chemistry" Ed. G. Nickless, Elsevier (1968). 83. H.W. Thompson and W.T. Cave, Trans. Farad. Soc. , 47, 951 (1951). 84. G.L. Cunningham, A.W. Boyd, R.J. Myers and W.D. Gwinn, J. Chem. Phys., 19_, 676 (1951). 85. E. Hirota, Bull. Chem. Soc. Japan, 31, 130 (1958). 86. R.F. Lake and H.W. Thompson, Proc. Roy. Soc. Lond. A., 314, 323 (1970). 87. H. Yamabe, H. Kato, and T. Yonezawa, Bull. Chem. Soc. Japan, 44, 604 (1971). 88. D.W.J. Cruikshank, J. Chem. Soc, 5486 (1961). 89. L.S. Bartell, L.S. Su and H. Yow, Inorg. Chem., 9_, 1903 (1970). 90. l.H. H i l l i e r and V.R. Saunders, Inter. J. Quant. Chem., IV, 203 (1970). 91. K.A.R. Mitchell, Chem. Rev., 69_, 157 (1969). 92. R.J. Gillespie and E.A. Robinson, Can. J. Chem., 41, 2074 (1963). 93. R.J. Gillespie and E.A. Robinson, Can. J. Chem.,39, 2171 (1961). 94. J.H.D. Eland and C.J. Danby, Int. J. Mass Spectrom. ion phys., 1^, 111 (1968). 95. M. Jonathan, D.J. Smith and K.J. Ross., Chem. Phys. Letters, 9_ 217 (1971). 96. H. Lynton and M.R. Trotter, J. Chem. Soc, 5112 (1960). 97. R.M. Fristrom, J. Chem. Phys., 20, 1 (1952). - 180 -98. (a) D.R. Lide, D.E. Mann and R.M. Fristrom, J. Chem. Phys., 26, 736 (1957); (b) D.R. Lide, D.E. Manne and J.J. Comeford, Spect. Acta, 21, 497 (1965). 99. W. Moffitt, Proc. Roy. Soc. Lond. A., 200, 409 (1949). 100. K.J. Palmer, J. Amer. Chem. Soc., 60, 2360 (1938). 101. D. Kivelson, J. Chem. Phys., 22, 904 (1954). 102. S. Rothenburg and H.F. Schaefer, J. Chem. Phys.,53, 3014 (1970). 103. A.B. Cornford, D. Chadwick, D.C. Frost, F.G. Herring, A. Katrib, CA. McDowell and R.A.N. McLean, "International Conference on  Electron Spectroscopy", California (1971). 104. R.C. Ferguson, J. Amer. Chem. Soc., 7_6_, 850 (1956). 105. A.J. Merer and R.S. Mulliken, Chem. Revs., 69, 639 (1969). 106. N. Jonathan, K. Ross and V. Tomilinson, Int. J. Mass Spectrom. ion phys., 1, 111 (1968). 107. M. Klasson and R. Manne, "International Conference on Electron  Spectroscopy" California (1971). 108. D.C. Frost, F.G. Herring, A. Katrib and CA. McDowell (to be published) . 109. G.J. Martin and M.L. Martin, J. Chim. Phys., 61, 1222 (1964). 110. P. Torkington, J. Chem. Phys.,17, 1279 (1949). 111. R. Bralsford, P.V. Harris and W.C. Price, Proc. Roy. Soc, 459 (1960). 112. E.A.V. Ebsworth in "Organometalic Compounds of the Group IV  Elements" Ed. A.G. McDiarmied, Marcel Dekker, New York (1968), p. 1. 113. S. Cradock and E.A.V. Ebsworth, Chem. Commun., 57 (1971). 114. CR. Brundle, M.B. Robin and H. Basch, J. Chem. Phys., 53, 2196 (1970). 181 -115. R.P. Hollandsworth, W.M. Ingle and M.A. Ring, Inorg. Chem., 6_, 844 (1967). 116. D.C. Frost, F.G. Herring, A. Katrib, R.A.N. McLean, J.E. Drake and N.P. Westwood, Chem. Phys. Letters, 10, 347 (1971). 117. D.C. Frost, F.G. Herring, A. Katrib, R.A.N. McLean, J.E. Drake and N.P.C. Westwood, Can. J. Chem.,49, 4033 (1971). 118. R.H. Herber, Inorganic Isotopic Syntheses, Benjamin, New York (1962), p. 31. 119. H.J. Emeleus and A.G. Maddock, J. Chem. Soc, 293 (1944). 120. J.E. Drake and J. Simpson, Inorg. Nucl. Chem. Letters, _2> 2 i 9 (1966), 121. G.J. Janz and Y. Mikawa, Bull. Chem. Soc Japan, 34_, 1495 (1961). 122. E.A.V. Ebsworth and J.J. Turner, J. Chem. Phys. , 36, 2628 (1962). 123,. E.A.V. Ebsworth and J.J. Turner, J. Phys. Chem. , 67, 805 (1963). 124. A.H. Sharbaugh, Phys. Rev., 78, 64 (1950). 125. V.W. Laurie, J. Chem. Phys., 26, 1359 (1957). 126. A.H. Sharbaugh, Phys. Rev., 74, 1870 (1948). 127. L.O. Brockway and I.E. Coop, Trans. Farad. Soc. , 34^ , 1429 (1938). 128. R.C. Mockler, J.H. Bailey and W. Gordy, J. Chem. Phys., 2JL, 1710 (1953). 129. S. Cradock, J. Chem. Phys., _5_5, 980 (1971). 130. J.C. Green, M.L.H. Green, P.J. Joachim, A.F. Orchard and D.W. Turner, P h i l . Trans. Roy. Soc. Lond. A, 268, 111 (1970). 131. C.C.J. Roothaan, Rev. Mod. Phys., 23, 69 (1951). 132. D.R. Hartree, Proc. Cambridge Phil. Soc, 24_, 89 (1928). 133. V. Fock, Z. Physik., 61, 126 (1930). 134. R.K. Nesbet in Advances in Quantum Chemistry, V. 3 (P.O. Lowdin, ed.) Academic Press, New York (1967). - 182 -135. A.D. McLean and M. Yoshimine, Tables of Linear Molecule Wave  Functions, International Business Machines Corporation, San Jose, California (1967). 136. John A. Pople and David L. Beveridge, Approximate Molecular  Orbital Theory, McGraw-Hill Book Company, New York (1970), Ch. 3. 137. J.A. Pople, D.P. Santry and G.A. Segal, J. Chem. Phys.,43, 5129 (1965). 138. M. Goeppert-Mayer and A.L. Sklar, J. Chem. Phys., 6^, 645 (1938). 139. J.A. Pople and G.A. Segal, J. Chem. Phys., 44, 3289 (1966). 140. J.A. Pople and G.A. Segal, J. Chem. Phys., 43, 5136 (1965). 141. J.A. Pople, D.L. Beverdige and P.A. Dobosh, J. Chem. Phys.,47, 2026 (1967). 142. Frank L. Pil a r "Elementary Quantum Chemistry", McGraw-Hill Book Company, New York (1968), Ch. 13. 143. T. Koopmans, Physica., 1, 105 (1933). 144. W.G. Richards, J. Int. J. Mass Spectrom. ion phys., 2, 419 (1969). 145. W.G. Richards and J.A. Horsley "Ab. Initio Molecular Orbital Calculations for Chemists", Oxford Science Research Papers. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items