INNER SHELL AND VALENCE SHELL ELECTRON EXCITATION OF GASEOUS MOLECULES STUDIED BY ELECTRON ENERGY LOSS SPECTROSCOPY by RANA N.S. SODHI B.Sc. (Hon.), University of Reading, 1975 M.Sc, University of Alberta, 1980 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES Department of Chemistry We accept t h i s thesis as conforming to the required standard The University of B r i t i s h Columbia August, 1984 © Rana N.S. Sodhi, 1984 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e 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 r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e head o f my d e p a r t m e n t o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g 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 g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . D e p a r t m e n t o f C H t n i S T R V The U n i v e r s i t y o f B r i t i s h C o l u m b i a 1956 Main M a l l V a n c o u v e r , Canada V6T 1Y3 Date \% Oc-To$£R. H'Sii-E - 6 (3/81) Abstract Electron energy loss spectroscopy has been used to obtain the inner s h e l l electron e x c i t a t i o n spectra of several d i f f e r e n t series of gaseous molecules. The spectra were a l l recorded under small momentum transfer conditions (usually 2.5 keV impact energy and small angle (~1°) scattering) and a l l spectral regions (both c e n t r a l atom and ligand) accessible by the present instrumentation (<1000 eV) have been measured. The series of molecules investigated include nitrogen containing molecules (NF 3, NH3 and the methyl amines) and several phosphorus compounds (PX 3, X = H, F, C l and CH 3; PF 5, OPF 3 and 0PC1 3). In addition the spectra of SiCCHj)^ have been obtained and compared with published spectra of related Si containing compounds. A l l of the inner s h e l l spectra show continuum structures which i n many cases can be reasonably * assigned to a shape-resonances. However, comparison of the inner s h e l l electron e x c i t a t i o n spectra of NF 3 with the X-ray photoelectron spectra (also reported here) show that continuum structure can also be ascribed, i n some cases, to onsets of "shake-up" continua. The r e l a t i o n s h i p of shape-resonance p o s i t i o n and bond length i s also examined i n the systems studied here. The valence s h e l l electron energy loss spectra of many of the above molecules are also reported. The assignment of these spectra i s shown to be greatly f a c i l i t a t e d by a comparison with the inner s h e l l spectra. - i i i -F i n a l l y , the inner s h e l l and valence s h e l l electron energy loss spectra of trans-1,3-butadiene and allene are also reported and assigned. In p a r t i c u l a r , the spectral assignment of the inner s h e l l spectrum of allene allows c l a r i f i c a t i o n of i t s complex and c o n t r o v e r s i a l valence s h e l l spectrum. - i v -Table of Contents Page Chapter 1 Introduction 1 A. General Introduction 1 B. Description of Various Processes i n Electron Spectroscopy 4 I) Electron E x c i t a t i o n 4 II) Photoionisation 7 III) "Shake-up" and "shake-off" 12 IV) X-ray Fluorescence, Auger Decay and Autoionisation 14 C. Fundamental concepts i n Electron Impact and the Relationship to Photoabsorption 18 D. The Relative Merits of Electron Energy Loss and Photoabsorption Spectroscopies 27 E. Inner S h e l l Electron E x c i t a t i o n Spectra 33 I) Discrete Portion 33 II) Continuum Features 38 III) Comparisons of Inner She l l E x c i t a t i o n Spectra with Valence S h e l l E x c i t a t i o n Spectra 39 F. Pote n t i a l Barrier and Shape-Resonance E f f e c t s I) P o t e n t i a l Barriers 41 II) Shape-Resonances 44 III) Relationship of Shape-Resonance Pos i t i o n with Bond-lengths 52 Chapter 2 Experimental 55 A. Experimental Methods 55 I) The Spectrometer 55 II) Sample Handling 61 III) Spectral A c q u i s i t i o n , C a l i b r a t i o n and Spectrometer Performance 64 IV) Other Measurements 67 B. Reference Energies for Inner She l l Electron Energy Loss Spectroscopy 69 - V -Page Chapter 3 Inner S h e l l E x c i t a t i o n , Valence E x c i t a t i o n and Core Ionisation i n NF 3 studied by Electron Energy Loss and X-ray Photoelectron Spectroscopies 83 Chapter 4 Electron Energy Loss Spectra of the S i l i c o n 2p, 2s, Carbon Is and Valence Shells of Tetramethylsilane 120 Chapter 5 E l e c t r o n i c Excitations i n Phosphorus containing Molecules. I. Inner Sh e l l Electron Energy Loss Spectra of PH 3, P(CH 3) 3, PF 3 and PC1 3 150 Chapter 6 E l e c t r o n i c Excitations i n Phosphorus containing Molecules. I I . Inner Sh e l l Electron Energy Loss Spectra of PF 5, 0PF 3 and 0PC1 3 192 Chapter 7 E l e c t r o n i c Excitations i n Phosphorus containing Molecules. I I I . Valence Sh e l l Electron Energy Loss Spectra of P(CH 3) 3, PC1 3, PF 3, 0PC1 3 and PF 5 225 * Chapter 8 Inner S h e l l Electron Energy Loss Spectra of the Methyl Amines and Ammonia 255 Chapter 9 High Resolution Carbon Is and Valence S h e l l E l e c t r o n i c E x c i t a t i o n Spectra of Allene and Trans-1,3-Butadiene studied by Electron Energy Loss Spectroscopy 275 Chapter 10 Concluding Remarks 304 References 306 - v i -- v i i -Table Description Page 5.1 Transitions from the *A^ ground state for C^ v symmetry 154 5.2 Energies, term values and possible assignments for the P 2p, 2s spectra of PH 3 159 5.3 Energies, term values and possible assignments for the P 2p, 2s spectra of P ( C H 3 ) 3 162 5.4 Energies, term values and possible assignments for the P 2p, 2s spectra of PF 3 165 5.5 Energies, term values and possible assignments for the P 2p, 2s spectra of PC1 3 169 5.6 Transitions from the *Aj^ ground state f o r D-j^ symmetry 176 5.7 Resonance term values and (P-X) bond lengths 181 5.8 Energies, term values and possible assignments f o r the F Is spectrum of PF 3 184 5.9 Energies, term values and possible assignments for the C Is spectrum of P(CH 3) 3 187 5.10 Energies, term values and possible assignments for the Cl 2p,2s spectra of PC1 3 190 6.1 Transitions from the ^A^ ground state i n D^ n symmetry 196 6.2 Energies, term values and possible assignments f o r the P 2p,2s spectra of PF 5 199 6.3 Energies, term values and possible assignments for the P 2p,2s spectra of 0PF 3 203 6.4 Energies, term values and possible assignments for the P 2p.2s spectra of 0PC1 3 206 6.5 Resonance positions above the mean i o n i s a t i o n edge and bond lengths 209 - v i i i -Table Description Page 6.6 Energies, term values and possible assignments f o r the F Is region of PF 5 212 6.7 Energies, term values and possible assignments for the 0 Is, F Is regions of OPF3 217 6.8 Energies, term values and possible assignments for the 0 Is region of 0PC1 3 219 6.9 Energies, term values and possible assignments f or the Cl 2p,2s regions of 0PC1 3 222 7.1 Term values for phosphorus L - s h e l l spectra and the calculated s o r b i t a l quantum defect 228 7.2 Molecular o r b i t a l s and experimental i o n i s a t i o n potentials for the valence o r b i t a l s of PH 3, PF3, PC1 3, P(CH 3) 3, 0PC1 3 and PF 3 230 7.3 Energies, term values and possible assignments f or the VSEELS spectrum of P(CH 3) 3 232 7.4 Energies and possible assignments for the VSEELS spectrum of PC1 3 236 7.5 Energies, term values and possible assignments for the VSEELS spectrum of PF 3 240 7.6 Energies and term values for the VSEELS spectrum of 0PC1 3 246 7.7 Dipole allowed/forbidden t r a n s i t i o n s for PF 5 In ^3h -y- n- e- ry -48 7.8 Energies and possible assignments for the VSEELS spectrum of PF 5 250 7.9 Term values from ISEELS and VSEELS for PH 3, P(CH 3) 3, PC1 3, PF 3, PF 5 and 0PC1 3 252 8.1 Energies, term values and possible assignments for the C Is region of the methyl amines 260 8.2 Energies, term values and assignment for the N Is energy loss spectrum of NH3 265 - ix -Description Energies, term values and possible assignments f o r the N Is region of the methyl amines Resonance energy positions from edge and C-N bond lengths for the methyl amines Energies, term values and possible assignments for the C Is energy loss spectrum of butadiene Energies, term values and possible assignments for the C Is energy loss spectrum of allene Energies of the features i n the valence s h e l l electron energy loss spectrum of butadiene Estimated t r a n s i t i o n energies from the valence o r b i t a l s of butadiene assuming constant term values for the valence s h e l l Energies, term values and various assignments f o r the valence electron e x c i t a t i o n spectrum of allene up to the 1st IP Energies and term values for the features above the 1st i o n i s a t i o n p o t e n t i a l i n the valence s h e l l electron energy loss spectrum of allene - X -- x i -Figure Description Page 4.1 Wide range inner s h e l l electron energy loss spectrum of Si(CH 3),, 123 4.2 S i 2p electron energy loss spectra of S K C H ^ ^ 125 4.3 S i 2p e x c i t a t i o n spectra of various s i l i c o n containing compounds with SI i n a tetrahedral environment 129 4.4 Relationship of bond length to shape-resonance term values for s i l i c o n containing compounds 137 4.5 C Is electron energy loss spectrum of SKCH^i^ 139 4.6 Valence s h e l l electron energy loss spectrum of SiCCHg)^ 143 5.1 P 2p,2s wide range electron energy loss spectra of PF 3, PCI 3, PH 3 and P ( C H 3 ) 3 152 5.2 P 2p and 2s electron energy loss spectra of PH 3 158 5.3 P 2p and 2s electron energy loss spectra of P(CHi) 3 161 5.4 P 2p and 2s electron energy loss spectra of PF 3 164 5.5 P 2p and 2s electron energy loss spectra of PC1 3 168 5.6 Inner s h e l l electron energy loss spectra of PC1 3 and PF 3 at various scattering angles 172 5.7 -Plot of the r a t i o (peak height of feature X)/(peak height of feature 2) for various t r a n s i t i o n s (X) i n the PC1 3 spectra of F i g . 5.6 as a function of (momentum t r a n s f e r ) 2 173 5.8 Expanded plot of the PH 3 P 2p ISEELS continuum structure. The 2p s a t e l l i t e structure from XPS measurements [175] i s shown below plotted on the same r e l a t i v e energy scale references to the 2p (mean) edge 178 5.9 Long-range and d e t a i l e d inner s h e l l e l e c t r o n energy loss spectra of the F Is region of PF 3 183 - x i i -Figure Description Page 5.10 Long-range and detai l e d inner s h e l l e lectron energy loss spectra of the C Is region of P(CH 3) 3 186 5.11 Long-range and detai l e d electron energy loss spectra of the C l 2p and 2s regions of PC1 3 189 6.1 P 2p,2s wide range electron energy loss spectra of PF 3, 0PF 3 and 0PC1 3 194 6.2 P 2p and 2s electron energy loss spectra of PF 5 198 6.3 P 2p and 2s electron energy loss spectra of 0PF 3 201 6.4 P 2p and 2s electron energy loss spectra of 0PC1 3 205 6.5 F Is electron energy loss spectrum of PF 5 211 6.6 Wide range electron energy loss spectra of the 0 Is and F Is regions of 0PF 3 215 6.7 Detailed electron energy loss spectra of the 0 Is and F Is regions of 0PF 3 216 6.8 Electron energy loss spectrum of the 0 Is region of 0PC1 3 218 6.9 Electron energy loss spectrum of the C l 2p,2s region of 0PC1 3 221 7.1 Valence s h e l l electron energy loss spectrum of P( C H 3 ) 3 231 7.2 Valence s h e l l electron energy loss spectrum of PCI 3 235 7.3 Valence s h e l l electron energy loss spectrum of PF 3 239 7.4 Valence s h e l l electron energy loss spectrum of 0PC1 3 243 7.5 Valence s h e l l electron energy loss spectrum of PFc 247 - x i i i -Figure Description Page 8.1 Long range electron energy loss spectra of the C Is region of the methyl amines 258 8.2 Short range, high re s o l u t i o n electron energy loss spectra of the C Is region of the methyl amines 259 8.3 Long range electron energy loss spectra of the N Is region of the methyl amines 262 8.4 Short range electron energy loss spectra of the N Is region of ammonia and the methyl amines 263 9.1 Inner s h e l l electron energy loss spectra of butadiene 278 9.2 Inner s h e l l electron energy loss spectra of allene 284 9.3 Valence s h e l l electron energy loss spectra of butadiene 290 9.4 Valence s h e l l e l ectron energy loss spectra of allene 295 - x iv -Acknowledgement s I wish to express my sincere thanks to my research supervisor, Dr. C.E. Brion. I t has been a pleasure to have worked with him and h i s support, d i r e c t i o n and assistance w i l l always be greatly appreciated. Thanks are also due to the various members of Dr. Brion's research group for contributing to an enjoyable working environment. Special thanks go to Dr. Suzannah Daviel for recording the valence s h e l l spectra of NF 3 and SiCCHg)^ on the new spectrometer and to Tong Leung for some assistance with computing. Dr. D.P. Chong, Dr. M.C.L. Gerry and Dr. A.J. Merer of th i s department are thanked for h e l p f u l discussions, as are Dr. A.P. Hitchcock (McMaster) and Dr. P.L. Langhoff (Indiana). Dr. R.G. Cavell and Anna-Marie Venezia-Floriano (Alberta) are thanked for supplying a number of i o n i s a t i o n potentials of the molecules presented here. Much appreciation i s due to T i l l y Schreinders for the typing of th i s t h e s i s . The s t a f f In the departmental workshops must also be thanked for th e i r capable assistance i n the maintenance of the spectrometer. F i n a l l y I wish to express my great appreciation to Marilyn Daniels for a l l her patience and encouragement. This thesis i s dedicated to her. - 1 -CHAPTER 1 INTRODUCTION A. General Introduction In the l a s t twenty years there has been a growing i n t e r e s t i n the use of electron impact spectroscopy to probe e l e c t r o n i c e x c i t a -tions i n atomic and molecular systems. Electron energy loss spectros-copy (EELS) [1-3] Is a well established technique for the study of e l e c t r o n i c t r a n s i t i o n s In the valence region. More recently [4] the technique has been extended to the study of inner s h e l l (core) electron t r a n s i t i o n s . In the past decade inner s h e l l electron energy loss spectroscopy (ISEELS) has produced much new and Interesting spectroscopic data [5-9] at resolutions comparable to, or i n some cases better than, that achievable with photoabsorption techniques i n the soft X-ray range. Much of the work performed on inner s h e l l electron e x c i t a t i o n (both ISEELS and photoabsorption) up to 1982 has been summarised i n a recently published bibliography [10]. In EELS, a monoenergetic beam of electrons i s used to exc i t e various e l e c t r o n i c t r a n s i t i o n s i n the sample. The energy required for such a t r a n s i t i o n can be found by measuring the energy loss of the scattered electrons ( i . e . the electrons which caused the t r a n s i t i o n s ) . The system may be described as:-- 2 -e(E ) + M • M* + e(E -E') o v o ' where E q i s the energy of the incident beam and (E Q-E') i s the energy of the scattered e l e c t r o n . Thus E', the energy loss (analogous to photon energy), i s the energy required for the sample to go from the M ground state to the M excited s t a t e . It can be seen that information akin to photoabsorption i s obtained, however, the use of electrons has some differences which i n some cases lead to d i s t i n c t advantages. 1) Photoabsorption i s a resonant process, the photon energy must exactly match the t r a n s i t i o n energy, whereas with electron impact the process i s non-resonant since the excess energy i s car r i e d o f f by the scattered e l e c t r o n . 2) D i f f e r e n t sources and hence techniques are required i n photoab-sorption i n order to cover a wide range of energies. E l e c t r o n impact, however, allows coverage of a range extending from zero energy loss to the X-ray region with a single spectrometer. 3) In the soft X-ray region (~200 eV - 1000 eV), the re s o l u t i o n of electron impact e x c i t a t i o n i s comparable or better than that achieved so far with photoabsorption [8]. 4) With low momentum transfer (high incident energy and small scattering angle) a fast electron beam provides a dipole e x c i t a -t i o n mechanism which i s an e f f e c t i v e a l t e r n a t i v e to the use of a tuneable photon source. A v i r t u a l photon f i e l d i s induced i n the - 3 -target by the passing e l e c t r o n . The target sees the e l e c t r i c f i e l d associated with the beam as a sharp impulse i n time, which, with high enough energy, approaches a delta function. Fourier transforming this into a frequency domain yi e l d s a continuous range of frequencies (assuming an i d e a l d elta function) of uniform i n t e n s i t y . Thus spectra obtained i n this manner w i l l c l o s e l y resemble o p t i c a l spectra and i n essence the o p t i c a l (dipole) s e l e c t i o n rules w i l l apply. On lowering the impact energy and/or increasing the scattering angle, the momentum transfer i s increased and o p t i c a l l y forbidden t r a n s i t i o n s become more important. In this way the use of electrons not only compliments the information obtained by photoabsorption but i s also able to extend i t . The primary concern of the work described i n t h i s thesis i s the study of dipole allowed t r a n s i t i o n s i n a v a r i e t y of molecular systems as obtained by using high energy electron impact and small scattering angle. The major focus w i l l be on inner s h e l l electron e x c i t a t i o n , however, valence s h e l l electron e x c i t a t i o n for many of the systems w i l l also be presented. The inner s h e l l e x c i t a t i o n spectrum i s often simpler and less ambiguous to assign than the valence s h e l l e x c i t a t i o n spectrum since the o r i g i n a t i n g o r b i t a l i s usually well separated from the other o r b i t a l s and can be p o s i t i v e l y i d e n t i f i e d . Thus knowledge of the term values (difference of the e x c i t a t i o n feature from the i o n i s a t i o n l i m i t ) from the inner s h e l l spectrum can aid i n the i n t e r -pretation of the valence s h e l l spectrum. Before discussing some of - 4 -the t h e o r e t i c a l considerations relevant to electron impact, the r e l a t i o n s h i p with photoabsorption and the features observed i n electron e x c i t a t i o n spectra i n more d e t a i l , i t i s useful to b r i e f l y review some of the aspects of electron spectroscopy i n general and how they r e l a t e to electron e x c i t a t i o n . B. Description of various processes i n electron spectroscopy The various types of processes which can occur are i l l u s t r a t e d i n Figure 1.1. Note that both X-ray (K, L, M, ...) and o r b i t a l ( I s , 2s, 2p, 3s, 3p, ....) notation for electron energy l e v e l s w i l l be used interchangeably throughout t h i s t h e s i s . I) Electron E x c i t a t i o n . As this i s the main topic of t h i s work, this i s introduced f i r s t , however, only a b r i e f d e s c r i p t i o n i s presented here. A more detail e d and complete d e s c r i p t i o n follows i n the subsequent sections. The process i s i s i l l u s t r a t e d i n Figures 1.1B and 1.1C In terms of a one electron p i c t u r e . Interaction of the i n i t i a l ground state of a molecule or atom with either a photon or electron can cause one of the electrons of the target species to be promoted to an unoccupied l e v e l . Use of a photon requires the t r a n s i t i o n energy (E') to be exactly equal to the photon energy whereas an incoming electron can impart whatever energy i s required. Figure 1.1B shows the promotion of an inner s h e l l electron while Figure 1.1C shows that of a valence el e c t r o n . The t r a n s i t i o n energies for the former (1.1B) l i e - 5 -t 11 I t J, t 1 — 11 11 IT 4 1 E k J t Ci) (il) photon unocc. L 3 . 1t L2 i t L, 2s i t K 1s D Figure 1.1 Processes involved i n e l e c t r o n spectroscopy. A. I n i t i a l s t a t e B,C. E l e c t r o n e x c i t a t i o n by e l e c t r o n impact or photoabsorption ( s e c t i o n B.I) D. P h o t o i o n i s a t i o n ( i ) XPS ( i i ) UPS ( s e c t i o n B.II) E. P h o t o i o n i s a t i o n r e s u l t i n g i n excited ion states -"shake-up" - "shake-off" ( s e c t i o n B.III) F. Core vacancy r e s u l t i n g from ion formation allows G and H ( s e c t i o n B.IV) G. X-ray fluorescence H. Auger process I. Core vacancy r e s u l t i n g from formation of an excited n e u t r a l species allows J ( s e c t i o n B.IV) J . A u t o i o n i s a t i o n . - 6 -( t y p i c a l l y ) i n the XUV or soft X-ray regions of the electromagnetic spectrum while for the l a t t e r (1.1C) the t r a n s i t i o n energies are i n the vacuum or far UV. Thus i n photoabsorption several d i f f e r e n t mono-chromators are usually needed to cover the UV and X-ray region. Transitions to the unoccupied l e v e l s from a p a r t i c u l a r l e v e l i n atoms w i l l r e s u l t i n an e x c i t a t i o n spectrum consisting of a series of li n e s converging upon the p a r t i c u l a r i o n i s a t i o n l i m i t . The serie s f i t s the following Rydberg formula. T = E I P - E ' = 7 ^ = 7 - 7-2- (1-B.l.) (n ) ( n~^$) where E' Is the t r a n s i t i o n energy; E _ p i s the i o n i s a t i o n l i m i t ; R i s the Rydberg constant (13.605eV); n i s the e f f e c t i v e quantum number and i s obtained by correcting the p r i n c i p a l quantum number, n, by a quantum defect, 6^, which i s c h a r a c t e r i s t i c of the A quantum number ( i . e . whether the electron i s promoted to an s, p, d etc. type of o r b i t a l ) . For the hydrogen atom 6^ = 0, thus the quantum defect can be thought of as a measure of the deviation from simple hydrogen-like behaviour. The term value, T, i s the difference of the Rydberg feature from the i o n i s a t i o n l i m i t . In essence i t can be thought of as the binding energy of the newly promoted electron occupying the Rydberg o r b i t a l . The i n t e n s i t y of t r a n s i t i o n s to the Rydberg o r b i t a l s f a l l s as n~ 3 [11,12], thus i t becomes increasingly d i f f i c u l t to observe the higher l y i n g Rydberg states. - 7 -The concept of the Rydberg o r b i t a l can be extended to mole-cule s . The Rydberg o r b i t a l s are large and d i f f u s e and hence w i l l i n c r e a s i n g l y see the molecule as one large core. However, i n addition to Rydberg o r b i t a l s , molecules w i l l also have unoccupied v i r t u a l valence o r b i t a l s which a r i s e out of the MO scheme. These o r b i t a l s are usually the antibonding counterparts of the bonding MOs. Thus they are delocalised and of s i m i l a r size to the occupied outer valence o r b i t a l s of the molecule. Depending on the p a r t i c u l a r molecule, the v i r t u a l valence o r b i t a l s can be low l y i n g , i n which case t r a n s i t i o n s to them w i l l be seen i n the disc r e t e portion of the e x c i t a t i o n spectrum, however, they may be high l y i n g and thus occur i n the continuum. The above aspects of Rydberg and valence o r b i t a l s w i l l be expanded upon In section E. A good review of the o r b i t a l concept i n molecular spectroscopy has been given by W i t t e l and McGlynn [13]. Robin [12] also gives a det a i l e d discussion of Rydberg and valence states. II) Photoionisation. This i s i l l u s t r a t e d i n Figure 1.1D. In this the i n i t i a l state of a molecule or atom i s bombarded with photons of a c h a r a c t e r i s t i c energy, hv, a t r a n s i t i o n occurs i n which the f i n a l state i s an ion plus a free e l e c t r o n . Conservation of energy requires that hv = E R + E _ p ( I . B .2) - 8 -where E^ i s the k i n e t i c energy of the ejected electron and i s the energy required to form the ion. Since the mass of the ion i s several thousand times that of the electron, conservation of momentum dicta t e s that e s s e n t i a l l y a l l the k i n e t i c energy i s taken up by the ejected e l e c t r o n . Thus by measuring and knowing hv, E.^, which i n the one electron example shown i n F i g . 1.1D i s the i o n i s a t i o n p o t e n t i a l or binding energy of an electron from a p a r t i c u l a r o r b i t a l , can be obtained. The i o n i s a t i o n p o t e n t i a l can be equated to the di f f e r e n c e i n the t o t a l energies of the ion state and the ground state of the species: E I P = E f S ^ N - 1 ) " E G ( N ) (1.B.3) g where E T_ i s the i o n i s a t i o n p o t e n t i a l of electron s; E (N) i s the IP d t o t a l energy of the ground state and E^ (N-l) i s the t o t a l energy of the ion which i s formed when electron s i s removed. Thus the i o n i s a -t i o n energies can be obtained t h e o r e t i c a l l y by rigorous c a l c u l a t i o n s of the appropriate t o t a l energies. Quantum mechanically the p r o b a b i l i t y of a t r a n s i t i o n from the i n i t i a l ground state (<|/') to the f i n a l state (<\>' = ion + free electron) i s proportional to the square of the t r a n s i t i o n moment i n t e g r a l M -<«|>" | Sp| c|>'> (1.B.4) - 9 -where p Is the dipole moment operator. Application of the Born-Oppenheimer approximation separates the wave function into a product of e l e c t r o n i c (4^) and nuclear (v* (RHv'(R)dR/c|>e (r,R)|2 P j c | / ( r , R ) d r (I.B.5) where the nuclear function (^M) bas been further pa r t i t i o n e d into v i b r a -t i o n a l (<1>V) and r o t a t i o n a l components. In the majority of cases r o t a -t i o n a l structure cannot be resolved and i s ignored (as i n (I.B.5)). For a photoelectron t r a n s i t i o n to be allowed, the integr a l s i n the above equation must be non-zero. Since the f i n a l state includes a free electron there i s always a non-zero value and as a consequence a l l one electron t r a n s i t i o n s are allowed. The v i b r a t i o n a l part of equation (I.B.5) i s termed the Franck-Condon factor and gives the i n t e n s i t i e s and shape of the v i b r a t i o n a l envelope. This shape can be i n d i c a t i v e of the type of electron being ionised. For example, i o n i s a t i o n of a non-bonding electron would lead to very l i t t l e change i n the nuclear coordinates of the ion from that of the ground state molecule and thus the v" = 0 •> v' = 0 (where v denotes the v i b r a t i o n a l quantum number) t r a n s i t i o n would have the strongest overlap and there would only be a short progression. In other words the adiabatic and v e r t i c a l i o n i s a t i o n energies coincide. An adiabatic t r a n s i t i o n i s defined as the least energy required to eject a p a r t i c u l a r - 10 -electron from a molecule i n i t s ground state, i . e . i t corresponds to the v" = 0 •* v* • 0 t r a n s i t i o n . A v e r t i c a l t r a n s i t i o n i s defined as the most intense, i . e . a v" = 0 •*• v f = n t r a n s i t i o n , where n i s the v i b r a t i o n a l function with the largest overlap with the ground state. T r a d i t i o n a l l y photoelectron spectroscopy has been divided into two sections depending upon the photon source being used. One major branch i s Photoelectron Spectroscopy (PES) or U l t r a v i o l e t Photo-electron (UPS). In t h i s case u l t r a v i o l e t r a d i a t i o n , t y p i c a l l y He I (21.22 eV) provides the photon source and hence only the valence o r b i t a l s are ac c e s s i b l e . Use of the He(II) resonance l i n e at 40.81 eV i n p r i n c i p l e allows the inner valence o r b i t a l s to be probed but i n practice this i s precluded by the He I "shadow". An excellent book by Rabalais [15] covers the various aspects of UPS. The other ^ branch i s Electron Spectroscopy for Chemical Analysis (ESCA) or X-ray Photo-electron Spectroscopy (XPS). In t h i s case the photon source i s an X-ray. The most common sources used are the Mg Ka (1253.64 eV) and A l Ka (1486.58 eV). Thus the core l e v e l s can be accessed. There has been much work on XPS ever since the pioneering work of Siegbahn [16,17], The primary a p p l i c a t i o n arises from the fact that the inner s h e l l electron shows a chemical s h i f t i n i t s binding energy which r e f l e c t s a change i n the molecular environment. A perspective of i t s impact can be obtained from Siegbahn's Nobel laureate address [18]. The work from th i s Swedish group has culminated i n a new high-resolution multipurpose XPS spectrometer which features X-ray monochromation [19]. - 11 -Ionisation energies obtained from photoelectron spectroscopy are invaluable to the i n t e r p r e t a t i o n of electron e x c i t a t i o n spectra i n that an accurate and unambiguous value i s obtained for the binding energy (or i o n i s a t i o n l i m i t ) of a p a r t i c u l a r e l e c t r o n . In theory, analysis of a Rydberg series should y i e l d the IP (see equation ( l . B . l ) ) . However, only a few l e v e l s can usually be i d e n t i f i e d i n the valence s h e l l spectrum due to such factors as the n~ 3 drop i n i n t e n s i t y , other overlapping t r a n s i t i o n s as well as deviations of the lower Rydbergs from the series due to possible valence-Rydberg o r b i t a l i nteractions [20]. These f a c t o r s , the lack of r e s o l u t i o n and l i f e t i m e considerations make matters even worse for inner s h e l l t r a n s i t i o n s . Thus i t i s of far more use to i n t e r p r e t the e x c i t a t i o n spectrum knowing the IP accurately from UPS or XPS and hence being able to e s t a b l i s h the term values with considerable accuracy. These term values are, i n general, c h a r a c t e r i s t i c of p a r t i c u l a r t r a n s i t i o n s (see section E). A further aid i n i n t e r p r e t i n g electron e x c i t a t i o n spectra can come from the s p e c t r a l shape of the photoionised band. In i o n i s a t i o n only the nature of the i n i t i a l o r b i t a l has to be considered, however, i n e x c i t a t i o n both l e v e l s must be considered. Thus i f the s p e c t r a l shape (and i n p a r t i c u l a r the v i b r a t i o n a l envelope at high resolution) i s s i m i l a r i n the i o n i s a t i o n and e x c i t a t i o n spectra, the t r a n s i t i o n i s suggestive of one to a Rydberg l e v e l . This arises from the large, d i f f u s e and non-bonding nature of the Rydberg o r b i t a l and so the upper (Rydberg) state i s very s i m i l a r to the ion state. However, care has - 12 -to be exercised since some of the lower Rydberg states may have some anti-bonding character due to Rydberg-valence mixing [20]. I l l ) "Shake-up" and "Shake-off". Upon photoionisation an excited ion state may be formed i n which the photoelectron has been emitted along with the simultaneous e x c i t a t i o n of an outer electron either to an excited bound state or into the continuum. The former Is termed "shake-up" while the l a t t e r i s termed "shake-off". The process i s i l l u s t r a t e d i n F i g . 1.1E. Upon i o n i s a t i o n the resultant ion can be i n one of a number of states with the i o n i s a t i o n energy of the photoelectron being given by a modified form of equation (1.B.3) E I p S , t = E f S , t - E G(N) (1.B.6) when t = 0, the ion i s i n i t s ground state and th i s gives r i s e to the major photoionisation peak. The s a t e l l i t e l i n e s are defined by t = 1, 2 etc. and since they denote excited ion states the s a t e l l i t e s w i l l appear on the low k i n e t i c energy side of the major peak. Equation (1.B.6) implies that there i s no fundamental d i f f e r -ence between the major peak and the s a t e l l i t e s and that each state i s reached by a one-step, "one-electron" dipole t r a n s i t i o n [18] (note: photoionisation follows dipole s e l e c t i o n r u l e s - see equation (1.B.5)). On a s i m p l i s t i c model, "shake-up" can be thought of as the emission of - 13 -a photoelectron with the attendant e x c i t a t i o n of a valence electron. Since the allowed, excited ion states must have the same symmetry as the ion i n i t s ground state, the valence electron t r a n s i t i o n must i n essence follow monopole s e l e c t i o n r u l e s . Thus i t might be thought that there i s a d i r e c t r e l a t i o n s h i p between XPS s a t e l l i t e structure and valence electron e x c i t a t i o n as observed by photoabsorption spectroscopy or EELS. However, care must be taken i n any comparison since EELS or photoabsorption spectroscopy leads to an excited neutral state which can e s s e n t i a l l y be described by a simple one electron d e s c r i p t i o n whereas the XPS s a t e l l i t e structure leads to an excited ion state. Martin et a l . [21,22] have shown that many electron theory including configuration i n t e r a c t i o n must be included i n both hole and ground state i n order to describe the s a t e l l i t e structure accompanying photoionisation adequately. An a l t e r n a t i v e approach which also accounts for c o r r e l a t i o n e f f e c t s i s the many-body Green function method used by Cederbaum and co-workers [23,24] to describe inner valence s h e l l i o n i s a t i o n processes. In addition to providing information on valence electron e x c i t a t i o n , knowledge of the XPS s a t e l l i t e spectrum can a s s i s t i n the i n t e r p r e t a t i o n of continuum structure i n ISEELS and photoabsorption spectra. The XPS spectrum provides information on the excited ion states and i n p a r t i c u l a r gives the v e r t i c a l energy for the i o n i s a t i o n processes. The production of these excited ion states would be mani-fested i n ISEELS by (adiabatic) onsets of new continua. Thus i t should be possible to i d e n t i f y which features i n the ISEELS spectrum - 14 -above the i o n i s a t i o n edge (the major i o n i s a t i o n peak) ar i s e from the onsets of excited ion states and by inference which features are due to other types of phenomena such as resonances or double e x c i t a t i o n s . Continuum structure i n ISEELS or photoabsorption spectroscopy w i l l be discussed i n section E. IV. X-ray Fluorescence, Auger Decay and Autoionisation. Following the creation of a hole state, either as an ion ( F i g . 1.1F), or as an excited neutral state (Figure 1.11), secondary processes can occur which allow the system to achieve a lower energy state. For the ion state i n which there i s a core vacancy, the two major decay modes are X-ray fluorescence and the Auger process. In the case of X-ray fluorescence the excess energy i s released i n the form of a photon whereas i n the Auger process the excess energy i s given to an emitted electron. The two processes are i l l u s t r a t e d i n Figures 1.1G and 1.1H respectively. The respective y i e l d s for atomic K s h e l l fluorescence and Auger decay are given by where WR i s the K s h e l l fluorescence y i e l d , a R i s the Auger y i e l d ; P^ and P^ are the t r a n s i t i o n p r o b a b i l i t i e s for X-ray fluorescence and the Auger process respectively. For the l i g h t e r elements the Auger P + P f A and a. P + P f A process dominates [25]. For example, the fluorescent y i e l d for the K s h e l l of the nitrogen atom i s 5.2 x 10~ 3 and i t s Auger y i e l d i s 0.995 [25]. For the (2s) l e v e l of phosphorus, the fluorescent y i e l d i s 3.9 x 10~ 5 and for the L (2p) l e v e l i t i s 6.2 x I0~h [25]. In spite of the low y i e l d for l i g h t elements various X-ray fluorescence studies have been reported u t i l i s i n g high r e s o l u t i o n spectrometers [26-28] . X-ray t r a n s i t i o n s are governed by dipole s e l e c t i o n rules and so the L^ ^ (-P) "* K (Is) t r a n s i t i o n (as shown i n F i g . 1.1) i s allowed but the L^ (2s) -*• K (Is) t r a n s i t i o n i s forbidden This factor can aid i n the i d e n t i f i c a t i o n of valence states and provide complimentary information to UPS [26]. In other studies, Nordgren et a l . [27] have reported the C K-shell X-ray fluorescence spectra of C0 2 and from a Franck-Condon f i t of harmonic o s c i l l a t o r s have determined the natural width of the C Is state to be 0.07 ± 0.02 eV. A further a p p l i c a t i o n stems from the combination of X-ray fluorescence energies with UPS i o n i s a t i o n potentials to estimate core-electron binding energies [28]. The Auger process i s the dominant decay mode for inner s h e l l vacancies of l i g h t atoms. The energy of the Auger electron i s given by the difference between the t o t a l energy of the i n i t i a l hole state and that of the two hole state: E A(XYZ) = E w f ( x ) - E m + + ( y z ) (1.B.7) where E.(XYZ) i s the k i n e t i c energy of the emitted XYZ Auger electron - 16 -E^+^j i s the t o t a l energy of the i n i t i a l hole state, a singly ionised species with a hole i n l e v e l X; -•jy-H-^yz) - s t n e t o t a - energy of the f i n a l two-hole state, a doubly ionised species with holes i n l e v e l s Y and Z. Thus i n F i g . 1.1H the emitted Auger electron would be designated as the KL^L 3 Auger t r a n s i t i o n . Unlike X-ray fluoresence, the Auger process involving an i n i t i a l K hole state can involve the s h e l l . I t s se l e c t i o n rules are AL = AS = AJ = 0 and pa r i t y unchanged. Thus the Auger process i s not a dipole t r a n s i t i o n followed by ej e c t i o n of an electron but rather i t arises from a coulombic rearrangement of the two electrons involved [29]. C l e a r l y i t i s the much faster process for l i g h t elements ( i t i s ~10 3 times faster for C and 0 K-shells than fluorescence [11]). A s p e c i a l type of Auger decay i s termed a Coster-Kronig t r a n s i t i o n . This i s a very rapid process and involves an Auger t r a n s i t i o n i n which the primary vacancy i s f i l l e d by a higher l y i n g electron within the same s h e l l . Thus a L ^ L j 3 M Coster-Kronig t r a n s i t i o n would involve the f i l l i n g of a L^ hole by an electron from the L 2 3 sub-shell with the r e s u l t i n g emission of a M-shell electron. Those Auger spectra which only involve inner s h e l l electrons are atomic-like i n nature. As i n the case of XPS spectra, inner s h e l l Auger l i n e s also show a chemical s h i f t . Extensive series of KL2L 3 Auger electron chemical s h i f t s have been reported for Si [30], P [31] and S [32] containing compounds. A combination of Auger and XPS chemical s h i f t s allows an estimate of the extra-atomic relaxation ( i . e . a t t r a c t i o n of the higher l e v e l s towards a hole) upon creation of - 17 -a core-hole. With this i t i s possible to estimate the r e l a t i v e p o l a r i s a b i l i t y of the various ligands [31,33]. Auger spectra which involve the valence s h e l l are very complex due to the many overlapping t r a n s i t i o n s . However, various gas phase studies as well as c a l c u l a -tions have been performed on Auger spectra [34-36]. Similar types of decay modes w i l l occur when the hole state a r i s e s from electron e x c i t a t i o n giving r i s e to an excited neutral state ( F i g . 1.11). The non-radiative process akin to Auger decay i s termed autoionisation ( F i g . 1.1J). As i n Auger decay, the process involves the f i l l i n g of the hole with the concomitant e j e c t i o n of an electron. The emitted electron may be the i n i t i a l l y excited electron or some other electron. These decay modes are important to photoionisation and electron e x c i t a t i o n as they govern the l i f e t i m e of the hole state and hence the natural l i n e width of the feature. The natural l i n e width (AE) i s related to the l i f e t i m e (At) v i a the Heisenberg uncertainty p r i n c i p l e : AE • AT ~ ft = 6.582 x 10~ 1 6 eV.sec (1.B.8) For example l e v e l s which can undergo rapid Coster-Kronig t r a n s i t i o n s have a short l i f e t i m e and hence are wide. For instance the natural linewidth f o r the (2s) l e v e l i n P i s 1.26eV whereas i t i s only 0.033 eV and 0.032 eV for the L 3 (2p 3 /, 2) and L 2 ( 2 p x ^ ) l e v e l s r e s p e c t i -vely and 0.53 eV for the K (Is) l e v e l [37]. In general, the natural linewidth (energy) for a given subshell increases with increase i n Z, - 18 -along the periodic Table [37]. The above discussions have dealt mainly with inner s h e l l vacan-c i e s . Obviously s i m i l a r processes can occur for valence hole state below the f i r s t IP. Another form of mechanism which can occur reducing l i f e t i m e s of a p a r t i c u l a r state i s pr e - d i s s o c i a t i o n . C. Fundamental Concepts i n Electron Impact and the Relationship to Photoabsorption To understand the processes which produce the features i n an electron energy loss spectrum and to re l a t e them to the o p t i c a l (photoabsorption) spectrum, i t i s necessary to discuss some of the fundamental concepts behind the c o l l i s i o n process. This can only come from a quantum mechanical de s c r i p t i o n such as was f i r s t given by Bethe i n 1930 [38]. A thorough review of Bethe's treatment has been given by Inokuti [39]. Wight [5], using the ideas discussed by Lassettre [40], has also discussed i n d e t a i l the Bethe theory for electron scattering by the hydrogen atom including a generalisation to more complex species. This approach [5,40] w i l l be followed i n the present work. In view of the e a r l i e r referenced works, a detailed treatment w i l l not be given and only the pertinent points and d e f i n i t i o n s w i l l be discussed. When an incident beam of electrons interacts with a target molecule, the l a t t e r may be excited from i t s i n i t i a l state to some excited state. The p r o b a b i l i t y of such a t r a n s i t i o n i s known as the - 19 -D i f f e r e n t i a l Cross-Section (DCS). This i s the number of incident electrons scattered per second through an angle 9 into a s o l i d angle dQ a f t e r e x c i t i n g the target to i t s n**1 excited state, divided by the number of electrons i n the incident beam which crossed unit area i n one second. If the incident beam i s approximated by a plane wave the DCS for i n e l a s t i c s c a ttering o f f an H atom i s given by [ 5 ] : | g (9) dQ - ^ _ | f n ( 9 ) | 2 dQ ( l . C . l ) o where k and k are the wave numbers for the incident and scattered o n beams resp e c t i v e l y and f n(®) - s t n e scattering amplitude. The square 2 of the scattering amplitude, | f n ( 9 ) | , i s the number of electrons/unit volume at unit distance and angle 9 which have excited the atom to i t s th n s t a t e . Since the wave number i s proportional to momentum, 2 k n | f n ( 9 ) | i s proportional to the number of scattered electrons crossing unit area i n one second at angle 9, and k Q i s proportional to the t o t a l number of electrons i n the incident beam crossing unit area per second. Thus the DCS can be experimentally determined by measuring these q u a n t i t i e s . In order to calculate the DCS knowledge of If (9)1 i s required. n This w i l l come out of the solu t i o n to the Schrodinger equation for the c o l l i s i o n , which for the prototype case of electron-H atom s c a t t e r -ing, gives an i n f i n i t e set of coupled d i f f e r e n t i a l equations of the - 20 -type [41] [V 2 + k 2 - V ) F (r, ) = ^ [-+/] V F (r. ) (1 .C.2) v r, n ^1. nn' n b L , mn m b b n Tl m+n where the matrix element, V , i s defined as mn * _ 2 2 V m n = Ju (r ) (--= f - ) U (r )dr (1.C.3) mn n a v r, r, m a a ba b and r & and r ^ are the co-ordinates of the target and incident (projec-t i l e ) electrons r e s p e c t i v e l y ; F (r, ) represents the wave function of the m b incident electrons and U ( F ) are the complete set of eigenfunctions of the unperturbed H atom. As equation (1 .C.2) stands i t cannot be solved and various approximations have to be made. One of the most common ones applied i n c o l l i s i o n theory i s the Born approximation [42], The basic assumption behind t h i s approximation i s that there i s l i t t l e i n t e r -a ction between the p r o j e c t i l e and the target. Thus the incident wave i s undistorted by the i n t e r a c t i o n and can be represented by an undis-torted plane wave; the e x c i t a t i o n i s due to a d i r e c t t r a n s i t i o n from the ground to the excited state and so a l l terms are zero except for v Q n » the p o t e n t i a l energy of the i n t e r a c t i o n between the scattered electron and the atom i n i t s f i n a l state i s small and so the d i s t o r t i o n of the scattered wave can be neglected, therefore V =0. nn The Born approximation Is only v a l i d for high incident electron energy - 21 -(> 5-7 times the e x c i t a t i o n energy) and hence w i l l be v a l i d i n the majority of the cases reported i n the present work. Application of the Born approximation s i m p l i f i e s equation (l.C.2) to ( V 2 r + k 2 ) F (7 ) = ^ V e i k o ' r b (l.C.4) r, n n D _ z on b •fi which can be solved by the method of Green's functions to y i e l d an expression for f n ( 9 ) . Thus the DCS can be written as dc< 9)dQ = m2 k n , f i ( k -k ) »F - ,2,_ dQ 7 X 4 i T l J e ° n Von d r I d Q ( 1 ' C - 5 ) (note r = r, ) F i n a l l y by defining a momentum transfer v a r i a b l e , K, where: I K | 2 = I k - k | 2 = k 2 + k 2 - 2 k k cos9 (l.C.6) 1 1 ' o n 1 o n o n ' generalising to a N electron system (whose coordinates are given by r ) and using the inte g r a t i o n formula derived by Bethe [38,39], namely: - 22 -iK. r . ,rr c e ,— 4it i K . r ( . -,N J dr = —-r e s (l. C . 7 ) I I K r - r 1 s 1 the following expression f o r the DCS i s obtained: 2 4 k dc( 9) = 4m e n dQ K e (K) | 2 on ' (l.C.8) where the matrix element e (K) i s on * N - -e (K) = /U E e s U dx„ (l.C.9) on J n , o N s=l where dx^ indicates i n t e g r a t i o n over a l l coordinates of the N el e c t r o n system. A useful concept which predates quantum mechanics i s that of the o s c i l l a t o r strength. The c l a s s i c a l p i c t ure defined the o s c i l l a t o r strength, f, to be the number of electrons i n free o s c i l l a t i o n at a p a r t i c u l a r frequency. The t o t a l o s c i l l a t o r strength was the number of electrons i n the target. The concept of t o t a l o s c i l l a t o r strength has been retained i n quantum theory as a useful means of defining t r a n s i -t i o n p r o b a b i l i t y . In e f f e c t the t r a n s i t i o n p r o b a b i l i t i e s are being normalised to the t o t a l number of electrons i n the system. This i s known as the Thomas-Reiche-Kuhn (TRK) sum rule [39]. - 23 -In the dipole approximation (which covers o p t i c a l s e l e c t i o n rules) the general form for the o p t i c a l (dipole) o s c i l l a t o r strength i s f o n ( 0 ) = 2 EJ < ( |rJ E r s I V (LC.10) s=l Bethe [38,39] has defined a generalised o s c i l l a t o r strength (GOS ) for p a r t i c l e c o l l i s i o n f ^ ^ K ) given by 2E N - -K s=l (l.C.11) 2E _ = ^ r l E o n W l substituting (l.C.11) into (l.C.8) and switching to atomic units the Bethe-Born r e l a t i o n s h i p [38,39] i s obtained, namely: n This form of o s c i l l a t o r strength i s only v a l i d i f the Born approxima-t i o n holds, however, Bethe has also defined the apparent GOS, f N ' ( K , E Q ) , where E q i s the impact energy. This i s calculated from the - 24 -experimental parameters. The o p t i c a l (dipole) o s c i l l a t o r strength i s a sp e c i a l case of the GOS. Expanding the exponential function i n equation (l.C.11) as a power series i n K, i . e . e i K . r _ x + 1 R > r + ( l K > r ) 2 > > # ( i K . r ) m 2! M! e becomes on e Q n = 0 + e x ( i K ) + e 2 ( i K ) 2 + ... ^ ( i K ) * 1 (l.C.13) where e = — - < > and orthogonality dictates that e = < 6 >=0. m m! n' s 1 o J o n 1 o s For m = 1, th i s matrix element becomes e, - <(|» |E r U > (l.C.14) 1 n' s • o which i s nothing more than the t r a n s i t i o n moment term i n the dipole approximation. Substituting (l.C.13) back into (l.C.11), the expanded form of the GOS i s obtained, namely: f(K) = 2 E R {c\ + (e2-2e 1e 3)K 2 + 0(K 4)} = f(0) + f ( l ) K 2 + f ( 2 ) K 4 + (l.C.15) - 25 -Thus as the momentum transfer, K, approaches zero, the GOS, f ( K ) , approaches f(0) which i s the o p t i c a l o s c i l l a t o r strength ( i . e . Limit f(K) = f ( 0 ) ) . The higher terms, f ( l ) , f(2) etc., represent quadrupole, octopole and higher t r a n s i t i o n s . Lassettre et a l . [43] have shown that the GOS approaches the o p t i c a l l i m i t as K approaches zero regardless of whether the Born approximation applies or not. From the above discussion i t can be seen that i f the experi-mental conditions are selected such that the momentum transfer approaches zero [44] then there exists a d i r e c t r e l a t i o n s h i p between the DCS and the o p t i c a l o s c i l l a t o r strength. This i s governed by the Bethe-Born r e l a t i o n s h i p (l.C.12). To achieve low momentum transfer, high incident electron energies ( E Q) and small angle (9) sca t t e r i n g ( i d e a l l y zero degrees) are required. If the incident energy i s very much greater than the energy transfer ( i . e . t r a n s i t i o n energy, E n ) and 9 i s small, s u b s t i t u t i o n of these quantities into equation (l.C.6) 2 2 (note: k = 2E and k = 2(E -E ) i n atomic units) gives: o o n o n K 2 = 2E Q [\{- E I L) 2 + 9 2) (l.C.16) o thus If 9 = 0 , the momentum transfer i s proportional to the energy transfer and hence equation l.C.12 indicates that: 35 " E „ " 3 '<°> (l.C.17) - 26 -Thus dipole spectra produced i n this manner by electron impact should be q u a l i t a t i v e l y s i m i l a r to the i r photoabsorption counterparts and only d i f f e r by a r e l a t i v e decrease i n i n t e n s i t y which i s proportional to the inverse of the energy loss cubed. This factor presents l i t t l e problem for a simple q u a l i t a t i v e comparison between o p t i c a l and electron energy loss spectra, e s p e c i a l l y over a short range. However, the above correction (l.C.17) has to be applied i f a quantitative comparison i s required [45]. A review of continuum o s c i l l a t o r strengths as obtained by electron impact spectroscopy has recently been given by Brion and Hamnett [45]. A l t e r n a t i v e l y dipole o s c i l l a t o r strengths can be obtained by measuring a se r i e s of electron energy loss spectra at various momentum transfers and extrapolating back to zero momentum transfer. This can be done i n two ways: either by f i x i n g the impact energy and varying the scattering angle, or by f i x i n g the scattering angle (usually at 0°) and varying the impact energy. The former method has been u t i l i s e d by Lassettre et a l . [2] while the l a t t e r method has been used by Hertel and Ross [46,47]. One aspect of these methods i s that non-dipole t r a n s i t i o n s can be studied and i d e n t i f i e d since the o s c i l l a t o r strengths w i l l extrapolate back to zero. However, for the study of o p t i c a l l y allowed t r a n s i t i o n s by electron impact i t i s far superior and much less 2 tedious to work d i r e c t l y ( i . e . as close as possible) to the K = 0 l i m i t by appropriate s e l e c t i o n of experimental conditions rather than employing the above mentioned extrapolation techniques. - 27 -D. The Relative Merits of Electron Energy Loss and Photoabsorption Spectroscopies In the previous section the r e l a t i o n s h i p between the o s c i l l a t o r strengths obtained by EELS and photoabsorption spectroscopies was established. I t was shown that the two techniques could produce the same information with regard to dipole ( o p t i c a l ) t r a n s i t i o n s . In t h i s section the r e l a t i v e merits of using electrons or photons to produce " o p t i c a l " ( i . e . dipole) spectra w i l l be discussed. No further d i s c u s -sion w i l l be given of the a b i l i t y of electron impact techniques to probe dipole forbidden t r a n s i t i o n s [1-3]. The obvious question to ask i s why would anyone wish to simula-te o p t i c a l spectra using electrons when presumably the spectra could be obtained d i r e c t l y with a photon source? C l e a r l y each technique must have i t s own advantages and disadvantages. These w i l l a r i s e from the experimental methods required to obtain the spectra and hence the various l i m i t a t i o n s inherent i n these techniques. To obtain a photoabsorption spectrum a suitable continuum l i g h t source i s required. U n t i l the advent of synchrotron r a d i a t i o n there was no e f f e c t i v e means of obtaining tuneable r a d i a t i o n i n the far UV and X-ray regions which would give a continuum source of high f l u x beyond 20 eV. Conventional l i g h t sources u t i l i s i n g hydrogen and noble gas continua have provided a u s e f u l , a l b e i t weak, and often structured source extending up to ~20 eV [48] while the X-ray region was l i m i t e d to the use of weak bremmstrahlung continua sources. However, even - 2b -with the a v a i l a b i l i t y and use of synchrotron sources there s t i l l e x i s t various experimental l i m i t a t i o n s to the use of photons i n c e r t a i n regions of the electromagnetic spectrum. In order to use a continuum l i g h t source one must select the region of ra d i a t i o n that i s required by suitable o p t i c a l monochroma-tio n and then transmit this selected l i g h t to the sample. This neces-s i t a t e s the use of windowless (for energies >10 eV) grating vacuum spectrometers. Since t h i s i s a dispersive technique and the amount of l i g h t absorbed i s measured as a function of wavelength (X,), d i f f e r e n t grating and monochromator designs are required to optimise the running conditions for the various parts of the electromagnetic spectrum. Brown [ 4 9 ] shows i l l u s t r a t i o n s of three types of monochromator i n use. Unlike photoabsorption, EELS i s a non-resonant technique. On passing through the sample, some of the primary beam i s scattered and the amount of energy required for a p a r t i c u l a r t r a n s i t i o n i s simply transferred to the system. This amount can e a s i l y be obtained by adding an equivalent voltage back to the analyser system thereby allowing the scattered electrons to reach the detector. Thus a sin g l e spectrometer i s able to cover a wide spectral range (from the IR through to the X-ray). By use of suitable retarding voltages i t i s possible to run the spectrometer i n a "constant analyser pass energy mode" and so the energy res o l u t i o n can be kept constant throughout the whole sp e c t r a l range. I t i s this f i n a l feature which presently gives a major - 29 -advantage to EELS over photoabsorption for the 200 - 1000 eV energy loss .region. This arises from the inverse r e l a t i o n s h i p of energy and wavelength. The larger the t r a n s i t i o n energy, the shorter the wave-length and hence the worse the energy r e s o l u t i o n . Thus above ~200 eV EELS becomes increasingly advantageous. For example, the "grass-hopper" monochromator shown by Brown [49] i s quoted as having a r e l a t i v e bandwidth (AX/X) of better than 10" 3 at 40 A ( i . e . a r e s o l u t i o n , AX, of 0.04 A ) . This corresponds to a FWHM of 0.3 eV for a t r a n s i t i o n at 310 eV. Recently Shaw et a l . [50] have reported the ISEELS spectrum for the C Is it t r a n s i t i o n (287.40 eV) of CO with a res o l u t i o n of 0.055eV. It should be noted, however, that the same wavelength r e s o l u t i o n (AX = 0.04 A) would be equivalent to 0.03 eV at 100 eV. Perhaps the best example remains the ISEELS N Is •> n t r a n s i t i o n recorded at 0.075eV resol u t i o n [51-53] i n which six v i b r a t i o n a l l e v e l s (centred at 401.10 eV) can be seen (see F i g . 2.5). This structure has not thus far been resolved i n any published o p t i c a l experiment. This spectrum would require a wavelength r e s o l u t i o n of 0.006 A. Figure 1.2 (taken from r e f . [6]) shows the equivalent value of A\(A) for fixed EELS resolutions (0.01 eV {state of the art with electron monochromation [50,53]} to 0.5 eV {unmonochromated electron beam from an oxide cathode}) as a function of energy l o s s . Thus with regard to energy resolution, o p t i c a l methods are superior below ~200 eV whereas EELS i s presently superior above ~200 eV. The r e s o l u t i o n c r i t e r i a i s not p a r t i c u l a r l y important above ~1000 eV due to the natural linewidths of the t r a n s i t i o n s . I t becomes advantageous again to use synchrotron Figure 1.2 Wavelength resolution plotted against e x c i t a t i o n energy f o r fixed values of energy resolution (taken from r e f . [ 6 ] ) . - 31 -methods, c e r t a i n l y beyond 2.0 - 2.5 keV where o p t i c a l monochromators are again more e f f i c i e n t . This r e s u l t s from a combination of the high f l u x c a p a b i l i t i e s of the synchrotron against the increasing i n t e n s i t y problems experienced with EELS a r i s i n g from the E n ~ 3 i n t e n s i t y factor (see equation l.C.17). However, i n a l l o p t i c a l monochromators the high synchrotron f l u x i s severely attenuated by the low r e f l e c t i v i t y of the gratings and mirrors at UV and X-ray energies. Various problems occur with the u t i l i z a t i o n of synchrotron r a d i a t i o n and these can often r e s u l t i n a d d i t i o n a l "spectral features" or incorrect o s c i l l a t o r strengths being observed. These include the p o s s i b i l i t y of higher energy r a d i a t i o n due to order overlapping, the presence of stray l i g h t and contamination of the monochromator mirrors and gratings. This l a t t e r problem i s p a r t i c u l a r l y serious i n the C Is region since the contamination w i l l a r i s e from carbon deposits formed from the decomposition of d i f f u s i o n pump o i l s ( for example) and cause a v a r i a t i o n i n the i n t e n s i t y of the l i g h t beam due to absorption. Surface contamination by carbon has even been seen under UHV (<10 - 9 t o r r ) conditions [54]. These same problems do not a r i s e i n ISEELS and indeed the true s p e c t r a l shape ( a f t e r applying the Bethe-Born correc-ti o n i . e . E ~ 3 f a c t o r ) i s obtained. A s t r i k i n g example of incorrect n s p e c t r a l i n t e n s i t y d i s t r i b u t i o n i s the valence s h e l l photoabsorption spectrum of N 2 reported by Gurtler et a l . [55]. The EELS spectrum, which shows the correct r e l a t i v e i n t e n s i t y , i s shown i n F i g . 2.4. The difference arises from l i n e saturation e f f e c t s i n the o p t i c a l work [55] caused by the band width of the l i g h t being larger than the - 32 -natural l i n e width. Since EELS i s a non-resonant process t h i s problem cannot ar i s e i n electron impact studies. This aspect has also been discussed by Inokuti [39]. Thus, somewhat i r o n i c a l l y , electron impact techniques may provide a more accurate means of obtaining o p t i c a l o s c i l l a t o r strengths than o p t i c a l methods! Indeed the experimental method i n EELS should be i n t r i n s i c a l l y more accurate than photo-absorption since the number of electrons which have been scattered i s d i r e c t l y measured whereas i n photoabsorption a difference i s taken between the incident and transmitted l i g h t . Methods which use l i g h t to excite the sample but use electron y i e l d methods to obtain the "absorption" spectrum [56] avoid the "difference" error, however, they grossly d i s t o r t the true s p e c t r a l shape. This arises from secondary processes (Auger or multiple Auger e f f e c t s ) and so more than one electron per photon i s produced. The large and variable e f f e c t of these secondary processes i n the d i s c r e t e and continuum portions of the inner s h e l l e x c i t a t i o n spectra has been c l e a r l y demonstrated i n dipole coincidence experiments [57-59]. Since the EELS method allows the true spectral shape to be obtained, accurate r e l a t i v e o s c i l l a t o r strengths can be determined. An absolute o s c i l l a t o r strength scale can be obtained either by norma-l i s i n g to a known feature i n an absolute i n t e n s i t y o p t i c a l spectrum or more r e a d i l y by a p p l i c a t i o n of the TRK sum rule (see the previous s e c t i o n ) . The TRK sum rule normalisation can be applied i n a s t r a i g h t forward fashion due to the f l a t (equal i n t e n s i t y ) v i r t u a l photon f i e l d provided by a fast electron (see above disc u s s i o n ) . - 33 -A major advantage o p t i c a l methods have over EELS i s the a b i l i t y to obtain spectra of s o l i d s or condensed phases v i a t h i n f i l m transmission techniques. E. Inner Sh e l l Electron E x c i t a t i o n Spectra In t h i s section the features of inner s h e l l electron e x c i t a t i o n spectra o r i g i n a l l y introduced i n section B w i l l be discussed i n more d e t a i l . They w i l l also be contrasted with t h e i r valence s h e l l counter-parts. As well as considering the discrete region of the spectrum, features i n the continuum spectrum w i l l also be discussed. Since only the dipole t r a n s i t i o n s w i l l be considered, the following discussion w i l l be appropriate to spectra produced by either o p t i c a l methods or by fast electron impact i n conjunction with small angle sc a t t e r i n g . I. Discrete Portion In inner s h e l l e lectron e x c i t a t i o n , features i n the d i s c r e t e portion ( i . e . below the i o n i s a t i o n l i m i t ) can usually be adequately described (semi-quantitatively) i n terms of a one-electron picture. In other words, as the promotion of an electron from a c o r e - l e v e l to a vacant l e v e l (see F i g . 1.1B). Transitions w i l l either be allowed or forbidden on the basis of dipole s e l e c t i o n r u l e s . As stated i n section B.I, the vacant o r b i t a l s are either Rydberg or v i r t u a l valence i n o r i g i n . The Rydberg o r b i t a l s are large, d i f f u s e , atomic-like and extend - 34 -well beyond the bounds of the ground state molecule. Thus the d e t a i l s of the molecular structure become less and less important with the Rydberg electron seeing e f f e c t i v e l y a single charged core [13]. The high l y i n g Rydberg states, which are e s s e n t i a l l y non-bonding, w i l l therefore be s i m i l a r i n geometry to the ionised state r e s u l t i n g from photoejection of the same electron and hence the respective features i n the e x c i t a t i o n and i o n i s a t i o n spectra should have a s i m i l a r spectral shape. The low l y i n g Rydberg o r b i t a l s , however, may show some antibonding character due to valence-Rydberg mixing [20,60,61] and thus may exhibit v i b r a t i o n a l broadening. As i n the atomic case, t r a n s i t i o n s to molecular Rydberg o r b i t a l s w i l l f i t a Rydberg series (equation l . B . l ) converging to the i o n i s a t i o n l i m i t . By knowing the term values ( i . e . the binding energy of the Rydberg el e c t r o n ) , i t i s possible to estimate the quantum defect ( 6 ^ ) which i s i t s e l f a measure of the deviation from simple hydrogen-like behaviour. ,The quantum defect i s c h a r a c t e r i s t i c of the p a r t i c u l a r A quantum number and r e f l e c t s the amount of penetration of the p a r t i c u l a r type of o r b i t a l into the core. Since the penetration decreases as s>p>d etc, so does the quantum defect. For second row atoms t y p i c a l values for s, p and d quantum defects are ~1, 0.6 and 0.1 respectively [12]. However, these values should only be used as a guide and not be applied too s t r i c t l y . Robin has surveyed many e x c i t a t i o n spectra [12] and has estimated empirical l i m i t s of 0.7 to 1.3 for s o r b i t a l s , 0.5 to 0.7 for p o r b i t a l s and -0.2 to 0.2 for d o r b i t a l s . On moving to the t h i r d row elements, the s h i e l d i n g of the core i s not as e f f e c t i v e and so the quantum defects of the more penetrating l e v e l s - 3 5 -should increase markedly. Typical values suggested for the s, p and d Rydberg l e v e l s of the t h i r d row are ~2, 1.6 and 0.0 r e s p e c t i v e l y . Since the t r a n s i t i o n s studied i n the present work are i n general governed by dipole s e l e c t i o n rules the following t r a n s i t i o n s : s •* s, p •*• p, d •* d and s «-> d would be formally forbidden i n the purely atomic case. However, i n the molecule each Rydberg o r b i t a l w i l l transform as one of the i r r e d u c i b l e representations of the molecular point group and hence one or more of the above t r a n s i t i o n s may be dipole allowed. Transitions to l e v e l s which would be formally dipole forbidden i n the atomic case, though, could be expected to exhibit less i n t e n s i t y because of the atomic contribution to the overlap. This has been discussed by Schwarz [60,61] i n r e l a t i o n to the 2p absorption spectra of Ar and the i s o - e l e c t r o n i c t h i r d row hydrides. In a l l cases the 2p •* 4p t r a n s i t i o n , as assigned by Schwarz [60,61] i s weaker than the t r a n s i t i o n s to the 4s and 3d l e v e l s . Thus a combination of the above considerations ( i . e . s e l e c t i o n r u l e s , term values and i n t e n s i t i e s ) can prove to be most h e l p f u l i n the i n t e r p r e t a t i o n of the Rydberg portion of inner s h e l l e x c i t a t i o n spectra. The v i r t u a l valence o r b i t a l s a r i s e out of the MO scheme and are usually the antibonding counterparts of the bonding o r b i t a l s . Thus they are strongly c h a r a c t e r i s t i c of the molecule. Transitions to the v i r t u a l valence l e v e l s may occur r e s u l t i n g i n states which may or may not l i e i n the discrete region of the e x c i t a t i o n spectrum. Transitions to bound v i r t u a l states are expected to be r e l a t i v e l y strong since the v i r t u a l - 36 -valence o r b i t a l s are delocalised around the molecular framework and have a comparable s p a t i a l extent to that of the ground state molecule [13]. These features are also expected to be somewhat broader than those due to Rydberg t r a n s i t i o n s because of t h e i r antibonding nature which can r e s u l t i n a large geometry change i n the upper state and hence r e s u l t i n considerable v i b r a t i o n a l e x c i t a t i o n of the f i n a l state. Depending on the molecule, intense t r a n s i t i o n s to the v i r t u a l valence le v e l s may precede the Rydberg t r a n s i t i o n s and so have larger term values than those c h a r a c t e r i s t i c for Rydberg t r a n s i t i o n s . For example the 2p ISEELS spectrum of C l 2 [62] shows two broad t r a n s i t i o n s * to the a l e v e l with term values of 9.50 eV from t h e i r respective (2p^/2 and 2 p ^ ^ ) edges. These are well separated from the Rydberg features which are c h a r a c t e r i s t i c a l l a y sharp. The term values for the 4s and 3d t r a n s i t i o n s are 3.56 eV and 1.72 eV giving quantum defects of 2.04 and 0.19 respectively, i n good accord with the "expected" values discussed above. The C Is ISEELS spectra of the methyl halides [63,64] also * provide examples where broad features can be ascribed to C Is + a * t r a n s i t i o n s preceding the Rydberg t r a n s i t i o n s . In this case the a o r b i t a l s are closer i n energy to the lowest Rydberg l e v e l s and of the same symmetry. Thus there ex i s t s the p o s s i b i l i t y of Rydberg-valence mixing [20,65]. This i s most l i k e l y i n CH 3F where the features overlap [53,64], however, for the remainder of the molecules the a and ns Rydberg l e v e l s (n = 3 for F, 4 for Cl etc) are c l e a r l y separated. The term values for the a l e v e l vary from 4.65 eV to 5.7 eV while those for the ns l e v e l go from 4.05 eV to 3.65 eV for CH 3F through to CH 3I - 37 -respectively. In contrast to the methyl halides, methane shows no feature i n the d i s c r e t e portion a t t r i b u t a b l e to a a t r a n s i t i o n and a l l the features are ascribed to Rydberg t r a n s i t i o n s [66,67]. * The above examples have a l l been for systems with a type v i r t u a l valence o r b i t a l s . V i r t u a l valence o r b i t a l s of the i t type also e x i s t i n many molecules. Molecules which contain it-bonds a l l show strong, low l y i n g features i n t h e i r electron e x c i t a t i o n spectra which can be * assigned to t r a n s i t i o n s to i t antibonding o r b i t a l s . For example, the * ISEELS spectrum of C 2H l t shows an intense C Is •+ i t t r a n s i t i o n with a term value of 6.3 eV whereas the spectrum of C 2H 6 only shows Rydberg tr a n s i t i o n s [68]. The difference between v i r t u a l valence and Rydberg o r b i t a l s and t h e i r extent i s r e a l l y emphasized i n molecules i n which "inner-well" and "outer-well" states can e x i s t . These "well" states a r i s e from the existence of some sort of p o t e n t i a l b a r r i e r (see subsequent section) and can lead to the enhancement of "inner-well" (valence) t r a n s i t i o n s at the expense of t r a n s i t i o n s to "outer-well" (Rydberg) states. An example i s provided by the S 2p spectrum of SF g [69,70]. This aspect w i l l be discussed i n greater d e t a i l i n section F. The extent of the Rydberg l e v e l s compared to valence l e v e l s i s also well i l l u s t r a t e d by comparing absorption spectra of gaseous and condensed phases [12]. The v i r t u a l valence o r b i t a l s , being l o c a l i s e d around the molecular framework w i l l only be s l i g h t l y perturbed on going from the gas to a s o l i d state. However, the Rydberg l e v e l s w i l l be very much affected and the features i n the spectrum w i l l not be v i s i b l e . F r i e d r i c h et a l . [65] have used - 38 -t h i s method to investigate the valence-Rydberg character of the 2p absorption spectrum of SiH^ amd PH 3. This technique provides one means of ascertaining Rydberg or valence character i n spectra where the features may be overlapped. II) Continuum Features In inner s h e l l electron e x c i t a t i o n spectra, features can often be seen above the i o n i s a t i o n edge i n the continuum. The features can a r i s e from various processes including double e x c i t a t i o n , onsets of "shake-up" continua and t r a n s i t i o n s to quasi-stationary state i n the continuum. This f i n a l process can give r i s e to prominent, l o c a l i s e d structure(s) which can be observed as much as ~20eV above the i o n i s a t i o n edge and can be described i n terms of the out-going electron being trapped i n a quasi-stationary state by some form of p o t e n t i a l b a r r i e r . Such a b a r r i e r would i s o l a t e the v i r t u a l l e v e l from the i o n i s a t i o n continua thereby giving the state an increased l i f e t i m e . A good example i s i n the ISEELS [69] or photoabsorption [70] spectrum of the S 2p region i n SFg. This shows two prominent features above the i o n i s a t i o n l i m i t which * can be assigned i n t„ and e a l e v e l s or a l t e r n a t i v e l y to t , and e 2g g 2g g shape-resonances. As t h i s topic i s going to be discussed i n more d e t a i l i n section F, no further mention w i l l be made of i t here. Other features i n the continuum w i l l r e s u l t from mu l t i - e l e c t r o n e x c i t a t i o n and/or i o n i s a t i o n . Double e x c i t a t i o n involves the simulta-neous e x c i t a t i o n of the core-electron along with a valence e l e c t r o n . In e f f e c t i t i s the ISEELS or photoabsorption analogue of "shake-up" i n XPS (see section B.III). These features can be quite prominent and have - 39 -been observed, for example, i n the inner s h e l l electron e x c i t a t i o n spectra of it-bonding systems such as 0 2 ^ , C 2H 2, C gH 6 [68], CO [4,45,71] and N 2 [4,71]. Simultaneous e x c i t a t i o n and i o n i s a t i o n (ISEELS or photoabsorption analogue of XPS "shake-off") w i l l r e s u l t i n an ion with the same configuration as an XPS "shake-up" state. Thus the ( v e r t i c a l ) features appearing i n XPS s a t e l l i t e spectra assigned to "shake-up" tr a n s i t i o n s ( i . e . i o n i s a t i o n and exci t a t i o n ) w i l l be manifested i n ISEELS/photoabsorption spectra as (adiabatic) onsets of "shake-up" continua. I l l ) Comparison of Inner S h e l l E x c i t a t i o n Spectra with Valence S h e l l E x c i t a t i o n Spectra Inner s h e l l electron e x c i t a t i o n spectra are generally r e l a t i v e l y simple to assign due to the energy i s o l a t i o n of the i n i t i a l core hole which unambiguously defines the i n i t i a l o r b i t a l from which the t r a n s i -t i o n a r i s e s . This i s i n contrast to the much more complex s i t u a t i o n that w i l l usually e x i s t i n valence s h e l l spectra due to the numerous c l o s e l y spaced valence o r b i t a l s . Thus Inner s h e l l spectra can often y i e l d more d e f i n i t e information on the previously unoccupied l e v e l s and this i n turn can be used to help c l a r i f y the assignments i n the more complex valence s h e l l spectra. Obviously there w i l l be differences between the spectra of the two s p e c t r a l regions ( i . e . core and valence) i n addition to the over-lapping t r a n s i t i o n s often present i n valence electron e x c i t a t i o n spectra. F i r s t consider the promotion of an electron to a v i r t u a l valence o r b i t a l . On the promotion of an inner s h e l l electron the f i n a l - 40 -state w i l l have a l o c a l i s e d core hole whereas the promotion of a valence electron w i l l r e s u l t i n a more delocalised valence hole. Therefore the newly promoted electron w i l l see almost a whole extra unit of charge i n the inner s h e l l case since the loss of shielding of the nucleus should be greater than that from a delocalised valence hole. Thus the term value ( i . e . binding energy) from the inner s h e l l spectrum should be larger than the term value from the valence s h e l l spectrum. In essence one i s saying that upon the creation a core hole, the occupied valence l e v e l s (including the newly occupied l e v e l ) w i l l relax more than i n the case of the creation of a valence hole. In contrast to the valence o r b i t a l s , the Rydberg o r b i t a l s are large and d i f f u s e . Therefore a Rydberg electron w i l l see one large core with unit p o s i t i v e charge. It w i l l be less influenced by whether the hole i s i n the valence s h e l l or at the core. Consequently there w i l l be l i t t l e increase i n term value on going from the valence s h e l l spectrum to the inner s h e l l spectrum for the Rydberg t r a n s i t i o n s . Wight and Brion [72] have noted differences of upto 0.5 eV only i n Rydberg term values whereas differences well i n excess of 2 eV can occur with v i r t u a l valence l e v e l s , as i s also c l e a r l y seen i n the present work. Knowledge of the term values from inner s h e l l spectra w i l l thus provide upper bounds for term values for valence s h e l l spectra and also help i d e n t i f y Rydberg and valence t r a n s i t i o n s r e s p e c t i -v e l y . C l e a r l y the above discussion assumes that a one electron picture i s adequate i n both cases to describe the t r a n s i t i o n s between the l e v e l s . While t h i s i s probably s a t i s f a c t o r y for inner s h e l l e xcitations - 41 -i t might not be for valence electron excitations and e s p e c i a l l y so for valence-valence t r a n s i t i o n s [13]. A t r a n s i t i o n from one degenerate l e v e l to another l e v e l which i s also degenerate w i l l r e s u l t more than one state. For valence-valence t r a n s i t i o n s the energies may be very d i f f e r e n t . For instance the f i r s t it -»• it t r a n s i t i o n i n benzene, which i s e, -*• e„ , r e s u l t s i n three state, - , ^B, and 1E. with lg 2u 2u l u l u energies of 4.8, 6.2 and 6.7 eV r e s p e c t i v e l y [13]. F. P o t e n t i a l Barrier and Shape-Resonance E f f e c t s . I) P o t e n t i a l B a r r i e r s For c e r t a i n molecules an "anomolous" i n t e n s i t y d i s t r i b u t i o n i s seen i n the d i s c r e t e and/or continuum region of the inner s h e l l e l e c t r o n e x c i t a t i o n spectra. This manifests i t s e l f i n an enhanced p r o b a b i l i t y of core to valence d i s c r e t e t r a n s i t i o n s at the expense of t r a n s i t i o n s to the Rydberg l e v e l s . Furthermore other strong and somewhat broad l o c a l i s e d features are often seen above the i o n i s a t i o n edge. These e f f e c t s were f i r s t noted for molecules containing highly electronegative ligands [73]. It was postulated that a strong repulsive force would act on the escaping electron near the electronegative ligands and so an e f f e c t i v e (charge) p o t e n t i a l b a r r i e r would ex i s t i n the v i c i n i t y of the ligands [73]. The b a r r i e r would separate the f i e l d Into an "inner-well" region and an "outer-well" region. The e f f e c t s of such a b a r r i e r are well i l l u s t r a t e d by the inner s h e l l electron e x c i t a t i o n spectra of SF g [69,70] which provide, perhaps, the clearest example of such e f f e c t s . The ISEELS spectra for the - 42 -F Is, S 2s and 2p regions are shown i n F i g . 1.3 (taken from r e f . [6]). Within a minimal basis set (s and p o r b i t a l s ) t r a n s i t i o n s to two v i r t u a l valence l e v e l s of a ^ and symmetries would be expected. Addition of S 3d o r b i t a l s would extend these to include two further l e v e l s of t„ -g and e symmetry re s p e c t i v e l y . The S 2p spectrum shows a strong s absorption i n the disc r e t e region which can be assigned to a t r a n s i t i o n to the v i r t u a l valence l e v e l . This i s followed by very weak structure assigned to a mixture of overlapping Rydberg t r a n s i t i o n s and a (dipole forbidden) t r a n s i t i o n to the l e v e l (note that both the F Is and S 2s spectra show a strong feature corresponding to an allowed t r a n s i t i o n to t h i s l e v e l ) . Two very prominent features are seen above the edge. These can be ascribed to t r a n s i t i o n s to the d - l i k e states of t£g and e^ symmetries. I t should be noted that the photoabsorption spectrum of the L s h e l l of SF 6 i s i d e n t i c a l (within experimental error) for both the gas and s o l i d phase [70] thereby providing evidence that the t r a n s i t i o n s described above are to f i n a l state l e v e l s within the molecular core i . e . inner-well (valence) states and not Rydberg states. Thus the e f f e c t of the p o t e n t i a l b a r r i e r i s seen to be two-fold. F i r s t i t can support quasi-stationary states above the i o n i s a t i o n edge which are e f f e c t i v e l y decoupled from the continuum by the b a r r i e r . Secondly, the v i r t u a l states within the inner-well w i l l have a strong overlap with the i n i t i a l state wave-function r e s u l t i n g i n s p e c t r a l features with greatly enhanced i n t e n s i t i e s . Conversely, the outer-well states ( i . e . d i f f u s e Rydberg o r b i t a l s ) , being i s o l a t e d from the molecu-l a r core by the b a r r i e r , w i l l have l i t t l e overlap with the i n i t i a l state - 43 --10 0 1-oH A ii IS / 0-5-1g Hu (0 • a m o o o Lit I -< CC H z D o o 0-30-20-10-120H 8 0 H 40H 1 r 6 8 0 Liu — i 1 r 230 / \ (R y d.) a 1Q •10 1 I .' 1 •' \ \ 1 \ +20 +30 eV i ' 1 r F 1S •2g 1 r 7 0 0 2g e g -i 1 1 r 7 2 0 eV S 2 s 2 5 0 •2g — i 1 r 2 7 0 eV S 2 p 170 i r 190 1 r 210 eV Figure 1.3 Energy loss spectra of SF g i n the F Is, S 2s and S 2p e x c i t a t i o n regions (taken from r e f . [6]). - 44 -wave function and t h i s w i l l lead to a d r a s t i c reduction i n the corres-ponding t r a n s i t i o n p r o b a b i l i t i e s . Features i n inner s h e l l electron e x c i t a t i o n spectra such as those described above for SF 6 are also present i n the spectra of molecules such as N 2 and CO [4,71], Obviously a b a r r i e r formed by a repulsive i n t e r a c t i o n of the electronegative ligands on the escaping electron (as postulated for SF 6) i s not applicable i n t h i s case and an a l t e r n a t i v e model i s required. Such a model has been proposed by Dehmer et a l . [74-77] involving a shape-resonance des c r i p t i o n . II) Shape Resonances The N Is ISEELS spectrum of N 2 [4,71] i s dominated by a very intense t r a n s i t i o n at 401.10 eV which has a term value of 8.8 eV with respect to the N Is i o n i s a t i o n edge (IP = 409.9 eV) and i t may be assigned to a Is -*• n t r a n s i t i o n . A few, very weak features a t t r i b u t -able to Rydberg t r a n s i t i o n s are also present i n the d i s c r e t e region. At about 9 eV above the edge ( i . e . at ~419 eV) a broad r e l a t i v e l y intense maximum i s seen [4,71]. Thus e f f e c t s are occurring i n the inner s h e l l spectrum of N 2 s i m i l a r to those observed for SFg, i . e . enhancement of i n t e n s i t y for t r a n s i t i o n s to v i r t u a l valence l e v e l s at the expense of those to the Rydberg l e v e l s , and also prominent features i n the c o n t i -nuum. Dehmer and D i l l [74,75] have described these observations i n terms of a c e n t r i f u g a l b a r r i e r which again separates the molecular f i e l d into inner-well and outer-well regions. The strong feature at 401.10 eV and the continuum feature are then assigned to shape-resonances caused - 45 -by the s c a t t e r i n g of the photoelectron ( i . e . the excited electron) by the anisotropic molecular f i e l d . A shape-resonance i s a quasi-stationary state which can be supported i n an inner-well p o t e n t i a l . The photoelectron ( i n t h i s case) can be temporarily trapped at a p a r t i c u l a r resonance energy before tunnelling through the p o t e n t i a l ( c e n t r i f u g a l ) b a r r i e r and escaping. Thus a b a r r i e r concept i s s t i l l needed. In t h i s case the b a r r i e r i s supplied by a competing i n t e r a c t i o n between the repulsive c e n t r i f u g a l terms (~Jl( A+l)/r 2, where r i s measured from the molecular centre) and the a t t r a c t i v e coulomb term ( ~ r - 1 , where r i s measured from outer nucleus) i n the t o t a l e f f e c t i v e p o t e n t i a l [77,78]. This balance can r e s u l t i n a c e n t r i f u g a l b a r r i e r at the periphery of the molecule. With the correct dimensions t h i s well can support quasi-stationary states which are eigenfunctions of the p o t e n t i a l and l o c a -l i s e d within the molecular core. Using an X^ multiple scattering approach, Dehmer and D i l l [74,75] have calculated the p a r t i a l cross-sections of the four dipole allowed e x c i t a t i o n channels (it , it , a and o ) for the N Is e x c i t a t i o n / i o n i s a -g u' g u' t i o n i n N 2. On t h i s basis the shape resonances observed at 401.10 eV and 419 eV are assigned to the n and channels respectively. A p a r t i a l wave expansion of the n wave function indicates that i t has s i g n i f i c a n t d-wave (Jl=2) character while the o" channel i s seen to have appreciable amounts of p-wave (A=l) and f-wave (A=3) character at low k i n e t i c energies ( i . e . photoelectron energy above the i o n i s a t i o n energy) [75]. E x c i t a t i o n of the N Is electron w i l l produce a p-wave which i s then scattered by the anisotropic molecular f i e l d into the higher A components. These w i l l contribute to the allowed a and it - 4 6 -i o n i s a t i o n channels thereby making the above t r a n s i t i o n s possible. The i n t e n s i t y of the t r a n s i t i o n s can now be explained i n terms of resonant trapping of the photoelectron by the appropriate c e n t r i f u g a l b a r r i e r . The i n t e n s i t y of the t r a n s i t i o n to the u state can be a t t r i b u t e d to the l o c a t i o n of t h i s highly l o c a l i s e d state within a c e n t r i f u g a l b a r r i e r formed by an A=2 e f f e c t i v e p o t e n t i a l . Dehmer and D i l l [75] have noted that t h i s state Is analogous to the n shape-resonance observed at ~2.3 eV i n e-N 2 s c a t t e r i n g experiments [79]. Molecular o r b i t a l c a l c u -l a t i o n s [80] have shown that the e-N 2 resonance i s due to the resonant trapping of an incident d-wave electron which penetrates the A=2 c e n t r i -fugal b a r r i e r and attaches i t s e l f to the LUMO (lowest unoccupied mole-cular o r b i t a l ) it valence o r b i t a l . Since a core-hole would e x i s t i n the * g case of inner s h e l l electron e x c i t a t i o n the shape-resonance would be s h i f t e d to a lower energy and, i n t h i s case, appear In the d i s c r e t e region [81]. The continuum shape-resonance i s explained [74,75] by the trapping of the excited electron by an A=3 c e n t r i f u g a l b a r r i e r i n the 0 u channel. In t h i s case the inner-well p o t e n t i a l can support a quasi-stationary state at ~9 eV above the i o n i s a t i o n edge. In other words the p a r t i a l f-wave (A=3) component of the 0*^ wavefunction can overcome i t s c e n t r i f u g a l b a r r i e r and r a p i d l y penetrate the molecular core thereby giving r i s e to the t r a n s i t i o n at ~419 eV [74,75]. Thus these resonances are a t t r i b u t e d to a high A c e n t r i f u g a l b a r r i e r e f f e c t . I t should be noted that the A=l component of the 0"^ - 47 -channel does not show resonant behaviour [82], i . e . the J!=l p o t e n t i a l i s not s u f f i c i e n t to form an e f f e c t i v e b a r r i e r . C entrifugal b a r r i e r s are well known for c e r t a i n atoms and can lead to prominent features. For example, cerium has a c e n t r i f u g a l b a r r i e r for the A=3 e f f e c t i v e p o t e n t i a l which separates the 4f wavefunction (which i s therefore an inner-well state) from the 5f and higher wavefunctions (outer-well s t a t e s ) . Thus a strong 3d •*• 4f t r a n s i t i o n i s seen whereas t r a n s i t i o n s to the higher f l e v e l s are suppressed [11,73]. For l i g h t e r atoms (eg. N) the wavefunction for the higher JI components are not able to penetrate into the core-region of atoms, however, i n the diatomic molecule the added molecular dimension ( i . e . making the well wider) allows the p o t e n t i a l to support these quasi-stationary states, i . e . i t allows penetration of the d- and f-waves into the molecular core. This dimensional aspect i s emphasized i n that the u channel, which acts perpendicular to the molecular axis, has an Jl=3 component but does not support a resonance [75]. Furthermore, since the a b i l i t y of a system to support a resonance depends on the dimensions of the anisotropic f i e l d , the spectral p o s i t i o n of the resonance should provide a s e n s i t i v e probe of inter-nuclear separation. This aspect w i l l be discussed i n more d e t a i l l a t e r . - 48 -Absolute o s c i l l a t o r strength measurements for the K - s h e l l electron e x c i t a t i o n and i o n i s a t i o n of N 2 have been made by Kay et a l . [71] using ISEELS. An electron impact energy of 8 keV was used to generate the spectrum with the i n e l a s t i c a l l y scattered electrons being samples at zero degree scattering angle. Kay et a l . [71] have compared th e i r r e s u l t s with the c a l c u l a t i o n s of Dehmer and D i l l [74,75]. Only a semi-quantitative agreement i s found between the experiment and the theory. However, a l l the major features are explained except for those a r i s i n g from double-excitation processes which are not accounted for i n the one-electron scattering model of Dehmer and D i l l [74,75]. The ca l c u l a t i o n s overestimate the peak areas of the resonances with the continuum shape-resonance i n p a r t i c u l a r disagreement. The positions are also displaced upwards by ~3 eV i n the c a l c u l a t i o n s . A better agreement i s obtained with the techniques employed by Langhoff and co-workers [83-88]. The resonance phenomena i s described more i n terms of a MO terminology. In this approach, conventional Hartree-Fock c a l c u l a t i o n s u t i l i s i n g gaussian basis sets are performed on the ground state of the molecule. From t h i s c a l c u l a t i o n a non-local p o t e n t i a l for the i o n i s a t i o n channel of i n t e r e s t can be constructed which w i l l be appropriate to describe the motion of the excited e l e c t r o n i n the frozen f i e l d of the remaining N-l electrons. The p o t e n t i a l has the following form: - 4 9 -( N - l ) = N G ( l . F . l ) where J and K are the usual coulomb and exchange operators. The subscript j denotes the doubly-occupied o r b i t a l s while G r e f e r s to the o r b i t a l from which the electron has been excited. This p o t e n t i a l , i n conjunction with the k i n e t i c energy operator (T) and the nuclear frame-work p o t e n t i a l (V) form the necessary Hamiltonian required to describe the p a r t i c u l a r e x c i t a t i o n channel. Consequently the following one-electron SchrOdinger equation can be solved. This r e s u l t s i n the so c a l l e d "improved v i r t u a l o r b i t a l s " which provide v a r i a t i o n a l l y correct approximations for the excited state o r b i t a l s within the approximation of a frozen core [84,89]. The bound functions w i l l provide adequate representations of the discrete e x c i t a t i o n s while the unbound functions provide a pseudo-spectrum of t r a n s i t i o n s which contains a l l the necessary physical information to describe the i o n i s a t i o n continuum. Langhoff and co-workers then use Stieltjes-Tchebycheff moment theory [83,88] to convert the pseudo-spectrum into a correct representation of the o s c i l l a t o r strength. ((T + V + V G ( N - l ) ) - e) - 0 ( 1 . F . 2 ) Using t h i s approach, Rescigno and Langhoff [84] have calculated - 50 -the p a r t i a l photoionisation cross-sections for the four accessible dipole-allowed e x c i t a t i o n channels i n the K-shell electron e x c i t a t i o n spectrum of N 2. They assign the discrete shape-resonance at 401.10 eV to a l a •+ l u core to valence t r a n s i t i o n . The calculated energy i s u g approximately 4 eV too low and the o s c i l l a t o r strength i s overestimated. This i s attributed [84] to the neglect of core-relaxation. The t r a n s i -tions to the Rydberg o r b i t a l s , which are less s e n s i t i v e to r e l a x a t i o n e f f e c t s , agree to within 0.5 eV with experiment. The calculated c r o s s -section for the continuum i s i n better agreement with experiment than that obtained from the multiple scattering c a l c u l a t i o n [74,75]. The continuum shape-resonance i s ascribed to a 1 a -*• 3 a core to valence g u t r a n s i t i o n . Thus the resonances are both attributed to t r a n s i t i o n s to * * v i r t u a l valence o r b i t a l s which can be equated with the it and a a n t i -* bonding o r b i t a l s [84] . In t h i s p a r t i c u l a r case the a o r b i t a l i s i n the continuum. The MO picture and the scattering model present two compliment-ary ways of looking at electron e x c i t a t i o n and the resonance phenomena. The success of both models i n describing the phenomena, at least semi-q u a n t i t a t i v e l y , should not be s u r p r i s i n g i n that the scattering picture places a l o c a l i s e d , quasi-stationary state (resembling a bound state) within the confines of the molecular p o t e n t i a l . As such a MO d e s c r i p -t i o n should also be able to describe the s i t u a t i o n . The r e l a t i o n s h i p between the MO c a l c u l a t i o n and multiple scattering picture for N 2 i s further emphasised i n that the lit and 3o o r b i t a l s correlate with the g u 3dit and 4 f a atomic o r b i t a l s within the united atom l i m i t [84], c o n s i s -- 51 -tent with the d-wave and f-wave natures found i n the x c a l c u l a t i o n a [75]. The above discussion has focussed on N 2. A si m i l a r comparison has been made for CO [71,86]. Consideration of other species, such as SFg and BF 3 lend support to the p a r a l l e l i n t e r p r e t a t i o n of shape-reso-nance phenomena using either a MO or multiple s c a t t e r i n g perspective. For example, the inner s h e l l e x c i t a t i o n spectra of SFg have been i n t e r -preted with LCAO-MO ca l c u l a t i o n s [90] as have those for BF 3 [91,92]. Mul t i p l e scattering c a l c u l a t i o n s have also been performed on the e x c i t a -t i o n spectra of BF 3 [93] and on the analogous e-SF 6 s c a t t e r i n g system [94]. The r e s u l t s of the multiple scattering c a l c u l a t i o n on BF 3 [93] should be p a r t i c u l a r l y noted with regard to the nature of the trapping mechanism i n these molecules. While the angular momenta i s found to be large enough (Jl>2) i n the diatomics to form a c e n t r i f u g a l b a r r i e r s u f f i -cient to support quasi-stationary states, t h i s i s not found i n the case of BF 3 [93]. It has been concluded that the b a r r i e r i n BF 3 r e s u l t s from a combination of c e n t r i f u g a l forces and strong electron repulsion i n the neighbourhood of the ligands [93]. Similar conclusion have been suggested for GeCl^ [95]. Regardless of how the p o t e n t i a l b a r r i e r a r i s e s , features associated with shape-resonances can be recognised i n many systems, as w i l l be seen i n many of molecules studies i n the present work. In l i g h t of the above discussion these shape-resonances can be often i d e n t i f i e d with knowledge of the MO scheme of the system. - 52 -III) Relationship Between Shape-Resonance Po s i t i o n and Bond Length As stated e a r l i e r , the spectral p o s i t i o n of the shape-resonance should provide a probe for inter-nuclear separation. This can be under-stood i n that the b a r r i e r , which w i l l define the inner-well p o t e n t i a l , i s located on the periphery of the molecule and hence r e f l e c t s the mole-cular dimensions. If a s i m p l i s t i c " p a r t i c l e i n a box" analogy i s made, the energies of the stationary states i n such a p o t e n t i a l should vary with 1/R2, where R represents the dimension of the well ( i . e . bond length). A more quantitative way of considering this arises from a multiple scattering treatment. Gustafsson and Levinson [96] have pointed out Siat within a multiple scattering picture, the wave vector of the photoelectron on resonance (k_) should be inversely proportional to the distance (R) between the scattering centres, i . e . k R = constant (1.F.3) r Thus i f the phase-factors can be neglected, i t then follows that the k i n e t i c energy of the photoelectron on resonance should be proportional to 1/R2. Gustafsson and Levinson [96] have plotted the difference of resonance energy from threshold ( 6 ) against 1/R2 for several diatomics and have obtained a reasonable c o r r e l a t i o n . Recently, N a t o l i [97] has demonstrated t h e o r e t i c a l l y , within a complete multiple scattering p i c t u r e , that for c e r t a i n conditions equation (1.F.3) i s v a l i d . More s p e c i f i c a l l y , the wave vector (k_) i s given by [97,98] - 53 -k = / ( 6 - V ) (l.F.A) r o where V q i s a mean intramolecular p o t e n t i a l r e l a t i v e to the vacuum l e v e l [98,99] and the constant depends on atomic phase s h i f t s [97-99]. The phase s h i f t s are energy dependent, however, Bianconi et a l . [98] have pointed out that i f the phase-shift dependence with energy i s smooth then (1.F.3) i s v a l i d . Bianconi et a l . [98] suggest that t h i s should generally be the case for the energy regions where the continuum o shape-resonances are located though not for n resonances [97-99]. The .term V i s p a r t i c u l a r to the system under i n v e s t i g a t i o n -o bond order as well as atomic pair [98-100]. Thus Hitchcock et a l . [100] have simply investigated the r e l a t i o n s h i p of resonance p o s i t i o n and bond length e m p i r i c a l l y . They have shown [100] that for a series of hydrocarbons a simple l i n e a r r e l a t i o n s h i p of the type 6 = aR + b (1.F.5) i s adequate to r e l a t e the known C-C distance to the c shape-resonance p o s i t i o n . Sette et a l . [99] have also applied this l i n e a r r e l a t i o n s h i p (1.F.5) to investigate a shape-resonance positions and bond lengths i n a large v a r i e t y of molecules. Good l i n e a r correlations are seen for classes of molecules characterised by the sum of the atomic numbers of - 54 -the atom pair involved i n the scattering process the Z dependence a r i s e s i n the V q term. As Z increases the a t t r a c t i v e part of the p o t e n t i a l w i l l increase and th i s r e s u l t s i n the d i f f e r e n t l i n e a r dependence observed for shape-resonance p o s i t i o n with bond length for each set [78,99]. The findings of Sette et a l . [99] also indicate that the phase-shift dependence can be assumed constant, at least for Z = 12 -18, for the a resonances. From the above discussion i t i s seen that a convincing argument for shape-resonance p o s i t i o n and bond lengths can be made. This allows a further means of p o s i t i v e l y i d e n t i f y i n g shape-resonance features i n the continuum i n a series of related molecules. Indeed with c a r e f u l considerations of the l i m i t a t i o n s i t should be possible to reverse the procedure to obtain bond lengths [98-100], - 55 -CHAPTER 2 EXPERIMENTAL In t h i s chapter the experimental methods used to obtain the el e c t r o n energy loss spectra presented i n this work w i l l be described. During the course of t h i s work, a d i g i t a l voltmeter (DVM) of high stated accuracy (Datron model 1071 [101]) was obtained. This permitted measurement of a set of accurate reference energies for inner s h e l l electron e x c i t a t i o n spectroscopy. This l a t t e r work i s also presented i n th i s chapter. A. Experimental Methods I) The Spectrometer An e x i s t i n g high r e s o l u t i o n ISEELS spectrometer [6,63] was used for a l l the electron energy loss measurements reported i n the present work with the exception of the valence s h e l l spectra of NF 3 and Si( C H 3 ) 1 + (see l a t e r i n t h i s s e c t i o n ) . The construction and operation of t h i s instrument has already been described i n d e t a i l and so only a b r i e f d e s c r i p t i o n i s presented here. For a f u l l d e scription the reader i s referred to the Ph.D. thesis of A.P. Hitchcock [6] or the de s c r i p t i o n given i n r e f . [63]. A schematic of the spectrometer (taken from r e f . [6]) i s shown in F i g . 2.1. Electrons produced i n the gun region (G) by a heated thoriated tungsten ribbon or wire are accelerated to the desired impact energy ( t y p i c a l l y 2.5 keV) to form a focussed electron beam. The beam Figure 2.1 Schematic of the Inner She l l Electron Energy Loss Spectrometer (taken from ref. [6]). - 57 -i s then retarded by a two-element lens ( L l ) to the required pass-energy of a hemispherical e l e c t r o s t a t i c analyser which acts as the monochro-mator (M). Following t h i s the beam i s accelerated back to the impact energy by a second two-element lens (L2) to pass through the i n t e r a c t i o n region (CC) after which i t i s retarded to the selected pass-energy of a second e l e c t r o s t a t i c analyser (A) by a three-element lens (L3) and transmitted to the detector (CEM). Various d e f l e c t i o n plates (D1-D4) allow minor corrections to the beam path to be made. The beam can be monitored throughout the spectrometer by measuring the current on various apertures (P1-P3) with an electrometer. I n i t i a l l y the instrument i s set up on the primary unscattered beam (zero energy loss) at zero degree scattering angle. This i s done by c a r e f u l l y adjusting the deflector plate voltages and monitoring the beam current on plates P l to P3. The current i s maximised and then minimised on each plate i n turn and then f i n a l l y maximised on the detector cone which, i n t h i s case, acts as a Faraday cup. The middle element of the three-element lens i s set at the voltage which maximises the transmission of the scattered current for the energy loss region about to be studied. Figure 2.2 shows a plot of the experimentally determined optimum focus voltages (with respect to ground) for the middle element of L3 as a function of energy loss for d i f f e r e n t analyser pass energies at 2.5 keV impact energy. Figure 2.3 shows the focus voltage as a function of analyser pass energy for the energy loss appro-priate to the C Is region (~ 290 eV), the N Is region (~ 400 eV) and the 0 Is region (~ 540 eV). - 58 -Pass Energy 3 0 0 4 0 0 5 0 0 6 0 0 ENERGY LOSS (eV) Figure 2.2 Optimum focussing voltages for the middle element of the three-element lens (L3) (see Figure 2.1) as a function of energy loss for different analyser pass energies at 2.5 keV Incident energy. - 59 -N1s i ' i 1 r 1 0 2 0 3 0 PASS ENERGY (eV) Figure 2.3 Optimum focussing voltages for the middle element of the three-element lens (L3) (see Figure 2.1) as a function of analyser pass energy for selected energy losses at 2.5 keV incident energy. - 6 0 -The r e s o l u t i o n of the spectrometer depends on the pass energies of the two analysers. The t h e o r e t i c a l resolution of th i s spectrometer i s given by [ 6 ] ^FWHM = °' 0 1 ^ M + V ( 2-A a ) where V_, and V. are the monochromator and analyser pass energies respec-M A t i v e l y . The actual r e s o l u t i o n can be obtained by measuring the p r o f i l e of the primary beam. When running a sample, part of the primary beam i s scattered e l a s t i c a l l y or i n e l a s t i c a l l y by the gas i n the i n t e r a c t i o n region. To obtain a spectrum a voltage, equivalent to the energy loss corresponding to the i n e l a s t i c s c a t t e r i n g , i s added on top of the small voltage already applied to the complete analyser system. Thus the scattered electrons can regain t h e i r energy loss and be transmitted through the analyser. By using a suitable o f f s e t voltage and scanning the energy loss region of i n t e r e s t a f u l l spectrum can be obtained. A high count rate i s attainable with valence s h e l l electron energy loss spectroscopy and so the spectra can be measured at zero degree scattering angle. However, i t i s not possible to obtain inner s h e l l spectra at zero degree scattering angle on th i s spectrometer due to the r e l a t i v e l y low count rate r e l a t i v e to the small but not i n s i g n i -f i c a n t background produced by the backscattering of the intense main beam on the analyser surface. To avoid t h i s , the main beam i s passed - 61 -through the centre of the gas c e l l at a small angle by use of a "double-d e f l e c t i o n " system (DD). This system consists of two sets of d e f l e c t i n g p l a t e s , operating i n the energy dispersing planes of the analysers, whose f i e l d s act i n opposite d i r e c t i o n . The plates i n the second set are twice the length of the plates i n the f i r s t set and are placed equidistant from the f i r s t set of plates and the centre of the gas c e l l . The f i e l d s are generated by a single voltage source. Thus the beam i s deflected by the f i r s t set of plates i n one d i r e c t i o n through angle 0 and then deflected back i n the opposite d i r e c t i o n by the second set of plates through angle 28. The net r e s u l t i s that the beam passes through the centre of the gas, where the concentration of gas i s the greatest, at ^ n angle 0. Table 2.1 shows the d e f l e c t i o n angle (9) obtained at incident beam energies of 1.5 keV and 2.5 keV for various voltages applied to the "double-deflection" system. T y p i c a l operating conditions employed i n the present work are 2.5 keV impact energy and a scattering angle of ~ 1°. This angle, while small enough to ensure dipole-dominated spectra, allows interception of the main beam by plate P3 before i t can reach the analyser. During the set-up procedure and for valence-shell spectra the long plates are grounded out and the short plates (along with t h e i r perpendicular counterparts) act i n the same manner as the other (x,y) def l e c t o r plates. II) Sample-Handling A second sample i n l e t system, completely constructed of s t a i n -less s t e e l , was added to the o r i g i n a l brass i n l e t system [6] which was - 62 -TABLE 2.1: D e f l e c t i o n angle (9) for various voltages (V y) applied to the double-deflection system for impact energies ( E Q ) of 1.5 keV and 2.5 keV' Voltage across plates (V y) Deflection angle (9) for impact energies (E Q) 1.5 keV 2.5 keV 52 0.90 0.54 106 1.84 1.10 161 2.79 1.68 218 3.78 2.27 277 4.80 2.88 Deflection angle given by 9 = t a n - 1 V y«x where x = 1.0 cm, length of short plate y = 1.1 cm, gap between p a r a l l e l plates (see any introductory physics text eg. F.W. Sears, M.W. Zemansky and H.D. Young, "University Physics", Addison Wesley (1980) p. 450-451). - 63 -retained for the reference gas l i n e . The gas pressure i n the new system was co n t r o l l e d by means of a G r a n v i l l e - P h i l l i p s s e r i e s 203 leak valve while that from the reference system was controlled with a Varian leak valve (model 951-5100). The two l i n e s (sample and reference) were connected on the low pressure side of the leak valves p r i o r to being fed into the gas c e l l . The ambient pressure of the spectrometer, which was monitored by an i o n i s a t i o n gauge, was allowed to r i s e from a base pressure of 4 x l 0 ~ 7 t o r r to 5xl 0 ~ 5 t o r r on sample introduction. Under these conditions only single scattering processes are observed [102]. A l l the samples were obtained commercially and were of high stated p u r i t y . The gas samples were taken d i r e c t l y from the cylinder using the appropriate regulator and a l l the connections made to the i n l e t system with 1/4" tubing (copper or s t a i n l e s s s t e e l ) and "swagelock" f i t t i n g s . L i q u i d samples were transferred from t h e i r container to an evacuated glass v i a l equipped with a t e f l o n valve to which was also attached a short piece of 1/4" diameter glass tubing thereby allowing the sample to be connected d i r e c t l y to the "swagelock" f i t t i n g s of the i n l e t system by using t e f l o n f e r r u l e s . The l i q u i d samples were degassed by repeated freeze-thaw cycles. Valence-shell spectra were run to check that the samples were free of any obvious v o l a t i l e impurities and also to ensure that the system was a i r t i g h t , the l a t t e r being an e s p e c i a l l y important consideration for inner s h e l l spectra i n the N and 0 K-shell regions. A leak could e a s i l y be detected by observing the intense N 2 (X •* b 1!! ) valence s h e l l feature at 12.93 eV [103]. From i t s valence-shell spectrum, the sample of PC1 3 was seen - 64 -to contain a small amount of HCl impurity which was removed by continuous pumping on the sample cooled down with a dry-ice/methanol mixture. A s i m i l a r process was also performed with O P C I 3 as a precautionary measure, even though no HCl was apparent i n the spectrum. No further p u r i f i c a t i o n was performed on any of the samples since the spectra indicated that they were e s s e n t i a l l y free of impurities. I l l ) Spectral A c q u i s i t i o n , C a l i b r a t i o n and Spectrometer Performance The spectra were obtained i n the following manner. The gaseous sample (or sample plus c a l i b r a n t ) was fed into the spectrometer which was set up i n the manner described above for the required spectral reso-l u t i o n . I t was always necessary to retune the spectrometer upon sample introduction. In p r a c t i c e , i t was found that the actual r e s o l u t i o n was very close to the t h e o r e t i c a l resolution (see equation (2.A.1 ) ) for pass energies > 5 eV, however, for high resolution valence s h e l l studies the resolution was best obtained by measuring the He(I) resonance l i n e at 21.218 eV [104], The spectral region of i n t e r e s t was selected by adding a voltage, corresponding to the required i n i t i a l energy, on top of the voltage already applied to the analyser system. The region was then successively scanned by voltage programming a power supply (Kepco P X 1 0 0 ) which was i n series with the energy-loss/analyser power supply (Fluke 410B), using the ramp output from the multichannel analyser (Fabritek 1064). However, for the continuous wide range scan shown i n Chapter 4, the ramp voltage was monitored by a Digitec voltmeter and i t s mechanical readout was connected to the shaft of a potentiometer which - 65 -could resistance programme the energy-loss/analyser power supply (see r e f . [6] for f u l l d e t a i l s ) . The voltage applied to the analyser system was measured i n the majority of cases with the Datron 1071 DVM, however, for a few of the spectra a Data P r e c i s i o n 3500 DVM was also used. The pulses from the detector (channeltron) were processed by standard pre-amp/amplifier/discriminator units and the output s i g n a l was stored i n the multichannel analyser. The signa l could also be monitored vi a a ratemeter. The channel address advances synchronously with the ramp voltage output and so a complete spectrum can be signa l averaged. The spectral range was determined by channel step size and the number of points. Any number up to 1000 points could be selected with the aid of a program control unit (Ortec 4610). Spectral a c q u i s i t i o n time depended on the number of points, r e s o l u t i o n and the energy region studied (due to the E ~ 3 drop-off (see equation (l.C.17)). A low r e s o l u t i o n (0.35 eV) inner s h e l l spectrum t y p i c a l l y took between 6 - 1 2 hours for the long range spectra i n the 100 ~ 350 eV energy loss regions while those for higher energy losses (N Is, 0 Is and F Is) required c o l l e c t i o n times between 24 and 48 hours. A c q u i s i t i o n times for the high r e s o l u t i o n i n n e r - s h e l l spectra varied between 12 and 48 hours whereas high r e s o l u -ti o n valence s h e l l spectra could be obtained i n a matter of minutes though the spectra were t y p i c a l l y run for 1 to 2 hours to optimise the signal/noise r a t i o . The data, once obtained, was plotted on a X-Y point p l o t t e r (Hewlett Packard 7004B). A l l measurements were taken from t h i s p l o t . For data manipulation (background subtraction, scale expansion etc.) the - 6 6 -data was transferred to f i l e s on the UBC computer by use of a d i g i -t i s e r . In p r i n c i p l e , once the spectrometer i s set-up on the primary beam for zero energy l o s s , a l l that i s required to put a spectrum on an absolute scale i s an accurate reading of the voltage added onto the analyser system (see next s e c t i o n ) . However, since the spectrometer was set-up using the cone as a Faraday cup and observing the (analogue) current on an electrometer whereas the spectra were run i n a pulse count mode, i n p r a c t i c e , the spectra obtained i n this work were put on absolute scales by comparison with a known feature run under the same operating conditions. The process involved c a l i b r a t i n g a feature i n the sample spectrum by using an external reference and then using the newly established standard to i n t e r n a l l y c a l i b r a t e the rest of the sample spectral features i n other energy ranges. To ensure the same operating conditions when using an external reference, reference and sample were run as a mixture. The reference was chosen so that i t s s p e c t r a l features and those of the sample did not overlap. The c a l i b r a t i n g feature (external or i n t e r n a l ) was run both before and a f t e r the feature being c a l i b r a t e d to ensure that there was no voltage d r i f t . This procedure was repeated several times. Since the count-rate drops approximately as the inverse-cube of the energy l o s s , the middle element of the three element lens (L3) was set for the feature with the largest energy loss so as to maximise i t s count rate. S p e c i f i c c a l i b r a t i o n d e t a i l s for the inner s h e l l spectra are presented i n the appropriate chapters. The valence s h e l l spectra were a l l c a l i b r a t e d against the He(I) resonance l i n e (21.218 eV - 67 -[104]) except for PC1 3 i n which case the HC1 impurity was used to c a l i -brate the spectrum before i t was pumped away. Most of the compounds presented i n t h i s work quickly degraded the high res o l u t i o n performance c a p a b i l i t i e s of the spectrometer. The spectrometer consists of only one chamber pumped by a single d i f f u s i o n pump with no d i f f e r e n t i a l pumping systems to i s o l a t e the gas c e l l from the rest of the spectrometer or the electron gun. Reaction of. reactive gases with the gun filament produced a considerable amount of contamina-t i o n which affected the c h a r a c t e r i s t i c s of the electron beam and thus frequent cleaning of the spectrometer was necessary to bring i t back to optimum performance. The effectiveness of t h i s can be seen i n the valence shell* and inner s h e l l spectra of N 2 which i l l u s t r a t e the capa-b i l i t i e s of the spectrometer. Figure 2.4 shows the N 2 (X •*• b^II ) valence-valence transiton obtained with a spectral r e s o l u t i o n of 0.017 eV which compares very favourably with the spectrum obtained at s l i g h t l y higher res o l u t i o n by Geiger et a l . [103b]. This feature provides a stringent test for the c a p a b i l i t i e s of the spectrometer as does the N 2 (Is •*• TI ) core-valence t r a n s i t i o n shown i n the next section ( F i g . 2.5). IV) Other Measurements The valence s h e l l spectra of NF 3 and S i ( C H 3 ) ^ presented i n t h i s work were obtained by Dr. Suzannah Daviel on a new ISEELS spectrometer which has recently come into operation i n t h i s laboratory. This new instrument [53], though s i m i l a r i n p r i n c i p l e to the ISEELS spectrometer used i n the present work employs a number of s i g n i f i c a n t new features including / N 2 valence AE=0.017eV ON 0 0 0 -12.4 12.6 12.8 13.0 ENERGY LOSS(eV) 13.2 Figure 2.4 High resolution electron energy loss spectrum of N 2 i n the region of the X + b 1 ^ t r a n s i t i o n . - 69 -1) separate d i f f e r e n t i a l pumping of the gun, monochromator, i n t e r a c t i o n and analyser regions 2) large radius (20 cm) hemispherical analysers and 3) c a r e f u l l y designed and highly e f f i c i e n t electron o p t i c s . These features have produced s i g n i f i c a n t improvements i n r e s o l u t i o n , i n t e n s i t y and s t a b i l i t y as well as permitting a background free operation at 6 = 0°. This l a t t e r point i s not only s i g n i f i c a n t for inner s h e l l spectra but also for valence s h e l l spectra as i t sometimes proved very d i f f i c u l t to tune out "ghosting" e f f e c t s produced by the back scattering of the main beam i n valence s h e l l spectra obtained on the old instrument (see Chapter 7). A f u l l d e s c r i p t i o n of the construction and performance of the new instrument i s given i n r e f . [53]. The XPS "shake-up" spectra of NF 3 presented and discussed i n Chapter 3 were recorded several years ago by the author on a MacPherson ESCA 36 photoelectron spectrometer situated at the University of Alberta. The experimental d e t a i l s are given i n Chapter 3. B. Reference Energies for Inner S h e l l E lectron Energy Loss Spectroscopy Spectral measurements of atomic and molecular energy l e v e l s generally r e l y on s u f f i c i e n t l y accurately known reference values against which measurements can be cal i b r a t e d and i n this regard ISEELS i s no exception. In this section a set of reference energies determined by ISEELS for the energy range 100 - 1000 eV are presented. I t should be - 70 -noted that electron energy loss spectroscopy (EELS) provides an excellent d i r e c t means of measuring the energies of inner s h e l l t r a n s i -tions since, due to the inherent nature of the technique i t s e l f , i t only involves the d i r e c t measurement of a voltage or a voltage d i f f e r e n c e . This i s i n contrast to o p t i c a l methods which are i n d i r e c t since they r e l y on a grating equation which i s l i a b l e to lead to greater energy errors at shorter wavelengths (higher energies). This i s well i l l u s -trated i n the case of the chlorine L - s h e l l e x c i t a t i o n spectrum of HC1 where two sets of independent ISEELS measurements have recently conclu-s i v e l y shown [105,106] that the energy scales of e x i s t i n g o p t i c a l absorption spectra are i n error by ~ 0.5 eV. A major source of error i n determining higher energy l e v e l s i n electron spectroscopy p a r t i c u l a r l y EELS i s the often l i m i t e d accuracy of commonly ava i l a b l e d i g i t a l voltmeters (DVM). For example, i f a DVM of only 4-1/2 d i g i t c a p a b i l i t y was used to measure a voltage, up to say 1000 eV, the reading alone could only be determined to ±0.1 eV. With a t y p i c a l manufacturer's quoted accuracy of ±0.1% the reading at 1000 eV would be no more accurate than ±1 v o l t ! As a r e s u l t even though somewhat more accurate DVM's are usually used many published excited state and i o n i s a t i o n energies maybe of somewhat lim i t e d accuracy. An examination of the l i t e r a t u r e reveals inconsistencies i n some published values p a r t i c u l a r l y i n XPS measurements. However, Lee [107] has shown that i n the case of XPS non l i n e a r i t i e s i n spectrometer energy response are the p r i n c i p a l determining f a c t o r . However, i n ISEELS spectrometers run at constant pass energies the considerations are d i f f e r e n t and DVM - 71 -accuracy can be a major fa c t o r . In order to make s u f f i c i e n t l y accurate measurements, up to 1000 eV, at least 6-1/2 d i g i t c a p a b i l i t y i s desirable with an accuracy of at least ±0.001% or ±0.01 eV. Such instruments are not ro u t i n e l y employed on most electron spectrometers. I f highly accurate c a l i b r a t i o n values are a v a i l a b l e then new inner s h e l l spectra may be put on accurate energy scales by measuring with reference to these c a l i b r a t e d l e v e l s using mixed sample gases. Provided that the calibr a n t i s close by i n energy then a DVM of lower accuracy than that used for the o r i g i n a l c a l i b r a t i o n determination w i l l s u f f i c e to obtain a s u f f i c i e n t l y accurate energy scale. It i s the purpose of the work presented i n t h i s section to provide such a range of reference c a l i b r a t i o n values for this purpose. Of the r e l a t i v e l y few commercially a v a i l a b l e DVM which come near to the sp e c i f i c a t i o n s needed a DATRON model 1071 DVM was selected and the values reported here were measured with t h i s instrument. The DATRON [101] model 1071 6-1/2 d i g i t DVM has a stated accuracy of ±0.001% (90 day) or ±0.002% (1 year) i n normal operation. In addition an averaging mode i s selectable which y i e l d s 7-1/2 d i g i t c a p a b i l i t y and then the accuracy i s claimed to be further increased by a factor of two. The DVM also features a u t o - r e c a l i b r a t i o n v i a a b u i l t - i n microprocessor which compensates for short-term aging. The measurements i n the present work were completed within 90 days of a c e r t i f i e d factory c a l i b r a t i o n to the above s p e c i f i c a t i o n s . The energy losses corresponding to several known inner s h e l l e l e c t r o n i c t r a n s i t i o n s have been redetermined using the ISEELS spectro-- 72 -meter described i n the previous section i n conjunction with the DATRON model 1071 DVM. Each of the t r a n s i t i o n s was measured at several d i f f e -rent energy resolutions to check for any change i n peak envelope or p o s i t i o n . This i s an e f f e c t i v e test since change of resolutiom changes the pass energy of the analyser and the voltages on the three-element lens i n the energy loss part of the spectrometer. The fact that peak energies are independent of r e s o l u t i o n (Table 2.3) shows that any changes i n t r a j e c t o r y i n the analyser do not a f f e c t the energy scale within the stated u n c e r t a i n t i e s . In choosing suitable t r a n s i t i o n s the following requirements were considered. (1) Possession of a s u f f i c i e n t l y intense, sharp and d i s t i n c t feature i n the energy region of i n t e r e s t . The sharpness of s p e c t r a l l i n e s i s inc r e a s i n g l y l i m i t e d by natural width considerations as the atomic number increases. (2) Ready a v a i l a b i l i t y of the c a l i b r a t i n g atom or molecule. (3) Ease of sample handling and introduction. (4) Inertness of the substance with respect to decomposition on the hot cathode or on the spectrometer and i n l e t surfaces. (5) A range of su i t a b l y spaced t r a n s i t i o n s to span the 100 - 1000 eV region. I t should be noted that i n EELS the i n t e n s i t y i n a spectrum varies approximately as the inverse cube of the energy loss (see equa-t i o n l.C.17) and hence t r a n s i t i o n s at higher energy losses have an i n t r i n s i c a l l y lower count rate. This and the fact that higher r e s o l u -- 73 -t i o n i s obtained at the expense of decreased count rate means that a compromise between count rate and res o l u t i o n must often be made. However, for ISEELS a s a c r i f i c e i n count rate to obtain high r e s o l u t i o n i s often unnecessary and indeed undesirable since the t r a n s i t i o n s at higher energy losses are s i g n i f i c a n t l y broadened by l i f e t i m e considera-tions (AE*At = 7 x 1 0 - 1 6 eV sec). Thus nothing i s gained i n such cases by taking the spectra at unnecessarily high r e s o l u t i o n . Once the sample has been introduced and the spectrometer set up for zero energy loss a l l that i s required for determining the energy scale i s a s u f f i c i e n t l y accurate reading of the applied voltage to the electron analyser system. In order to ensure that the spectrometer was accurately set up for zero-energy loss the e l a s t i c a l l y scattered peak was measured i n the case of the highest res o l u t i o n (0.070 eV FWHM). It i s possible to do t h i s because the primary unscattered beam does not enter the analyser system. However, at lower resolutions the signal at the channeltron for normal operating conditions was too large to allow d i r e c t measurement of the e l a s t i c peak and so under these conditions the * value obtained for the C0(C Is -»• it (v=0)) t r a n s i t i o n by d i r e c t measure-ment at high res o l u t i o n was used as an a l t e r n a t i v e reference point. In these measurements CO was introduced simultaneously with the other sample gas. A l t e r n a t i v e l y , a well known lower energy valence or core t r a n s i t i o n i n the sample i t s e l f was used as an i n t e r n a l c a l i b r a n t . As can be seen from Table 2.3 (see below) the measured t r a n s i t i o n energy i s independent, within experimental error, of the r e s o l u t i o n , and whether an e l a s t i c or an i n e l a s t i c reference was used. Thus over the range of - 74 -energy resolutions used, which corresponds to normal operations, i t can be concluded that energies are independent of any instrumental response function. * Figure 2.5 shows the v i b r a t i o n a l l y resolved N 2 (N Is n ) t r a n s i t i o n measured at high r e s o l u t i o n . The spectral features are i n excellent agreement with e a r l i e r published works by Hitchcock and Brion [52] and also by King et a l . [51]. This t r a n s i t i o n provides a very stringent test of the performance of the spectrometer as well as being a most useful c a l i b r a t i o n point i n the median of the energy range up to 1000 eV. Nitrogen K s h e l l excited states have a s u f f i c i e n t l y long l i f e -time that the natural l i n e width (~ 0.1 eV, see r e f . [51] and [52]) merits the use of high r e s o l u t i o n and the energy values obtained i n t h i s study as well as those from e a r l i e r work [51,52] are l i s t e d i n Table 2.2. The r e s o l u t i o n achieved here i s comparable to that i n the e a r l i e r work while the s i g n a l to background r a t i o i s somewhat superior. The presently obtained values are seen to be i n excellent agreement with those reported by King et a l . [51] which were also obtained using a highly accurate voltage measuring system. The present values are consistently higher than the e a r l i e r values reported on the same i n s t r u -ment by Hitchcock and Brion [52]. This, as noted below, i s due to the l i m i t e d accuracy of the less sophisticated DVM used i n the e a r l i e r work. Carbon monoxide i s a very suitable secondary standard since, the C Is -*• u t r a n s i t i o n i s dominated by the v=0 transiton [50,52]. Thus there i s no reason to expect any s i g n i f i c a n t s h i f t i n the energy of the 400 401 402 E N E R G Y L O S S ( E V ) 403 Figure 2.5 High resolution electron energy loss spectrum of N 2 i n the * region of N Is + i t e x c i t a t i o n . The s o l i d l i n e through the data points represents the sum of s i x Lorentzian l i n e shapes. - 76 -TABLE 2.2: Transition Energies (eV) in the high resolution N 2 (N Is -*• n ) electron energy loss spectrum Final State Vibrational Level Transition energy (eV) This Work Hitchcock et a l a King et a l b 0 400.88(2) 400.70(5) 400.86(3) 1 401.10(2) 400.93(1) 401.09(1) 2 401.33(2) 401.16(1) 401.31(1) 3 401.56(5) 401.39(1) 401 .54(1) 4 401 .77(7) 401 .60(2) 401.76(1) 5 401.98(10) 401.82(3) 401.98(2) 6 - - 401.19(2) a - reference [52] b - reference [51] ^ - In the present work a l l estimated errors are absolute whereas the errors for v=l and above in references [51] and [52] are relative to the energy for the v=0 peak. - 77 -peak po s i t i o n with changes i n r e s o l u t i o n . The peak value for the v=0 t r a n s i t i o n has been found to be 287.40(2) eV regardless of the r e s o l u -t i o n used. This was established by running a mixture of He and CO with the He(I) t r a n s i t i o n (21.218 eV) [104] being used to set the absolute energy s c a l e . The value obtained for the C0(C Is u (v=0)) t r a n s i t i o n at both high and low resolutions was found to be the same within experi-mental error . Measurements have been made for selected t r a n s i t i o n s i n a series of atoms and molecules whose spectroscopy had previously been studied using ISEELS. Spectral features observed were i n a l l cases I d e n t i c a l to those observed i n the e a r l i e r reported studies of SF 6 [69], Ar [108], CO [50,52], N 2 [51,52] and Ne [109] . Table 2.3 summarises the data obtained for the selected t r a n s i t i o n s at various energy resolutions. I t might be expected that only the SF 6 (S 2p^^ 2 •* t 2 g ) and N 2 (Is ->• * a (v=l)) positions would vary with res o l u t i o n as these alone posses s u f f i c i e n t l y intense and resolvable neighbouring t r a n s i t i o n s ( i n the case of N 2, the other v i b r a t i o n a l components and for SF 6, the S 2p^y 2 spin-orbit component). In fact the separation of the spin-orbit compo-nents i n SF 6 i s large enough that the measured po s i t i o n of the (S 2p^^ 2 •»• c 2 g ) peak maximum does not vary within experimental error over the range of resolutions employed. For N 2 the curve f i t t i n g i n Figure 2.5 shows that there i s no s i g n i f i c a n t s h i f t of peak maxima for v=0, 1 or 2 at high res o l u t i o n due to overlap from neighbouring peaks. The small possible s h i f t i n peak maximum observed with decrease i n re s o l u t i o n (see - 78 -TABLE 2.3: Measured energy levels of the reference lines as a function of resolution Inner-shell Transition Energy lot ;s (eV) at quoted resolutions (eV) tr a n s i t i o n 3 0.070b 0.105 0.140 0.210 0.350 SF 6 S 2 p 1 / 2 * t 2 g - 184.54(5) - - 184.51(5) Ar 2p2/2***s - 244.37(2)c - - -CO C ls-»n*(v-0) 287.40(2) - 287.40(2)d - -N 2 ls*n*(v-l) 401.10(2) - 401.10(4) 401.08(4) 401.05(5) CO 0 ls-m* 534.21(9) - 534.19(10) 534.20(10) 534.12(10) SF 6 F l s * a l g - - - - 688.27(15)e Ne ls-»3p - - - - 867.13(8) a A l l values quoted with the energy loss scale established from the C0(C Is •*• it (v»=0)) transition except where indicated otherwise. The gases were run as mixtures. b Zero-energy loss determined from the elastic peak. c Internal calibration against Ar(I) (11,828(5) eV) - reference [104]. d Measured with respect to He(l) (21.218(1) eV) - reference [104]. e Internally calibrated against SF 6 (S2p 1^ + t2g^* - 79 -Table 2.3) i s consistent with the shape of the i n t e n s i t y d i s t r i b u t i o n i n the broad v i b r a t i o n a l envelope. However i t can be seen that the values quoted for each of the various t r a n s i t i o n s at the d i f f e r i n g resolutions are the same within experimental e r r o r . Thus i t i s possible, with the selected t r a n s i t i o n s , to use the c a l i b r a t i o n values for the peak p o s i -tions i n a given spectrum independent of reso l u t i o n (at least below ~ 0.35 eV FWHM) A summary of the best values obtained here together with l i t e r a t u r e values from both ISEELS [50-52,69,108,109] and photoabsorp-t i o n [110-114] are shown i n Table 2.4 together with the t o t a l estimated unc e r t a i n t i e s . The major possible sources of systematic error i n t h i s work include (a) DVM accuracy (b) differences i n t r a j e c t o r i e s for e l a s t i c and i n e l a s t i c a l l y scattered electrons due to electron o p t i c a l e f f e c t s i n the lens system (c) d i f f e r e n t instrumental response functions f or e l a s t i c and i n e l a s t i c a l l y scattered electrons (d) spectrometer s t a b i l i t y (e) spe c t r a l linewidth. The errors due to (a) should be less than 0.01 eV according to the manufacturers' s p e c i f i c a t i o n s . R e s e t a b i l i t y was not a problem due to the fact that the DVM had 7-1/2 d i g i t s . Errors due to (b) and (c) are considered to be n e g l i g i b l e compared to the stated uncertainties i n Table 2.4 since the measurements at varying r e s o l u t i o n (Table 2.3) show no s i g n i f i c a n t s h i f t s i n measured energies. In t h i s regard i t should also be noted that peak p o s i t i o n s h i f t s were not observed between measurements using e l a s t i c and i n e l a s t i c reference peaks. Change i n re s o l u t i o n of the spectrometer involves change of pass energies i n the analysers, and thus changes i n the decelerating three - 80 -TABLE 2.4: Reference energies for inner shell electron energy loss spectroscopy Inner-shell Transition Transition energy (eV) b ISEELS Optical This work3 Literature Ref Literature Ref SF 6 S 2 p 1 / 2 + t 2 g 184.54(5) 184.27(10) [69] 184.55 [110] Ar 2p3/2 ••• 4s 244.37(2) 244.39(1) [108] CO C Is + n*(v=0) 287.40(2) 287.31(5) 287.40(2) [52] [50] N 2 Is * it*(v=l) 401.10(2) 400.93(6) 401.09(4) [52] [51] CO 0 Is + n* 534.21(9) 534.11(8) [52] 534.2(3) [111] SF 6 F Is - a l g 688.27(15) 688.0(2) [69] 687.5 687.8 [112] [113] Ne Is •+ 3p 867.13(7) 867.05(8) [109] 867.13(5) [114] a Average values taking into account a l l the data from the different resolutions (see Table 2.3). b Errors are shown in brackets e.g. 184.52(6) means 184.52 + 0.06 eV. - 81 -element energy loss lens r a t i o s . This would reveal i f any s i g n i f i c a n t energy scale s h i f t s a r i s i n g from changes i n t r a j e c t o r i e s were occurring. Spectrometer s t a b i l i t y (item (d)) i s of necessity very high by design since long s i g n a l averaged scans are necessary for high r e s o l u t i o n operation [52]. Any d r i f t or i n s t a b i l i t y more than ~ 0.01 eV would be immediately apparent as a loss of r e s o l u t i o n and a b l u r r i n g of peak p o s i t i o n . S i m i l a r l y a l l power supplies were selected with the r i p p l e being less than 0.002 eV peak to peak and the s t a b i l i t y corresponding to better than 0.01 eV. The voltage ramp for the energy loss scans was derived from the Fabritek Multichannel Analyser. The l i n e a r i t y of t h i s ramp over the range employed was within the stated uncertainties. This was established by double checking the energy scale by ( i ) point counting over the number of channels i n question and ( i i ) d i r e c t measurement off the spectra. The s p e c t r a l linewidth (e) for higher energy t r a n s i t i o n s i s often quite large due to natural linewidth (uncertainty p r i n c i p l e ) considerations. This i s r e f l e c t e d i n the l a r g e r uncertainties for the respective peaks. While the present values agree well with some of the previously reported data (see Table 2.4) there are some notable discrepancies which l i e well outside the boundaries of stated experimental error. The e a r l i e r obtained energy values produced using t h i s ISEELS spectrometer [52,69,109] are consistently s l i g h t l y low compared to the present work. However t h i s i s d i r e c t l y a t t r i b u t a b l e to the l i m i t e d accuracy of the less sophisticated 5-1/2 d i g i t DVM used i n the e a r l i e r work since the discrepancies are well within the manufacturer's stated error l i m i t s f or - 82 -the instruments used. Agreement of the present work for N 2, CO and Ar with the other published ISEELS data [50,51,108] i s e x c e l l e n t . The most serious discrepancy (> 0.5 eV) i s with the o p t i c a l values for the F Is e x c i t a t i o n i n SF g [112,113]. In t h i s regard i t should be noted that the separations between the ISEELS values for SF 6 (S 2 p ^ 2 t^) and SF 6(F Is •*• a, ) are i d e n t i c a l i n the present and e a r l i e r [69] ISEELS work. l g The electron energy loss measurement i s much more d i r e c t than the energy c a l i b r a t i o n of o p t i c a l instruments p a r t i c u l a r l y at short wavelengths. This together with the generally excellent self-consistency of the other measurements lends confidence to the present ISEELS measurements for SFg. It i s possible that the o p t i c a l peak shape and therefore peak pos i t i o n could be d r a s t i c a l l y affected by l i n e saturation e f f e c t s which can occur i n o p t i c a l spectra due to the resonant nature of the t r a n s i t i o n [39,115,116]. Thus a consistent set of c a l i b r a t i o n energies has been obtained i n the energy loss range below 900 eV. The set of values (Table 2.4) provides convenient reference points for c a l i b r a t i o n purposes and these values are used throughout the present work. - 83 -CHAPTER 3 INNER SHELL EXCITATION, VALENCE EXCITATION AND CORE IONISATION IN NF, STUDIED BY ELECTRON ENERGY LOSS AND X-RAY PHOTOELECTRON SPECTROSCOPIES In t h i s chapter the ISEELS and XPS spectra of the N Is and F Is regions of NF 3 are presented and examined i n d e t a i l together with the VSEELS spectrum. The information obtained by each technique i s d i f f e r e n t yet complimentary and the i n t e r p r e t a t i o n of the r e s u l t s should be f a c i l i t a t e d by a j o i n t consideration of a l l three spectroscopies (VSEELS, ISEELS and XPS) together with previously published photoelectron data [117-119]. The VSEELS spectrum of a molecule i s often complex and ambiguous i n i t s assignment because of the p o s s i b i l i t y of many overlapping t r a n s i t i o n s a r i s i n g from the close proximity and number of valence o r b i t a l s as well as the numerous manifolds of Rydberg l e v e l s . The ISEELS spectra, however, are generally r e l a t i v e l y simple and s t r a i g h t -forward to assign since the i n i t i a l core l e v e l i s usually unambiguous, being well separated i n energy from other l e v e l s . Thus ISEELS can often give more d e f i n i t e information on the previously unoccupied o r b i t a l and t h i s i n turn may be useful i n c l a r i f y i n g the assignments i n the VSEELS spectrum. This depends on the extent to which term values are transferable and an attempt to address t h i s consideration i s made i n t h i s Chapter. The XPS spectrum provides information on the excited ion - 84 -states and and i n p a r t i c u l a r gives the v e r t i c a l energy for the i o n i s a t i o n processes. The production of these excited ion states would be expected to be manifested i n ISEELS by onsets (adiabatic energies) of new continua. Thus the XPS s a t e l l i t e data provides information on which of the features i n the ISEELS spectrum above the i o n i z a t i o n edge a r i s e from the onsets of excited ion states and by inference which features are due to other types of phenomena such as shape-resonances, or "inner w e l l " states trapped i n the continuum [77], The molecule NF 3 i s the f l u o r i n a t e d analogue of NH 3 and i s pyramidal with C^ v symmetry. Studies of f l u o r i n a t e d compounds are of p a r t i c u l a r i n t e r e s t due to the highly electronegative nature of the F ligands, since "anomolous" i n t e n s i t y d i s t r i b u t i o n s are frequently observed i n the electron e x c i t a t i o n spectra. In p a r t i c u l a r for core spectra, a very high r e l a t i v e p r o b a b i l i t y of core to v i r t u a l valence t r a n s i t i o n s has often been observed at the expense of t r a n s i t i o n s to Rydberg l e v e l s . Other d i s t i n c t i v e features beyond the i o n i s a t i o n edge are also often observed i n such molecules. These e f f e c t s can be explained i n terms of shape-resonances caused by the formation of an e f f e c t i v e p o t e n t i a l b a r r i e r [73,77] either by electron repulsive forces i n the neighbourhood of the ligands [73,93] or a c e n t r i f u g a l b a r r i e r caused by the anisotropic nature of the molecular f i e l d [75,77]. While much work has been done both t h e o r e t i c a l l y and experimentally on the e x c i t a t i o n spectra of other fl u o r i n a t e d compounds, such as BF 3 [73,91-93,120-122], CF^ [66,92,123,124] and SF 6 [69,70,73,77,90] very - 85 -l i t t l e work has been done on NF 3. To date no ISEELS spectra have been reported for either the N or F Is regions of NF 3. There has, however, been l i m i t e d discussion of these regions i n e a r l i e r reported photo-absorption studies [120,121,125] obtained at lower res o l u t i o n than that used i n the present work. Some apparent inconsistencies i n t h i s e a r l i e r work are investigated i n the present more detailed ISEELS measurements. The only VSEELS spectrum i s that of unpublished work referred to i n the book by Robin [12], while the only UV photoabsorption spectra of the valence s h e l l [126] does not extend beyond 10 eV and i s featureless. To further aid i n the assignment of the VSEELS spectrum knowledge of the occupied l e v e l s are required. The i o n i z a t i o n potentials of these can be obtained from photoelectron spectroscopy. Only the outermost valence s h e l l o r b i t a l s (below 21.2 eV) are accessible with He(I) radia t i o n and such spectra hae been reported for NF 3 by Potts et a l . [117] as well as Bassett and Lloyd [118] who have also used He(II) r a d i a t i o n . The valence-shell photoelectron spectra u t i l i z i n g both Zr (151.4 eV) and A l Ka (1486.58 eV) have also recently been recorded [119] and these spectra provide the i o n i z a t i o n potentials for the inner valence electrons. Some e a r l i e r measurements of the Is core electron i o n i z a t i o n potentials of NF 3 have been reported for both F and N [127,128]. A v a r i e t y of MO ca l c u l a t i o n s have also been reported f o r NF 3. The only a b - i n i t i o c a l c u l a t i o n so far has been that of Unland et a l . [129]. Other less sophisticated c a l c u l a t i o n s have been used including CNDO [118,130,131], MNDO [132] and the Xa method [133,134]. - 86 -Experimental D e t a i l s 1) EELS measurements The inner s h e l l spectra were recorded on the ISEELS spectrometer described i n the previous chapter. An impact energy of 2.5 keV was used and the spectra were sampled at ~1° scattering angle. The N Is spectrum * * was calibrated against both the CO (C Is •*• it ,v=0) and N 2 (N Is ^ it v=l.) features (see Table 2.4). The newly established N Is value of N F 3 was used to i n t e r n a l l y c a l i b r a t e the F Is spectrum. The valence s h e l l spectrum was recorded on the new ISEELS spectrometer [53]. An impact energy of 3 keV was used with the scattered electrons sampled at zero degrees. The He(I) l i n e (21.218 eV [104]) was used to c a l i b r a t e the spectrum. 2) XPS Measurements The Is core l e v e l s and associated s a t e l l i t e structure were recorded using a McPherson ESCA 36 photoelectron spectrometer situated at the University of Alberta. The spectra were obtained by i r r a d i a t i n g the sample with A l Ka (hv = 1486.58 eV) X-rays which entered the sample c e l l through an aluminum f o i l window of 0.0001" thickness. The sample • was fed into the c e l l v i a s t a i n l e s s s t e e l feed l i n e s . The pressure was controlled by a G r a n v i l l e - P h i l l i p s series 203 variable leak valve and was monitored using a MKS Baratron pressure meter. Ty p i c a l pressures - 87 -used were between 150-200 microns. The major (Is) l i n e for each region was calibrated separately with the gas and c a l i b r a n t being mixed on the low pressure side of the i n l e t system p r i o r to introduction into the sample c e l l . The pressure of the reference l i n e was also controlled by a Gr a n v i l l e P h i l l i p s leak valve. Sample and reference gases were of approximate equal pressures with a t o t a l pressure somewhere between 150-200 microns. Repeated rapid cycles were run (calibrant-sample-calibrant) to minimize any possible error between reference and sample peaks due to s l i g h t pressure d r i f t s . The N Is l i n e was c a l i b r a t e d against a mixture of Ne and Ar from which the Ne KLL Auger l i n e (804.56(2)eV- k i n e t i c energy) and the Ar 2s core l e v e l (326.37(5)eV-binding energy) were used as the reference values [135]. The F Is l i n e was c a l i b r a t e d against the Ne KLL Auger l i n e and the Ne Is l e v e l (870.31(2)eV-binding energy) [135]. In order to try to d i s t i n g u i s h which, i f any, of the s a t e l l i t e structures i n the N Is and F Is XPS spectra are due to i n e l a s t i c s c a t tering of the outgoing photoelectron, a mixture of Ne and NF 3 at equal pressures was run and the low k i n e t i c energy ( i . e . , high binding energy) side of the Ne Is peak was recorded. Since the f i r s t "shake-up" feature i n the Ne Is spectrum does not occur u n t i l 37.3 eV [17], any feature between t h i s and the major Ne Is component would be due to energy loss processes involving i n e l a s t i c a l l y scattered photoelectrons. Energy loss contributions from NF 3 would be within 25 eV (see VSEELS spectra, f i g . 3.5) of the major Is component. There would also be a prominent energy loss feature at 16.85 eV [136], due to the intense - 88 -2p-*3s t r a n s i t i o n i n Ne. The NF 3 valence features w i l l of course be present i n the N Is and F Is XPS spectra i f the corresponding energy loss processes are occurring. Results and Discussion NFj possess symmetry and i t s electron configuration i s : ( lap 2 (le)4 ( 2 a L ) 2 O a ^ 2 ( 2 e ) 4 (4a x> 2 F Is N Is F 2s N 2s ( 3 e ) 4 ( 5 3 l ) 2 ( 4 e ) 4 ( 5 e ) 4 ( l a , , ) 2 ( S a ^ 2 ( 7 a i ) ° (6e)° v i r t u a l valence The experimental and calculated i o n i s a t i o n potentials are presented i n Table 3.1. The ordering of the occupied l e v e l s i s that obtained from the X^ calcu l a t i o n s [133,134]. However, the ordering of the v i r t u a l o r b i t a l s i s not c l e a r . A MNDO c a l c u l a t i o n suggests that the 7a 1 o r b i t a l i s at the lower energy, whereas HAM/3 and Hartree-Fock [137] calc u l a t i o n s give the unoccupied o r b i t a l s v i r t u a l l y the same energy with the e o r b i t a l being s t a b i l i z e d more r e l a t i v e to the a^ o r b i t a l upon creation of a N Is hole. Since a l l of these methods are not expected - 89 -Table 3.1 Experimental and calculated i o n i z a t i o n potentials for NF Or b i t a l Expt Calculated IP (ev) l p ( a ) (eV) X ( D V ) ( b ) a X ( M S )( c ) a MNDO HF CND0-M0^ CND0 ( f } Unoccupied V -(a) Valence 6a j l a 2 5e 13.73 13.82 13.97 14.55 14.74 13.95 14.18 •• 16.15 15.91 15.45 16.77* 18.23* 16.94* 18.05* •• 16.55 16.08 15.63 16.61* 17.96* 16.18* 16.34* •• Ae 17.52 17.19 16.71 17.16 19.63 17.84 18.89 •• 5 a l 3e 19.71 19.45 18.74 19.A3 22.63* 21.14 22.71* •• 21.14 21.30 20.62 21.17 22.57* 22.47 22.42* N 2s S 2e 26.A9 25.87 25.89 27.86 30. A2 27.47 F 2s 39.62 A3.69 45.49 41.76 „ 3 a l 2 a l A3.06 A9.71 A9.45 A8.63 N Is A14.36 F Is le l a l 693.2A (a) Experimental values from ref. 118, Hel, H e l l UPS (valence); r e f . 119, Zr M XPS ( 4 8 ^ ; r e f . 119, A l Ka XPS (2e, 3aj); t h i s work (N Is and F Is) see Table 3.3. (b) ref. 133. (c) ref. 134. (d) ref. 129. (e) ref. 131. (f ) ref. 118. * order d i f f e r e n t from l i s t e d . + order i s uncertain; MNDO suggests o^* i s 7a^, and o^* i s 6e, but t h i s i s not considered conclusive—see discussion i n text. - 90 -to give a r e l i a b l e or meaningful value to the unoccupied l e v e l s , no conclusion about t h e i r ordering w i l l be drawn and they w i l l be designated as a 1 and o*2 • A l l spectra w i l l be discussed with respect to this designation and to the experimental IP's and t r a n s i t i o n s shown i n Tables 3.1, 3.2 and 3.3. 1. Inner S h e l l Ionization by XPS Figures 3.1 and 3.2 show the XPS spectra obtained for the N Is and F Is regions r e s p e c t i v e l y . The results are summarized i n Tables 3.3a and 3.3b. The N Is IP (414.36 eV) i s i n f a i r agreement with the value (414.2 eV) previously reported by Finn et a l . [127]. However, there i s serious disagreement between the value reported i n the present work for F Is (693.24 eV) and that given by Davis (694.45 eV) [128]. In the present work c a l i b r a t i o n was achieved by sandwiching the peak i n question between the two c a l i b r a t i o n values (as outlined i n the previous s e c t i o n ) . This c a l i b r a t i o n procedure gives a check on the l i n e a r i t y of the spectrometer.t I t should also be noted that the k i n e t i c energy of the F Is electron produced by A l Ka X-rays i s 793.34 eV which i s only t The k i n e t i c energy of the photoelectron i s determined by K.E. = kAV+C where AV i s the p o t e n t i a l difference between the analyser plates, k i s an experimentally determined machine constant and C accounts for contact p o t e n t i a l s . Use of two c a l i b r a t i o n values allows an experimental determination of k for each run. Table 3.2 Transitions for C_ Symmetry Tr a n s i t i o n F i n a l State Dipole Allowed* a l ~ a l A l yes «-* e E yes a l ~ a2 A 2 no a 2 •«-> e E yes e <-> e A x + A 2 + E yes * Transitions from ground state (A^) to f i n a l states of A^ and E symmetry are allowed - 92 -RELATIVE ENERGY (eV) -10 0 10 20 3 0 40 50 6 0 0 -410 4 2 0 4 3 0 4 4 0 4 5 0 4 6 0 4 7 0 BINDING ENERGY (eV) Figure 3.1: The X-ray photoelectron spectrum for the nitrogen Is l e v e l and associated s a t e l l i t e structure of NF-j obtained with AA Koj 2 (1486.58 eV) r a d i a t i o n . The features have been f i t t e d using a gaussian'line shape. The shaded parts are the contributions from higher X-ray (Kcu and Ko^) components. Also shown i s the spectrum obtained of the low ( k i n e t i c ) energy side of the neon Is spectrum from a mixture of Ne and NF-j (see text for d e t a i l s ) . The p o s i t i o n of the prominent VSEELS t r a n s i t i o n s are also indicated for Ne and N F 3 . - 93 -RELATIVE ENERGY (eV) 0 10 20 30 4 0 50 6 0 i — 1 — i — 1 — i — 1 — i — 1 — i — • — i — " — i — 1 — r 50 7 0 0 710 7 2 0 730 740 7 5 0 7€ BINDING ENERGY (eV) Figure 3.2: The X-ray photoelectron spectrum for the f l u o r i n e Is l e v e l and associated s a t e l l i t e structure of NF 3 obtained with AJl Ka^ 2 (1486.58 eV) r a d i a t i o n . The features have been f i t t e d using a gaussian'line shape. The shaded parts are the contributions from higher X-ray (Kaq and Ka^) components. Also shown i s the spectrum obtained of the low ( k i n e t i c ) energy side of the neon Is spectrum from a mixture of Ne and NF-j (see text for d e t a i l s ) . The p o s i t i o n of the prominent VSEELS t r a n s i t i o n s are also indicated for Ne and N F 3 . - 94 -Table 3.3(a) Peak Energies i n the N Is XPS Spectrum of NF 3 Feature Energy (eV) Difference energy (eV) from main l i n e Assignment Lite r a t u r e This work X 414.2 a 414.36 0 Is hole A 421.47 7.11 B 427.13 12.77 Sh 430.39 ~16.03 C 433.26 18.90 D 437.20 22.66 BAND E 445-475 shake up & shake off a Ref. 127. Table 3.3(b) Peak Energies i n the F Is XPS Spectrum of NF Feature Energy (eV) Difference energy (eV) Assignment Lite r a t u r e This work from main l i n e Y a 694.45 693.24 0 Is hole P 698.45 5.21 Q 702.64 9.40 R 709.77 16.53 S -719.7 -24.5 T -722.3 -29.1 U -737.2 -44.0 3 Ref. 128. - 95 -11.22 eV below the Ne KLL Auger peak used for c a l i b r a t i o n . Thus the present value i s thought correct. Both the N Is and F Is regions exhibit r i c h s a t e l l i t e structure extending to binding energies at least 60 eV above ( i . e . , at lower k i n e t i c energies than) the major Is component. The peaks are broad and t h i s i n part r e f l e c t s the use of unmonochromated Al Ka X-rays (FWHM of the Ne Is under the conditions used here was ~ 1 eV). This X-ray source contains a va r i e t y of lower i n t e n s i t y l i n e s as well as the main Ka^ ^ (unresolved) components. Contributions from the higher X-ray components (mainly Ka 3 and Ka^) have been l a r g e l y accounted for i n the f i t t i n g procedure that used a modified version of SUNDERE [138]. The Ka 3 and Ka^ X-ray contributions are shown shaded i n Figs. 3.1 and 3.2. Only the most prominent s a t e l l i t e structure has been f i t t e d with no attempt made to f i t the extensive lower i n t e n s i t y , unresolved structure extending to higher binding energies (lower k i n e t i c energies). Also shown i n each of Figs. 3.1 and 3.2 (uppermost trace) i s the spectrum of the high binding energy (low k i n e t i c energy) side of the Ne Is peak obtained form the Ne/NF3 mixture. The features are i n d i c a t i v e of i n e l a s t i c s c attering (energy-loss) of the emitted photoelectron caused by NF 3 and Ne. It was assumed that the energy loss features would be quite s i m i l a r i n the Ne, N, and F Is regions since the impact energies of the photoelectrons (616, 1072, and 793 eV r e s p e c t i v e l y ) are a l l well above the e x c i t a t i o n energies. The experiment should give an estimate of the extent of energy-loss contributions to the XPS spectra. The major energy loss components i n the (low resolution) Ne/NF3 spectrum i n the Ne Is region - 96 -were found to be at 13.4 eV and 16.8 eV (see Figs. 3.1 and 3.2, i n s e r t s , which of course are the same spectrum). The former peak i s s o l e l y due to energy loss contributions from NF 3 and agrees well with features d i -r e c t l y observed i n the high res o l u t i o n valence s h e l l spectrum of NF 3 obtained on the EELS spectrometer (see peaks 5-6 (12.81-13.75 eV) of F i g . 3.5). A second NF 3 valence feature of approximately equal i n t e n -s i t y would be expected at ~ 16.2 eV (corresponding to peak 8 of F i g . 3.5). However, what i s observed i s a peak at 16.8 eV of approximately double the expected i n t e n s i t y from NF 3, alone. This i s at t r i b u t e d to equal contributions a r i s i n g from both NF 3 and Ne energy loss features (the Ne 2p->3s energy loss feature occurs at 16.85 eV [136]. Structure at ~ 16.0 eV appears i n both the N Is (shoulder (sh) i n F i g . 3.1) and the F Is (peak R i n F i g . 3.2) XPS spectra. The energy loss peak at 13.4 eV i s e s s e n t i a l l y absent i n the F Is XPS spectrum, while some contribu-t i o n i s probably occurring i n the N Is spectrum. However, a comparison of the r e l a t i v e i n t e n s i t i e s of the spectra i n each of figures 1 and 2 leads to the conclusion that a large percentage of the XPS " s a t e l l i t e structure" i s due to "shake-up" processes rather than from i n e l a s t i c s c a t t e r i n g . I t can be seen that the s a t e l l i t e structures i n the N Is and F Is regions are quite d i f f e r e n t . The N Is region has a peak A at 7.11 eV above the major Is component. This i s followed by a peak B at 12.77 eV above the main l i n e . This along with a p a r t i a l l y resolved shoulder (sh) at 16.0 eV i s thought to be at least p a r t i a l l y due to the energy-loss features. However the r e l a t i v e i n t e n s i t y of the 12.77 eV feature - 97 -suggests that some contribution from genuine s a t e l l i t e structure may be present. The most intense features (C,D) i n the N Is s a t e l l i t e spectra l i e i n the 18-25 eV region. This i s then followed by la r g e l y unresolved bands extending out to ~ 60 eV binding energy above the main Is l i n e . The F Is region shows a doublet (P,Q) i n the region 5-10 eV above the major Is component. This i s followed by a large peak (R) at 16.6 eV which coincides with the second energy-loss feature. The e s s e n t i a l absence i n the XPS spectrum of any intense feature at ~ 13 eV (the lower energy-loss feature) leads to the conclusion that the peak at 16.6 eV i s lar g e l y due to "shake-up" rather than energy-loss processes. At higher k i n e t i c energies there follows a broad region of l a r g e l y unresolved structure stretching out to ~ 65 eV above the main Is l i n e . Other structures at S, T, and U are also apparent. The various features of the XPS spectra w i l l be further considered i n the i n t e r p r e t a t i o n of the electron energy loss spectra ( i e : the ISEELS and VSEELS measurements) i n the following sections. Inner s h e l l e x c i t a t i o n by ISEELS The electron energy-loss spectra for the N Is and F Is regions of NF 3 are shown resp e c t i v e l y i n figures 3.3 and 3.4. Figures 3.3a and 3.3b show d e t a i l s of major N Is tr a n s i t i o n s recorded at resolutions of 0.14-0.28 eV FWHM while F i g . 3.3c shows a long range scan recorded at resolutions of 0.36 eV FWHM. In a si m i l a r way F i g . 3.4a shows the long range spectrum of the F Is spectrum and F i g . 3.4b the d e t a i l . Both spectra i n F i g . 3.4 were recorded at a resolution of 0.36 eV FWHM. - 98 -a) tz CO -z. LU LU > < _ l LJ CC i r r r A 2 3 4 AE=028eV ! 1 I I i I 11 f 'I N 2 IS-7T* J Impurity i / 402 4 0 6 410 414 b ) N 2 IS-TT* 4 0 0 4 0 4 NF, N K - S H E L L 2 3 4 4 0 8 412 N F 3 + N 2 "AE=0.28eV ,NF 3 AE=O.I4eV C) 1 0 -fc to UJ I-LU LU I 2 3 4 N 2IS— 7 T * I Impurity / N K - E D G E i B Sh C N v. NF, N K-SHELL XPS A E = 0 . 3 6 e V — i > 1 ' 1 • i • i • r 400 410 420 430 440 450 ENERGY LOSS (eV) Figure 3.3: Nitrogen Is electron energy loss spectra of NF-j. 3.3b) also shows the spectrum of a ^/NF-j mixture. The positions of the nitrogen K-edge and the XPS structure i n Figure 3.3c) are taken from Figure 3.1 and Table 3.3(a). - 99 -lo-r a) A \ NF 3 FK-SHELL AE = 0.36 eV 5 4 I FK-EDGE >-CO 6 8 0 ^ 1 0 -b) LU cr 700 720 740 A / X P S — I 2 FK - E D G E P r Q 3 R 760 5 H 685 695 705 ENERGY LOSS (eV) 715 Figure 3.4: Fluorine Is electron energy loss spectra of NF^. The positions of the fl u o r i n e K-edge and the XPS s a t e l l i t e structure are taken from Figure 3.2 and Table 3(b). - 100 -The energies of the s p e c t r a l features and possible assignments for the spectra shown i n Figs. 3.3 and 3.4 are given i n Tables 3.4 and 3.5 r e s p e c t i v e l y . Both e x c i t a t i o n spectra have been studied e a r l i e r by Vinogradov et a l [125] using soft X-ray photoabsorption. While the present ISEELS work i s i n generally good agreement with the F Is spectrum reported i n the photoabsorption work [125] there i s a serious discrepancy i n the case of the N Is spectrum. In p a r t i c u l a r Vinogradov et a l . [125] report a prominent unassigned feature at 400.9 eV which i s e s s e n t i a l l y absent i n the ISEELS spectra of NF 3 reported here. The energy of t h i s peak corresponds c l o s e l y to that of the N Is i t * t r a n s i t i o n i n molecular N 2 ([4] and Table 2.4) and i s thus almost c e r t a i n l y due to an N 2 impurity. This view i s confirmed by the ISEELS spectrum we have obtained for a mixture of N 2 and NF 3 which i s shown i n figure 3.3b (upper t r a c e ) . I t should be noted that there does appear to be a very s l i g h t trace of N 2 impurity i n a l l the N Is spectra of NF 3; however the magnitude of the dominant N 2 (Is •> i t ) t r a n s i t i o n [4] i s at most very small. In t h i s regard i t should be noted that there i s no sharp feature at 12.93 eV i n the valence s h e l l spectrum ( F i g . 3.5), and t h i s i s i n d i c a t i v e of the absence of N 2, which has i t s most intense valence s h e l l t r a n s i t i o n at 12.93 eV [103]. Thus i t i s reasonable to conclude that the present NF 3 N Is spectrum i s e f f e c t i v e l y free from any contributions above 402 eV a r i s i n g from nitrogen impurities. The N Is ISEELS spectrum of NF 3 (Figure 3.3) i s dominated by an intense Is •*• a t r a n s i t i o n at 407.10 eV (peak 1). This can be i n t e r -- 101 -Table 3.4 Energies, term values and possible assignments i n the N K-shell energy loss spectrum of NF, Feature Energy Loss (eV) Term Value Possible Assignment Photoabsorption(eV) v ' 400.9 ( b ) 1 407.10 7.26 Is + 0* 406.6 2 411.02 3.34 Is •+ 3s 3 411.99 ( c ) 2.37 Is -* 3p 411.9 4 413.24 1.12 Is > 4p 413.0 414.36 ( d ) 413.4 K-edge 0 Is + 414.1 5 425 (a) Ref. 125. (b) This feature was not explained. I t i s thought to be due to impurity N 2 (Is -* it*). See text. (c) This feature c a l i b r a t e d against CO (Cis -* it*) 287.40 eV and N 2 (Nls •* it*) 401.10 eV. (d) XPS th i s work, see Table 3.3. - 102 -Table 3.5 Energies, term values and possible assignments i n the F K-shell spectrum of NF 3 electron energy loss Feature Energy Loss(eV) Term Value Possible Assignment (a) Photoabsorption(eV) 1 687.42 ( b ) 5.82 Is (o*) 686.4 K-edge 693.24 ( c ) 0 Is •+ » 2 697 3 -709 (a) Ref. 125. (b) Internally c a l i b r a t e d against feature 3 of NF 3 (N Is) energy loss spectrum. (c) XPS - see Table 3.3. - 103 -preted as an enhanced inner-well f i n a l (valence) state trapped by the p o t e n t i a l b a r r i e r created by the highly electronegative F ligands. This p o t e n t i a l b a r r i e r model concept has been discussed by Dehmer et a l . [73,77]. On the basis of t h i s model the p r o b a b i l i t y of t r a n s i t i o n s * to unoccupied a type valence o r b i t a l s (Inner-well) would be enhanced at the expense of excitations to Rydberg o r b i t a l s (outer-well). In accord with this view the spectrum does show some Rydberg strucure (peaks 2, 3 and 4) of much lower r e l a t i v e i n t e n s i t y leading up to the K-edge. S i g n i f i c a n t , lower i n t e n s i t y , structure also appears i n the continuum. The pre-edge spectrum i s q u a l i t a t i v e l y s i m i l a r to that observed e a r l i e r for the S 2p spectrum of SF 6 [69] which was also attributed to intense inner well valence excitations and weak outer well Rydberg structure. It i s of i n t e r e s t to compare the N Is energy loss spectrum of NF 3 with that of the i s o e l e c t r o n i c molecule N(CH 3) 3 (see F i g s . 8.3 and 8.4). The spectra are very d i f f e r e n t and this i s a t t r i b u t a b l e to the behaviour of the electronegative F ligand as compared to the electron donating CH 3 l i g a n d . This aspect i s discussed i n Chapter 8. F i n a l l y , i n the N Is ISEELS spectrum of NF 3, a broad maximum i s observed i n the continuum at '>425 eV. The F Is ISEELS spectrum of NF 3 ( F i g . 3.4) also shows a strong Is -»• a band located at 687.42 eV. There appears to be no Rydberg structure, but this may be masked by the expected large natural width of * the a band and i t s proximity to the F Is edge. As i n the case of the N Is energy loss spectrum, there e x i s t s d e f i n i t e structure i n the c o n t i -- 104 -nuum with maxima (peaks 2 and 3) at ~ 697 and ~ 709 eV re s p e c t i v e l y . However, i n the case of the F Is spectrum t h i s structure i s on a r e l a t i -vely more intense background compared to the s i t u a t i o n i n the N Is region. This i s to be expected, however, since the vacancy i s on an F atom which i s on the periphery of the molecule. The electrons o r i g i n a -t i n g from here would l i k e l y have much less of a po t e n t i a l b a r r i e r to overcome, and thus would have s i g n i f i c a n t p r o b a b i l i t y of going to outerwell states as we l l . Thus there would l i k e l y be less enhancement * of the a band r e l a t i v e to the continuum. * The term value for the N Is •+• a t r a n s i t i o n i s 7.26 eV, whereas that for the F Is -»• a t r a n s i t i o n i s 5.82 eV (Tables 3.4 and 3.5). Thus i t would appear that the term values are not transferable between the two core-hole centres. This raises two questions; (a) what i s the make * up of the a envelope i n each case, and (b) what i s the e f f e c t of the core-hole being on the periphery of the molecule as opposed to being at the centre? * In NF 3 the unoccupied a o r b i t a l s are of a^ and e symmetry, and since the N Is and F Is o r b i t a l s transform as a^ and a ^ e res p e c t i v e l y , t r a n s i t i o n s to both a l e v e l s are dipole allowed from each centre (see Table 3.2). Both peaks are broad and could be composed of various unresolved components. In an attempt to see what contributions might be expected, MO ca l c u l a t i o n s using HAM/3 [139] were performed. Since HAM/3 i s primarily parametrised for u- systems i t i s not expected to necessa-r i l y give good r e s u l t s for a type systems. However, whereas the eigen-values obtained might not be s a t i s f a c t o r y , the eigenvectors should give - 105 -a reasonable i n d i c a t i o n [140] of the make up of the valence o r b i t a l s . Calculations were performed on the molecule with respective core electron vacancies and with half an electron ( T r a n s i t i o n State Formalism [141]) d i f f u s e l y added to the v i r t u a l o r b i t a l s . The eigenvectors of the v i r t u a l o r b i t a l s i n the molecule with an N Is hole indicate the c o n t r i -bution from the N atom to the 6e o r b i t a l to be predominantly 2p x and 2p^ (implying a strong s •*• p dipole allowed t r a n s i t i o n ) while the 7a^ o r b i -t a l i s comprised of approximately equal 2s and 2p^ character. On t h i s * basis the a envelope i n the N Is spectrum would have contributions from both 7a ^ and 6e o r b i t a l s but with a larger contribution from the 6e. On creating an F Is hole the symmetry of the molecule i s reduced from C„ to C since the vacancy w i l l be l o c a l i s e d on the one F centre. 3v s In this case the v i r t u a l e and a^ o r b i t a l s become a', a", and a' respec-t i v e l y . The eigenvectors of these o r b i t a l s for the F atom with the vacancy show v i r t u a l l y no p o r b i t a l contribution to one of the a' and the a" o r b i t a l s , leaving the only s •*• p contribution coming from the remaining a* o r b i t a l . Thus the difference i n term values may simply r e f l e c t the difference i n the contribution from the various possible t r a n s i t i o n s to the broad envelope. As well as possible e f f e c t s on symmetry the removal of core electrons ( i . e . , N Is or F Is) w i l l determine the p o t e n t i a l i n which the electron i n the newly occupied o r b i t a l finds i t s e l f . The o o r b i t a l s i n the p o t e n t i a l b a r r i e r model are expected to be mainly within the inner well are hence l o c a l i s e d around the (central) N atom. The removal of an N Is electron increases the core charge by one and so the a o r b i t a l - 106 -energies w i l l be determined by a Z + 1 ( i . e . , an 0 atom) cen t r a l core. The e f f e c t of creating a peripheral hole would not have as much e f f e c t since the v i r t u a l o r b i t a l s are l o c a l i s e d around the central atom. Hence an electron i n a v i r t u a l o r b i t a l would be harder to remove ( i . e . , i t would have an increased binding energy (term value)) when there i s a c e n t r a l atom vacancy as opposed to a ligand vacancy. This e f f e c t was e a r l i e r noted i n ISEELS core spectra SF 6 [69], i n which the v i r t u a l o r b i t a l s are well separated, thereby removing any ambiguity with regard to the f i n a l states. The difference i n term values between the species with a S 2p hole and that with an F Is hole was ~ 1.8 eV. Therefore no d e f i n i t e conclusions can be drawn about the r e l a t i v e separation of the 7a^ and 6e o r b i t a l s in.NF 3 except that they are probably within 1 - 1 . 5 eV ( i . e . , the difference i n the two term values) of one another. The N Is spectrum c l e a r l y shows sharp Rydberg structure. Features 3 and 4 (411.99 eV and 413.24 eV) have term values of 2.37 eV and 1.12 eV r e s p e c t i v e l y . This i s i n excellent agreement with the calculated term values (2.36 eV and 1.18 eV) obtained for n = 3 and 4 respectively when using the approximate quantum defect (6 = 0.6) expected for a p s e r i e s . Accordingly features 3 and 4 have been assigned as t r a n s i t i o n s to the 3p and 4p l e v e l s . Feature 2 has been assigned as the N Is -»• 3s ( a ^ Rydberg t r a n s i t i o n . It has a term value of 3.34 eV which gives an estimated quantum defect of 0.98, i n agreement with the expected magnitude of the quantum defect (6 = 1.0) associated with an s s e r i e s . The positions of the most dominant s a t e l l i t e structures from the - 107 -XPS data (Figures 3.1 and 3.2) have also been indicated on the ISEELS spectra (Figure 3.3, A-D and figure 3.4, P-S). Features a t t r i b u t a b l e to i o n i z a t i o n and e x c i t a t i o n ("shake-up") i n XPS should appear as new continua beyond the i o n i z a t i o n edge i n ISEELS. I t should be noted that (a) the values reported from the XPS data are the v e r t i c a l positions of the broad envelopes whereas the new continua i n ISEELS w i l l appear from the onset ( i e : a d i a b a t i c ) ; (b) the reso l u t i o n obtained i n the ISEELS spectra i s about 3 times better than that of the XPS data, thus more onsets might be apparent i n ISEELS; (c) no d e f i n i t e inference can be drawn from comparison of the observed cross-section i n the one process compared with the other since XPS deals with a photoelectron with k i n e t i c energy several hundreds of eV above the threshold whereas ISEELS w i l l r e f l e c t threshold behaviour. Even with the foregoing considera-tions i t should be noted that a comparison of the data from XPS and ISEELS spectra suggests that much of the ISEELS continuum structure can l i k e l y be interpreted as being due to the onset of "shake-up" continua ( i e : Is i o n i z a t i o n and valence s h e l l e x c i t a t i o n ) . Possible general forms of such continua, s t a r t i n g at apparent d i s c o n t i n u i t i e s are i n d i -cated by dashed l i n e s on Fi g s . 3.3 and 3.4. Between the N Is edge and po s i t i o n A i n Figure 3.3 there i s an i n d i c a t i o n of considerable complex structure. This can be att r i b u t e d to various electron excitations (simultaneous Is and valence s h e l l e x c i t a t i o n ) . Similar type of s t r u c -ture i s also present i n the F Is spectrum. In the N Is ISEELS spectrum (F i g . 3.3) there i s a broad maximum i n the continuum (feature 5) at ~ 425 eV. Since no continuum resonances are expected for NF 3 t h i s struc-- 108 -ture i s presumably due to i o n i z a t i o n plus e x c i t a t i o n on top of the d i r e c t i o n i z a t i o n continuum. This feature i s associated with a con-tinuum due to "shake-up" A i n the XPS spectrum ( F i g . 3.1). S i m i l a r l y , feature 3 ( F i g . 3.4) i s l i k e l y the apparent maximum of a "shake-up" continuum corresponding to the broad peak R (see F i g . 3.2). Valence She l l E x c i t a t i o n by VSEELS The valence s h e l l electron energy-loss spectrum for NF 3 i s shown i n Figure 3.5 and summarized, along with tentative assignments, i n table 3.6. The spectrum i s q u a l i t a t i v e l y s i m i l a r to that reported by Robin [12], however there are some va r i a t i o n s i n r e l a t i v e i n t e n s i t y . For instance features 7 and 8 are much less intense than i n the spectrum reported i n the present work. This difference i s consistent with the d i f f e r e n t impact energies used. The spectrum shown by Robin [12] was excited by 100 eV electrons whereas the impact energy used here was 3000 eV. Both spectra were obtained at 0° scattering angle. The spectrum shown by Robin also has an unfortunate break i n the data j u s t where feature 9 appears i n figure 3.5. Robin [12] has made a l i m i t e d assignment i n which most t r a n s i t i o n s are a t t r i b u t e d to Rydberg f i n a l l e v e l s . However i t would be expected that the valence o r b i t a l s would have a stronger i n t e r a c t i o n with the v i r t u a l valence o r b i t a l s than with the more d i f f u s e Rydberg l e v e l s e s p e c i a l l y given the possible existence of a p o t e n t i a l b a r r i e r which may also a f f e c t the valence s h e l l spectrum. In t h i s regard the broadness of some of the observed bands e.g., 1, 2, and 8) i s more - 109 -3 s " T -3 p 3 s RYDBERG 3 s \ r n 3 P 4 e —I m 3 p y, 5 e i r rr f 3 P la, J? 3 s I NF, VALENCE A E = 0.035 eV 6 a , . J rrrT/ *\ 3 P 3 s 5 a « n rn 3 P 3 e • ,1 1 4 e I 2 — 5 e -6a, • , Q 2 3 4 5 6 I I I I •3e —I 5a, 7 8 9 10 I I I I i r i 1 1 r 1 1 1 r ~i 1 1 r 10 15 2 0 25 E N E R G Y LOSS (eV) Figure 3.5: NF-j valence s h e l l electron energy loss spectrum. The top manifold shows the positions of Rydberg series estimated from term values and the Rydberg formula. The i o n i z a t i o n l i m i t s are taken from photoelectron spectroscopy. The bottom manifold shows the estimated positions for valence-valence t r a n s i t i o n s (see text for d e t a i l s ) . Table 3.6 Poss ible Assignments in the Valence S h e l l Spectrum of NF Rydberg Trans i t ions (eV) Major Observed Valence-Valence Feature Energy (eV) Assignment 6a j la2 5e 4e 5 a l 3e 1 8.64 6a^ Oj* 6a ^ -*• a 2* 4. 2 9.45 10.39(3s) 3 11.21 Ia2 • Oj* 11.36(3p) 5e > Oj* la2 + o 2 * 5e * 02* 4e > o^* 4 12.37 12.61(4p) 5 12.81 4e -> ©2* 13.03(5p) 13.26(6p) 13.21(3s) 6 13.75 13.78(3p) 14.18(3p) 14.13(3s) 7 15.01 5aj > Oj* 5a j + o 2 * 15.03(4p) 15.15(3p) 15.45(5p) 15.43(4p) 15.68(6p) 15.85(5p) 8 16.20 3e + Oj* 3e + ©2* 16.08(6p) 16.40(4p) 16.85(5p) 17.05(6p) 16.37(3s) 17.34(3p) 17.80(3s) 9 18.12 4a ^ ->• o^* 18.59(4p) 18.77(3p) 10 -19 4a j + ©2* 19.01(5p) 19.24(6p) 20.02(4p) 20.44(5p) 20.67(6p) la~ -> 7a, , i s d ipole forbidden. Since It i s not known which of a * or o * i s the 7a , both are g iven . - I l l -suggestive of valence-valence rather than Rydberg t r a n s i t i o n s . For t h i s reason the spectrum has been assigned as being composed of valence-valence t r a n s i t i o n s with some Rydberg t r a n s i t i o n s superimposed. The spectrum can be divided into three sections; ( i ) t r a n s i t i o n s a r i s i n g from the N lone-pair o r b i t a l (6a±); ( i i ) t r a n s i t i o n s a r i s i n g from the F lone-pairs (la 2» 5e and 4e) and t r a n s i t i o n s a r i s i n g from the predominantly N-F bonding o r b i t a l s (5a^, 3e). Since a l l the MO schemes (see Table 3.1) give the 6a ^ as the highest occupied molecular o r b i t a l the most l i k e l y assignment of features 1 and 2 i s to the 6a 1 -*• o"^ and 0*2 valence-valence t r a n s i t i o n s since t r a n s i t i o n s to both the 6e and 7a^ o r b i t a l s are allowed. I t should be noted that the band comprising features 1 axtd 2 i s broad. A further contribution to the width could possibly come from the 6aj^ •*• 3s(a^) Rydberg t r a n s i t i o n . However, using the 3s term value (3.34 eV) from the ISEELS data discussed above, the predicted p o s i t i o n of this feature O s a ^ ) would be at 10.4 eV i n the region between peaks 2 and 3. In order to predict the positions of other possible valence-valence t r a n s i t i o n s the experimental separations of the remaining o r b i t -t a l s from the ba^, as derived from photoelectron spectroscopy [118], have been added to the assigned positions of the suggested 6a ^ -*• o^ and * o"2 valence-valence t r a n s i t i o n s . The positions are shown i n the lower part of F i g . 3.5. I t should be r e c a l l e d that the l a 2 -*• 7a^ t r a n s i t i o n i s forbidden (see Table 3.2); however, since there i s doubt as to the exact ordering of the unoccupied cr o r b i t a l s , both t r a n s i t i o n s are - 112 -ind i c a t e d . The ordering of the occupied o r b i t a l s i s that given by the X f f c a l c u l a t i o n s [133,134]. Other cal c u l a t i o n s (see Table 3.1) have reversed the order of the l a ? and 5e o r b i t a l s . The X ca l c u l a t i o n s use 1 a t r a n s i t i o n state formalism and therefore take into account relaxation accompanying i o n i s a t i o n . This ordering also agrees with that obtained by K e l l e r e r et a l . [130] i n which they calculate a Koopman's theorem defect and combine i t with values predicted by CND0/2 formalisation. With the exception of the X^ c a l c u l a t i o n a l l cal c u l a t i o n s presented i n Table 3.1 apply Koopman's theorem. The agreement obtained between the above predictions of the ener-gies of a d d i t i o n a l valence-valence t r a n s i t i o n s and major features i n the spectrum shown i n F i g . 3.5 i s quite good. The l a 2 , 5e and 4e o r b i t a l s are v i r t u a l l y non-bonding and l o c a l i s e d on the F atoms [134]. Features 4 and 5 along with much of the i n t e n s i t y under feature 6, which comprise the second section, are attributed mainly to t r a n s i t i o n s from these * o r b i t a l s to the a o r b i t a l s . The r e l a t i v e narrowness of the bands comprising features 4-6 i s consistent with t r a n s i t i o n s coming from non-bonding o r b i t a l s . This i s i n contrast to the higher energy section comprising features 7-9. The width of th i s section i s quite broad and the bulk of the i n t e n s i t y i s assigned to t r a n s i t i o n s to the a o r b i t a l s a r i s i n g from the 5a^ and 3e o r b i t a l s , which are predominantly N-F bonding. The structure to higher energy (>18 eV) probably arises from t r a n s i t i o n s from the 4a^ o r b i t a l . The valence s h e l l spectrum shown i n F i g . 3.5 i s thus l i k e l y to contain contributions from a number of valence-valence t r a n s i t i o n s . - 113 -However, t r a n s i t i o n s to Rydberg l e v e l s w i l l also be present. In t h i s regard close examination of the spectrum ( F i g . 3.5) shows evidence for fi n e structure on top of the broader l e v e l s a t t r i b u t e d to valence-valence t r a n s i t i o n s . These p a r t i a l l y resolved shoulders were repeatedly seen on d i f f e r e n t scans of the same spectrum. In addition the somewhat narrower bands (3, 6, and 7) have energies corresponding to expected Rydberg t r a n s i t i o n s (see below). In order to assign possible Rydberg structure i t i s assumed that the term values for Rydberg l e v e l s are transferable between the ISEELS and the VSEELS spectra. Before doing t h i s , however, i t i s appropriate to consider why Rydberg term values might be transferable whereas v i r t u a l valence o r b i t a l term values are not necessarily so. From the VSEELS data the term value for the LUMO o r b i t a l (o^ - F i g . 3.5, feature 1) i s 5.09 eV, whereas the term value corresponding to the peak of the broad envelope encompassing the N Is •+• o* t r a n s i t i o n s i n the N ISEELS data ( F i g . 3.3, feature 1) i s 7.26 eV. * Therefore the o^ term value i s l i k e l y to be even higher. In essence the term value gives the "binding energy" of the excited electron i n the previously unoccupied o r b i t a l . Thus i n the former case the term value * gives the binding energy of the electron i n the o^ o r b i t a l when the hole exists i n the valence s h e l l and i n the l a t t e r case the binding energy of the electron when an N Is core hole e x i s t s . The loss of sh i e l d i n g by the removal of the valence electron to the v i r t u a l valence o r b i t a l should be very much less than that caused by the removal of the l o c a l i s e d core electron. Hence, i n the ISEELS case, the electron i n the 0* o r b i t a l should see almost a whole extra unit of charge and so i t - 114 -should be harder to remove. I t w i l l therefore have an increased binding energy (term value). In contrast to the valence o r b i t a l s the Rydberg o r b i t a l s are large and d i f f u s e and hence w i l l see the molecule as one large core. Thus they should be less affected by where the vacancy occurred and so have transferable term values, whereas the valence o r b i t a l s , being much more l o c a l i s e d , w i l l be more susceptible to l o c a l v a r i a t i o n s i n s h i e l d i n g . The expected positions of the valence-Rydberg t r a n s i t i o n s were calculated for n = 3 and 4 using the term values obtained for these l e v e l s from the N ISEELS spectrum ( i . e . , 3.34, 2.37, and 1.12 eV for the 3s, 3p, and 4p o r b i t a l s r e s p e c t i v e l y ) . The positions of the 5p and 6p Rydbergs were estimated using a quantum defect of 0.6 (which applies also to the 3p and 4p l e v e l s ) . These assignments and energies are shown i n Table 3.6 and F i g . 3.5 (upper portion). The predicted values of the 6a±, l a 2 , and 4e •*• 3p t r a n s i t i o n s (11.36, 13.78, and 15.15 eV respectively) are i n agreement with the narrow features 3, 6, and 7, which are at 11.21, 13.75, and 15.01 eV respectively, and these have been assigned accordingly. The t r a n s i t i o n s to the higher Rydberg l e v e l s w i l l not be as intense and weak features i n the spectrum can be at t r i b u t e d to these. The spectrum as a whole i s consistent with the i n t e r p r e t a t i o n given as that of predominantly valence-valence t r a n s i t i o n s with valence-Rydberg t r a n s i t i o n s on top. The assignment of feature 9 i s not clear since i t s width and p o s i t i o n i s not consistent with an assignment to a Rydberg l e v e l . I t may also a r i s e , along with the i n t e n s i t y at ~ 19 eV, from t r a n s i t i o n s to - 115 -the o* le v e l s o r i g i n a t i n g from the 4a^ o r b i t a l . This would give i t a higher term value than the other valence-valence t r a n s i t i o n s but one that i s more consistent with that from the N Is ISEELS spectrum. This i s not unexpected i n view of ideas discussed e a r l i e r since the 4a± o r b i t a l i s e s s e n t i a l l y due to the N 2s o r b i t a l and i s thus more l o c a l i s e d than the outer valence l e v e l s . Comparison of VSEELS and ISEELS Spectra with the XPS S a t e l l i t e Structure The process that occurs i n VSEELS i s the e x c i t a t i o n of an outer ( i . e . , valence s h e l l ) electron to an excited bound state ( v i r t u a l valence or Rydberg l e v e l ) . The s a t e l l i t e structure i n XPS i s due to the formation of excited ion states i n which the process, at least i n a simple model, can be thought of as the emission of a photoelectron with the attendant e x c i t a t i o n of an outer electron to either an excited bound state ("shake-up") or to the continuum ("shake-off"). A comparison of the r e s u l t s of the two spectroscopies may y i e l d information on the valence manifold and the types of processes that may occur upon photo-i o n i s a t i o n . Care has to be taken i n any such comparison since VSEELS can e s s e n t i a l l y be described i n terms of one electron process, whereas "shake-up/shake-off" i s at least a two electron process. Martin et a l . [21] have shown that many-electron theory including configuration i n t e r -action has to be included i n both the hole and ground states i n order to describe adequately the s a t e l l i t e structure accompanying photoionisa-t i o n . Using t h i s approach they have successfully analysed the s a t e l l i t e structure of HF [22]. An inspection of the data, however, reveals (as - 116 -they have noted) that the four most intense peaks can be interpreted, at least to a f i r s t approximation i n terms of one-electron e x c i t a t i o n s . Thus an attempt has been made to analyse the NF 3 XPS spectra i n terms of one-electron excitations i n order to see whether any meaningful informa-ti o n can be obtained without resorting to complex c a l c u l a t i o n s . This i s aided by comparison with the VSEELS and ISEELS spectra along with a consideration of the p o t e n t i a l b a r r i e r phenomenon. The N Is XPS s a t e l l i t e spectrum has been assigned (see Table 3.3) as being dominated by t r a n s i t i o n s to Rydberg l e v e l s (peaks C and D, F i g . 3.1) with the lower i n t e n s i t y structure (A, B, E, F, etc.) a r i s i n g p r i m a r i l y from t r a n s i t i o n s to the unoccupied valence l e v e l s and energy loss features. These tentative assignments have been made on the basis of two considerations. F i r s t l y , Creber et a l . [142] i n a study of the XPS s a t e l l i t e spectra of the second row hydrides (CH^, NH3, H 20, HF) i s o e l e c t r o n i c with Ne have suggested that Rydberg-like o r b i t a l s relax less than valence-type o r b i t a l s when a core hole i s created i n XPS ( t h i s i n c i d e n t -a l l y i s consistent with the previous assertion that term values are transferable from ISEELS to VSEELS for Rydberg t r a n s i t i o n s , but not necessari l y for t r a n s i t i o n s to unoccupied valence l e v e l s ) . The energies of peaks A, B, and sh r e l a t i v e to the main Is l i n e ( F i g . 3.1, Table 3.3) are close to the valence s h e l l e x c i t a t i o n energies i n the VSEELS spectrum ( F i g . 3.5, Table 3.6). Further peaks E and F (of F i g . 3.1) would correspond to valence t r a n s i t i o n s from the 3 a a n d 2e inner valence o r b i t a l s to the o* l e v e l s . However, peaks C and D are at - 117 -r e l a t i v e energies of 18.87 eV and 23.36 eV respectively, which are much greater than the estimated energies of any t r a n s i t i o n s a r i s i n g from outer valence o r b i t a l s to the v i r t u a l valence l e v e l s . In view of these observations, and taking into account the conclusions of Creber et a l . [142], peaks C and D are assigned as being p r i m a r i l y due to t r a n s i t i o n s to Rydberg le v e l s from occupied valence l e v e l s ( i . e . , t h i s assumes that both occupied and unoccupied valence o r b i t a l s relax to s i m i l a r extents, but both relax more than Rydberg l e v e l s ) . Secondly, assuming the above assignment of Rydberg and valence t r a n s i t i o n s i n the XPS spectrum ( F i g . 3.1, Table 3.3) i s correct, a comparison of i n t e n s i t i e s between XPS ( F i g . 3.1), ISEELS. ( F i g . 3.3), and also apparently VSEELS ( F i g . 3.5) shows an i n t e r e s t i n g reversal i n the r e l a t i v e i n t e n s i t i e s of t r a n s i t i o n s to Rydberg and valence l e v e l s . In ISEELS, t r a n s i t i o n s to valence l e v e l s predominate over those to Rydberg l e v e l s , whereas the reverse s i t u a t i o n seems to occur i n XPS. This behaviour i s not e n t i r e l y unexpected considering the p o t e n t i a l b a r r i e r attributed to the three F ligands i n NF 3. In NF 3 the p o t e n t i a l b a r r i e r (as can be seen from the ISEELS spectra and to some extent the VSEELS spectra here) separates the Rydberg and valence o r b i t a l s , causing the ISEELS (and possibly the VSEELS) to be valence dominated. However, i n the XPS of NF 3 there i s an a d d i t i o n a l hole (compared to the ISEELS and VSEELS), i . e . , one electron l e s s , and t h i s would surely extensively reduce the p o t e n t i a l b a r r i e r by siphoning of e l e c t r o n i c charge from the surrounding F atoms. In these circumstances we might expect to see a r e l a t i v e increase i n the r a t i o of - 118 -Rydberg to valence e x c i t a t i o n In the XPS "shake-up" spectrum. The F Is "shake-up" spectrum of NF 3 i s rather d i f f e r e n t from that of the N Is. This difference can be immediately a t t r i b u t e d to the reduction of symmetry from C^ v to C g upon removal of one of the F Is electrons. In this case the a^ o r b i t a l s become a", a 2 becomes a", and the e symmetry o r b i t a l s reduce to a' and a". Thus a large number of a' -»• a' and a" a" valence t r a n s i t i o n s become possi b l e . In keeping with the arguments used i n the case of N Is, features P, Q, S, and T are probably due to t r a n s i t i o n s to valence l e v e l s , whereas R i s due to t r a n s i t i o n s to Rydberg l e v e l s . F i n a l l y , a consideration of the various types of spectra and the i r r e l a t i v e energies for NH 3 and NF 3 lends some support to the preceeding conclusions. In p a r t i c u l a r the N Is ISEELS spectrum of NH 3 ([51] and Chapter 8) i s dominated by t r a n s i t i o n s to Rydberg-type l e v e l s since there i s no b a r r i e r i n NH 3. S i m i l a r l y , the XPS shake-up spectrum of NH 3 [142] appears to be Rydberg-like, with the major features being at s i m i l a r r e l a t i v e energies to the main Is l i n e as i n the case of NF 3. In the l i g h t of the present work i t would be of i n t e r e s t to study the various spectra of the molecules NHF 2 and NH2F i n comparison with those of NH 3 and NF 3« Conclusions It has been shown i n t h i s chapter that while each spectroscopy (ISEELS, VSEELS, and XPS) y i e l d s separate information on the e l e c t r o n i c structure of molecules, a consideration of a l l three i n conjunction with - 119 -each other can lead to a further understanding of each process and the molecular e l e c t r o n i c structure. A l l three spectroscopies give informa-t i o n on valence-valence and valence-Rydberg l e v e l e l e c t r o n i c t r a n s i -t i o n s . The ISEELS spectra were found to be t y p i c a l examples of molecules with highly electronegative ligands i n that they showed a strongly enhanced t r a n s i t i o n to d* l e v e l s , low-intensity Rydberg structure and features i n the continuum. A comparison with the XPS s a t e l l i t e structure indicates that the continuum structure can be associated with "shake-up" phenomena. The VSEELS spectrum show much structure that has been attributed to both valence-valence and valence-Rydberg t r a n s i t i o n s . Comparing the term values obtained from the ISEELS and VSEELS spectra indicates that those associated with Rydberg l e v e l s are transferable, whereas those associated with the v i r t u a l valence l e v e l s are generally not transferable. While more s p e c i f i c conclusions must await a d e t a i l e d and sophis-ti c a t e d t h e o r e t i c a l treatment of NF 3, the present studies have c l e a r l y indicated that the combined use of various electron spectroscopies provide more detai l e d i n s i g h t into fundamental i o n i s a t i o n and e x c i t a t i o n processes. In addition, further insights have been gained into the p o t e n t i a l b a r r i e r model. - 120 -CHAPTER 4 ELECTRON ENERGY LOSS SPECTRA OF THE SILICON 2p,2s, CARBON Is and VALENCE SHELLS OF TETRAMETHYLSILANE In the previous chapter the ISEELS and VSEELS spectra of a compound with highly electronegative ligands were presented and discussed. It would be i n s t r u c t i v e to now consider the spectra of a compound which does not have electronegative ligands. In this chapter the ISEELS spectra of tetramethyl silane (TMS), ( C H 3 ) 4 S i , i n the C Is, Si 2s and 2p regions as well as the VSEELS spectrum are presented. The Si 2p spectra are compared and contrasted with published photoabsorption spectra of SiF^, SiH^ and other related S i containing molecules with varying ligands to further examine the ef f e c t s of the ligand on i n t e n -s i t y d i s t r i b u t i o n within the spectra. TMS i s a substance of fundamental and p r a c t i c a l importance and i s used as a cal l b r a n t i n NMR spectroscopy. TMS i s very stable, however, to date only very l i m i t e d studies have been made of e l e c t r o n i c e x c i t a t i o n processes of TMS i n the gas phase. For example Roberge et a l . [143] have measured the valence s h e l l photoabsorption of TMS and some related molecules up to 85,000 cm - 1 ( i . e . up to an energy of 10.5 eV) which i s close to the upper l i m i t of l i g h t transmission by windows and lenses i n conventional o p t i c a l spectrometers. Above ~10 eV there are few sources of continuum r a d i a t i o n s u f f i c i e n t l y intense for obtain - 121 -ing d e t a i l e d photoabsorption spectra, p a r t i c u l a r l y i n the carbon and s i l i c o n core e x c i t a t i o n region (>100 eV). Synchrotron r a d i a t i o n provides a suitable l i g h t source, but to date no such study of TMS has been made, although s i m i l a r studies of the i s o e l e c t r o n i c molecule, SiF t t [144,145] and also s i l a n e , SiH l t [65,146] have been reported. Dehmer [73] has further discussed the Si 2p absorption spectrum of S i F 4 with reference to the e f f e c t i v e p o t e n t i a l b a r r i e r model. The Si 2p ( i . e . LJJ. e l e c t r o n i c e x c i t a t i o n of TMS and i t s chloro derivatives has been studied by Fomichev et a l . [147] using f i l t e r e d brehmsstrahlung rad i a t i o n from a tungsten anode i n the li m i t e d energy range from 102 -109 eV. No photoabsorption spectrum of either the carbon Is or the s i l i c o n Is regions of TMS has been published to date. In contrast to the li m i t e d studies made of electron e x c i t a t i o n i n TMS much work has been reported on the photoelectron spectra of both valence and core regions. These w i l l greatly a s s i s t i n the i n t e r p r e t a -t i o n of the electron e x c i t a t i o n spectra. He(I) photoelectron spectra of TMS have been reported by several groups [148-150]. The valence s h e l l PES spectrum of the i s o e l e c t r o n i c molecule, SiF t +, has also been published [149], S i 2p binding energies of TMS have been measured i n several laboratories [30,151-154]. Experimental D e t a i l s The inner s h e l l spectra were recorded using an impact energy of 2.5 keV with the scattered electrons sampled at ~1° scattering angle. The energy scales were established f o r both regions ( S i L - s h e l l and C - 122 -K-shell) with respect to the Ar (2p •*• 4s) t r a n s i t i o n at 244.37 eV. The valence s h e l l spectrum was obtained on the new spectrometer [53]. An impact energy of 3 keV was used with the scattered electrons sampled at zero degree sc a t t e r i n g angle. The spectrum was cal i b r a t e d with the He(I) resonance l i n e at 21.218 eV [104]. Results and Discussion The TMS molecule, SKCHg)^ i s of T j symmetry and the electron configuration and unoccupied valence o r b i t a l s may be written as [73, 148], ( l a 1 ) 2 ( 2 a 1 ) 2 ( l t 2 ) 6 ( 3 a 1 ) 2 ( 2 t 2 ) 6 ( 4 a 1 ) 2 ( 3 t 2 ) 6 ( 5 a 1 ) 2 ( 4 t 2 ) 6 ( l e ) ' + ( l t 1 ) 6 ( 5 t 2 ) 6 S i l s < C i s — ' Si2s Si2p » valence ' (6a 1)0(6t 2)°(2e)0(7t 2)0 i unoccupied ' The various spectra are u s e f u l l y discussed with respect to th i s configu-r a t i o n . 1. Inner S h e l l Spectra Figure 4.1 shows a continuous wide range scan of the S i 2p,2s and C Is regions of the e l e c t r o n i c e x c i t a t i o n spectrum of TMS between 40 and 380 eV. This survey spectrum was obtained at a re s o l u t i o n of 0.36 eV FWHM. The more prominent features are numbered on the figure c o n s i s -tent with the designations on the more de t a i l e d spectra shown i n Figures 3 O o >-C O L U Si 2s, 7 8 'v V A L E N C E \ A l _ S H E L L \ : : \ (CH3)4Si SILICON 2p, 2s CARBON Is c J T - T 235 AE = 0.36eV 100 2 0 0 250 E N E R G Y LOSS (eV) Figure 4.1: Wide range inner s h e l l electron energy loss spectrum of tetramethylsilane. - 124 -4.2 and 4.5. The assignments of the various i o n i z a t i o n edges are also discussed l a t e r . In the region up to ~100 eV the steeply f a l l i n g " t a i l " of the valence s h e l l i o n i z a t i o n continuum i s c l e a r l y v i s i b l e . The S i 2p spectrum i s dominated by two large peaks, one (feature 3) below and one (feature 6) above the i o n i z a t i o n l i m i t . Similar but less intense s t r u c -tures (features 7,8) are seen below and above the 2s i o n i z a t i o n l i m i t . A sharp increase i n cross section i s seen at ~285 eV, consistent with the onset of carbon Is ( i . e . K-shell) e x c i t a t i o n . A number of dis c r e t e states and continuum structures are apparently present and these are discussed i n d e t a i l below i n the section on C Is e x c i t a t i o n . The o v e r a l l spectrum (Figure 4.1) displays clear evidence of the separate subshells of S i and C at the expected positions. Since no large s h i f t s i n o s c i l l a t o r strength or s i g n i f i c a n t delayed onsets are observed, i t i s apparent that no strong i n t e r - s h e l l electron c o r r e l a t i o n e f f e c t s are occurring. Likewise we may expect that the S i Is spectrum i s even more atomic l i k e due to i t s comparative "energy i s o l a t i o n " at -1844 eV [155]. T h e - s i l i c o n (2p,2s) and carbon (Is) spectra have been further examined separately at high r e s o l u t i o n . Figure 4.2 (lower section) shows the S i 2p and 2s e x c i t a t i o n between 100 and 170 eV i n greater d e t a i l at an energy r e s o l u t i o n of 0.36 eV FWHM. The det a i l e d structure i n the large peak i n the region of ~105 eV i s shown at higher r e s o l u -tions (0.18 eV and 0.10 eV FWHM respectively) i n the two upper spectra of Figure 4.2. At least f i v e features (numbered 1-5) are c l e a r l y present i n t h i s part of the spectrum. The 2p_,~ i o n i z a t i o n edge (105.94 125 -2P3/22P|/2 rrnV-I 2345 AE=O.I8eV "c. Z3 o 3 o| >-V) z UJ I • I . I • I • I • I • I 2P3/2 2P|/2 r - n — r 12 3 4 5 (CKlSi SILICON 2p,2s i L J i L AE = O.IOeV J i I i L 102 106 110 114 100 102 104 2p3/22p rm 12349 1/2 106 106 A E = 0.36 eV 110 112 ~2s 7 t I ' • » • ' i I i I i I i I 1 L_ 100 110 120 130 KO 150 160 170 ENERGY LOSS ( e V ) Figure 4.2: S i l i c o n 2p electron energy loss spectra of tetramethylsilane. D e t a i l s are given i n Table 4.1. - 126 -ev) has been assigned using the XPS value for TMS reported by Kelfve et a l . [30]. Very s i m i l a r values have been recorded by Perry and J o l l y [151] and also by Drake et a l . [152,153]. A value of 100.7 eV reported by Gray et a l . [154] i s c l e a r l y grossly i n erro r . This i s not sur p r i s i n g i n view of the fact that an a r b i t r a r y (and i n c o r r e c t ) value of 285.0 eV was assigned to the carbon Is binding energy of TMS and t h i s was used [154] to c a l i b r a t e the scale (the correct value [152,153] should be 289.78 eV, as i s discussed below). A value of 0.61 eV for the 2 p3/2 1/2 s P * n o r b i t s p l i t t i n g has been used i n assigning the 2 p j y 2 i o n i z a t i o n edge i n accord with the findings of Kelfve et a l . [30] from t h e i r XPS work on a va r i e t y of s i l i c o n - c o n t a i n i n g compounds as well as photoabsorption studies on SiH^ and S i F ^ [65,144,146] and c r y s t a l l i n e s i l i c o n [156]. It i s to be expected that the magnitude of t h i s s p l i t t i n g i s l a r g e l y independent of the molecular type. The SI 2s i o n i z a t i o n edge (157.31 eV) has been assigned using the XPS value obtained by Venezia-Floriano and Cavell [157]. This gives a 2p-2s s p l i t t i n g of ~51.2 eV , i n close agreement with the estimated s p l i t t i n g of 51.4 eV suggested by Fomichev et a l . [147] for S i . A value of 41.5 eV for the 2p-2s s p l i t t i n g i n SiH^ reported by Hayes and Brown [146] i s almost c e r t a i n l y i n c o r r e c t . The energies, term values and possible assignments of the Si(2p, 2s) spectrum are shown i n Table 4.1. The energy values of features 1-5 obtained by Fomichev et a l . [147] using soft X-ray absorption are also shown and are consistently lower than the present ISEELS values by approximately 0.5 eV. The shapes and r e l a t i v e i n t e n s i t i e s of both the present ISEELS and soft X-ray spectra [147] are almost i d e n t i c a l i n the - 127 -Table 4.1 Energies and term values of features in the Si(2p,2s) spectrum of Si(CH 3)^ Feature Energy" Term Value Possible Assignment Photoabsorption^ (eV) 2 p 3 / 2 2 ? l / 2 2 p 3 / 2 2 p 1 / 2 (eV) 1 103.60 2.34 2.94 2p-*a1, t 2/4p - 103.15 2 104.19 1.75 2.35 2p*5s, 3d 2p-*a1> t£/4p 103.65 3 104.73 1.21 1.81 2p-5p 2p->5s, 3d 104.15 4 105.45 0.49 1.09 2p-4d 2p-5p 104.80 5 105.97 — 0.57 - 2p+4d 105.7 105.94a 0 — 2p-»=> - 105.7 ^Pl/2^ n l*' t 106.55 b — 0 - 2p-.» 6 124.1 - Resonance (o*(7t2)) 7 155.07 2. 2s->t2 2s l i m i t -157.31 c -157 .6 d 0 -6 -173 Resonance (o*(7t2)) a. XPS values from Reference [30]. References [151-153] give values of 105.83 eV and 106.02 eV respectively. b. The spin orbit s p l i t t i n g of 0.61 eV has been assigned following data i n reference [30]. Similar s p l i t t i n g s are reported elsewhere [65,144,146,156], c. A. Venezia-Floriano and R.G. Cavell, reference [157]. d. Foraichev et a l . [147] suggest a Si 2p-2s s p l i t t i n g of 51.4 eV. + With respect to the 2s l i m i t (157.31 eV, reference [157]). * Estimated uncertainty ±0.05 eV. - 128 -energy range 102-108 eV. An examination of the term values for peaks 1-5 with respect to both ^?^/2 a n c * ^ p l / 2 e c * S e s s n o w s that the peak spacing i s approximately the same as the spin-orbit s p l i t t i n g at the i o n i z a t i o n l i m i t s (~0.6 eV — see preceeding dis c u s s i o n ) . It might therefore be possible that a double overlapping Rydberg series converg-ing to the respective 2 p ^ ^ a n Q i° ni z ation l i m i t s could be c o n t r i -buting to the o v e r a l l 2p e x c i t a t i o n band shown i n Figure 4.2. However ap p l i c a t i o n of the Rydberg formula and the 2p i o n i z a t i o n l i m i t s i n d i c a -tes that no simple consistent assignment can be made on the basis of term values and quantum defects. Since only one broad structured band i s observed below the i o n i z a t i o n edge i t i s highly l i k e l y that t h i s encompasses (mixed) valence and Rydberg states. In p a r t i c u l a r from consideration of the expected term value i t i s evident that the 4s Rydberg i s missing i n the Si(2p,2s) spectra. It i s i n s t r u c t i v e at ths point to compare the Si 2p spectrum of TMS ( i . e . Si(CH 3)^) with the soft X-ray absorption spectrum [144-146] of the i s o e l e c t r o n l c molecule Si F ^ and also with the spectra of the species SiH^, S i C l 4 and ( s o l i d ) S i 0 2 . The soft X-ray spectra of some of these species are conveniently shown on one diagram i n the p u b l i c a t i o n by Dehmer [73]. These, as well as the spectrum of SiH^ [65] and the presently obtained Si 2p spectrum of TMS have been reproduced on the same energy scale and are shown i n Figure 4.3. Four peaks ( l a b e l l e d A, B, C, and D as i n reference [73]) occur In each of the S i 2p spectra of SiFj t, S i C l ^ and S10 2 whereas I t would seem (Figures 4.1 and 4.2) that only two prominent bands ex i s t i n the Si 2p spectrum of TMS. The - 1 2 9 -ENERGY LOSS (eV) Figure 4.3: S i l i c o n 2p e x c i t a t i o n spectra of various S i containing compounds with S i i n a tetrahedral environment; SiCCHj)^ t h i s work, S i C l ^ , S i 0 2 and S i F H as shown i n ref. [73]. Deta i l s are given i n Table 4.2. - 130 -difference i n behaviour between TMS and SiF t t can be a t t r i b u t e d to the high e l e c t r o n e g a t i v i t y of the F ligand whereas the CH 3 ligand i s l i k e l y to be electron donating [31,33]. Thus p o t e n t i a l (charge) b a r r i e r e f f e c t s [73] are l i k e l y to be much greater i n SiF t t than i n TMS, and indeed there may be no e f f e c t i v e b a r r i e r i n Si(CH 3) 1 +. If t h i s i s the case, then i n SiF^ the band C above the Si 2p edge can be a t t r i b u t e d to be an inner-well state trapped by the p o t e n t i a l b a r r i e r due to the surrounding f l u o r i n e ligands. Peaks of type C are also seen [73] i n S i C l ^ and S i 0 2 presumably because these species also have highly electronegative ligands. Very big differences were observed e a r l i e r i n the carbon K-shell spectra of CF^ [66,92,133] and CI^ [66,72], there being an obvious p o t e n t i a l b a r r i e r e f f e c t i n CF^ while CH^ has a more normal Rydberg type of spectrum. In t h i s regard i t i s of i n t e r e s t to compare the wide range Si(2p) spectra of TMS with those of SiH^ [65,146] and elemental Si [146,156] as well as with those of S i F ^ and the other related s i l i c o n containing species discussed above. The observed spectrum of Si(CH 3) l + i s very s i m i l a r to those of SiH l + [65,146] and S i [146,156] and t h i s lends further credence to the suggestion that poten-t i a l b a r r i e r e f f e c t s i n SKCHg)^ are e f f e c t i v e l y absent. Considering now d e t a i l s of the pre-edge structure i n the Si 2p spectrum of TMS (Figure 4.2) the large structure (103-110 eV) contains at least f i v e f a i r l y evenly spaced bands at a separation of approximately 0.5 eV. As has been discussed e a r l i e r , i t has not been possible to f i t the peaks to any obvious Rydberg states and furthermore the spacing i s not compatible with v i b r a t i o n a l structure. Similar conclusions were drawn by Dehmer - 131 -[73] for the corresponding band (B) i n SiF^, although Hayes and Brown [146] report having f i t t e d Rydberg series to the same band i n S i F ^ . Considering further the region of the TMS spectrum below the S i 2p i o n i z a t i o n edge only a single broad band (containing maxima 1-5, Figure 4.2) can be seen and this i s l i k e l y to encompass a mixture of valence and Rydberg states. In contrast for S i F ^ [144-146], S i 0 2 and S i C l ^ [73] two bands, A and B [73], are c l e a r l y present. S i m i l a r l y i n SiH^ [65,146] two bands are present below the S i 2p edge. However, a detail e d t h e o r e t i c a l analysis of a high res o l u t i o n spectrum of S i H 4 [146] shows that the f i r s t band i s due to the overlapping a (a^) and a ( t 2 ) valence states expected for molecules of Tj symmetry (TMS, SiH^) for which the unoccupied o r b i t a l s are, i n order of increasing energy; ( a 1 ) ( t 2 - p l i k e ) ( e ) ( t 2 - d l i k e ) * It i s also suggested [65] that the a (a^) has some 4s Rydberg character. The second band i n SiH^ i s found [65] to consist of several series of (sharp) Rydberg l i n e s . The nature of these assignments i n SIH^ have been confirmed by running the absorption spectra for s o l i d SiH^ [65], i n which case the Rydberg l e v e l s are suppressed. In S i F ^ the two v i r t u a l o r b i t a l s have been found to be well separated [73,144]. Peak A i n S i F ^ has been assigned as 2p -*• a (a^) and peak B i s a composite of 2p -* a ( t 2 ) and the various Rydberg l e v e l s [144]. Again the valence charac-t e r i s t i c s have been confirmed [144] by running the spectrum of the - 132 -s o l i d . As noted i n reference [144] the term value ( i n e f f e c t the * "binding energies" of electrons i n the v i r t u a l o r b i t a l s ) for the a (a^) l e v e l i s higher i n S i F ^ than i n SiH^ whereas the reverse i s true for the t 2 o r b i t a l . This i s d i r e c t l y a t t r i b u t a b l e to the e l e c t r o n e g a t i v i t y of the F ligand and therefore i t can be expected that as the electronega-t i v i t y of ligands decrease for a series of s i l i c o n compounds bands A (a^) and B ( t 2 ) should converge. This can c l e a r l y be seen i n going from S i F ^ •* S i C ^ •*• S i h ^ . Since -CH 3 i s even less electronegative and i n fact electron donating [31,33] i t i s suggested that the a± and t 2 l e v e l s i n TMS overlap and indeed i t i s possible that the order may even be reversed. A s i m i l a r trend of the merging of bands A ( a ^ l i k e ) and B(t 2-lik.e) can be seen i n the Si 2p absorption spectra [147] of (CH,) S i C l , , . as x goes from 0 to 4. Given a l l the considerations d x (4-x) discussed above, i t i s considered that feature 1, and the corresponding part of 2 are due mainly to the Y/2 c o m P o n e n t s °f t n e overlapped a± and t 2 valence bands. In view of the term value (2.35 eV) i t i s also possible that the t 2 i s mixed with the 4p Rydberg l e v e l . The S i 2s o r b i t a l i n TMS i s of aj^ symmetry and so only t r a n s i -tions to the t 2 o r b i t a l are dipole allowed. The term value for feature 7 with respect to the 2s edge (see Table 4.1) i s 2.24 eV. This feature i s therefore assigned as the 2s -*• a ( t 2 ) t r a n s i t i o n with possibly some contribution from the 4p Rydberg i n view of the term value. This i s i n accord with the term value for feature 1 of the Si 2p spectrum which lends further support to the suggestion that the a^ and t 2 states are overlapped or even reversed i n the S i 2p spectrum as discussed above. - 133 -The close proximity of the S i 2p edge would lead to overlap of the t r a n s i t i o n s to the valence and Rydberg l e v e l s , thus precluding any simple Rydberg analysis as has also been suggested i n the case of band B ( t 2 ) for S i F ^ [73]. However, using approximate quantum defects of 2.0 ( s ) , 1.6 (p), and 0.0 (d) applicable for t h i r d row (Na-Ar) atom contain-ing molecules, the following term values for the 2p •»• Rydberg t r a n s i -tions would be expected: 4s (3.40 eV), 4p (2.36 eV), 5s and 3d (1.51 eV), 5p (1.18 eV) and 4d (0.85 eV). C l e a r l y the 4s t r a n s i t i o n i s not observed i n the Si L - s h e l l spectra (but see discussion on C Is spectrum). Features 1 and 2 could be i d e n t i f i e d with a t r a n s i t i o n to the 4p Rydberg l e v e l ; however, this involves a p •> p t r a n s i t i o n , which i s dipole forbidden i n the purely atomic case. Thus a strong t r a n s i t i o n ( e s p e c i a l l y considering that the 4s t r a n s i t i o n i s seemingly absent) i s not expected and features 1 and 2 are better ascribed to the a o r b i t a l s as discussed above. This i n t e r p r e t a t i o n i s supported by those given on other s i l i c o n 2p spectra [65,144] where no t r a n s i t i o n s to p Rydberg le v e l s have been assigned. In any case, the 4p Rydberg l e v e l , being of * t 2 symmetry, i s l i k e l y to mix with the a ( t 2 ) o r b i t a l to form a * o" ( t 2 ) / 4 p mixed valence-Rydberg state [65]. The width and lack of structure c l e a r l y Indicate the presence of valence character. Tentative assignments based upon a l l of the above considerations are given i n Table 4.1. Moving on to the structure i n the region of and beyond the 2p edge i t would appear that the feature C i n the S i F 4 and S i C l ^ spectrum [6] Is absent i n TMS. As stated e a r l i e r peak C i s l i k e l y to be a a (2e) inner well state, resonance enhanced [73,75,77]. This - 134 -i n t e r p r e t a t i o n i s supported by the absence of such a peak i n SiH^ [65,146] and c r y s t a l l i n e S i [146,156]. A recent study [158] of the asymmetry parameter, 6, for photoemission from the S i 2p l e v e l of S i F ^ and S i ( C H 3 ) 4 ( i . e . TMS) shows marked differences i n 8 for the two species i n the region 5-16 eV above the respective 2p i o n i z a t i o n edges -exactly where peak C i s situated i n S i F ^ [144-146] but apparently absent i n S i ( C H 3 ) ^ (see Figure 4.2). The state corresponding to peak C i n S i F ^ i s expected to quantum mechanically i n t e r f e r e with the underlying 2p d i r e c t i o n i z a t i o n continua thus causing v a r i a t i o n s i n 8 compared with the simpler continuum processes i n t h i s region i n the case of TMS. A broad continuum feature can be seen i n the S i 2p spectra shown * i n F i g . 4.3 (designated as D). This can be assigned to a d - l i k e a ( 7 t 2 ) shape-resonance feature [73,159]. As was discussed i n section F of Chapter 2, the p o s i t i o n of t h i s feature should be related i n some way to the bond distance between the ionised atom and i t s neighbour(s). Indeed, within a simple k i n e t i c scattering picture the resonance p o s i -t i o n from the i o n i s a t i o n edge ( 6 ) should vary l i n e a r l y with R~ 2 where R i s the bond distance [96] (see equation (1.F.3)). N a t o l i [97,98] has placed this r e l a t i o n s h i p on firmer t h e o r e t i c a l grounds but has indicated that the k i n e t i c energy of the electron ( 6) should be referenced to an intramolecular p o t e n t i a l (V Q) - see equation (1.F.4), which depends on the atoms involved. Hitchcock et a l . [100] have shown that a simple l i n e a r r e l a t i o n s h i p (equation 1.F.5) i s adequate to r e l a t e bond distance and resonance p o s i t i o n i n a series of hydrocarbons. In view of these - 135 -discussions, the r e l a t i o n s h i p of the resonance p o s i t i o n (D - F i g . 4.3) for these Si-containing compounds with bond length has been examined. Table 4.2 summarises the relevant data from F i g . 4.3 and other l i t e r a t u r e sources. The energies of the features D ( i . e . the t 2 compo-nent) as a function of S i - l i g a n d bond distance give points that l i e close to straight l i n e s (see F i g . 4.4) of the forms 6 = 38.0 I T 2 + 6.4 (r = 0.973) 6 = -14.7 R + 45.1 (r = -0.965) It can be seen that at least for tetrahedral type s i l i c o n - c o n t a i n i n g species the i n t u i t i v e l y more reasonable [96-98] R - 2 v a r i a t i o n gives at least as good a l i n e a r f i t as the more empirical R dependence selected by Hitchcock et a l . [100] i n the case of carbon-carbon s i t e containing molecules. Also shown i n F i g . 4.4 are the corresponding points for R and R~2 dependencies for the features C, which are i n the continuum ( i . e . , S i F ^ and S i 0 2 ) and also that for S i C l ^ , which i s a discrete t r a n s i t i o n below the S i 2p edge ( 6 = -0.8 eV). It Is also of i n t e r e s t that the p l o t s for features C and D are e s s e n t i a l l y p a r a l l e l i n each representation. From the l i n e a r dependencies estimates have been made of the expected p o s i -tions of type C features i n SiH^ and S i ( C H 3 ) i + the predicted p o s i t i o n ( 6 = +1 eV) i s at ~107 eV i n the very broad peak ( F i g . 2) around the S i 2p edge. The feature C i s generally weaker than D and t h i s i s probably why i t i s not apparent i n SiH^ where D i s already rather weak. The lack of - 136 -Shape-resonance (features C and D, F i g . 4.3) p o s i t i o n s , resonance term values ( 6 ) and bond lengths for rel a t i o n s h i p s shown i n F i g . 4.4. Molecule p o s i t i o n of shape-resonance ( e V ) ^ a ' S i 2 p ( b ) IP (eV) Resonance term value 6 = E - IP (eV) Bond length R ( A ) ( c ) D C D C R-2 R S l ( C H 3 ) k 124.1(5) 107.1+ 106.1(1) 18.0(6) -1.0+ 0.284 1.875(2) SiH\ 131.9(10) ( d ) 113.5+ 107.3 24.6(10) 8.1+ 0.456 1.481(1) S1C1„ 125.8(10) 109.6 110.4(1) 15.4(11) -0.8 0.245 2.019(4) S i 0 2 129.5(5) 115.0 108.5(10) 21.0(15) 6.5 0.386 1.61 SiF^ 133.0(5) 117.2 111.9(1) 21.1(6) 5.3 0.414 1.554(4) (a) (b) (c) (d) Sit C H j ) ^ t h i s work; SiF^ - Ref. [144]; S i H 4 , S i C l 4 , S i 0 2 - ref . [73]. S K C H ^ , S i H 4 , S i C l M S i F 4 , - re f . [30]. 0.2 eV has been added to the Si 2p-jy2 values given i n re f . [30] to give an 'average' S i 2p value for the IP. S i 0 2 - ref . [160]. 5.0 eV has been added to the values given i n ref. [160] to bring the IP to the vacuum l e v e l see r e f . [161] for a discussion of t h i s correction procedure. SiCCHj)^, SiH^, S i C l 4 , SiF^ - re f . [162], S i 0 2 - ref.. [163]. This pos i t i o n was estimated from the spectrum of Si H 4 given i n re f . [146]. Estimated from F i g . 4.4 using the R~2 dependence. Figure 4.4: Relationship of bond length to shape resonance (features C and D on Figure 4.3) term value (6) for s i l i c o n containing molecules. Variations are shown of 6 with 1/R2 ( l e f t hand side) and R (r i g h t hand s i d e ) ; observed values are shown as s o l i d c i r c l e s with the open c i r c l e s being estimated values from the l i n e a r plots as shown. - 138 -prominent features i n the continuum of the SiH^ spectrum i s not s u r p r i s -ing due to the low scattering power of the (small) H ligands. This observation lends further general support to the atom-atom sca t t e r i n g viewpoint for understanding resonance features i n molecular spectra. The above results for tetrahedral ligand systems surrounding s i l i c o n c l e a r l y support the existence of a simple r e l a t i o n s h i p between resonance p o s i t i o n and the ionized (excited) atom-ligand internuclear separation. The carbon Is (K-shell) spectrum of TMS i s shown i n Figure 4.5 at an energy res o l u t i o n of 0.36 eV FWHM. The i n s e r t shows a more de t a i l e d view of the lower energy region at higher resolutionm (0.21 eV FWHM). To date no other carbon Is spectrum of TMS has been reported. The ener-gies, term values and possible assignments are shown i n Table 4.3 together with the carbon Is binding energy as determined by XPS [152, 153]. The spectrum i s s t r i k i n g l y s i m i l a r i n appearance to the C Is spectrum of CH^ [67,72] and i n t h i s regard the analysis could be consi-dered i n terms of a substituted methane. The term values for the f i r s t two peaks i n TMS ( i . e . (1+2) and 3) are 3.40 eV and 2.47 eV whereas those i n are 3.70 eV and 2.70 eV. The t r a n s i t i o n s i n CH^ have been assigned as going to the lowest Rydberg l e v e l s which are of 3s (a^) and 3p ( t 2 ) symmetry r e s p e c t i v e l y . The former t r a n s i t i o n ( i . e . to 3s ( a p i s allowed i n CH^ only by vibronic coupling [67]. However i n TMS the C Is o r b i t a l s transform as both a^ and t 2 and so dipole allowed t r a n s i -tions to both a^ and t 2 o r b i t a l s can be expected. A l t e r n a t i v e l y i f the spectrum i s analysed i n terms of a substituted methane, the l o c a l symmetry i s C, and t r a n s i t i o n s from the Is (a^) o r b i t a l are dipole - 139 -is n r 23 5 A \ (CH3)4S CARBON Is -•s n—i-i r^ 1-12 3 4 5 A E = 0.2leV i • i i I i i — i 1 — i — 264 286 2B8 290 292 294 A E = 0.36eV i i I I I I I I L _ 290 3 0 0 310 320 3 3 0 ENERGY LOSS(eV) Figure 4.5: Carbon Is electron energy loss spectrum of tetramethylsilane. Details are given i n Table 4.3. - 140 -Table 4.3 Energies, term values and possible assignments of features in the C(ls) spectrum of Si(CH 3) 4 Feature Energy Term Value Possible Assignment eV 286.26 3.52 Is + 4s (v=0) 286.51 3.27 Is -»• 4s (v-1) 287.31 2.47 Is + 4p Is + o*(a 1,t 2) -289 .6 Is limit 289.78a 0 Is ~303 shape resonance (o*(7t 2)) a References [152,153]. * Estimated uncertainty ±0.05 eV for peaks 1-3. - 141 -allowed to le v e l s of a^ or e symmetry. Thus peaks (1+2) and 3 can be assigned as t r a n s i t i o n s from the C Is to a Rydberg 4s (a^) o r b i t a l and a Rydberg 4p (a^,3) l e v e l ^ . The increase of i n t e n s i t y for the former t r a n s i t i o n as compared to that i n C r ^ i s a t t r i b u t a b l e to i t being d i r e c t l y allowed instead of v i a a vibronic coupling mechanism. A closer look at the f i r s t feature shows that i t c l e a r l y consists of two components (1+2) separated by ~0.25 eV. This type of phenomenon has also been observed i n halogen mono-substituted methanes [63,64] where a separation of 0.30 eV was observed. These were assigned as the v=0 and v=l components of the corresponding C Is •* ns (a^) t r a n s i t i o n s and therefore features 1 and 2 i n the C Is spectrum of TMS have been assigned i n a s i m i l a r manner. Compared to the other substituted methanes feature 3 i s broader and lacks the v i b r a t i o n a l structure which can be c l e a r l y seen i n the others [63,64]. I t i s suggested that feature 3 consists of the Is •> 4p Rydberg t r a n s i t i o n which l i e s on top of broader C Is ->• a t r a n s i t i o n (a^ and t 2)« The term value obtained from the Si 2p spectrum (see Table 4.1) agrees with this assignment. The r e l a t i v e i n t e n s i t i e s and spectral shapes of feature 3 and features (1+2) are consistent with such an i n t e r p r e t a t i o n . The large peak marked 5 on Figure 4.5 i s mainly due to unresolved higher Rydberg l e v e l s converging on the C Is i o n i s a t i o n edge. A low i n t e n s i t y broad feature (6) i n the i o n i s a t i o n continuum may be assigned as a shape resonance (probably o * ( 7 t 2 ) ) . ^ In the case of TMS the lower Rydberg lev e l s are designated as 4s and 4p since the c e n t r a l atom Is s i l i c o n . - 142 -It i s of i n t e r e s t to note the complete absence of a Si 2p -*• 4s (a±) Rydberg t r a n s i t i o n since no features appear i n the spectrum at the expected term value (~3 eV). This appears to be an example of a case where the f i r s t Rydberg l e v e l belongs to the ligands as opposed to the molecule as a whole. Examples have been seen [69] where the p r o b a b i l i -t i e s of t r a n s i t i o n s to Rydberg l e v e l s are low compared to valence o r b i t a l s . This i s the case when there e x i s t s an e f f e c t i v e p o t e n t i a l b a r r i e r leading to inner well and outer well states. This i s not the case for TMS as can be seen from the C Is spectrum which shows a normal Rydberg type structure unlike the F Is spectra of SF 6 [69] and SiF l t [145], both of which exhibit r e l a t i v e l y intense inner well type sta t e s . 2. Valence She l l Spectrum The valence s h e l l spectrum of TMS between 6 and 29 eV i s shown i n Figure 4.6 and summarized, along with tentative assignments, i n Table 4.4. The i o n i z a t i o n l i m i t s shown on Figure 4.6 are taken from measure-ments made by photoelectron spectroscopy [149,164]. Previously published valence s h e l l e x c i t a t i o n spectra, obtained by UV absorption, extend only as far as ~9.2 eV [167], ~10.5 eV [143] and ~11.3 eV [168]. The spectrum reported here shows four d i s t i n c t bands with p a r t i a l l y resolved f i n e stucture c l e a r l y evident i n each band. These features a r i s e from t r a n s i t i o n s to unoccupied v i r t u a l valence l e v e l s and/or Rydberg l e v e l s . In order to i d e n t i f y which features a r i s e from Rydberg t r a n s i -tions two assumptions have been made. F i r s t l y i t has been assumed that -fc 'c 3 o 1 5 CO LU 1 a be m ns I •. L |10 8 11 4 3 I 56 12 13 2 a be m nd L ? . 3 4 P 3 4 nd |_ ns I 3 4 4t, ndi 3 4 J L -5t, 5 6 nd(t2) i 5 6 np 3 4 Y 5 a f J L 4 5 6 J i L j i i (CH3)4S, VALENCE AE = 0.035 eV 10 15 20 ENERGY LOSS (eV) Figure 4 . 6 : Valence s h e l l electron energy loss spectrum of tetramethylsllane. Estimated positions of the f i r s t few members of the Rydberg series are shown below the spectrum. Details are given i n Table 4 . 4 . - 144 -Table 4.4 Energies and Possible Assignments of Valeria? Shell Transitions of S K O ^ Feature Observed* Energy (eV) Predicted Energies of Rydberg transitions (eV) IP (ev) a 4s 4P 3d 4d a 1 b c a 2 b c 10 11 7.04 7.34 7.64 8.60 8.93 9.20 10.05 10.61 11.00 11.50 11.81 12.14 12.86 13.26 13.7 14.7 15.7 6.90. 7.22k 7.5 ' 5t 2) 10.4 10.7 (4t 2) 13.1 (53^ 8.62, 8.95 K5t 2) 9.2 ' 8.78 9, 9.4 78 llf(5t 2) 11.0 ) ( l t l ) 11.2) 11.5 (le) 12.1 12.4 (4t,) 12.3 12.6 (4t,) 13.9 (Sa^ 9.37, 9.70K5t2) 10.0 ' 10.29 10.62 10.9 (5c,) ( l t r ) 11.8 12.0 12.3 (le) 12.7 I 12.9 13.2 12.9 ^ l t . ) (4t 7) 14.4 (Sa^ 14.7 (5sL) 13.2 (le) 13.8 ) (4t 2) 14.1) 15.6 (Sa^ ^ Originating orbitals are given In parenthesis after the estimated transition energies. a Binding energies from ref. [149] except le [164], Order of orbitals as per ref. [164-165]. * Estimated uncertainty +0.03 eV. - 145 -the term values for the Rydberg l e v e l s are transferable between the VSEELS and ISEELS spectra. This was discussed i n the previous chapter. From the C Is spectrum of TMS (Figure 4.5, Table 4.3) the mean term value for the ( l o c a l i s e d - see following discussion of C Is spectrum) 4s Rydberg l e v e l (feature (1+2)) i s 3.40 eV and that for the ( l o c a l i s e d ) 4p Rydberg l e v e l (feature 3) Is 2.47 eV. Using these to calculate the quantum defect by means of the Rydberg formula (equation ( l . B . l ) ) gives r i s e to term values T of 1.51 eV and 1.21 eV for the 5s and 5p Rydberg le v e l s r e s p e c t i v e l y . The S i 2p spectrum (Figure 4.2) should i n p r i n c i -ple give the 3d term value. However, the 3d l e v e l i s not resolved from the 5s and so the value of the 3d term value can only be estimated to have an upper l i m i t of -4.8 eV. Robin [170] has noted that the term value for the lowest d l e v e l Is close to 13,500 cm-* (1.67 eV) regard-less of the compound's chemical nature and so 1.67 eV and 0.92 eV (obtained from the 3d quantum defect calculated using the Rydberg formula) have been used for the 3d and 4d term values r e s p e c t i v e l y . Secondly, since TMS i s of T^ symmetry and the Rydberg o r b i t a l s transform as a^ for the s l e v e l s , t 2 for the p l e v e l s , and e + t 2 for the d l e v e l s , the following valence-Rydberg t r a n s i t i o n s are dipole allowed on the basis of symmetry considerations: t 2 ->• n s ( a 1 ) *1» e» c l > fc2 + n p ( t 2 ) t l f t 2 •*• nd(e) a L , e, t l f t 2 -»• n d ( t 2 ) - 146 -However, t r a n s i t i o n s which would be formally dipole forbidden i n the case of atomic systems ( i . e . s ->• s, p p, d + d , s d) have been assumed to exhibit less i n t e n s i t y than those which are dipole allowed i n the atomic case (see Chapter 1, section E). This i s most l i k e l y to be the case where the MO's are made up of predominantly one (A) type of heavy atom atomic o r b i t a l [164]. The HOMO o r b i t a l i s the 5 t 2 l e v e l which i s a bonding o r b i t a l comprised mainly of C 2p and Si 3p parentage [149,164] and so t r a n s i -tions to the ns(a^) l e v e l s should be seen. The UPS spectrum [149] shows the 5 t 2 l e v e l to be Jahn-Teller s p l i t into three components with i o n i z a -t i o n energies of 10.29, 10.62 and 10.90 eV. Thus the 5 t 2 -»> ns(a 1) t r a n s i t i o n s could be expected to show three Jahn-Teller components separated by ~0.3 eV with estimated energies of 6.90, 7.22 and 7.50 eV respectively for the f i r s t Rydberg l e v e l . The f i r s t feature i n the VSEELS spectrum of TMS does indeed show evidence of three components separated by ~0.3 eV with energies of ~7.0, ~7.3 and ~7.6 eV and there-fore these have been assigned accordingly. The centre of the second feature has an energy of 8.93 eV which gives an estimated term value of 1.69 eV ( i . e . 10.62 eV - 8.93 eV). Feature 2 i s thus assigned as being due predominantly to 5 t 2 * 3d t r a n s i t i o n s . Again the feature i s seen to consist of several components ( t h i s i s clearer i n the UV spectrum reported i n r e f . [167]). This may be due to Jahn-Teller s p l i t t i n g and/or to t r a n s i t i o n s to both the 3d(e) and 3d(t 2) l e v e l s . However, thi s feature (2) also coincides with t r a n s i t i o n s from the 5 t 9 l e v e l to - 147 -the 5s(aj^) l e v e l . The symmetry allowed 5 t 2 •*• 4 p ( t 2 ) t r a n s i t i o n would be expected to occur at ~6.1 eV. No sharp structure i s observed i n t h i s region, consistent with the assumptions that the 5 t 2 o r b i t a l i s mainly of p character and hence has l i t t l e p r o b a b i l i t y of t r a n s i t i o n s to p l e v e l s . The rest of the spectrum has been assigned i n a s i m i l a r manner as indicated i n Table 4.4 and Figure 4.6. With the exception of the le and 5a ^ o r b i t a l s , the binding energies of the outer valence electrons have been taken from the UPS spectrum reported by Jonas et a l . [149]. Their assignment [149], based upon a CNDO/2 c a l c u l a t i o n , does not agree with the proposed assignment of the X-ray photoelectron spectrum reported by Perry and J o l l y [164] based upon i n t e n s i t y considerations and extended HUckel c a l c u l a t i o n s as well as various ab i n i t i o c a l c u l a t i o n s [165,166]. The feature at 15.6 eV i n the UPS spectrum o r i g i n a l l y assigned to the le o r b i t a l [149] has subsequently been at t r i b u t e d to the 5a ^ o r b i t a l , i n accord with the other assignments [164-166]. The value of the le i o n i z a t i o n p o t e n t i a l given by Perry and J o l l y [164] has been used i n the present work. Following the 5 t 2 o r b i t a l are the It , le and 4 t 2 o r b i t a l s which have been considered to have mainly C-H bonding character [149,164] and thus to possess a large C 2p component. The 4 t 2 o r b i t a l , however, does have some S i 3p character attributed to i t [149,164], The intense feature centred at p o s i t i o n 4 i s a t t r i b u t a b l e to the 4 t 2 •* 4s(ai) t r a n s i t i o n . Features 5-8, which are much less intense, can be attributed to t r a n s i t i o n s from these three o r b i t a l s ( l t ^ , le and 4 t 2 ) to the 3d l e v e l s . The symmetry allowed t r a n s i t i o n s to the n p ( t 2 ) l e v e l s - 148 -have been assumed to have l i t t l e or no i n t e n s i t y since they would involve predominantly p •*• p t r a n s i t i o n s . The f i n a l outer valence o r b i t a l i s the 5a! which consists mainly of Si 3s and C 2p atomic o r b i t a l s [149,164]. The 5aj^ -* 4 p ( t 2 ) t r a n s i t i o n coincides with the very intense structure around features 9 and 10 which i s also where the i o n i -zation l i m i t s of the l e , l t j and 4 t 2 o r b i t a l s occur. Transitions from the 5a^ o r b i t a l to the 3 d ( t 2 ) l e v e l s may also contribute. It can be seen that the t r a n s i t i o n energies and structures observed i n the spectrum a r i s i n g from the various o r b i t a l s support the assumptions made at the beginning of this section and i n Chapter 1, section 5 concerning the r e l a t i v e i n t e n s i t i e s of allowed and forbidden "atomic-like" t r a n s i -tions i n the l i g h t of the suggested atomic o r b i t a l compositions of the various molecular o r b i t a l s [147,164], Generally the agreement between the observed features and the predicted positions of the valence-Rydberg t r a n s i t i o n s are quite good. This give more support to the concept of t r a n s f e r a b i l i t y of Rydberg term values between inner s h e l l and valence electron excitations as discussed i n Chapter 3. So far nothing has been said about valence-valence t r a n s i t i o n s . While much of the structure can be attributed to valence-Rydberg t r a n s i t i o n s the p o s s i b i l i t y of co n s i -derable underlying i n t e n s i t y due to broad valence-valence t r a n s i t i o n s cannot be ruled out. This conclusion was reached i n the assignment of the VSEELS spectrum of NF 3. This i n t e r p r e t a t i o n would also be consis-tent with that for the S i 2p s h e l l spectra of TMS which has been assigned as con s i s t i n g of overlapping t r a n s i t i o n s to both valence and Rydberg l e v e l s . Thus valence-valence t r a n s i t i o n s a r i s i n g from MO's with - 149 -S i character ( i . e . : 5 t 2 , 5a^) could be expected. The i n t e r p r e t a t i o n of Rydberg t r a n s i t i o n s on top of valence t r a n s i t i o n s i s also consistent with the spectral i n t e n s i t y d i s t r i b u t i o n . However, no d e f i n i t e conclu-sions can be made u n t i l good qu a l i t y c a l c u l a t i o n s have been made for the TMS molecule. - 150 -CHAPTER 5 ELECTRONIC EXCITATIONS IN PHOSPHORUS CONTAINING MOLECULES» I. INNER SHELL ELECTRON ENERGY LOSS SPECTRA OF PH,, P(CH,),, PF q AND PCI,. It was seen i n the previous chapters that the ligand has a profound e f f e c t on the i n t e n s i t y d i s t r i b u t i o n s observed i n inner s h e l l electron e x c i t a t i o n spectra. This was c l e a r l y seen when contrasting the Si L - s h e l l spectrum of SiCCHj)^ with other substituted silanes (see F i g . 4.3). As a continuation of these studies, the energy loss spectra of several phosphorus containing compounds are now reported. To date there have been no electron impact studies on the inner s h e l l e x c i t a t i o n spectra of any phosphorus compounds and only l i m i t e d photoabsorption studies on PC1 3 [171-173], 0PC1 3 [173,174], SPCI3 [173] and PH 3 [65,146]. However, there have been several t h e o r e t i c a l discussions on the P 2p spectrum of PH 3 [60,61,65,170], In t h i s chapter, the ISEELS spectra of the t r i v a l e n t phosphorus compounds PX 3 (X = H, F, C l and CH 3) for the P L - s h e l l regions (F and C K - s h e l l , C l L - s h e l l ) are presented. Following chapters w i l l deal with the ISEELS spectra of the higher coordinate phosphorus compounds PF 5, 0 P F 3 and 0PC1 3 and with the VSEELS spectra of some of these compounds. Experimental D e t a i l s . The spectra were a l l recorded on the ISEELS spectrometer - 151 -described i n Chapter 2. Unless otherwise stated i n the text, a l l the spectra were obtained using an impact energy of 2.5 keV with the scattered electrons sampled at ~1° sca t t e r i n g angle. A l l the P 2p spectra were c a l i b r a t e d against the N 2 (N Is -*• i t , v=l) feature at 401.10 eV. The other spectra were i n t e r n a l l y c a l i b r a t e d against th e i r respective P 2p features. RESULTS AND DISCUSSION Phosphorus L-Shell (2p and 2s) Spectra - General Features The long range spectra of the P 2p,2s (L s h e l l ) region are shown i n F i g . 5.1. The spectra were recorded at a resolution of 0.36 eV FWHM. The assigned i o n i s a t i o n edges are taken from XPS values [31,175]. Only the P 2p average positions of the Zp-^/Z 1/2 ^ o u ^ ^ e t w e r e reported [31] due to the l i m i t e d energy r e s o l u t i o n . Therefore an estimated spin-or b i t s p l i t t i n g of 0.90 eV has been used [61], along with a s t a t i s t i c a l weighting of 2:1 to predict the positions of the respective 2p 3/ 2 (2p - 0.30 eV) and 2 p ^ 2 (2p + 0.60 eV) edges. Using t h i s procedure the values obtained i n the case of PH^ agree well with those reported by Schwarz [61]. Before examining each spectrum i n d e t a i l , i t i s of i n t e r e s t to note the general s i m i l a r i t i e s between the long range spectra presented here ( F i g . 5.1) and the spectra of the corresponding s i l i c o n compounds (see F i g . 4.3). Each phosphorus L s h e l l spectrum ( F i g . 5.1) shows a broad continuum structure (at ~150 - 160 eV i n the phosphorus s e r i e s ) . The feature was att r i b u t e d to a d - l i k e shape-resonance [73,77] i n the s i l i c o n series and presumably an i n t e r p r e t a t i o n of a s i m i l a r structure - 152 --i 1 T 1 1 1 1 1 1 1 1 1 1 i i i r 10 5 10 n r 3 5 14 P 2p edge A E = 0.36eV PHOSPHORUS 2p,2s REGION P F , P C I , £ I O < 5 H 10 H 5 H P 2p edge P H , P ( C H 3 ) 3 ~ r — i — i — ' — i — ' — i — ' — i — ' — i — • — i — • — i — 1 — n ~ 130 140 150 160 170 180 190 200 210 ENERGY LOSS (eV) Figure 5.1: Phosphorus 2p,2s wide range electron energy loss spectra of PF 3, PC1 3, PH 3 and P(CH 3) 3. A l l spectra were obtained with an impact energy of 2500 V, a scattering angle ~1°, and a res o l u t i o n of 0.36 eV FWHM. - 153 -can be given here. S i m i l a r i t i e s between the spectra of respective S i and P containing compounds i n the discre t e part of the spectrum are: ( i ) the hydrides both show a broad feature followed by Rydberg structure leading to the edges ( i i ) the f l u o r i d e s show two well separated major bands, the second with apparent Rydberg structure on top ( i i i ) the chlorides both show merging bands well below the edge ( i v ) a l l the discre t e structures i n the methyl compounds are v i r t u a l l y on top of one another and very close to the edge (v) Both the f l u o r i d e s and chlorides show a strong inner well trapped [73] or resonance state [77]. In the f l u o r i d e s t h i s i s j u s t above the edge, while i n the chlorides i t i s r i g h t at the edge. It can thus be seen that the ligands have a very s i m i l a r e f f e c t i n both s e r i e s . These ideas w i l l be further discussed i n a l a t e r section. The assignment of the detai l e d spectra for each molecule are now considered. Phosphorus 2p and 2s Spectra - Discrete Regions The molecules, PX 3, are of C^ v symmetry. In a minimum basis set (d o r b i t a l s excluded) the empty molecular o r b i t a l s are of a^ and e symmetry (the -CH 3 group has been considered as one u n i t ) . Table 5.1 l i s t s the dipole allowed t r a n s i t i o n s to these l e v e l s . Since the P 2p o r b i t a l s transform as a^ and e, and the P 2s o r b i t a l i s of a± symmetry, tr a n s i t i o n s from both these o r b i t a l s to both v i r t u a l o r b i t a l s are allowed. In order to ascertain the positions of the t r a n s i t i o n s to the - 154 -TABLE 5.1 T r a n s i t i o n s from the 1 A 1 Ground State f o r C 3 v Symmetry F i n a l C o n f i g u r a t i o n * F i n a l State Dipole Allowed from ground stat e hole st a t e occupied o* o r b i t a l a l a l A l Yes a l e E Yes e a l E Yes e e Aj + E Yes e e A 2 No The (2p3/2)~ » ( ^ p ^ ^ ) " a n <* (2s)"~ holes are of e, a, and a^ symmetry r e s p e c t i v e l y . The o* o r b i t a l s are of a^ and e symmetry. - 155 -v i r t u a l o r b i t a l s i n the P L - s h e l l ( i . e . , 2p and 2s) spectra two assumptions have been made. Namely: ( i ) The major feature i n the 2s energy loss i s assumed to be 2s •+• * a (e) t r a n s i t i o n ( i i ) The term values for features i n the P 2s spectrum are assumed to be transferable to the P 2p spectrum. The f i r s t assumption i s based upon the make-up of the a (e) * * o r b i t a l as compared to the a ( a ^ . The a (e) o r b i t a l should have a larger proportion of phosphorus 3p o r b i t a l character (mainly 3p , 3p ) x y and l i t t l e phosphorus 3s o r b i t a l character i n comparison with the a (aj) o r b i t a l , which i s mainly 3s, 3p z*» This i s supported by CNDO/2 calc u l a t i o n s we have performed, as well as by Xa cal c u l a t i o n s [176]. Thus t r a n s i t i o n s from an s o r b i t a l to the a ( a ^ l e v e l should be weaker than those to the a (e) l e v e l since the former would have a larger s •*• s component which i s formally dipole forbidden i n the case of atomic systems. The second assumption concerning term value t r a n s f e r a b i l i t y has been discussed e a r l i e r (Chapter 1, section E; Chapter 3). The 2s and 2p core-hole vacancies are on the same atom and therefore the electrons should see v i r t u a l l y the same cen t r a l core p o t e n t i a l . However, small differences i n term value may occur due to the d i f f e r e n t s h i e l d i n g Here and i n a l l cases the p r i n c i p a l axis of the molecule has been designated as the z ax i s . - 156 -c a p a b i l i t i e s of the s and p o r b i t a l s . The 2p and 2s regions of the spectra shown i n F i g . 5.1 are shown i n d e t a i l i n Figs. 5.2-5.5, which are discussed and presented for each molecule i n the following sections. The P 2p regions i n Figs. 5.2-5.5 were run at a reso l u t i o n of 0.18 eV FWH while the 2s regions have been extracted from the long range spectra ( F i g . 5.1) which are at a r e s o l u -t i o n of 0.36 eV. The r e l a t i v e energy scales i n Figs. 5.2-5.5 are the same for the respective 2p and 2s spectra and these have been aligned according to t h e i r respective i o n i z a t i o n edges which were determined from XPS measurements as described above. The 2s spectra as shown were obtained by subtracting a l i n e a r ramp background from the t o t a l spectrum (F i g . 5.1) so as to more c l e a r l y display the spectral features. In a l l cases (Figs. 5.2-5.5 i t can be seen that the spectral features are much broader for the 2s spectra than for the 2p case. This broadening i s far beyond that a t t r i b u t a b l e to the differences i n energy res o l u t i o n (0.18 vs. 0.36 eV) which would i n any case appear n e g l i g i b l e on the energy scale of the f i g u r e s . Furthermore, there seem to be few i f any Rydberg t r a n s i t i o n s apparent i n the 2s spectra. The extremely broad peak (assigned predominantly to the a (e) l e v e l ) i n each of the 2s spectra i s at t r i b u t a b l e to the occurrence of a fa s t autoionisation process analo-gous to an L 1 L 2 3 M Coster-Kronig Auger t r a n s i t i o n ( i . e . , an i n i t i a l vacancy f i l l e d from within the same s h e l l plus e j e c t i o n of a valence e l e c t r o n ) . In the ISEELS spectra the process would involve an i n i t i a l excited state with a 2s vacancy being f i l l e d by a 2p electron with auto-i o n i s a t i o n of a valence e l e c t r o n . The r e l a t i v e l i f e t i m e broadening - 157 -(1.2 eV extra width) observed i n the case of argon for the 2s XPS peak r e l a t i v e to that for the 2p peak (see ref. [17], page 4) lends support to t h i s argument. (See also Chapter 1, section B IV). Phosphine (PH,) PH 3 i s the simplest molecule presented here and the d e t a i l e d r e s u l t s for the P 2p and 2s spectra are shown i n F i g . 5.2 and Table 5.2. It has been the subject of several e a r l i e r studies and discussions [60,61,65,146,170] which have mostly focussed on the 2p region with l i t t l e or no treatment on the 2s region. The spectrum recorded i n the present work i s i n good agreement with the s l i g h t l y higher r e s o l u t i o n XUV spectra reported by Hayes and Brown [146] and also by F r i e d r i c h et a l . [65]. F r i e d r i c h et a l . [65] have compared the various assignments reported i n the l i t e r a t u r e [60,61,146,170]. Their own c a l c u l a t i o n s [65] concur with the conclusions reached by Schwarz [60,61], Thus peaks 1-3 ( F i g . 5.2, Table 5.1) can be assigned as 2p •*• a t r a n s i t i o n s followed by the various 2p •+ Rydberg t r a n s i t i o n s leading up to the edge. The order * * of the a l e v e l s , according to these studies [60,61,65] i s a (a^) followed by a (e). This i s i n agreement with e a r l i e r Xa-SW c a l c u l a t i o n s [177] but contrary to the more recent Xoc-DV c a l c u l a t i o n [176]. I t i s clear i n any case that both l e v e l s are very close. However, the term value (5.21 eV) for feature 13 which i s assigned as 2s + a (e) i n the 2s spectrum ( F i g . 5.2) i s close to that for peak 1 (5.05 eV) and there-fore indicates that the recent Xa-DV ordering i s l i k e l y c o rrect. This - 158 -T E R M VALUE (eV) 8 0 -4 T T i—i—r 1 2 3 P 2 p A e d g e s « , R — 4 56 78 9 10 PH, P2p 11 AE=O.I8eV 132 13 3 6 140 P 2 s edge PH, P2s . .* * AE=0.36eV "1 ' 1 ^ — I 1 T ~ 188 192 196 2 0 0 E N E R G Y LOSS(eV) Figure 5.2: Phosphorus 2p and 2s electron energy loss spectra of PH 3. The P 2p spectrum (upper trace) i s at high r e s o l u t i o n (0.18 eV FWHM). The P 2s spectrum (lower trace) i s extracted from F i g . 5.1. The spectra are aligned with respect to the 2p(mean) and 2s i o n i s a t i o n edges. - 159 -TABLE 5.2 Energies, Term Values, and Possible Assignments for the P 2p,2s Spectra of PH3 Feature Energy Loss (eV) (a) Term Value (eV) Possible Assignments^^ 2 p 3 / 2 2 p 1 / 2 5.05 4.27 5.17 3.59 4.49 2.38 3.28 1.93 2.83 1.69 2.59 0.99 1.89 0.62 1.52 0.99 0.53 0 0 2 p 3 / 2 2 p 1 / 2 1 132.00 2 132.78 3 133.46 4 134.67 5 135.12 6 135.36 7 136.06 8 136.43(12) 9 136.96 137.42(12) limit b limit( b> 137.05 137.95 11 141.4(4) 12 156.7(5) 2p 3 / 2, 2 p 1 / 2 l 1 3 l N. 2s l i m i t ( ' c ; 189.67(15) 194.88 -4. -19. 2s 5.21 0 4s 3d 5s, 4d 6s etc a (aj) 4s 3d 5s, 4d 6s etc "shake-up" "shake-up" predominantly 2s •+ a (e) (a) Estimated uncertainty in energy-ljss values is ± 0.08 eV except where stated. Spectra are calibrated against N 2 (Is •* n , v • 1) at 401.10 eV. (b) The spin-orbit splitting of 0.90 eV [61] has been used to estimate the 2p 3/ 2 and 2pjy 2 spin-orbit components from the 2p (mean) values [31], see text for details. (c) Ref. [175]. (d) Final occupied orbital with either 2p 3/ 2 or 2pjy 2 hole state. t See Fig. 5.8. * With respect to the 2p (mean) edge [31]. - 160 -assignment cannot be considered as completely conclusive due to the assumptions discussed above. I t should be noted that the feature a t t r i -buted to the 2s edge by Hayes and Brown i n the spectrum of PH 3 [146] i s i n fact the discr e t e pre-edge feature (13) observed i n the present ( F i g . 5.2). Trimethyl Phosphine (P(CH3),) The deta i l e d 2p and 2s spectra of P(CH 3) 3 are shown i n F i g . 5.3 and the spectral positions and possible assignments summarised i n Table 5.3. The 2p spectrum consists of a number of overlapping t r a n s i t i o n s , a l l within 3.5 eV of the edge, and i s thus d i f f i c u l t to assign unambiguously. There are no obvious Rydberg s e r i e s , whch may indi c a t e valence-Rydberg mixing [20,65]. The term value for the 2s •+• a (e) t r a n s i t i o n (feature 10, F i g . 5.3) i s 3.38 eV. This corresponds very cl o s e l y with that (3.30 eV) for the f i r s t feature i n the 2p spectrum which has accordingly been assigned as the 2p •> ^ ^2/2^~ a t r a n s i t i o n . Thus the ordering of the v i r t u a l o r b i t a l s i s indicated to * * be a (e) followed by the a (a^) as was found for PH 3 (see preceeding discussion). This i s i n agreement with the recent Xoc c a l c u l a t i o n s [176] which finds the lowest unoccupied o r b i t a l with any s i g n i f i c a n t phospho-rus contribution to be e i n character. The remainder of the spectrum i s d i f f i c u l t to assign. However, i n many ways the sp e c t r a l shape i s i n keeping with what might be expected i f the 2p + a features were superimposed on the Rydberg - 161 -T E R M VALUE (eV) 8 0 -4 • II I I I -,\:/ 1 2 34 5 6 7 8 P2p , 2p . edges * i r r P(CHJ 3'3 P2p AE=O.I8eV 32 10 136 P 2s edge 140 P(CHJ 3'3 P2s AE=0.36eV 188 192 196 E N E R G Y LOSS(eV) Figure 5.3: Phosphorus 2p and 2s electron energy loss spectra of P(CH 3) 3. The P 2p spectrum (upper trace) i s at high r e s o l u t i o n (0.18 eV FWHM). The P 2s spectrum (lower trace) i s extracted from F i g . 5.1. The spectra are aligned with respect to the 2p(mean) and 2s i o n i s a t i o n edges. - 162 -TABLE 5.3 Energies, Term Values, and Possible Assignments fo r the P 2p,2s Spectra of P ( C H 3 ) 3 Feature Energy L o s s ^ a ^ Term Value Possible Assignments^^ (eV) (eV) 2 p 3 / 2 2 p 1 / 2 2 p 3 / 2 2 P l / 2 . 1 132.65 3.30 a*(e) 2 133.09 2.86 3.76 4s 3 133.79 2.16 3.06 A o-\e) 4 134.05 (12) 1.90 2.80 3d, B 4s 5 134.67 1.28 2.18 5s A 6 135.03 0.92 1.82 4d 3d, B 7 135.9 (2) 0 0.9 edge 4d 8 2 p 3 / 2 l i m i t ^ 2 Pj/2 H m i t ( b ) 136.8 (3) 135.95 136.85 0 0 0 edge 7 t 9 153.0 (5) -16. Shape-resonance 2s 2s l i m i t K c > 190.23 (15) 193.61 3.38 0 predominantly 2s •* o*(e) (a) Estimated uncertainty i n energy-lgss values i s ± 0.08 eV except where stated. Spectra are c a l i b r a t e d against N 2 (Is •* n , v =» 1) at 401.10 eV. (b) The s p i n - o r b i t s p l i t t i n g of 0.90 eV [61] has been used to estimate the 2 p 3 / 2 and 2 p ^ 2 s p i n - o r b i t components from the 2p (mean) values [31], see text for d e t a i l s . (c) The 2p(mean)-2s separation was.taken to be 57.36 eV based upon the average of other P 2p(mean)-2s separations [175]. This value i s within 0.17 eV of a l l the other separations [175]. (d) F i n a l occupied o r b i t a l with e i t h e r 2 p 3 / 2 or 2 p j / 2 hole s t a t e . Note that f o r peaks 3-6 two a l t e r n a t i v e assignments (A or B) are given (see text) for the two components with o ( a j ) as the f i n a l o r b i t a l . With respect to 2p (mean) edge [31]. - 163 -features i n the PH 3 spectrum (compare F i g . 5.3 with F i g . 5.2). On t h i s basis features 1, 3-4, and 6 or 1, 3, and 5 i n the P ( CH 3) 3 could be related to features 1, 2, and 3 i n the PH 3 spectrum. If t h i s i s the case, then features 1 and 3 ( F i g . 5.3) can be assigned as the two * s p i n - o r b i t components of the 2p -»• o (e) t r a n s i t i o n and either features 3 and 5 (Scheme A, Table 5.3), or features 4 and 6 (Scheme B, Table 5.3) to the two components of the 2p •*• a ( a ^ t r a n s i t i o n . The rest of the spectral i n t e n s i t y would then be due to various Rydberg t r a n s i t i o n s . Feature 2, for instance, would be due i n both schemes to the 2p •*• (2p_jy2) - 1^ 8 t r a n s i t i o n with feature 4 having a contribution from the concomitant 2p •*• (,2p^^)~^As t r a n s i t i o n . Based upon the quantum defect calculated from the term value of feature 2, the predicted 5s term value i s 1.34 eV, which allows feature 5 to also be assigned as the 2p •+• ( 2 p ^ 2 ) _ 1 5 s t r a n s i t i o n . Transitions to the d Rydberg l e v e l s would also be expected. Table 5.2 summarises the possible assignments. Phosphorus T r i f l u o r i d e (PF,) The 2p spectrum of PF 3 ( F i g . 5.4) i s rather d i f f e r e n t from that of the i s o e l e c t r o n i c molecule P(CH 3) 3 ( F i g . 5.3). The data for both the 2p and 2s spectra are shown i n F i g . 5.4 and summarised i n Table 5.4. There i s a c l e a r difference i n the 2p spectra which may be ascribed to the e f f e c t s on the valence-Rydberg separation by the electronegative F ligand as compared to the electron donating -CH 3 ligand. Features 1-3 are c l e a r l y t r a n s i t i o n s to v i r t u a l valence o r b i t a l s , as indicated by the large term values. The second band (features 4-7) i s probably comprised - 164 -TERM VALUE (eV) 8 4 0 - 4 i i i i i i i i i CO P 2 ^ 2 P , / 2 E D A E S I I I i i i-i i i i r I I i 1 2 3 4 5 67 8 9 10 11 1213 14 P F 3 P2p UNIT . . . . • • ITRARY i - v J v.— (ARB AE=O.I8eV > i | I | l | 1 136 140 144 INTENSI , , , \ i* 17 18 P 2 s e d g e P2s INTENSI UJ > * 4 i .« * * * V • «*-, • . * *> t«, • * - * * * * V 4V V . %1 « « ' » REl . . . » *•» • * .»»**^* V AE=0.36eV 192 196 2 0 0 2 0 4 ENERGY LOSS(eV) Figure 5.4: Phosphorus 2p and 2s electron energy loss spectra of PF 3. The P 2p spectrum (upper trace) i s at high r e s o l u t i o n (0.18 eV FWHM). The P 2s spectrum (lower trace) i s extracted from F i g . 5.1. The spectra are aligned with respect to the 2p(mean) and 2s i o n i s a t i o n edges. - 165 -TABLE 5.4 Energies, Term Values, and Possible Assignments fo r the P 2p,2s Spectra of PFj Feature Energy L o s s ^ 3 ^ Term Value Possible Assignments^^ (eV) U iV) 2 p 3 / 2 2 p 1 / 2 2P3/2 2 p i / 2 1 2 3 4 5 6 7 8 9 10 11 2p,/, l i m i t i b ( 2v\',\ limlt< b> 135.00 . 135.61 136.52 138.0 (2) 138.64 139.45 139.83 (10) 140.61 141.30 (12) 142.03 (15) 142.7 141.77 142.67 6.77 6.16 5.25 3.77 3.13 2.32 1.94 1.16 0.47 0 7.06 6.15 4.67 4.03 3.12 2.84 2.06 1.37 0.64 0 0 o*(e) o ( e ) o*(a 1) 4s 3d 5s, 4d edge o*(e) a ( a j ) 4s 3d 5s, 4d edge 12 13 14 15 143.74 (20) 144.25 (20) 145.9 (3) 157.3 (5) -3.8 f -15.2 T "Shake-up" inner-well state/shape-resonance shape-resonance 2s 16 17 2s l i m i t ( c ) 192.60 (15) 196.21 (20) 197.53 (20) 199.49 6.89 3.28 1.96 0 2s -» o^e) 2s •+ a (a^) 2s •»• 4p (a) Estimated uncertainty i n energy-loss values i s ± 0.08 eV except where stated. Spectra are c a l i b r a t e d against N 2 (Is •* n*, v = 1) at 401.10 eV. (b) The s p i n - o r b i t s p l i t t i n g of 0.90 eV [61] has been used to estimate the 2p-y 2 a n t * 2 P l / 2 s p i n - o r b i t components from the 2p (mean) values [31], see text for d e t a i l s . (c) Ref. [175]. (d) F i n a l occupied o r b i t a l with e i t h e r 2 p 3 / 2 or 2 p j / 2 hole s t a t e . *with respect to the 2p (mean edge) [31]. - 166 -of t r a n s i t i o n s to the various Rydberg l e v e l s . However, the broad shape of the leading edge (feature 4) and also the width of feature 5 suggests that t h i s band may consist of Rydberg t r a n s i t i o n s superimposed on top of a valence t r a n s i t i o n . This type of s i t u a t i o n was observed for the Si 2p spectrum of SiF^ [144] and i n view of the s i m i l a r ways i n which the ligands seem to a f f e c t the ce n t r a l atom i n the case of both P and S i , i t i s thought to be very l i k e l y here. The 2s spectrum shows one intense peak (feature 16) which i s assigned as 2s > a (e) (as i n the case of PH 3 and also P(CH 3) 3) and the term value (6.89 eV) for t h i s corresponds approximately to that for the f i r s t band (peaks 1-3) i n the 2p spectrum. This feature (16) i s followed by further structure (feature 17) which has a term value of 3.28 eV, which corresponds with that for features 4-5 i n the 2p spectrum. Since the 2s •> 4s Rydberg t r a n s i t i o n would be expected to be very weak ( i t would correspond to a dipole forbidden s •*• s t r a n s i t i o n i n the atomic core), t h i s feature (17) has been assigned as the 2s •+• a (a^) t r a n s i t i o n . Thus on t h i s basis the a (e) - a (a^) separation would be quite large (~3.5 eV). This agrees with the Xa c a l c u l a t i o n s of Xiao et a l . [176], which predict a ground state separation of 3 eV. Feature 18 i s presumably the 2s •+• 4p Rydberg t r a n s i t i o n . The f i r s t band of the 2p spectrum shows three components (peaks 1, 2, and 3). As stated e a r l i e r , t h i s has been t e n t a t i v e l y assigned to the various components of the 2p -»• a (e) t r a n s i t i o n . Peaks 1 and 2 are assigned to the two allowed f i n a l states of the 2p •> ( 2 p „ / 9 ) _ i a (e) - 167 -t r a n s i t i o n ( i . e . , e <-> e, see Table 5.1) and peak 3 to the 2p •+• ( 2 p j ^ ) 0 ( e) t r a n s i t i o n . The rest of the assignments of the features i n the spectrum are summarised i n Table 5.4. Phosphorus T r i c h l o r i d e (PCI,) The 2p spectrum of PC1 3 ( F i g . 5.5), as i n the analogous spectra for PF 3 and PH 3, shows intense structure well below the edge, which must be due to t r a n s i t i o n s to the v i r t u a l valence o r b i t a l s because of the large term values (Table 5.5). This i s followed by weaker Rydberg structure l y i n g on top of a broad resonance i n the region of the 2p edge. The electron impact excited spectrum reported here d i f f e r s s i g n i -f i c a n t l y from a previously reported o p t i c a l (soft X-ray) spectrum [171, 172] i n that a prominant extra peak (feature 4, F i g . 5.5) i s observed (compare with the o p t i c a l spectrum as shown i n F i g . 5.6). The l i k e l i -hood of t h i s being due to an impurity seems remote i n t h i s s p e c t r a l region since any phosphorus containing impurity a r i s i n g from PC1 3 i s l i k e l y to be extremely i n v o l a t i l e and apart from t h i s sharp feature the rest of the spectrum i s i n agreement with the o p t i c a l r e s u l t . However, the spectrum was c a r e f u l l y rechecked using a new sample of PC1 3 from a d i f f e r e n t source and found to be i d e n t i c a l . Another (non-spectroscopic) p o s s i b i l i t y i s that the extra peak (4) arises from a "ghosting" e f f e c t due to the primary electron beam eith e r h i t t i n g the anode or some other surface. The p o s s i b i l i t i e s of the peak being due to energy losses caused by scattering from the anode - 168 -8 TERM VALUE (eV) 4 0 - 4 m — r 12 3 4 P 2 P 3 / 2 2 P | / 2 e d 9 e S \ * 5 6 7 8 910 PCI. P2p >-cc < LT m cr < 00 LU AE=O.I8eV 32 13 136 140 JL P 2 s edge PCI, P2s UJ > _ J LU cr AE=0.36eV 188 192 196 2 0 0 ENERGY LOSS(eV) Figure 5.5: Phosphorus 2p and 2s electron energy loss spectra of PC1 3. The P 2p spectrum (upper trace) i s at high r e s o l u t i o n (0.18 eV FWHM) The P 2s spectrum (lower trace) i s extracted from F i g . 5.1. The spectra are aligned with respect to the 2p(mean) and 2s i o n i s a t i o n edges. - 169 -TABLE 5.5 Energies, Term Values, and Possible Assignments for the P 2p,2s Spectra of PCI3 Feature Energy Loss v ' (eV) Term Value (eV) Possible Assignments^ 2 p 3 / 2 2 p 1 / 2 2?3/2 2Pi/ 2 1 2 3 4 5 6 7 8 9 10 2 p 3 / 2 limit <|»> 2pj/2 H m i t ( b ) 11 12 13 2s l i m i t ( c ) 132.98 (12) 133.50 134.11 (10) 135.11 136.84 137.61 138.25 139.10 (12) 139.57 (10) 139.98 (12) 139.84 140.74 149.5 (5) 152-168 191.07 (15) 197.47 6.86 6.34 5.73 4.73 3.00 2.23 1.59 0.74 0.27 0 7.24 6.63 5.63 3.93 3.13 2.49 1.64 1.17 0.76 0 a (e) Symmetry For' 4s 5s, 3d 6s, 4d etc a Ce} bidden Transition 4s on top of 5s, 3d inner-well state/shape 6s, 4d resonance etc. 9 Shape resc + "Shake—i jnance JP" 2s predominantly 2s •+ a*(e) 6.40 0 (a) Estimated uncertainty in energy-lgss values is ± 0.08 eV except where stated. Spectra are calibrated against N 2 (Is -• 11 , v = 1) at 401.10 eV. (b) The spin-orbit splitting of 0.90 eV [61] has been used to estimate the 2p 3^ 2 and 2pj/ 2 spin-orbit components from the 2p (mean) values [31], see text for details. (c) Ref. [175] (d) Final occupied orbital with either 2p 3/ 2 or 2pjy 2 hole state. Assignment in table for pyramidal excited states. See text for discussion on planar excited states. *with respect to the 2p (mean) edge [31]. - 170 -can be eliminated since i t would occur at an energy loss equivalent to the anode voltage (~600 V, with respect to the cathode). As a further check the spectrum was also rerun using a lower impact energy of 1500 V instead of the o r i g i n a l 2500 V, a procedure that involves complete retuning of the spectrometer. Since the conditions would now be d i f f e r -ent, any e f f e c t s of "ghosting", i f present, would be expected to change. No change i n the number or energy of the spectral features was observed. In view of these observations, feature 4 can be assigned as a d i p o l e -forbidden t r a n s i t i o n , which i s observed i n electron impact spectroscopy where the momentum transfer i s f i n i t e . At the l i m i t of zero momentum transfer the o s c i l l a t o r strength of such a t r a n s i t i o n should go to zero; i . e . , the (dipole) o p t i c a l o s c i l l a t o r strength vanish at the o p t i c a l l i m i t [2,8]. Under the conditions used i n the present work the momentum transfer i s kept as small as possible so as to ensure that dipole processes strongly dominate the spectra. In some e a r l i e r spectra weak, non-dipole processes have been observed under these conditions [69,109], In the case of PC1 3, peak 4 i s r e l a t i v e l y much more intense than previously observed dipole-forbidden t r a n s i t i o n s . In order to confirm the o p t i c a l l y forbidden nature of t h i s t r a n s i t i o n the spectrum was run at 1500 V impact energy and several d i f f e r e n t d e f l e c t i o n (scattering) angles up to ~5°, which i s the l i m i t of the present beam d e f l e c t i o n system. At an impact energy of 1500 V i t was possible to tune the spectrometer to operate at an even smaller scattering angle (and hence closer to zero momentum transfer) than at 2500 V due to improved electron o p t i c a l focussing. Three spectra of PC1 3 and PF 3 obtained at - 171 -various momentum transfers ( i . e . , scattering angles) are shown i n F i g . 5.6. The o p t i c a l absorption spectrum (corresponding to zero momentum transfer) of PC1 3 [171] i s also shown for comparison and i t c l e a r l y indicates the absence of peak 4. In a l l other respects the ISEELS and o p t i c a l spectra are equivalent. In the case of the ISEELS spectra a l i n e a r function, extrapolated from the leading edge, has been subtracted i n order to f a c i l i t a t e the estimation of peak heights. The non-dipole nature of feature 4 i n PC1 3 i s c l e a r l y apparent since i t s r e l a t i v e i n t e n s i t y to the rest of the spectrum increases markedly with increase i n s c a ttering angle. From the dimensions of the de f l e c t o r plates and the magnitude of the applied voltage i t i s possible to estimate the de f l e c t i o n angle (9) for a given electron energy (see Table 2.1). These angles have been estimated for each of the spectra shown i n F i g . 6 and used to obtain the momentum transfer, K, i n atomic units (a.u.) (see equations l.C.6 and l.C.16). I t i s usual to consider the v a r i a t i o n of r e l a t i v e i n t e n s i t i e s of allowed (or forbidden) t r a n s i t i o n s with K 2. A plot of the r a t i o s (peak height of feature 4)/(peak height of feature 2) versus K 2 should extrapolate back to zero at zero momentum transfer i f feature 4 i s dipole forbidden [2,9]. This behaviour i s found for feature 4 as indicated i n F i g . 5.7, which also shows s i m i l a r p l o t s f o r some of the dipole-allowed features of the PC1 3 P 2p spectrum (features 3, 5, and 6 resp e c t i v e l y of F i g . 5.6) over the li m i t e d range of momentum transfer possible with t h i s instrument. In contrast to the behaviour for feature 4 i t can be seen that the r a t i o s for the dipole-allowed t r a n s i t i o n s remain almost constant. The steep r i s e i n i n t e n s i t y of - 172 -Figure 5.6: Inner s h e l l electron energy loss spectra of (a) PC1 3 and (b) PF3 at various scattering angles. The spectra were obtained with an impact energy of 1500 V. The dipole-forbidden t r a n s i t i o n i n PCI3 i s marked with an a s t e r i s k (*). Also shown with the PC1 3 ISEELS spectrum i s the o p t i c a l spectrum taken from ref. [171]. - 173 -Scattering Angle 8 (degrees) 1 2 3 4 5 feature 3 0 . 0 0.2 0.4 0.6 0 . 8 1.0 (Momentum Transfer)2 K2(a.u.)2 Figure 5.7: Plot of the r a t i o (peak height of feature X)/(peak height of feature 2) for various t r a n s i t i o n s (X) i n the PCI 3 spectrum ( F i g . 6) as a function of (momentum transfer) . S o l i d c i r c l e s -ISEELS data; s o l i d t r i a n g l e s - o p t i c a l data [171]. - 174 -feature 4 with increase i n i s notable i n that t h i s i s much more rapid than i s usually observed for forbidden t r a n s i t i o n s [2], The PF 3 spectra ( F i g . 5.6, right hand side) show no discernable changes over the same range of momentum tra n s f e r . The spectra of PH 3 and P ( C H 3 ) 3 were also checked i n the same way and again no discernable changes were apparent. The PF 3, P(CH 3) 3 and PH 3 were a l l run under the same conditions as the PC1 3 spectra and provide further evidence that peak 4 i s i n no way due to "ghosting" or other instrumental e f f e c t s . The range of momentum transfer i s quite small and the sudden and rapid emergence of peak 4 i s quite unusual and without precedent i n e a r l i e r ISEELS studies. Consider now the assignment of the spectrum P 2p, 2s ( F i g . 5.5) of PC1 3. Peaks 1-4 are c l e a r l y due to 2p •* a l e v e l s . The term value * for the a (e) l e v e l from the 2s spectrum (feature 13) i s 6.40 eV. The fact that the term value for feature 1 i s somewhat larger (6.86 eV) than 6.40 eV suggests t h i s peak arises from the process 2p ( ^ p ^ ^ ) " °* (*].)• If t h i s i s the case, the ordering of the v i r t u a l l e v e l s would be a (a^) followed by a (e). This i s the reverse of the assignment given by Topol et a l . [172] based on an Xa-SW c a l c u l a t i o n . However, since the c a l c u l a t i o n only gives a separation of 0.12 eV for the a (a^) and a (e) l e v e l s the r e s u l t i s inconclusive. Feature 4 i s c l e a r l y a symmetry-forbidden t r a n s i t i o n on the basis of the evidence presented above (see also Figs. 5.6 and 5.7). The only dipole-forbidden t r a n s i t i o n possible, within the C^ v point group of the pyramidal molecule, i s 2p(e) •*• (2p,. ) - 1 a (e), with an A 2 f i n a l state (see Table 5.1). The only other - 175 --1 possible explanation for feature 4 would ar i s e i f there i s a geometry change i n going to some of the upper states. If t h i s occurs i t would most l i k e l y be to states of ( i . e . , planar) symmetry. I t i s i n t e r e s t i n g to note that feature 4 i s approximately a spin-o r b i t s p l i t t i n g away from feature 3. This can be explained within the context of a D_. system. Table 5.6 shows the f i n a l configurations Jn possible here for symmetry and whether they would be o p t i c a l l y accessible from a D^^ ground state. I t i s seen that only the (.^2/2^ * a (e') (to an E 1 f i n a l state) i s accessible under dipole s e l e c t i o n r u l e s . Thus feature 3 could be assigned to the dipole allowed i * 2p •* (2P3/2) 0 ( e') t r a n s i t i o n with feature 4 as the concomitant l * (dipole forbidden) 2p •*• (^P^^^ a (e'-1 t r a n s i t i o n . From the spectra ( F i g . 5.6) and the angular v a r i a t i o n ( F i g . 5.7) three dipole allowed features (1-3) are c l e a r l y present. Hence the remaining two dipole l * allowed features (1 ,2) could be due to 2p (2p.j/2 1/2^ ° ^ a l ^ t r a n s i t i o n s with pyramidal upper and lower states. However, i t should be remembered that the i n i t i a l state i s of C^ v symmetry and regardless of the f i n a l state geometry ( i . e . , C ^ o r D ^ ) , t n e s e l e c t i o n r u les apply, but with the i n t e n s i t i e s of t r a n s i t i o n s being strongest for those which are allowed for both symmetries [178] . Examination of the o p t i c a l spectrum ( F i g . 5.6) shows at most a weak unresolved shoulder at p o s i t i o n 4 and more l i k e l y no i n t e n s i t y at a l l i n view of the t a i l i n g that i s apparent on a l l peaks i n the spectrum. The rapid emergence of feature 4 i s of i n t e r e s t and c l e a r l y t h e o r e t i c a l studies, as well as more systematic - 17,6 -TABLE 5.6 Transitions from the *AJ Ground State for Symmetry F i n a l Configuration * F i n a l State Dipole Allowed from ground state hole state occupied a* o r b i t a l e' e' AJ + A 2 No e' e' E' Yes e' a 2 E" No a 2 e' E" No a 2 a2* A' A l No ( 2 p 3 / 2 ) a n <* ^p^ / 2 ^ ~ holes are of e' and a 2 symmetry respectively. The o* o r b i t a l s are of a 2 and e' symmetry. - 177 -variable angle electron impact spectroscopy would be h e l p f u l i n e s t a -b l i s h i n g the i d e n t i t y of th i s t r a n s i t i o n . The remaining features (5-10) can be assigned to the various 2p -*• Rydberg t r a n s i t i o n s . The l a t t e r few being on top of an inner well trapped state or resonance evident by the large broad underlying peak at ~140 eV i n F i g . 5.5 (see following sec t i o n ) . The assignments are also summarised i n Table 5.5. Phosphorus 2p, 2s Spectra - Continuum Features A l l the phosphorus spectra show considerable i n t e n s i t y at or beyond the 2p edge ( F i g . 5.1). In the case of PH 3 t h i s manifests i t s e l f as an i n f l e c t i o n ~141.4 eV followed by a broad band with a maximum at ~157 eV. This has been at t r i b u t e d to a delayed onset caused by c e n t r i -fugal b a r r i e r e f f e c t s a r i s i n g from p -»• continuum d - l i k e state t r a n s i -tions [49,146]. This e f f e c t i s seen i n atoms (e.g., Si) as well as SiH^ [146]. A closer examination of the continuum spectrum of PH 3 reveals structure between the edge and the band maximum. This i s a t t r i b u t a b l e to i o n i s a t i o n plus e x c i t a t i o n ("shake-up"). In ISEELS (or photoabsorp-tion) t h i s would appear as onsets of new continua. An expansion of the PH 3 continuum spectrum i s shown i n F i g . 5.8 together with the s a t e l l i t e ("shake up") portion of the PH 3 (2p) XPS spectrum measured elsewhere [175]. Both spectra i n F i g . 5.8 are plotted on the same ho r i z o n t a l r e l a t i v e energy scale. Much of the i n t e n s i t y i n the continuum band of the ISEELS spectrum (upper trace) can be assigned to onsets of "shake-up" continua, evidenced by the peaks i n the XPS spectrum (lower t r a c e ) . - 178 -ENERGY LOSS (eV) — r 1 1 1 1 H O 1 1 150 1 1 1 1 | 1 1 160 * • • • >" 1 1 I 1 I 170 P H 3 its* .-.••/•• • ••-'•>.•-•. l , ISEELS "v /< 2p / ml •V - • *. XPS ] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 r 0 10 2 0 3 0 RELATIVE ENERGY (eV) ure 5.8: Expanded plot of the PH 3 P 2p ISEELS continuum structure. The 2p s a t e l l i t e structure from XPS measurements [175] i s shown below plotted on the same r e l a t i v e energy scale, referenced to the 2p (mean) edge. - 179 -A s i m i l a r explanation was used i n Chapter 3 to explain continuum s t r u c -tures i n ISEELS spectra of NF3. Considering now the other molecules whose spectra are shown i n F i g . 5.1, i t can be seen that the continuum structure i s somewhat d i f f e r e n t to that i n PH 3 with the o s c i l l a t o r strength being directed into more l o c a l i s e d channels. This i s d i r e c t l y a t t r i b u t a b l e to the ef f e c t s of the bulky, many electron ligands i n contrast to the s i t u a t i o n i n PH3. Indeed, compared to the other spectra, that for PH 3 i s very atomic l i k e i n the continuum (compare also S i and SiH^ [19]). The behaviour of th i s series of molecules i s very s i m i l a r to those seen In the S i series i n F i g . 4.3 and the l o c a l i s e d and intense continuum structure ( p a r t i c u l a r l y i n the case of PF 3 and PCI3) provide further examples of p o t e n t i a l b a r r i e r or shape-resonance phenomena [73,77]. Thus both PCI3 and P F 3 show an intense, sharp feature ( F i g . 5.1) at (PClj-feature 10) or just beyond (PF 3 - feature 14) the edge which can be described as a resonance enhanced t r a n s i t i o n to a trapped (probably d-type) state. A small peak i s also seen righ t at the edge i n P(CH 3) 3, however, i t i s not as intense as i n PF 3 or PCI3. The methyl ligand i s electron-donating [31,33] and so an e f f e c t i v e p o t e n t i a l b a r r i e r would not be expected. The lower i n t e n s i t y i s consistent with t h i s . A l t e r n a t i v e l y , the feature may simply be due to unresolved Rydberg le v e l s converging onto the edge, or a combination of both e f f e c t s . - 180 -A l l the three molecules show a broader structure between 10 and 20 eV above the edge. These can be at t r i b u t e d to higher d - l i k e shape resonances, although, as has been discussed above, the structure i n PH 3 can at least i n part be associated with shake-up processes as also observed i n XPS [175]. The shape of the PC1 3 structure i s i n t e r e s t i n g i n that i t shows a sharp r i s e followed by further l a r g e l y unresolved structure. I t i s possible that this shape i s due to "shake-up" phenomena on top of the resonance structure as suggested above i n the case of PH 3. I t would be i n t e r e s t i n g to compare the ISEELS spectrum with the XPS s a t e l l i t e spectrum, but as yet t h i s has not been reported. In view of the recent spate of i n t e r e s t between resonance p o s i t i o n and bond length [95-100] and the success of the l i m i t e d s e r i e s of Si containing compounds shown i n the previous chapter, i t i s of i n t e r e s t to examine whether such a r e l a t i o n s h i p also e x i s t s for these compounds. Table 5.7 summarises the relevant data for possible resonan-ces i n PF 3, PC1 3, and P(CH 3) 3. The lower energy resonance follows the trend observed for s i l i c o n , however, the good l i n e a r c o r r e l a t i o n found for the outer resonance i n the s i l i c o n series i s not apparent here. However, a deta i l e d consideration of t h i s phenomenan involves many factors such as d i f f e r e n t phase s h i f t s of the s c a t t e r i n g centre and varying geometries [99]. The reasonable c o r r e l a t i o n observed i n the s i l i c o n series may r e f l e c t the constant tetrahedral geometry of a l l the s i l i c o n compounds studied. The f a i l u r e of the c o r r e l a t i o n i n the case of the PX, species may be due to t h e i r variable geometry. C l e a r l y more TABLE 5.7 Resonance Term Values 6(eV) and (P-X) Bond Length R(A) Molecule R ( A ) ( a ) R-2 (A-2) 6 (lower) (eV)< b ) 6 (higher) ( e V ) ( c ) PF 3 1.563 0.4093 3.8 15.2 P(CH 3) 3 1.843 0.2944 0.5 16.7 PCI3 2.043 0.2396 -0.1 9.4 (a) From Landholdt-BHrnstein (New Series) II/7 "Structure Data of Free Polyatomic Molecules" Springer-Verlag, B e r l i n , 1976. (b) F i r s t resonance position. Data from Tables 5.3-5.5. 6 = Resonance Energy - I.P. (c) As (b), but for higher resonance. - 182 -systematic studies need to be made for a large group of molecules before more d e f i n i t e general conclusions can be drawn. Ligand Spectra (F Is, C Is, and C l 2p,2s) The inner s h e l l spectra of the various ligand regions (F Is, C Is, and C l 2p,2s) are each t y p i c a l of spectra associated with that p a r t i c u l a r edge. Each ligand spectrum w i l l be discussed i n turn and compared with the respective c e n t r a l atom spectra. The v i r t u a l valence o r b i t a l orderings as discussed i n the preceding P 2p,2s spectra have been used i n the assignment of the ligand spectra. The F Is spectrum of PF 3 i s shown i n F i g . 5.9 and summarised i n Table 5.8. The spectrum consists of a r e l a t i v e l y intense peak followed by a broad band with some structure both before and a f t e r the edge. The pos i t i o n of the edge i s taken from XPS measurements [179]. The F Is o r b i t a l s of PF 3 transform as and e symmetries and so t r a n s i t i o n s to a l l the v i r t u a l o r b i t a l s (a± and e) are possible. Therefore, the f i r s t * peak i s assigned as the F Is •*• a (e) t r a n s i t i o n using the v i r t u a l o r b i t a l ordering as assigned i n the P 2p,2s spectra. This peak has a term value of 5.2 eV, which i s approximately 1.5 eV lower than that f o r the P 2s or P 2p •*• a (e) t r a n s i t i o n . The difference i s of the same order as that found between the N Is and F Is •*• a t r a n s i t i o n s i n NF 3 (Chapter 3) and that for the S 2p and F Is + o t r a n s i t i o n s i n SF 6 [69]. As was previously noted, t h i s difference i s d i r e c t l y a t t r i b u t a b l e to the location of the core hole, since the creation of a core hole on the central atom e f f e c t i v e l y increases the core charge by one, and hence the - 183 -0 0 t -cr < cr CD cr < AE=0.36eV 0 FIs '.Writ FIs edge 1 2 3 4 5dt 7 0 0 i 1 1 1 1 1 i I I 725 7 5 0 L_ 10 LU > _1 . LU K cr D AE=0.36eV E F1s edge l — i — i — i — | — i — i — i — i — | — i — i — i — i — | — i — i — 1 1 I ' 1 1 r 690 695 700 705 ENERGY LOSS (eV) Figure 5.9: Long-range and de t a i l e d inner s h e l l electron energy loss spectra of the F Is region of PF 3. - 184 -TABLE 5.8 Energies, Term Values, and Possible Assignments for F Is Spectrum of PF, Feature (a) Energy Loss Term Value Possible Assignments (eV) (eV) 1 688.98 (15) 5.2 o*(e) 2 691.22 (20) 3.0 o*(a 1 )/Rydberg 3 693.64 (30) 0.6 Rydberg F Is l l m l t ( b ) 694.2 0 Ionisation edge 4 699.6 (5) -5.4 d-like shape resonance Estimated uncertainty given in brackets. Ref. [179]. - 185 -a o r b i t a l energies are determined by a Z+l cen t r a l core, whereas the creation of a peripheral core hole on the F ligand has less e f f e c t on * the energy of the l o c a l i s e d a o r b i t a l . Thus the electron i n a v i r t u a l o r b i t a l w i l l have a higher term value when there exists a cen t r a l hole as opposed to a peripheral hole. The rest of the features can be assigned i n a straightforward manner. Feature 2 i s assigned, on the basis of i t s term value, as the P Is + fl ( a p t r a n s i t i o n , but may also include Rydberg t r a n s i t i o n s . Feature 3 i s probably due to Rydberg series converging onto the edge. The weak continuum structure (4) at 699.6 eV i s presumably an inner-well state/shape-resonance. Figure 5.10 shows the C Is spectrum of P(CH 3) 3. The data i s summarised i n Table 5.9. The C Is edge i s determined from XPS [180]. The spectrum i s si m i l a r to that of CH^ [65] and to the C Is spectrum of SiCCHg)^ ( F i g . 4.5). In assigning the spectrum of S^CHg)^ i t was useful to compare i t with the C Is spectra of mono substituted methyl halides [64]. A s i m i l a r process w i l l be used here. Features 1 and 2 are assigned as the C Is * 4s(a!) and C Is + 4p(e) Rydberg t r a n s i t i o n s . In contrast to the present case for P(CH 3) 3, the analogous features i n the methyl halide spectra were c l e a r l y resolved. I t should be noted that the spectrum recorded here for P ( C H 3 ) 3 was run at a higher r e s o l u -t i o n than for the methyl halide spectra [64]. With this i n mind, and with the apparent r e l a t i v e closeness of the s and p l e v e l s , i t i s * suggested that a C Is a t r a n s i t i o n i s also contained within t h i s * Since the cen t r a l atom i s from the t h i r d row ( i . e . , P), the f i r s t Rydberg l e v e l s are the 4s and 4p l e v e l s . - 186 -P. tn 1 0 H g Cis edge n r 2 4 5 P(CHJ 3'3 C l s or < or CD or < >-5 H AE = 0.36eV CO LU LU > _ l LU cr T — i — r i r 10 290 i — i — i r 1 2 3 4 — I — i — i — i — i — i — i — i — i — i — i — i — i — i r 300 310 320 & C1s edge AE=O.I8eV 5 H 286 T T T 288 290 292 E N E R G Y L O S S ( e V ) 294 Figure 5.10: Long-range and d e t a i l e d inner s h e l l e l ectron energy loss spectra of the C Is region of P(CH 3) 3. - 187 -TABLE 5.9 Energies, Term Values, and Possible Assignments for C Is Spectrum of P(CH_). Feature (a) Energy Loss (eV) Term Value (eV) Possible Assignments 1 287.13 3.17 4s * on top of a 2 287.83 (8) 2.47 4p(e) 3 288.31 1.99 4p ( a i ) 4 289.24 1.06 5p, 3d C Is l i m i t ( b ) 290.30 5 293.5 (3) -3.20 "shake-up" 6 302.5 (5) -12.20 or resonance (a) Estimated uncertainty ± 0.10 eV except where stated. (b) Ref [180]. - 188 -feature. The term values from the P 2p spectrum are consistent with t h i s i n t e r p r e t a t i o n . The core hole being l o c a l i s e d on the C could very * e a s i l y lower the term value for the ( e ) . Indeed, with the close proximity of the (e) and 4p(e) states the spectrum might be better described i n terms of a mixed Rydberg-Valence state [20] as discussed e a r l i e r for simpler molecules. Feature 3 Is assigned as the C Is + 4p(a 1) t r a n s i t i o n and feature 4 to the C Is •* 5p and 3d Rydberg l e v e l s . Post-edge structure can also be seen and can be att r i b u t e d to either "shake-up" or resonances. Structure 6 (~303 eV) i s seen i n many hydrocarbon spectra and l i k e l y to be a a (C-H) resonance [100]. F i n a l l y the C l 2p,2s spectrum of PC1 3 i s considered. The spectra are shown i n F i g . 5.11 and summarised i n Table 5.10. The positions of the C l 2p^2 a n t * ^P]_/2 8 P i n ~ o r b i t components have been estimated from the reported C l 2p (mean) XPS value [179] using spin-orbit separation of 1.6 eV [181] along with a 2:1 spectral weighting. The C l 2s edge was taken from XPS data [179]. The spectrum i s i n good agreement with the previously published C l 2p o p t i c a l absorption spectrum of PCI 3[172,173] . This spectrum i s very s i m i l a r to other C l L - s h e l l spectra recorded for the various chloromethanes [182]. Features 1 and 2 are attributed to the spin-orbit components of the 2p •*• a ( a ^ t r a n s i t i o n s and features 2 and 3 to the 2p •+ a ( e ) . Since the a l e v e l s are close, as Indicated i n the P 2p spectrum, i t i s possible that 1 and 2 contain a l l of these t r a n s i t i o n s . The concept of using the C l 2s term value to estimate the positions of the a components i s not as straightforward here since the C l 2s o r b i t a l s - 189 -p K Cl 2p edges M i n i I I 12345 7 8 9 W W AE=0.36eV PCI 3 Cl 2p,2s 1 2s edge 10 —I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r 2 0 0 220 2 4 0 2 6 0 280 Cl 2p 2 2 p | / 2 edges 1 — r r 4 5 6 I 8 AE=O.I8eV -XL — j 1 1 1 1 ' 1 — ' 1 2 0 0 2 0 4 208 212 216 ENERGY LOSS (eV) Figure 5.11: Long-range and de t a i l e d electron energy loss spectra of the Cl 2p and 2s regions of PC1 3. - 190 -TABLE 5.10 Energies, Term Values, and Possible Assignments for the Cl 2p,2s Spectra of PC13 Feature Energy Loss^a^ (eV) Term Value (eV) Possible Assignments^^ 2 p 3 / 2 2 p 1 / 2 2P3/2 2 Pl /2 1 2 3 4 5 6 7 Cl 2 p 3 / 2 l i m i t ^ Cl 2 P j / 2 l i m i t ( b ) 8 9 1 0 t \ Cl 2s l i m i t u ; 200.41 201.66 203.43 (8) 205.14 (12) 206.05 206.65 208.32 206.1 207.7 212.24 (15) 215.43 (30) 271.7 (3) 278.2 5.7 4.4 2.6 0.9 0 -0.6 6.0 4.2 2.5 1.6 1.0 -0.7 < ( a l > o* (ej 4s etc. a*(a,) a*(e5 4s etc or 2p -<• a or 2p + 4s on top of resonance - 5 . 6 J -8.8 T "Shake-up" or resonance 2s * 2s •* a 6.5 0 (a) Estimated uncertainty in energy-loss values is ± 0.08 eV except where stated. (b) The spin-orbit sp l i t t ing of 1.6 eV [181] has been used to estimate the 2 p ^ 2 and 2p j / 2 spin-orbit components from the 2p (mean) values [179]. Same procedure as for the P 2p has been used. (c) Ref. [179] (d) Final occupied orbital with either 2 p 3 / 2 o r 2p j / 2 hole state. *with respect to 2p (mean) edge [179]. - 191 -transform as both a^ and e combinations. The term values for the t r a n s i t i o n s i n the C l 2p spectra are lower than i n the P 2p spectra, consistent with the peripheral l o c a t i o n of the core hole. Features 3 and 4 have term values of 2.6 eV and 2.5 eV from the ^3/2 a n C* ^1/2 e < * S e s r e s P e c t i v e l y « This i s consistent with a t r a n s i t i o n to the 4s Rydberg l e v e l and they have been assigned accordingly. Thus feature 3 i s either s o l e l y due to a Rydberg t r a n s i t i o n or to a combination of valence and Rydberg. The rest of the features can be assigned to various Rydberg t r a n s i t i o n s . The rapid and intense r i s e i n o s c i l l a t o r strength i n the region of the i o n i s a t i o n edge suggests that t h i s may be on top of a broad 2p •*• a (a^) resonance - as was found in*the P spectrum. The o s c i l l a t o r strength beyond the edge i s probably due to various d - l i k e shape-resonances or "shake-up". - 192 -CHAPTER 6 ELECTRONIC EXCITATIONS IN PHOSPHORUS CONTAINING MOLECULES. I I . INNER SHELL ELECTRON ENERGY LOSS SPECTRA OF PFc;,OPF, AND QPC1 ? In the preceding chapter, the ISEELS spectra of the t r i v a l e n t compounds PH 3, PF 3, PC1 3 and P ( C H 3 ) 3 were examined and also compared with the spectra of the analogous s i l i c o n s e r i e s . The ligands were seen to have a si m i l a r e f f e c t on the central atom core spectra for both s e r i e s , however, unlike the s i l i c o n series there did not seem to be a s i m i l a r l i n e a r r e l a t i o n s h i p for shape-resonance p o s i t i o n with bond length i n the l i m i t e d pshosphorus s e r i e s . In the analysis of the P 2p spectra i t was found that the term values obtained from the P 2s spectra were very u s e f u l . In t h i s chapter, the study of the ISEELS spectra of phosphorus compounds i s extended to include PF 5, OPF 3 and 0PC1 3. The r e l a t i o n s h i p of resonance p o s i t i o n with bond length w i l l be further examined as w i l l the assumption that term values from the same core atom centre are transferable. As before, a l l inner s h e l l regions accessible with the current instrumentation (P L - s h e l l , C l L - s h e l l , 0 and F K-shells) are presented. Experimental D e t a i l s The spectra were recorded and c a l i b r a t e d i n the same manner as for the t r i v a l e n t phosphorus compounds. I - 193 -RESULTS AND DISCUSSION Phosphorus Spectra The long-range P 2p,2s spectra are shown i n F i g . 6.1. Also shown on the spectra are the positions of the i o n i s a t i o n edges taken from XPS [31,175]. As only the mean 2p values were reported [31] the positions of the 2 p 3 ^ a n d spin-orbit components were estimated as before, using a spin-orbit s p l i t t i n g of 0.90 eV [61] i n conjunction with a spectral weighting of 2:1 ( i . e . , 2 p ^ 2 = 2p - 0.30 eV, 2 p ^ 2 = 2p + 0.60 eV). A comparison of the spectra reported here with the PF 3 and PC1 3 spectra ( F i g . 5.1) show a number of s i m i l a r i t i e s which can be associated with the e f f e c t s of the electronegative ligands. The discr e t e portions show r e l a t i v e l y strong t r a n s i t i o n s to the v i r t u a l valence o r b i t a l s at the expense of Rydberg t r a n s i t i o n s . These are followed by intense t r a n s i t i o n s which are at the edge i n the chloro-compounds, and j u s t beyond the edge i n the fluoro-compounds, and these features can be associated with inner-well trapped states/shape-resonances [73,77]. Following t h i s , the spectra show broad continuum features which are also associated with shape-resonances [77]. The nature of the edge and post-edge features w i l l be discussed In more d e t a i l l a t e r , but f i r s t the di s c r e t e portions of the spectra w i l l be discussed. The higher r e s o l u t i o n 2p spectra of these compounds have also been obtained and each spectrum i s discussed below i n turn i n comparison with the corresponding 2s region. The 2s spectra are plotted on the same hori z o n t a l scale as the 2p spectra. In order to render the features more apparent, a l i n e a r ramp, extrapolated from the leading - 194 -T i 1 r i 1 1 r i r g P 2 p edge 10 AE=0.36eV ~1 1 1 1 1 — 130 140 150 P H O S P H O R U S 2p,2s REGION O P F , OPCI. 1 1 1 ' 1 1 1 ' 1 1 1— 160 170 180 190 2 0 0 210 E N E R G Y LOSS (eV) Figure 6.1: Phosphorus 2p,2s wide range electron energy loss spectra of PF 5, 0PF 3, and 0PC1 3. A l l spectra were obtained with an impact energy of 2500 V, a sc a t t e r i n g angle ~1°, and a resolution of 0.36 eV FWHM. - 195 -edge, has been subtracted from the 2s spectra. The OPF 3 and 0PC1 3 spectra were extracted from the long-range spectra shown i n F i g . 6.1, whereas the PF 5 2s spectrum was rerun separately. The P 2p and P 2s spectra are l i n e d up on t h e i r respective i o n i s a t i o n edges. In the case of the 2p spectra the mean value of the Zp-^/Z -^jl i ° n * - s a t * - o n edges was used. The 2s spectra are considerably broader than those for the 2p spectra due to fast autoionisation processes of a Coster-Kronig type. Two assumption used i n Chapter 5 w i l l also be made here, namely: ( i ) the major i n t e n s i t y i n the 2s spectrum can be a t t r i b u t e d to t r a n s i t i o n s to o r b i t a l s with the largest p o r b i t a l content as they w i l l have the largest s -»• p (atomic) dipole allowed c o n t r i -bution. ( i i ) since the core-hole vacancy i s situated on the same atom, term values from the 2s spectra are transferable to the 2p region. Phosphorus Pentafluoride - PF R This molecule belongs to the point group. The possible t r a n s i t i o n s are l i s t e d i n Table 6.1. Using a minimal basis set, the vacant o r b i t a l s are of a J, e' and a£ symmetry. Since i n the case of PF 5 the 2s o r b i t a l i s of a^ symmetry and the p o r b i t a l s transform as e 1 and a'2, the following t r a n s i t i o n s should be dipole allowed i n the phosphorus L - s h e l l spectra: - 196 -TABLE 6.1 Transitions from the 1AJ Ground State in D-j^ Symmetry Final Configuration * Final State Dipole Allowed from ground state hole state occupied o* orbital a i a i A i No a i e' E' Yes a i a 2 A2 Yes e' a i E' Yes e 1 e' E' Yes e' e' A j ' + A£ No e' a 2 E" No a 2 a i A2 Yes a 2 e' E" No a 2 a 2 A i No (2s) , (2p-jy2) an<* (2pj/ 2) holes are of a j , e' and ajj symmetry respectively. The a* orbitals are of aJ, e' and a 2 symmetry. - 197 -2s(ap - o*(a" 2), o*(e«) 2p(e') > ( 2 p 3 / 2 r 1 a * ( a p , C ^ ^ ) " 1 a*(e« ) 2p(ap * ( 2 p 1 / 2 ) - l o * ( a p i * The t r a n s i t i o n 2p(a£) -• ( 2 p 1 y 2 ) - i a (e') i s formally dipole forbidden. From a consideration of the spectrum which i s shown i n F i g . 6.2, and the spin-o r b i t s p l i t t i n g s of the 2 p ^ 2 ^ 2 i o n i z a t i o n edges, i t seems reasonable to assign features 1 and 2 as the s p i n - o r b i t components of the 2p -»• a (a^) t r a n s i t i o n . Feature 3 i s the 2p(e') •+• ( 2 p 3 ^ 2 ) _ i o ' (e') t r a n s i t i o n . I t has a very s i m i l a r term value to feature * 11, which i s assigned as the 2s -*• a (e) t r a n s i t i o n . Feature 12, which has no counterpart i n the 2p spectrum, i s then the 2s •*• a ( a 2 ) t r a n s i -t i o n . The rest of the features 4-7, which are r e l a t i v e l y sharp, are assigned to various 2p -* Rydberg t r a n s i t i o n s . The assignments are summarised i n Table 6.2. Two points should be emphasised i n connection with the above assignments. F i r s t l y , the respective term values for the t r a n s i t i o n s to the a (e') o r b i t a l from the 2p(e') and 2s o r b i t a l s agree to within 0.1 eV. Secondly, as would be expected, the i n t e n s i t y of the t r a n s i t i o n assigned to the 2s electron to the doubly degenerate a (e') o r b i t a l (mainly of P 3p(x,y))^ i s much greater than that to the non-degenerate ^ The p r i n c i p a l axis of the molecule i n the present work i s designated as tne z axis. - 198 -TERM VALUE (eV) 8 4 " T 0 - 4 1 2 P 2p 3 l 22p l | 2 edges 3 4 5 6 7 P F 5 P2p from long-range spectrum AE=O.I8eV I 3 6 140 I 4 4 P 2s edge !L_ 12 PF 5 P2s AE=0.36eV I92 196 2 0 0 2 0 4 ENERGY LOSS(eV) Figure 6.2: Phosphorus 2p and 2s electron energy loss spectra of PF 5. P 2p spectrum (upper trace) i s at high resolution (0.18 eV FWHM). P 2s spectrum (lower trace) i s at 0.36 eV FWHM. The spectra are aligned with respect to the 2p (mean) and 2s i o n i s a t i o n edges. - 199 -TABLE 6.2 Energies, Term Values, and Possible Assignments for the P 2p,2s Spectra of PF 5 Feature Energy Loss (eV) (a) Term Value (eV) Possible Assignments (d) 2 p 3 / 2 2p 1/2 2 p 3 / 2 2 p 1 / 2 2P3/2 1 138.22 (12) 2 138.97 (10) 3 140.73 (10) 4 141.67 5 142.51 6 143.10 7 144.00 l i m i t b 144.38 l i m i t b 145.28 8 149.1 (3) 9 157.5 (5) 10 162.2 (5) 11 2s l i m i t c 12 198.12 (20) 201.87 202.60 6.16 5.41 3.65 2.71 1.87 1.28 0.38 0 6.31 4.55 3.61 2.77 2.18 1.28 0 (a}) a*(e') 4s 3d 5s, 4d a (a{) 4s 3d 5s, 4d -4.4' -12.8 -17.5 inner-well state/shape-rgsonance o^(P-F a x) shape resonance 0 (P~^eq) shape resonance 2s 3.75 0 -0.73 2s -» a (e') 2s •+ a'ia'p (a) Estimated uncertainty in energy-lgss values is ± 0.08 eV except where stated. Spectra are calibrated against N 2 (Is -»• it , v = 1) at 401.10 eV. (b) The spin-orbit sp l i t t ing of 0.90 eV [61] has been used to estimate the 2p$/2 a n i * ^7>i/2 spin-orbit components from the 2p edge (mean) values [31], see text for detai ls . (c) Ref. [175]. (d) Final occupied orbital with either 2p j / 2 o r n o ^ e state. * With respect to the 2p edge (mean) [31]. - 200 -* a (a^) o r b i t a l (mainly of P 3p(z)). This also lends support to the i n t e r p r e t a t i o n of the 2s spectra of the t r i v a l e n t phosphorus compounds where the maximum i n t e n s i t y of the broad envelope was ascribed to the degenerate a (e) o r b i t a l . Phosphoryl T r i f l u o r i d e - 0PF q The assignment of the P 2p spectrum of 0PF 3 ( F i g . 6.3) i s less clear than that of P F 5 due to the fact that i t i s a mixture of overlapping t r a n s i t i o n s . In this regard the s i t u a t i o n i s s i m i l a r to that for the c e n t r a l atom ISEELS spectra of P F 3 ( F i g . 5.4) and also the o p t i c a l absorption spectrum of S i F ^ [144]. Thus the spectrum presumably consists of a t r a n s i t i o n to a v i r t u a l o r b i t a l , followed by a t r a n s i t i o n to a higher v i r t u a l o r b i t a l with superimposed Rydberg t r a n s i t i o n s . The spectral shape i s c e r t a i n l y consistent with such an assignment. The ground state molecule, l i k e i t s counterpart PF 3, i s of C^ v symmetry for which the dipole-allowed t r a n s i t i o n s are l i s t e d i n Table 5.1. The vacant o r b i t a l s , using a minimal (no d) basis set, are of a^, e and a^ symmetry. Transitions to a l l these le v e l s are possible from the 2s and 2p o r b i t a l s . A CNDO/2 c a l c u l a t i o n indicates that the f i r s t a^ o r b i t a l i s predominantly P 2s i n character and so should show l i t t l e i n t e n s i t y i n the P 2s spectrum. The second aj^ o r b i t a l i s analogous to the a" o r b i t a l In P F 5 with a l o t of p(z) character. Thus the 2s spec-- 201 -TERM VALUE (eV) 8 4 0 - 4 1 1 1 1 1 1 1 1 1 00 P 2P3/2 2 P l|2 e d 9 e S 11 I I I I 1 2 3 4 5 6 0PF 3 P2p UNIT 1 7 ITRARY "•^w" *'""\ / (ARB AE=O.I8eV > 1 , , , , , . , 136 140 144 1 fENSI y P2s edge 1 OPF3 fENSI 10 P2s 1 LU > • •* ** •. • *. ' 11 * • • •*• . _ATI . ••».* •. •*• LU cr • •' • * AE=0.36eV 192 196 200 204 ENERGY LOSS(eV) Figure 6.3: Phosphorus 2p and 2s electron energy loss spectra of 0PF 3. P 2p spectrum (upper trace) i s at high resolution (0.18 eV FWHM). P 2s spectrum (lower trace) extracted from Figure 6.1. The spectra are aligned with respect to the 2p (mean) and 2s i o n i s a t i o n edges. - 202 -trum would be expected to be dominated by a t r a n s i t i o n to the a (e) o r b i t a l (as per preceding discussion for PF 5) and feature 10 i s assigned as such. The breadth of the peak i n the 2s spectrum and the i n t e n s i t y on the high energy side of feature 10 i s consistent with there being an * a d d i t i o n a l , less intense 2s •*• a (a^) t r a n s i t i o n underneath, which would be expected i n the case of 0PF 3 since there are two a^ type v i r t u a l o r b i t a l s , as indicated above. The t r a n s i t i o n to the lower a^ o r b i t a l i s l i k e l y under the low energy side of peak 10, corresponding to peaks 1 and 2 i n the 2p spectrum discussed below. A l l assignments are summarised i n Table 6.3. Applying these to the 2p spectrum, features 1 and 2 are assigned * as the spin-orbit components of the 2p to the f i r s t a (a^) o r b i t a l . The term values of peaks 3 and 4 with respect to the ^.P^/l a n c * ^1/2 e c * 6 e s are almost the same as the term value of feature 10 (see Table 6.3) thus these t r a n s i t i o n s are assigned to the spin-o r b i t components of the 2p •*• a (e) t r a n s i t i o n s . However, the i n t e n s i t y , sharpness and term value of feature 4 suggests that i t also contains a contribution from the 2 p 3 y 2 s p i n - o r b i t component of the 2p -»• 4s Rydberg t r a n s i t i o n . The term value i s also i n keeping with other phosphorus 4s values (see P F 5 ) . Features 5 and 6 are assigned to higher Rydberg t r a n s i t i o n s . These are evidently superimposed on top of broad underlying structure that can be assigned to the 2p t r a n s i t i o n to the second a ( a ^ o r b i t a l . This i n t e r p r e t a t i o n i s consistent with the proposed assignment of the 2s spectrum discussed above. - 203 -TABLE 6.3 Energies, Term Values, and Possible Assignments for the P 2p,2s Spectra of OPF3 Feature Energy Loss^ a^ Term Value Possible Assignments^^ (eV) (c sV) 2P3/2 2Pi/2 2 p 3 / 2 2 Pl /2 1 137.18 5.78 0 (aj) 2 3 4 5 6 j^>3/2 l i m i t b 2 p ^ 2 l i m l t b 7 137.86 139.60 140.37 (8) 141.53 142.44 142.96 143.86 5.10 3.36 2.59 1.43 0.52 0 6.00 4.26 3.49 2.33 1.42 0 o*(e) 4s 5s, 3d o*(a 1) a (e) 4s Ion tgp of 5s, 3d| p+a (aj) 148.3 (3) -5 inner-we11 state/shape-8 9 159.7 (5) 166.6 (5) -16.4? -23.3T resonance 0 (P-F) shape resonance a (P-0) shape resonance 2s 10 2s l i m i t c 11 196.92 (20) 200.47 206.1 (5) 3.55 0 -5.6 2s •»• o*(e) inner-well state/shape resonance (a) Estimated uncertainty in energy-lgss values is ± 0.12 eV except where stated. Spectra are calibrated against N 2 (Is -* it , v • 1) at 401.10 eV. (b) The spin-orbit sp l i t t ing of 0.90 eV [61] has been used to estimate the 2 p 3 / 2 and 2 p ^ 2 spin-orbit components from the 2p edge (mean) values [31], see text for detai ls . (c) Ref. [175] (d) Final occupied orbital with either 2p 3 / 2 o r 2 Pl /2 n o ^ - e state. ? With respect to the 2p edge (mean) [31]. - 204 -Phosphoryl T r i c h l o r i d e - 0PC1, The d e t a i l e d phosphorus 2p and 2s spectra of OPCI3 a r e s n o w n *-n F i g . 6.4 and the s p e c t r a l information i s summarised i n Table 6.4. The o p t i c a l absorption spectrum of 0PC1 3 i n the P 2p region has been previously reported [173,174]. However, i t i s important to note that the energy scale shown i n the preliminary o p t i c a l work [174] i s c l e a r l y i n error by about 9 eV. However, the spectrum shown i n the second paper [173] i s consistent with the spectrum reported here. The l a t t e r paper [173] also shows the r e s u l t s of an X -SW c a l c u l a t i o n , however, no a i n t e r p r e t a t i o n i s given. The molecule 0PC1 3 i s of C^ v symmetry and so i s governed by the same se l e c t i o n rules (Table 5.1) as OPF3. Proceeding as before, the major feature (15) i n the 2s spectrum i s assigned to the 2s •*• a (e) t r a n s i t i o n . There i s evidence of a shoulder on th low energy side of t h i s peak. This i s presumably due to a 2s •*• a (a^) t r a n s i t i o n . Applying the 2s term values so obtained (Table 6.4) to the 2p spectrum, features 1 and 2 may be assigned to the spin-orbit components of the 2p •*• a (a^) t r a n s i t i o n and features 3 and 4 to the two components of the 2p -»• a (e) t r a n s i t i o n . The term value for feature 5 with respect to the ^3/2 e ( * £ e * s 3*^7 eV. This i s somewhat high for the 4s term value when i t i s compared to those observed for the peaks assigned to t r a n s i t i o n s to the 4s l e v e l i n the other phosphorus spectra. These values l i e between ~2.4 and 3.1 eV for the assignments given here and i n Chapter 5. Furthermore, the i n t e n s i t y of peak 4 suggests i t contains contributions from more than one t r a n s i t i o n . Since the separation between - 205 -TERM VALUE (eV) 8 4 0 -4 oo >-or < or GO or < b 00 UJ LU > i r P 2 P 3 / 2 2 p i / 2 e d 9 e S rn—i—i i II I I—TTT 1 2 3 4 5 6 7 8 9 10 11 12 • 1 OPCI; P2p AE=O.I8eV LU 32 136 140 1 15 P 2s e d g e •V" OPCI. P2s AE=0.36eV 188 92 196 200 ENERGY LOSS(eV) Figure 6.4: Phosphorus 2p and 2s electron energy loss spectra of 0PC1 3. P 2p spectrum (upper trace) i s at high resolution (0.18 eV FWHM). P 2s spectrum (lower trace) extracted from Figure 6.1. The spectra are aligned with respect to the 2p (mean) and 2s i o n i s a t i o n edges. - 206 -TABLE 6.4 Energies, Term Values, and Possible Assignments for the P 2p,2s Spectra of OPCI3 Feature Energy Loss^ a^ (eV) Term Value (eV) Possible Assignments^^ | 2P3/2 2 p 1 / 2 2P3/2 2Pl/2 1 2 3 4 5 6 7 8 9 10 11 12 2 p 3 / 2 l i m i t < b ) 2pi/2 l i m i t ( b ) 13 14 1 5 ( ^ 2s limit K Q-> 134.06 134.85 135.81 (10) 136.67 137.49 138.00 138.27 (12) 139.03 139.61 140.59 (10) 141.09 (10) 141.73 (20) 141.06 141.96 152.4 (5) 164.3 (5) 193.67 (20) 198.86 7.00 6.21 5.25 4.39 3.57 3.08 2.99 2.03 1.45 0.47 -0.03 0 7.11 6.15 5.29 4.47 3.98 3.89 2.93 2.35 1.37 0.87 0.23 0 a*(a x) 0 -cr < t m rr < 5 H -i 1 1 1 r FIs edge i i r — i 1 2 3 4 - T — i — < — T 0 P F -F1s AE=0 .36eV __ CO L±J - i o H LU > L U L T 5 -J —|—-i 1 1 1 1 1 1 r 6 8 5 7 I 0 01s edge 1 _ _ : ">—r~ 7 3 5 T i 1 r 7 6 0 n T 1 23 4 0 P F 3 01s A E = 0 . 3 6 e V -i r L | 1 1 1 1 1 1 1 525 5 5 0 575 ENERGY LOSS (eV) - i — i — i — i 6 0 0 Figure 6.6: The wide range electron energy loss spectra of the 0 Is and F Is regions of 0PF 3. The spectra are aligned with reference to the respective Is edges. - 216 -CO t z or < cr GO < >-CO LU LU > io H 5 4 < 10 —I i LU cr io 8 T E R M V A L U E (eV) 4 2 0 -2 - 4 -6 1 1 r ^ F 1s edge 2 .V. OPF, F1s AE=0 .36eV 6 9 0 6 9 4 6 9 8 702 ^ 0 1s edge 7 0 6 2 3 OPF3 01s AE=0.36eV • • •."n-»-.-.v..«-l<..'-*"—•"• - c _ 532 5 3 6 5 4 0 5 4 4 E N E R G Y L O S S ( e V ) 548 Figure 6.7: Detailed electron energy loss spectra of the 0 Is and F Is regions of OPF 3. The spectra are aligned with reference to the respective edges. TABLE 6.7 Energies , Term Values , and Poss ib le Assignments for the 0 Is, F Is Regions of OPF 3 Feature Oxygen Is F luor ine Is Poss ible Assignment Energy Loss (eV) Term Value (eV) Energy Loss (eV) Term Value (eV) 1 533.87 (20) 5.4 690.30 (30) 5.5 Is •>• * (a^) 2 535.81 (15) 3.5 692.71 (20) 3.1 Is * a* (e) 3 -537 -694 Is * a* (a^) edge 3 539.3 695.8 4 544.3 (5) -5 .0 700.1 (5) -4 .3 inner we l l trapped state/shape resonance 5 555 - 565 710 - 720 shape-resonance s t ructures aRef. [179J. - 218 -00 ZD or < cr m or < >-00 LU h-LU > h-< _J LU LT 10 530 10 H 550 570 J 0 1 s edge T AE=0.36eV 530 —I 1 1 1 I 534 538 542 E N E R G Y L O S S ( e V ) 546 550 Figure 6 .8: Detailed electron energy loss spectra of the 0 Is region of 0PC1 3. The i n s e r t shows a wide range spectrum. - 219 -TABLE 6.8 Energies, Term Values, and Possible Assignments for the 0 Is Region of O P C I 3 Feature Energy Loss Term Value Possible Assignment (eV) (eV) 1 532.60 (15) 5.5 Is •* a* (a^) 2 533.86 (15) 4.2 Is •+ a* (e) 3 537.72 (20) 0.4 inner well trapped state/ shape-resonance 0 Is l i m i t 3 538.1 0 4 542 - 550 shape-resonance aRef. [179]. - 220 -the peripheral p o s i t i o n of the core hole. This difference i s much greater than i n the case of 0PF 3 (compares Tables 6.3 and 6.7) but more i n keeping with that observed for P F 5 and the molecules i n Chapter 5. Features 1 and 2 are assigned as t r a n s i t i o n s to the a (a^) and a (e) o r b i t a l s r e s p e c t i v e l y . There does not appear to be any s i g n i f i c a n t i n t e n s i t y for a t r a n s i t i o n to the second a (a^) o r b i t a l , however, there i s i n t e n s i t y on the high energy side of feature 2 which may be due to such a t r a n s i t i o n (compare P 2p spectrum, Table 6.4 and F i g . 6.4). Feature 3 i s assigned to a t r a n s i t i o n to an inner-well trapped state/ shape-resonance. The weak structure, l a b e l l e d 4, i s also a t t r i b u t e d to a shape-resonance. Figure 6.9 shows the C l 2p,2s spectra. The spectral data i s summarised i n Table 6.9. The spectrum i s i n good agreement with the previously published o p t i c a l spectrum [173] and i s s i m i l a r to other e a r l i e r reported C l 2p,2s spectra such as those for PC1 3 and the chloro-methanes [182]. Overlapping bands complicate the spectral assignment. Features 1 and 2 are assigned to the spin-orbit components of the 2p •*• a (aj) t r a n s i t i o n . Transitions to the other v i r t u a l o r b i t a l s ( a ( e ) , a (a^)) are also expected. Using the separation of the v i r t u a l * * * * o r b i t a l s from the P 2p spectrum ( a ( a 1 ) - a (e) ~ 1.75 eV, a ( a 1 ) - a (aj) ~2.61 eV) as a guide, features 3 and 5 are assigned to the 2p •*• a (a^) t r a n s i t i o n . The 4s Rydberg t r a n s i t i o n s are assigned to features 4 and 6. The rest of the Rydberg t r a n s i t i o n s are on top of the rapid r i s e i n o s c i l l a t o r strength, giving r i s e to features 7 and 8, which are assigned to a t r a n s i t i o n to the inner-well trapped state or shape-resonance. Feature 9 i s thought due to a a (P-Cl) shape-resonance since 6 (the - 221 -10 -Kfc Cl 2p edges 111HI i r — i I 235678 9 10 AE=0.36eV O P C I 3 Cl 2p,2s r r n 12 fcCI 2s edge n — I — r 2 2 5 I—r 2 6 5 2 0 5 2 4 5 2 8 5 2P34 2\£ C I2p edges 1 — n — I I I — 2 3 4 5 6 7 8 10 AE=O.I8eV 5 -T T 2 0 0 2 0 4 2 0 8 2 1 2 E N E R G Y L O S S (eV) 2 16 Figure 6.9: High reso l u t i o n (0.18 eV FWHM) electron energy loss spectrum of the C l 2p region of 0PC1 3 (lower t r a c e ) . The upper trace shows the combined C l 2p,2s region recorded with a resolution of 0.36 eV FWHM. - 222 -TABLE 6.9 Energies, Term Values, and Possible Assignments for the C l 2p,2s Regions of OPCI3 Feature Energy L o s s ^ 3 ^ (eV) Term Value (eV) Possible Assignments^^ 2P3/2 2 p i / 2 2 p 3 / 2 2P1/2 1 2 3 4 5 6 7 8 2 p 3 / 2 l i m l t f J J 2 p i / 2 l i m i t ( b ) 9 10 11 2s l i m i t ( c ) 200.82 202.35 (10) 203.62 204.28 205.27 (10) 205.93 (10) 207.52 (10) 209.22 (12) 206.9 208.5 216.65 223 272.6 275.7 278.26 6.0 4.5 3.2 2.6 1.6 0.9 -0.7 0 6.1 4.8 4.2 3.2 2.5 0.9 -0.8 0 o*(a,) a ( a ^ 4s et c . inner well state/shape resonance' o*(a, ) o*(eJ o*(a 1) 4s inner well s t a t e / shape-resonance -16 T o-*(Cl-P) Bht " shaV ape resonance ce-up" 2s 2s + 0 * 2s •* a 5.6 2.5 0 (a) Estimated uncertainty i n energy l o s s ± 0.08 eV unless otherwise stated. (b) 2p edge Ref. [31]; 2p 3 /, 2 = 2p (mean) - 0.53 eV; 2 p 1 / 2 = 2p (mean) + 1.07 eV; A s p i n - o r b i t s p l i t t i n g of 1.6 eV [18] has been assumed. (c) 2s edge r e f . [31]. (d) F i n a l occupied o r b i t a l with e i t h e r 2p-j/ 2 or 2 p ^ 2 hole s t a t e . *With respect to C l 2p (mean) edge [31]. - 223 -negative of the term value) i s ~9.3 eV, which i s reasonably close to the * average value of 10.8 eV observed (see Table 6.5) for a (P-Cl) resonances from the P 2p spectra. A weak feature 10 i s possible due to "shake-up". CONCLUSIONS The ce n t r a l atom and ligand ISEELS spectra measured for a series of t r i v a l e n t phosphorus compounds presented i n Chapter 5 have been extended to include the higher coordinated phosphorus compounds PF 5, 0PF 3, and OPClg. A number of questions raised i n Chapter 5 have been answered. The use of the P 2s term values to assign the P 2p spectra i s * further supported by the present work. The f i r s t two a l e v e l s are w e l l separated i n the presently studied molecules and hence the assignment of the core to valence t r a n s i t i o n s i s less ambiguous. The respective a (e) (or a (e') for PF 5) term values for the P 2s and 2p spectra were found to be within ~0.1 eV of each other, which lends confidence to the * ordering of the a l e v e l s as given i n the previous chapter, and also indicates that term values are transferable when the core hole i s located on the same atom. The continuum structure was s i m i l a r to that observed for PF 3 and PC1 3 i n that there was an intense feature at or just beyond the edge a t t r i b u t a b l e to an inner-well trapped state/shape-resonance followed by broad shape resonances. In addition, each spectrum here showed an extra resonance i n comparison with the PX 3 spectra. This i s d i r e c t l y a t t r i b u t a b l e to the fact that the molecules here each possess two d i f f e r e n t kinds of ligands. A comparison of the - 224 -resonance positions i n the P 2p spectra i n P C 1 3 and O P C I 3 confirmed the * a (P-Cl) nature of the resonance i n these molecules. A s i m i l a r conclu-sion was reached i n the assignment of the feature i n OPF 3 and PF 3 ( i . e . , * a (P-F)). The resonances i n the P 2p PF 5 spectrum were consistent with there being two types of f l u o r i n e s i n PF 5 ( i . e . , a x i a l and e q u a t o r i a l ) . * A comparison of 0 P C 1 3 and OPF 3 confirmed the a ( P - 0 ) nature of the second resonance i n those molecules. This lends support to the e x i s -tence of some sort of r e l a t i o n s h i p between resonance p o s i t i o n and bond length [ 9 6 - 1 0 0 ] . The absence of such a r e l a t i o n s h i p for the molecules studied i n Chapter 5 would appear to be due to a d i f f e r e n t phenomenon being responsible for the structure i n P(CH 3) 3. It i s possible that the continuum features i n P(CH 3) 3 are due to "shake-up" processes, as was found for PH 3. Obviously great care must be exercised i n the assignment of continuum features which can be ascribed to one or more of several e f f e c t s including trapped-inner well states, resonances, "shake-up" or double e x c i t a t i o n . - 225 -CHAPTER 7 ELECTRONIC EXCITATION IN PHOSPHORUS-CONTAINING MOLECULES. I I I . VALENCE SHELL ELECTRON ENERGY LOSS SPECTRA OF P(CH,),, PCI,, PF,, OPCI,, and Very l i t t l e Information on the valence s h e l l e x c i t a t i o n processes e x i s t s i n the l i t e r a t u r e on the molecules studied i n the previous two chapters. The only reported VSEELS spectrum for any of t h i s series of compounds i s that for PH 3 referred to i n the book by Robin [12]. However, there have been several photoabsorption studies on some of these molecules, but only over a rather l i m i t e d energy range (up to =10 eV) dictated by the use of conventional o p t i c a l spectrometers and l i g h t sources. These studies included spectra of PF 3 and PC1 3 [184-186], P(CH 3) 3, and 0PC1 3 [185], as well as PH 3 [178,184,185]. In thi s chapter the VSEELS spectra of P(CH 3) 3, PC1 3, 0PC1 3, PF 3, and PF 5 are presented up to 20 eV and beyond. The spectra are interpreted with the aid of the ISEELS r e s u l t s from Chapters 5 and 6. A comparison of these two techniques was useful i n t e n t a t i v e l y assigning the VSEELS spectra of, NF 3 and S i ( C H 3 ) 4 (see Chapters 3 and 4), since the ISEELS spectra are generally r e l a t i v e l y simple to assign due to the energy i s o l a t i o n of the I n i t i a l core hole, which unambiguously defines the i n i t i a l o r b i t a l of the t r a n s i t i o n . EXPERIMENTAL DETAILS The spectra were a l l obtained on the spectrometer described i n - 226 -Chapter 2. An impact energy of 2.5 keV was used to obtain the spectra with the scattered electrons sampled at zero degree scattering angle. The spectra were obtained with a t y p i c a l energy res o l u t i o n of 0.035 -0.050 eV. At zero degree scattering i t i s not always possible to obtain background-free operation i n the valence s h e l l region. This was e s p e c i a l l y the case for PC1 3 and PF 5. However, removal of the gas indicated that the background signal was smoothly varying and possessed no sharp features due to "ghosting" e f f e c t s a r i s i n g from r e f l e c t i o n of the primary beam, br secondary emission from electrode surfaces. With the exception of PC1 3 a l l spectra were cal i b r a t e d with the He(I) l i n e (21.218 eV). I n i t i a l l y the PC1 3 sample as supplied contained a small HCl impurity that was used to c a l i b r a t e the spectrum. A l l traces of HCl were then removed by continuous pumping on a PC1 3 sample that was cooled with a dry-ice/methanol mixture. A s i m i l a r procedure was used for 0PC1 3, as a precautionary measure even though no HCl was immediately apparent. RESULTS AND DISCUSSION Before discussing the VSEELS spectra i t i s useful to review some of the pertinent points from the discre t e portion of the P 2p spectra reported for these compounds. With the exception of P ( C H 3 ) 3 a l l P 2p spectra show strong core •*• v i r t u a l valence t r a n s i t i o n s well separated from the core -*• Rydberg t r a n s i t i o n s . I t i s possible to assign these t r a n s i t i o n s to those going to the f i n a l o r b i t a l s expected from a minimal basis set. The ordering of the previously unoccupied valence o r b i t a l s - 227 -was established by comparison of the P 2p spectra with the P 2s spectra. The remaining t r a n s i t i o n s can be assigned to Rydberg t r a n s i t i o n s . Since these originate from a 2p l e v e l , the dominant Rydberg t r a n s i t i o n s are assigned as those going to the s and d Rydberg l e v e l s . A p •*• p t r a n s i -t i o n i s formally dipole forbidden i n the pure atomic case and hence would be expected to have lower i n t e n s i t y , e s p e c i a l l y i n more highly symmetric molecules [61]. Table 7.1 summarizes the term values (T) obtained from Chapters 5 and 6 for t r a n s i t i o n s to the v i r t u a l valence o r b i t a l s and also to the assigned 4s Rydberg l e v e l . Also shown i s the quantum defect (6) obtained for the s Rydberg series from the 4s term values. These quantum defects a l l l i e between ~1.7-1.9, which i s only s l i g h t l y lower than the "expected" quantum defect of 2 for the s Rydberg serie s of the t h i r d (Na-Ar) row [12]. On moving from the core to the valence region the term values would be expected to be lower for the v i r t u a l valence l e v e l s . This i s e a s i l y r a t i o n a l i s e d i n terms of the l o c a t i o n of the hole. The loss of shie l d i n g caused by the removal of a valence electron should be less than that caused by the removal of a l o c a l i s e d core electron from the centre of the molecule and so the electron i n the newly occupied a o r b i t a l sees something approaching a whole extra unit of charge i n the case of a ce n t r a l core hole and hence has a higher term value ( i . e . , binding energy). This e f f e c t has been discussed previously with regard to NF 3 and i s s i m i l a r to the e f f e c t which occurs when the core hole i s located on a ligand as opposed to the central atom (see the previous chapters). The e f f e c t of the i n i t i a l o r b i t a l vacancy l o c a t i o n on - 228 -TABLE 7.1 Term values (T) for phosphorus L-shell spectra and the calculated s orbital quantum defects (6) Molecule Term Value ( e V ) ( a ) Quantum^ ^ Defect 6 Virtual O r b i t a l s ( b ) Rydberg •4s o*(l) - 3.12 1.91 PCI 3 6.9 6.4 - 3.06 1.89 P ( C H 3 ) 3 3.3 ~2.0 - 2.86 1.82 O P C I 3 7.0 6.2 4.4 3.08 1.90 P F 5 6.2 3.6 -0.7^) 2.71 1.76 From Chapters 5 and 6. (a) (b) Term values are with respect to the P 2p edge except where stated. Symmetries of the unoccupied virtual orbitals (as assigned in Chapters 5 and 6 are: o*(l) o*(2) o*(3) PH 3 , P ( C H 3 ) 3 > P F 3 P C I 3 0 P C 1 , P F c (c) e a l a l • i a l e e a l a2 (d) (e) (f) D i f f i c u l t to locate accurately, mean of 2 p 3 ^ and 2s term values used. Mean term value of the two possible assignments given. This term value is with respect to the P 2s edge. Calculated from T = 13.605/(n-6) where n = principal quantum number. - 229 -Rydberg term values should be somewhat less because ( i ) the term values fo r Rydberg o r b i t a l s are smaller than those for the (LUMO) valence o r b i t a l s ; ( i i ) the Rydberg o r b i t a l s are large and d i f f u s e and w i l l therefore tend to see the molecule as one large core. In assigning the VSEELS spectra an attempt w i l l f i r s t be made to locate the valence •* v i r t u a l valence t r a n s i t i o n s . In order to do t h i s i t i s i n i t i a l l y assumed that the term values are independent of the or i g i n a t i n g valence o r b i t a l . A l l expected term values are calculated from the experimental v e r t i c a l i o n i s a t i o n potentials as obtained from photoelectron spectroscopy. The v e r t i c a l i o n i s a t i o n potentials are summarised i n Table 7.2. Once the valence •»• v i r t u a l valence t r a n s i t i o n s are located, the possible valence ->• Rydberg t r a n s i t i o n s w i l l be assigned. The term values obtained from Chapters 5 and 6 w i l l be assumed to give upper bounds for those i n the valence spectra. Other evidence from the ligand ISEELS spectra w i l l be considered at appropri-ate junctions i n the following discussion of the various VSEELS spectra. Trimethyl Phosphine (P(CH-,)-,) The VSEELS spectrum of P ( C H 3 ) 3 from 4-25 eV i s shown i n F i g . 7.1. The energies, term values and possible assignments are summarised i n Table 7.3. The only previously reported valence s h e l l e x c i t a t i o n spectrum i s the UV absorption spectrum reported by Halmann [185] which only extends up to ~6.7 eV. This UV spectrum [185] shows a strongly absorbing, broad band centered at ~6.15 eV, i n very good agreement with - 230 -TABLE 7.2 Molecular orbitals^ 3^ and experimental ionisation potentials'^ (eV) for the valence orbitals of PH-j, PF3, PCI3, P(CH 3) 3, OPCI3, and PF 5 MO PH3 M0(c> PF 3 PCI3 P(CH 3) 3 MO OPCI3 M0 PF 5 5 a l 10.58 4 & 1 12.29 10.52 8.65 lie 11.93 2e" 15.54 2e 13.50 15.89 11.69 11.50 2a2 12.40 6e' 16.46 21.2 4e 16.31 11.99 11.25 lOe 12.94 5a 2 16.75 3e 17.45 12.94 13.25 13ax 13.48 8a I 17.13 3 a l 18.57 14.23 13.70 9e 13.86 la^ 17.79 2e 19.36 15.19 14.60 12 3 l 15.35 5e' 18.43 2 a l 22.60 18.81 16.65 8e 16.50 le" 19.1 l l a j 19.53 4a 2 19.5 4e' -21 ( a ) 0rder from [176] for PH3; [133,176,187,188] for PF 3; [188,189] for PCI3 [189,190] for OPCI3 and [191] for P F V ( b ) From PES, [192] for PH3; [118] for PF 3; [193] for PCI3; [190] for 0PC13; [193,194] for PFj and [195] for P(CH 3) 3. v 'Numbering ignoring core-levels. - 231 -H CO LU h-L J > LU 1 23 4 3d 3 7.46 1.19 4a j -»• 5s 4 8.24 3.01 4e •»• 4s 5 9.8 6 10.6 7 13.4 Term values are calculated with respect to the IP of the i n i t i a l o r b i t a l . These IP's are given i n Table 7.2. - 233 -feature 1 of the spectrum reported here. Robin has suggested [12] that the i n t e n s i t y of this feature indicates that i t has a large HOMO (4a 1)->o(a 1) valence component. This feature i s also l i k e l y to encompass the t r a n s i t i o n from the HOMO to the 4s Rydberg l e v e l ; indeed, i t i s * l i k e l y that the 0 (aj) and 4s w i l l form a strongly mixed valence-Rydberg pair as discussed by F r i e d r i c h et a l . [65]. The term value of feature 1 i s 2.44 eV and Robin [12] has indicated that this would be appropriate for a t r a n s i t i o n to a 4p l e v e l . However, ab i n i t i o c a l c u l a t i o n s [187] indicate that the 4a^ o r b i t a l (the phosphorus lone pair) i s l a r g e l y P 3p in character (65% P 3p, 14% p 3s). This result i s also borne out be recent X a c a l c u l a t i o n s [176] (60% P 3p, 11% P 3s). Thus a t r a n s i t i o n to 4p l e v e l should not be very intense [61]. Application of the Rydberg formula to the term value of feature 1 (2.44 eV) indicates that feature 3 could be the 4a^ •*• 5s Rydberg t r a n s i t i o n . Feature 2 can then be ascribed to the 4a^ •*• 3d Rydberg t r a n s i t i o n . The term value for t h i s feature i s 1.42 eV, which i s close to the 1.51 eV expected for a 3d term value calculated using a quantum defect of zero. The term values obtained for a l l the above t r a n s i t i o n s from the outer-most 4a^ o r b i t a l have been applied with respect to the other IP's i n order to predict the positions of the l e v e l s leading to the respect-ive l i m i t s . These estimated positions are indicated on F i g . 7.1 and i t can be seen that the p o s i t i o n of t r a n s i t i o n s leading to the 4e and l a 2 l i m i t s contributing to feature 4 are over estimated on th i s b a s i s . The term value for feature 4 from the next (4e) o r b i t a l i s 3.01 eV. The following o r b i t a l s (4e-3a,) have s i g n i f i c a n t C 2p character [176,187] - 234 -and the term value of 3.01 eV i s much more l i k e that observed for the C Is •*• 4s t r a n s i t i o n . This i s s i m i l a r to the findings for SKCHj)^ where i t was suggested that the f i r s t Rydberg l e v e l belonged to the ligand as opposed to the molecule as a whole. Pending a detai l e d theo-r e t i c a l treatment, nothing much can be said of the rest of the spectrum except that i t i s b u i l t up of a number of overlapping t r a n s i t i o n s to the a and Rydberg l e v e l s . Phosphorus T r i c h l o r i d e (PCI,) Figure 7.2 shows the VSEELS spectrum of PC1 3 from 4-20 eV. The data i s summarised i n Table 7.4. The spectrum i s much more complex than that of P(CH 3) 3, having many more c l e a r l y resolved t r a n s i t i o n s . The UV absorption spectrum reported e a r l i e r [184,186] at s l i g h t l y higher reso-l u t i o n shows v i b r a t i o n a l structure, but the spectra only extend as far as 10 eV. The term values for features 1 and 2 from the HOMO (4aj) o r b i t a l are 4.83 and 4.54 eV res p e c t i v e l y . Since these are both much higher than the term value for the 4s Rydberg l e v e l (3.06 eV - Table 7.1) from the P 2p ISEELS spectrum, they are assigned as t r a n s i t i o n s to * * the a (aj) and a (e) v i r t u a l valence l e v e l s , r e s p e c t i v e l y . The o r b i t a l ordering i s taken from that given i n Chapter 5. However, these l e v e l s are close together and the order could e a s i l y be i n fact reversed. Applying the term values thus obtained to the other valence lev e l s i t can be seen ( F i g . 7.2) that much of the spectrum can be reasonably assigned to valence-valence t r a n s i t i o n s . The features that remain unassigned are therefore at t r i b u t e d to Rydberg t r a n s i t i o n s . Thus - 235 -C O L U LU > L U cr o,e a,e P C I 3 Valence v.. o,e 4s a,e 4s "3d | 2 e 4s 4 P fir '3a, 4s 13d1 13d' 4 e 4s 3d 1 r - i — 1 — • — 4p 13d ?4a, I I 12 II III I I I I I I I 34 567 8 91011 12 13 14 n — 1 — 1 — 1 — 1 — 1 — r — n — 1 — 1 — 1 — 1 — 1 — 1 — r 10 15 20 ENERGY LOSS (eV) Figure 7.2: The valence s h e l l electron energy loss spectrum of PCI 3 . The i o n i s a t i o n edges are taken from photoelectron spectroscopy (see Table 7.2). Positions of the valence-valence ( t a l l bars), valence-Rydberg (short bars) t r a n s i t i o n s estimated from the term values are shown leading to each l i m i t (see Table 7.4). - 236 -TABLE 7.4 Energies and pos s ib le assignments for the VSEELS spectrum of PCI, Feature Energy Loss (eV) Valence - Valence Assignment (a) Valence - Rydberg Assignment (b) 1 5.68 2 5.97 3 7.03 4 7.43 5 8.25 6 8.63 7 9.05 8 9.85 9 10.48 10 10.76 11 11.40 12 11.96 13 12.64 14 13.42 4a j -»• c^Caj) la^ 4a 4e 3e ° * ( e ) 3e ->• a (e) 3a j 2e 2e • cr (a , ) • a (e) ° * ( a l ) a (e) 4e ^ ( a j ) 4a, 4a 1 4s 4p 4a j •*• 3d, 4e * 4s 3e -»• 4s l a 2 + 3d, 4e -»• 3d 3e -* 3d, 3ai •+ 4s 3a^ * 4p 2e + 4s, 3aj •+ 3d 2e ->• 3d (a) * o ^ ( a i ) term value = 4.83 eV, a (e* term value = 4.54 eV, I P ' s : see Table 7.2. (b) Only f i r s t member of se r ie s g iven : 4s term value = 2.95 eV, 4p term value = 2.26 eV and 3d term value = 1.46 eV. I P ' s : see Table 7 .2 . - 237 -the sharp feature 7 i s described as a combination of the 4a ^ •*• 3d and 4e •* 4s Rydberg t r a n s i t i o n s , the term values of which are 1.46 eV and 2.95 eV respectively, As with PCCrl^g, the 3d term value Is close to that which i s expected (1.51 eV) with a quantum defect of zero. The 4s term value (2.95 eV) i s only ~0.1 eV less than that i n the ISEELS spectrum. The 4s term value from the C l 2p spectrum i s ~2.6 eV. These term values (as well as those calculated for the 5s and 4d l e v e l s ) have been applied to the other valence o r b i t a l IP's and the positions of the corresponding l e v e l s are indicated on the spectrum ( F i g . 7.2). It can be seen that a l l the major features are reasonably accounted for by t h i s tentative assignment process. Feature 5, which shows a clear v i b r a t i o n a l progression i n the UV spectrum [186] (with the same v i b r a t i o n a l spacing as i n the f i r s t ionised state i n the photoelectron spectrum [117]) has been a t t r i b u t e d by Robin [12] to a 4a ^ -* 4p Rydberg t r a n s i t i o n . The term value (2.26 eV) i s i n between those of the 4s and 3d l e v e l s . The a^ o r b i t a l s should have some P 2s character. According to ab i n i t i o c a l c u l a t i o n s [189] the 4a x o r b i t a l has 14% P 3s character and 32% P 3p character. Thus a 4a^ •*• 4p t r a n s i t i o n might be expected. Feature 5 i s thought to have t h i s t r a n s i t i o n on top of the 3e •*• a (a^) t r a n s i t i o n . The c a l c u l a -t i o n also indicates that the 3a^ o r b i t a l has 3% P 2s character. Apply-ing the term value of 2.26 eV to the 3a± i o n i s a t i o n p o t e n t i a l y i e l d s a 3a ^ •* 4p t r a n s i t i o n energy of 11.97 eV, i n excellent agreement with feature 12 (11.96 eV). The rest of the outer valence o r b i t a l s should have no s character (as also indicated i n the c a l c u l a t i o n [189] and - 238 -hence l i t t l e Intensity to the p Rydberg l e v e l s . Phosphorus T r i f l u o r i d e (PF,) The VSEELS spectrum of PF 3 i s shown i n F i g . 7.3, and the data i s summarised i n Table 7.5. As i n the case of PC1 3, the UV absorption spectrum (up to ~10 eV) has been reported [184,186]. In the UV spectrum [186] feature 3 shows v i b r a t i o n a l structure and, l i k e PC1 3, has a s i m i l a r spacing to that of the f i r s t ionised state [192]. The spectrum i s quite d i f f e r e n t to that of PC1 3 with two intense bands (features 1 + 2 and feature 3) followed by overlapping broad features. There are no sharp features that could be obviously assigned to Rydberg t r a n s i t i o n s , i n contrast to the s i t u a t i o n i n PC1 3. In t h i s regard t h i s spectrum i s s i m i l a r to the i n n e r - s h e l l spectra of PF 3, which are t y p i c a l of molecu-les with highly electronegative ligands, i n that strong t r a n s i t i o n s to the valence l e v e l s are observed at the expense of Rydberg t r a n s i t i o n s . The term values for features 1 and 2 from the 4a 1 (HOMO) o r b i t a l are 4.38 eV and 4.12 eV r e s p e c t i v e l y . Robin [12] has assigned these In * turn as due to the t r a n s i t i o n s 4a ^ •*• 4s and Ua-^ -*• a (P-F). However, the P 2p ISEELS spectrum indicates that the 4s term value i s 3.13 eV and therefore the 4s term value i n the valence s h e l l spectrum would be expected to be smaller. Therefore, features 1 and 2 have been assigned i n the present work to a t r a n s i t i o n s . The ISEELS assignment gave the a (e) l e v e l a term value of 6.8 eV and that for the a (a^) of ~3.5 eV. Assuming that t h i s assignment i s correct, i t i s u n l i k e l y that the * narrowly separated features 1 and 2 are due to the 4a^ -•• o (e) and 4a^ •>• - 239 -1 1 — C O L U I -Z I L U r > 1 LU t r e a,/4s -1— 5s 3e P F 3 Valence e o,/4s 5s 4e I T 1a2 e ° . / 4 s 5s —r— 5s a,/4s e o,/4s l — 5s ;2e e a, /4s —I— 5s 3a, i :2a, | i I l I | i 1 i I | i I I I | 10 15 2 0 2 5 ENERGY LOSS(eV) Figure 7.3: The valence s h e l l electron energy loss spectrum of PF 3. The i o n i s a t i o n edges are taken from photoelectron spectroscopy (see Table 7.2). Positions of t r a n s i t i o n s from the 4a ^ o r b i t a l have been applied to the other IP's. The spectrum i s summarised i n Table 7.5. - 240 -TABLE 7.5 Energ ie s , term values and poss ib le assignments for the VSEELS spectrum of P F 3 Feature Energy Loss (eV) Term Value ( e V ) ( a ) Poss ib le Assignments 1 7.91 4.38 * 4aj •*• a (e) 2 8.17 4.12 3 9.49 2.79 * 4a^ -»• a ( a ^ M s 4 10.89 5 6 11.20 11.64 4.69 4.68 l a , -»• a (e) 4e -> a (e) 7 13.18 4.25 * 3e •»• o (e) (13.52)* (2.79) 4e •*• a (a^ )/4s 8 14.21 4.36 3a. a*(e) 1 * 3e + a (a^)/4s (14 .56) f (2.79) 9 14.93 4.43 2e •+ o*(e) 10 11 15.77 16.88 2.80 2.48 3a^ •* a ( a ^ M s 2e •> o* (a 1 ) /4s (?) 12 18.09 4.51 2al + a*(e) ^ 'Term values are c a l c u l a t e d with respect to the IP of the i n i t i a l o r b i t a l . These IP ' s are given i n Table 7.2. ^ P o s i t i o n estimated using term value from feature 3. - 241 -a ( a j ) t r a n s i t i o n s . Therefore, features 1 and 2 are assigned to Jahn-T e l l e r components of the 4a^ •+• a (e) t r a n s i t i o n which would lead to a iE degenerate f i n a l state. This conclusion concurs with that given previously by McAdams et a l . [186] for the valence-shell o p t i c a l absorption spectrum of PF 3. The difference of ~2.5 eV between the ISEELS and VSEELS term values for the a (e) LUMO o r b i t a l i s i n keeping with that found for NF 3 (see Chapter 3) and PC1 3 (compare Tables 7.1 and 7.3). Feature 3 i s then assigned as the 4a^ •* a (a^) (or very l i k e l y * the 4aj^ •+• a ( a ^ M s ) t r a n s i t i o n . Robin [12] has assigned t h i s feature (3) i n the UV spectrum to the 4a^ •+• 4p Rydberg t r a n s i t i o n . The term value of t h i s feature i s 2.79 eV which, as expected, Is s l i g h t l y l e s s than that for the ISEELS 4s (3.13 eV) or a ( a x ) (~3.5 eV) t r a n s i t i o n s . Both an ab i n i t i o c a l c u l a t i o n [187] and an X c a l c u l a t i o n [176] ind i c a t e a roughly equal P 3s and P 3p character to the 4a 1 o r b i t a l . This would imply that t r a n s i t i o n s to both 4s and 4p Rydberg l e v e l s might be seen, but the spectrum does not seem to r e f l e c t t h i s s i t u a t i o n . C l e a r l y more t h e o r e t i c a l work i s needed i n order to c l a r i f y t h i s s i t u a t i o n . The rest of the valence-valence assignments, based upon the term values of features ( 1 + 2 ) and 3 with respect to the 4a ^ l i m i t are shown i n F i g . 7.3. On t h i s basis the majority of the remaining features i n the spectrum of PF 3 can be reasonably assigned to these valence-valence t r a n s i t i o n s (Table 7.5). As stated above, there seems to be very l i t t l e Rydberg structure. Feature 4 has been ascribed to the 4a^ •* 3d t r a n s i t i o n by Robin [12]. This feature has a term value of 1.40 eV with respect to the 4a i l i m i t , which i s reasonable for such an assignment (a - 242 -quantum defect of zero would predict a term value of 1.5 eV). Phosphoryl Chloride (OPCI,) The VSEELS spectrum of 0 P C 1 3 ( F i g . 7.4) i s si m i l a r to that for P C I 3 i n that i t shows features that can be ascribed to both valence-valence and valence-Rydberg t r a n s i t i o n s . The only previously reported spectrum i s that by Halmann [185], which extends as far as ~6.7 eV. This spectrum [185] shows a weak plateau at ~6.5 eV that was ascribed to a forbidden n •*• i t t r a n s i t i o n . There i s evidence of some o very weak structure i n t h i s region i n the VSEELS spectrum. The spectrum has been assigned i n a si m i l a r manner to those f o r the preceding molecules. Features 1 and 2 are considered to a r i s e from valence-valence t r a n s i t i o n s since i t s respective term values (4.36 and 3.74 eV respectively from the H e o r b i t a l ) are larger than that ascribed to the 4s Rydberg l e v e l (3.08 eV) i n the ISEELS spectrum. Thus these features (1 and 2) are assigned to the H e •* a (a^) and H e •*> a (e) tr a n s i t i o n s r e spectively. The ordering i s that of the P 2p spectrum (term values of 7 .0 eV and 6.2 eV) and that of the 0 Is spectrum (term values of 5.5 eV and 4.2 eV). The reversal of i n t e n s i t y i n contrast to that observed i n the 0 Is ISEELS spectrum i s consistent with the o r i g i n a t i n g o r b i t a l i n the VSEELS spectrum, being of 0 2p character. The term values with respect to the H e l i m i t have then been applied to the other IP's (Table 7.2) and the estimated positions of the corresponding t r a n s i t i o n s are indicated i n F i g . 7.4. I t should be noted - 243 -4s I i r 3d 4s •9e 3d i 12a, OPCI3 Valence 4p L o, e v UL \ I./ 4s I 3 i r _ T _fl3a, \ 3d I r |10e 4 3 1 3 d i — r 3d I r 3 Be 4s _1_ 4s _1_ *11e 3d 1 2 3 4 5 6 7 8910 11—1415 1617 18 19 20 21 "i r n r H 1 r 1 0 15 E N E R G Y L O S S (eV) "T 1 1 r 2 0 Figure 7.4: The valence s h e l l electron energy loss spectrum of 0PC1 3. The i o n i s a t i o n edges are taken from photoelectron spectroscopy (see Table 7.2). Positions of the valence-valence ( t a l l bars), valence-Rydberg (short bars) t r a n s i t i o n s as estimated from the term values are shown leading to each l i m i t (see Table 7.6). - 244 -that feature 2 cannot be assigned to a t r a n s i t i o n to the LUMO o r b i t a l (a (a^)) from the 2a 2 l e v e l , as t h i s t r a n s i t i o n i s dipole forbidden under C^ v s e l e c t i o n rules. Thus the i n t e n s i t y leading up to feature 5 can be reasonably ascribed to the various valence-valence t r a n s i t i o n s indicated on F i g . 7.4, since the correspondence of the spectral features with the predicted positions i s good. The molecule 0PC1 3 has a second a (a^) v i r t u a l o r b i t a l (see Table 7.1) that so far has not been con s i -* dered. From a consideration of t h i s a (a^) ISEELS term value and that of the 4s ISEELS Rydberg l e v e l , i t i s l i k e l y to mix with the Rydberg l e v e l to give a a (a^)/4s conjugate [65] i n the valence spectrum. Feature 5, the f i r s t feature not f i t t e d by this proposed valence-valence scheme, has a term value of 1.48 eV with respect to the l i e o r b i t a l and can be assigned to the l i e •+• 3d Rydberg t r a n s i t i o n . There i s also evidence of a shoulder on the low energy side of feature 5. This has a term value ~2.6 eV ( s i m i l a r to the 4s term value i n the C l 2p spectrum) from the lOe o r b i t a l , and so i s assigned to the lOe •*• * 4s/a (a^) t r a n s i t i o n . These term values have been applied to the other IP's and the estimated positions of the t r a n s i t i o n s are Indicated on the spectrum. Transitions to the p Rydberg series might also be expeted. The a b - i n i t i o c a l c u l a t i o n [189] indicates that the 12a^ o r b i t a l has a 10% 0 2s component. Thus a 12a^ •*• 4p t r a n s i t i o n might be expected to have some i n t e n s i t y . Feature 15 has a term value of 2.21 eV with respect to the 12a x i o n i s a t i o n p o t e n t i a l . Since t h i s term value i s close to that given for the 4p Rydberg l e v e l i n PC1 3 (2.26 eV). Feature 15 i s assigned accordingly. As with PC1 3, t h i s term value i s applied - 245 -also to the other a^ symmetry o r b i t a l . Obviously the spectrum i s made up of many overlapping t r a n s i -t i o n s , and no c l e a r , unambiguous assignment can be made. The s p e c t r a l data show evidence of s i g n i f i c a n t contributions from valence-valence t r a n s i t i o n s . Table 7.6 summarises the positions of the features. Phosphorus Pentafluoride (PFc;) F i n a l l y , the spectrum of PF 5 w i l l be considered. The spectrum i s shown i n F i g . 7.5. To date there has been no UV absorption or VSEELS spectrum reported for t h i s molecule. The molecule PF 5 i s of symmetry. From a consideration of a minimal basis set, the v i r t u a l * * * o r b i t a l s are a ( a j ) , a (e')» and a (a£). However, only the f i r s t two l e v e l s w i l l be considered since the a (a£) l e v e l i s located right at the i o n i s a t i o n edge i n the ISEELS spectra. The experimental i o n i s a t i o n potentials are taken (Table 7.2) from photoelectron spectroscopy [193,194]. However, there i s a considerable lack of agreement between the various c a l c u l a t i o n s as to the o r b i t a l ordering. Some of these assignments have been summarised by Goodman et a l * [194]. The ordering used i n the present work (see Table 7.2) i s that of S t r i c h and V e i l l a r d [191], who performed an a b - i n i t i o LCAO MO SCF c a l c u l a t i o n with a medium size basis set. Table 7.7 summarises the dipole allowed/forbidden t r a n s i t i o n s for the outer valence region of t h i s molecule. From the ISEELS spectrum, the term value of the 4s Rydberg l e v e l i s 2.75 eV. Since feature 1 has a term value of 3.15 eV with respect to - 246 -TABLE 7.6 Energies and term values for the VSEELS spectrum of OPCI3 Feature Energy Loss (eV) Feature Energy Loss (eV) 1 7.57t 12 12.53 2 8.19f 13 12.66 3 9.35 14 12.83 4 10.18 15 13.14f 5 10.45t 16 13.59 6 10.73 17 13.94 7 11.11 18 14.54 8 11.37 19 15.16 9 11.61 20 16.03 10 11.87 21 17.00 11 12.34 Features used to estimate term values - see text. Term values: a!!( al) = A.36 eV o (e) = 3.74 eV a (a,)/4s = 2.6 eV 3d = 1.48 eV 4p = 2.21 eV These have been applied to the IP's listed in Table 7.2. - 247 -2 3 4 5 6 7 8 9 1 0 PF 5 Valence • a (e') and 6e' •+ a (aj) t r a n s i t i o n s . Thus the term values for the o ( e 1 ) and a (aj) are ~3.15 eV and 4.17 eV respec-t i v e l y . This compares with 3.6 eV and 6.2 eV for the P 2p spectrum and 2.7 eV and 5.0 eV for the F Is spectrum shown i n Chapter 6. The reduct-ion i n term value (~2 eV) for the LUMO o r b i t a l i n going from ISEELS to VSEELS i s i n keeping with those observed for other molecules with highly electronegative ligands (eg., NF 3, PF 3, PC1 3, and 0PC1 3). Thus the assignment of the 2e" to the HOMO o r b i t a l i s c l e a r l y supported by the VSEELS spectrum. The term values as obtained above for the 2e" and 6e l i m i t s are now applied to the other IP's. The positions of a l l the t r a n s i t i o n s so derived are indicated i n F i g . 7.5, and the assignments summarised i n Table 7.8. A dashed v e r t i c a l l i n e on the spectrum indicates the p o s i -tions of t r a n s i t i o n s that are dipole forbidden. The agreement between the spectra and the predicted valence-valence t r a n s i t i o n s i s very good i n almost a l l cases. A better agreement would be obtained i f the order (Table 7.2) of the 4a*2 and 4e' l e v e l s i s reversed. The c a l c u l a t i o n [191] indicates that the separation of these le v e l s i s only 0.16 eV, and therefore not r e l i a b l e for pred i c t i n g the ordering. The revers a l of thi s ordering i s supported by the work of Cox et a l . [193], who i n the analysis of the He(I) spectrum assigned the 21 eV feature to the 4a'I - 250 -TABLE 7.8 Energies and poss ib le assignments for the VSEELS spectrum of P F 5 Feature Energy Loss (eV) (a) Poss ib le Assignments * 1 12.29 2e" a ( e ' ) , * 6e' •+ a (aj), 5a£ * a (aj) 2 13.96 6e' + a ( e * ) , - o * (e ' ) 3 14.35 5e' ->• a*(ap l a ' ->• o* (e ' ) * , 4 15.13 5e' ->• o ( e ' ) * t 5 15.43 4 a2 a (a ! ) * , 6 15.97 l e " a ( e ' ) 7 16.51 t * t 8 16.86 4e' a (a ) 9 18.65 from 7aJ (?) 10 19.4 Exchanging 4a? and 4e' order 5 4 e ' + a (a[) 7 4e' * a * (e ' ) 8 4 a £ + o * ( a p (^Ass ignments based on applying fo l lowing term values to IP ' s i n Table 7.2. a * (e ' ) 3.15 eV Feature 1 from 2e" o*(a{) 4.17 eV Feature 1 from 6e' - 251 -l e v e l . The VSEELS spectrum ( F i g . 7.5) can be adequately described by just considering predominantly valence-valence t r a n s i t i o n s . This i s s i m i l a r to the s i t u a t i o n discussed above for PF 3(and also NF 3). SUMMARY AND CONCLUSIONS The VSEELS spectra of several phosphorus-containing compounds have been presented i n t h i s chapter and compared with t h e i r ISEELS spectra. The ISEELS spectra c l e a r l y indicate the dominant presence of t r a n s i t i o n s to a v i r t u a l valence l e v e l s , s p e c i a l l y those with e l e c t r o -negative ligands. Transitions to valence l e v e l s also seem to dominate i n the valence s h e l l spectra presented here. The term values, as expected, are lower than for the corresponding ISEELS t r a n s i t i o n (approximately 2-2.6 eV lower for those with highly electronegative ligands, e.g. F ) . The valence-valence nature of these t r a n s i t i o n s i s evidenced by the term values being even larger than those for the ISEELS 4s Rydberg t r a n s i t i o n s . The higher term values for the ISEELS spectra are explained by the e f f e c t of the l o c a l i s e d , central nature of the phosphorus core hole. The ligand ISEELS spectra, where the core hole i s situated on the periphery of the molecule, generally have term values that l i e i n between those of the central atom ISEELS spectra and the VSEELS spectra. Term values for the a l e v e l s a r i s i n g from the three regions of the molecules are summarised i n Table 7.9. Also shown are the term values for the 4s Rydberg l e v e l . In cases where these are only s l i g h t l y lower than those for a a o r b i t a l of the same symmetry i n the P 2p ISEELS spectra, the corresponding feature i n the VSEELS spectra cannot be s o l e l y assigned to one or the other. In these cases the - 252 -TABLE 7.9 Term values from ISEELS and VSEELS for P H 3 , P ( C H 3 ) 3 , P C 1 3 , P F 3 , P F 5 , and 0PC1 3 Molecule O r b i t a l ISEELS Term Values ( e V ) ( a ) VSEELS Term Values (eV) Phosphorus (2p) L i g a n d ( b ) PH 3 a * ( e ) a ( a , ) 4s 5.1 4.5 2.48 } 3.76 P ( C H 3 ) 3 o*(e) a ( a , ) 4s 3.3 ~2 2.86 3.17 j. 2.44 3.01 P C I 3 o* (a , ) o (e) 4s 6.9 6.4 3.06 5.7 2.6 4.83 4.54 2.95 P F 3 o*(e) 0" ( a , ) 4s 6.8 -3 .5 3.12 5.2 3.0 -4.4 ( 4 . 6 8 ) ( d ) J. 2.79 P F 5 o (e T ) 4s 6.2 3.6 2.71 5.0 2.7 4.17 3.15 C l 0 O P C I 3 a* (a , ) a*(e5 a ( a , ) 4s 7.0 6.2 4.4 3.08 6.0 4.5 3.2 2.6 5.5 4.2 4.36 3.74 J. 2.6 ^ ' F r o m Chapters 5 and 6. ' ^ F Is , 0 Is , C Is, and C l 2p as appropiate . ( c ) R e f . [12]. ^ W i t h reference to F l o n e - p a i r o r b i t a l . - 253 -feature i s probably best ascribed as a mixed valence-Rydberg ( a /4s) l e v e l [65]. I t should be noted that the difference i n term values i s much smaller for the Rydberg l e v e l s than for the lowest a l e v e l s on going from the ISEELS to the VSEELS spectra. Levels assignable to the 3d Rydberg l e v e l i n the VSEELS spectra were seen to have a constant term value (~1.4-1.5 eV). The quantum defect obtained (~0) i s consistent with the assignment to d l e v e l s . I t can be seen that, e s p e c i a l l y i n the case of PF 3 and PF 5, the VSEELS spectra are dominated by valence-valence t r a n s i t i o n s with very l i t t l e evidence of any Rydberg s e r i e s . The VSEELS spectra of PF 5 support the assignment of the HOMO o r b i t a l of P F 5 as being the 2e" o r b i t a l [191,194]. S i m i l a r l y 0PC1 3 and PC1 3 show many tr a n s i t i o n s assignable to valence-valence t r a n s i t i o n s ; however, valence-Rydberg t r a n s i t i o n s are also apparent. The spectrum of P ( C H 3 ) 3 ( F i g . 7.1) has a very intense f i r s t feature, and i n this respect i t i s quite s i m i l a r to that observed for PH 3 [12]. The i n t e n s i t y of the f i r s t feature i n P(CH 3) 3 c l e a r l y indicates a contribution from valence-valence t r a n s i -t i o n ( s ) . I t presumably contains t r a n s i t i o n s to the a (e) and o (a^)/4s l e v e l s . The term values for higher energy features, a r i s i n g from o r b i t a l s located on the ligand, are each si m i l a r to that for the f i r s t t r a n s i t i o n i n the C Is spectrum (C Is •*• 4s t r a n s i t i o n ) and are an i n d i c a t i o n of a t r a n s i t i o n to a " l o c a l i s e d " Rydberg o r b i t a l , as was suggested for S i ( C H 3 ) t t and i n the methylamines (see Chapter 8). In the present work i t can be seen that the ISEELS spectra can be used as an aid to the assignment of the more complex VSEELS spectra. - 254 -Spectra of condensed phases that r e s u l t i n the suppression of Rydberg t r a n s i t i o n s would be h e l p f u l i n c l a r i f y i n g the valence-valence and/or valence-Rydberg nature of the spectra. In t h i s regard the e f f e c t of the highly electronegative (F) ligand on the spectra seemingly p a r a l l e l s t h i s e f f e c t i n that the valence-valence t r a n s i t i o n s are trapped (and therefore enhanced) by a charge b a r r i e r e f f e c t of the type more usually involved i n the case of i n n e r - s h e l l spectra. - 255 -CHAPTER 8 INNER SHELL ELECTRON ENERGY LOSS SPECTRA OF THE METHYL AMINES AND AMMONIA In Chapters 5 and 6 the ISEELS spectra of several phosphorus compounds were presented and the e f f e c t s of the ligand on the r e l a t i v e s p ectral i n t e n s i t i e s contrasted. A s i m i l a r comparison was made between the ISEELS spectra of SiCCHg)^ presented i n Chapter 4 and e a r l i e r reported photoabsorption spectra of r e l a t e d s i l i c o n compounds. The ligand was seen to play an important role i n the observed i n t e n s i t y d i s t r i b u t i o n s . In contrast to ligands such as H and CH 3, highly e l e c -tronegative ligands, for example F, enhance the p r o b a b i l i t y of t r a n s i -tions to v i r t u a l valence l e v e l s at the expense of those to Rydberg l e v e l s . This was also seen i n the ISEELS spectra of NF 3 presented i n Chapter 3. In t h i s chapter, as a continuation of t h i s work, the ISEELS spectra of (CH 3) 3N, which i s i s o e l e c t r o n i c with NF 3, i s presented as well as the spectra of the other methyl amines ((CH 3) 2NH and CH 3NH 2) and NH3. The ISEELS spectra of NF 3, (CH 3) 3N and NH3 are also compared with the t h i r d row phosphorus analogues PF 3, P ( CH 3) 3 and PH 3. To date the only previously reported inner s h e l l electron e x c i t a -t i o n spectra of these molecules have been the ISEELS spectra of NH3 and CH 3NH 2 reported by Wight and Brion [72] and a recent X-ray absorption spectrum of NH3 [196]. However, there have been several studies on the valence electron e x c i t a t i o n spectra. For example, Tannenbaum et a l . - 256 -[197] have reported the UV absorption spectra of the methyl amines up to ~8 eV. There have also been numerous studies on NH3 [198]. The lowest excited states i n the valence and i n n e r - s h e l l spectra have been assigned to t r a n s i t i o n s to l e v e l s of mainly 3s and 3p Rydberg character [72,197]. Salahub [199] has performed semi-empirical MO-CT calcu l a t i o n s on these molecules and assigned the lowest valence excitations to HOMO (N lone * p a i r ) a t r a n s i t i o n s . However, ab i n i t i o c a l c u l a t i o n s on NH3 [200, 201] and (CH 3) 3N [201] support the Rydberg assignment. The spectra presented here provide further evidence for the assignment of the lowest e l e c t r o n i c t r a n s i t i o n s to Rydberg l e v e l s . EXPERIMENTAL DETAILS The spectra were obtained on the ISEELS spectrometer described i n Chapter 2. An impact energy of 2.5 keV was used and the scattered electrons were sampled at ~1° scattering angle. The C Is spectra of the amines were c a l i b r a t e d against the S 2 p ^ ^ t2g(184.54 eV) feature of SFg. These C Is spectra were used, except i n the case of (CH 3) 3N, to i n t e r n a l l y c a l i b r a t e the N Is regions of the methyl amines. The N Is * spectra of (CH 3) 3N and NH3 were c a l i b r a t e d against the N 2 (N Is •> u (v = 1), 401.10 eV) and CO (C Is •*• n* (v = 0), 287.40 eV) features r e s p e c t i v e l y . RESULTS & DISCUSSION Carbon Is spectra The long-range spectra of the C Is region of the methyl amines - 257 -are shown i n F i g . 8.1. These spectra were obtained with a resolution of 0.36 eV FWHM. More detai l e d short-range spectra, recorded at a higher resolution (0.18 eV FWHM), are shown i n F i g . 8.2. The assigned i o n i z a -t i o n edges for (CH 3) 3N and CH 3NH 2 have been taken from XPS [180,202]. As no value for the C Is IP of (CH 3) 2NH has been reported, i t s value was assumed to be the mean of the other two IP's. Table 8.1 summarises the spectral data. The spectra are s i m i l a r to that for CH^ [66,67,72] and the analysis can be considered i n terms of that for a mono-substituted methane. The C Is spectrum of CH 3NH 2 i s v i r t u a l l y i d e n t i c a l to that for C 2H 6 [68]. Similar observations were made i n the C Is spectra of P ( CH 3) 3 and S K C H ^ . Hitchcock and Brion have discussed the C Is spectra of the methyl halides [63,64]. It i s i n s t r u c t i v e to compare the analysis there with the spectra presented here. The methyl halides a l l possess a feature a t t r i b u t a b l e to low-lying a o r b i t a l [64]. However, the rest of the spectrum can be assigned s o l e l y to Rydberg t r a n s i t i o n s . In the amine spectra reported here there i s a complete absence of any feature a t t r i -butable to a low-lying a feature. The d i s c r e t e portions of the amine C Is spectra reported here are s i m i l a r to the remainder of the methyl halide spectra and can be assigned s o l e l y to Rydberg features i n an analogous manner. The assignments based upon these arguments [64] are shown i n Table 8.1 The post-edge behaviour i s d i f f e r e n t i n these two sets ( i . e . , the amines and methyl halides) of molecules. The methyl amines show a broad - 258 -i—I—i—r 10-rr-^f-1 2 6 r i — i — i — i — i — i — 1 — 1 — 1 — 1 — 1 r | Is edge Is edge n—i—i—I—I—I—I—i—i r (CH 3) XNH 3. X C1s ( x = 1 3 ) AE=0-36eV CH3NH2 > I O H < cc H CD < b 5-C O LJJ U J > i, 9 'i i "i |Is edge ioH 3 7 UJ cr (CH,)2NH (CH3)3N — I — i — i — i — i — | — i — i — i — i — | — \ — i — i — i — | — i — i — i i i i i i i i 285 295 305 315 325 335 ENERGY LOSS (eV) 8.1: Long range electron energy loss spectra of the C Is region of the methyl amines. - 259 -10 H 5H co => 10-> < cc H m cc < >- 5-CO U J I— U J 54 1 1 l i s edge I I I M 1 1 2 3 4 5 6 / A / 1 2 3 4 5 6 7 8 i'\ nr 12 34 5 6 A (CH S),NH 3.. C1s < " , - 3 ) 1s edge 1s edge AE=0-18eV C H 3 N H 2 286 288 290 292 294 E N E R G Y L O S S (eV) (CH3)2NH (CH3)3N 296 298 F i g . 8.2: Short range, high r e s o l u t i o n electron energy loss spectra of the C Is region of the methyl amines. TABLE 8.1 Energies, Term Values and Possible Assignments for the C Is Region of the Methyl Amines C H 3 N H 2 (CH 3) 2NH (CH 3) 3N Feature Energy j(a) Feature Energy T Feature Energy T P o s s l b l e ( b ) (eV) (eV) (eV) (eV) (eV) (eV) Assignment 1 287.70 3.90 1 287.61 3.82 1 287.84 3.42 3s 2 287.85 3.58 2 288.08 3.18 3s+v 3 288.3 3.1 2 288.78+ 2.82 4 288.67+ 2.76 3 288.80+ 2.46 3p(x,y) 5 288.95 2.48 4 289.07 2.19 3p(x,y)+v 3 289.46 2.14 6 289.46 1.97 5 289.51 1.75 3p(z) 4 290.21 1.39 7 290.03 1.40 6 290.06 1.20 4p(x,y)/3d 5 290.59 1.01 8. 290.35 1.08 1. 4p(z) IP* 291.60 0 IP* 291.43 0 IP* 291.26 0 Is l i m i t 6 291.8 -0.2 9 293.2 -1.8 7 294.3 -3.0 & a shape-resonance +Calibrated feature, estimated uncertainty ± 0.08 eV for CH3NH2, (CH 3) 2NH; ± 0.15 eV for (CH 3) 3N. *XPS CH3NH2 r e f . [202], (CH 3) 3N ref. [180]; (CH 3) 2NH mean of other two values. (a) T = I P _ Energy ( b) (CH 3) 2NH and CH3NH2: C g symmetry p(x,y) = pa', p(z) = pa". - 261 -and f a i r l y intense feature j u s t beyond the i o n i z a t i o n edge ( F i g . 8.1). This i s not the case for the methyl halides, nor for that matter methane [67,72]. However, ethane [68] shows s i m i l a r behaviour to the molecules here. The difference between the C Is spectra of methane [72] and ethane [68] i s n i c e l y i l l u s t r a t e d i n a recent paper by Hitchcock et a l . [100] and c l e a r l y indicates a peak j u s t beyond the edge for ethane. The continuum features can be att r i b u t e d to a a shape resonance [77] associated with the C-C bond i n ethane [100] and with the C-N bond i n the amines. In MO terms these can be thought of as excitations into a * a anti-bonding state [87]. Thus the major difference between the methyl halides and the molecules here i s the lo c a t i o n of the cr o r b i t a l . In the former they are low l y i n g and i n the d i s c r e t e portion, whereas i n the methyl amines they are i n the continuum. The a shape resonances w i l l be discussed further below. Nitrogen Is spectra Figure 8.3 shows the long-range N Is spectra of the methyl amines. The spectral r e s o l u t i o n i s 0.36 eV FWHM. More detai l e d short-range spectra are shown i n F i g . 8.4 along with a high-resolution spectrum (0.14 eV FWHM) of NH3. The Is i o n i z a t i o n edges are taken from XPS measurements [203]. On going from NH3 to (CH 3) 3N a gradual change in spectral features can be observed. The NH3 spectrum i s c h a r a c t e r i s -t i c of a dominant (atomic-like) Rydberg spectrum with a l a r g e l y s t r u c -tureless continuum of low i n t e n s i t y . In contrast, the (CH 3) 3N spectrum shows l i t t l e Rydberg structure and i s dominated by a broad continuum - 262 -icH 5H oo 3 or < or m CO LU r -LU > I-< LU LT 5 H io H i i i i i i 1s edge 1 2 3 4 I I I I I I I I I I I I AE=0.36eV ( C H 3 ) X N H 3 . X N1s ( x = 1 ' 3 ) !!! 1s edge 1 2 4 I 1s edge 1 2 3 CH.NH; (CH 3) 2NH (CH 3) 3N 5 H — r — i — i — i — i — | — i — i — i — i — | — i — i — i — i — r - i — i — I i i 1 1 1 1 I 3 9 5 4 0 5 4 1 5 4 2 5 4 3 5 4 4 5 E N E R G Y L O S S ( e V ) F i g . 8.3: Long range electron energy loss spectra of the N Is region of the methyl amines. - 263 -I O H 5 H to 1 0 > 10 cr < cr H m cr < >-t 5 GO UJ UJ I'o-l _J LU rr 5 H 10 5 4 -i 1 1 1 1 r r-, • • • • t l § - e d g e 12 3 j 4 5 6 1s edge 1 2 A 3 4 M f i i r 1 2 3 I 1s edge | 1s edge 1 2 ( C H 3 ) x N H 3 . x N1s (x = 0-3) NH 3 AE=O.I4eV C H 3 N H 2 AE=0.36eV (CH 3 ) 2 NH AE=0.28eV (CH 3) 3N AE=0.28eV 399 ^ 1 i 1 r 403 407 411 ENERGY LOSS (eV) 415 8.4: Short range electron energy loss spectra of the N Is region of ammonia and the methyl amines. The ( d i f f e r i n g ) spectral r e s o l u t i o n i s indicated on each spectrum. - 264 -feature that can be ascribed to a a (N-C) shape resonance. As stated above, the NH3 spectrum can be assigned to t r a n s i t i o n s to the Rydberg l e v e l s and th i s spectrum has been published e a r l i e r [72, * 196]. There i s no evidence of any d i s t i n c t a feature i n the d i s c r e t e portion, though Robin [12] has suggested that the HOMO •»• 3s feature i n the valence s h e l l UV spectrum i s on top of a continuous valence s h e l l t r a n s i t i o n , and Schwarz [61] suggests an admixture of valence a n t i -bonding character for the n = 3 Rydberg l e v e l s . Rather weak features i n the NH3 Is continuum (~14 eV above the edge) have been i d e n t i f i e d with a a (N-H) shape resonance [99]. The NH3 spectrum presented here i s summarized i n Table 8.2 along with possible assignments. The re s u l t s of the e a r l i e r ISEELS work [72] and the X-ray absorption spectrum [196] are also shown. The energies of the spe c t r a l features i n the d i f f e r e n t ISEELS spectra are i n good agreement with each other and with the X-ray absorption spectrum, though i n the l a t t e r the features are uniformly ~0.1-0.2 eV higher i n energy. The assignments In the present work d i f f e r only s l i g h t l y from the e a r l i e r ISEELS work [72]. This r e s u l t s from the much higher res o l u t i o n achieved here. Features 1 and 2 are assigned to the Is -*• 3s t r a n s i t i o n plus v i b r a t i o n a l component, thereby concurring with the suggestions of Wight et a l . [72], No v i b r a t i o n a l component i s discernable at the resol u t i o n employed i n the X-ray work [196]. Features 3 and 4 are assigned to the 3p(e) and 3p(a 1) Rydberg l e v e l s r e s p e c t i v e l y , with a separation of ~0.6 eV, which agrees well with separation observed (< 0.6 eV) i n the valence electron e x c i t a t i o n spectrum [204], No feature corresponding to feature 4 was observed Table 8.2 Energies , Term Values and Assignment f o r the N Is Energy Loss Spectra of NH^ ISEELS X-ray Absorpt ion Present Work Wight et a l . + + Akimov et a l . Feature Energy (eV) ± 0.08 eV Term Value (eV) Assignment Energy (eV) Assignment Energy (eV) Assignment 1 400.61 4.91 3s 400.6 3s 400.8 3s 2 400.92 4.60 3s + v — 3 402.29 3.23 3 P (e) 402.2 3p(e) 402.4 3p(e) 4 402.85 2.67 3 p ( 3 l ) — -403.0 3p(aj) 5 403.52 2.00 4s 403.5 3 p ( 3 l ) 403.6 4s 6 404.14 1.38 4p/3d 404.1 4s/3d/4p(e) 404.1 4p(e) IP* 405.52 0 + r e f . [72] * r e f . [196]. *XPS r e f . [203] - 266 -in the e a r l i e r ISEELS work due to poorer resolution [72], however, i t i s seen i n the X-ray spectrum [196]. On th i s basis feature 5 i s now re-assigned to the Is •*• 4s t r a n s i t i o n consistent with the i n t e r p r e t a t i o n given by Schwarz [61] and also by Akimov et a l . [196]. In comparing the ISEELS spectrum with the valence s h e l l spectrum of the three amines and ammonia, two points can be noted. F i r s t l y , there occurs a reversal i n the i n t e n s i t y of the features associated with the 3s and 3p Rydberg l e v e l s . This r e f l e c t s the s - l e v e l c h a r a c t e r i s t i c s of the o r i g i n a t i n g o r b i t a l i n the core spectra which favour an s •*• p t r a n s i t i o n (dipole allowed i n the pure atomic case) over the s •> s t r a n -s i t i o n (dipole forbidden i n the pure atomic case), e x p e c i a l l y i n a mole-cule as symmetrical as NH 3 [61]. The second point concerns the term values of the features. The term values are between 0.4 and 0.5 eV larger for the 3s and 3p features i n the ISEELS spectra than for the corresponding features i n the valence s h e l l spectra [198]. This arises from the l o c a l i z e d nature of the core-hole, and thus the newly promoted electron sees a centre approximating a (Z + 1) core. A c t u a l l y , t h i s difference i s quite small and i t r e f l e c t s the Rydberg nature of the f i n a l o r b i t a l In both spectra. Rydberg o r b i t a l s , being large and d i f f u s e , are less s e n s i t i v e to the loc a t i o n of the hole. In NF 3, where the lowest unoccupied o r b i t a l i s a a antibonding o r b i t a l , the d i f f e r e n -ce i n term values between core and valence spectra i s 2.17 eV. Table 8.3 summarizes the spectral features of the N Is region of the methyl amines. Of these only the spectrum of CH 3NH 2 has been previously reported [72]. As before, the f i r s t features of CH 3NH 2 are TABLE 8.3 Energies, Term Values and Possible Assignments for the N Is Region of the Methyl Amines CH3NH2 (CH 3) 2NH (CH 3) 3N Feature Energy T ( a ) Feature Energy T Feature Energy T Possible (eV) (eV) (eV) (eV) (eV) (eV) Assignment 1 400.78 4.39 1 401.04 3.89 1 (401.8) (3.0) 3s 2* 402.03 3.14 2 402.30 2.63 2 403.0 1.8 3p 3 403.55 1.62 3 403.24 1.69 - — — 3d etc. 4 404.8 0.3 4 405.8 -0.9 3 406.5 -1.7 a* shape-resonance IP 405.17 IP 404.93 IP 404.82 C a l i b r a t e d feature, estimated uncertainty ± 0.12 eV for CH3NH2, (CH 3) 2NH; ± 0.2 eV for (CH 3) 3N. *XPS ref. [203] ( a ) T = I P _ Energy - 268 -assigned as t r a n s i t i o n s to the 3s and 3p l e v e l s r e s p e c t i v e l y . The term values are lower than those i n NH3 which would be expected upon rep l a c -ing an H ligand by an a l k y l group [12] and once again the term values are ~0.5 eV higher than for the corresponding valence s h e l l spectra [198]. A further difference between the spectra of CH3NH2 and NH 3 i s that the r e l a t i v e i n t e n s i t y of the N Is •+• 3p t r a n s i t i o n to that of the N Is + 3s t r a n s i t i o n i s less i n CH 3NH 2. However, the former t r a n s i t i o n to the 3p i s s t i l l more intense. The major difference between CH 3NH 2 and NH3 i s the appearance of a broad feature at the edge. This previously unassigned feature [72] p a r a l l e l s that observed i n the C Is spectrum and can be assigned to a a (N-C) shape resonance [99]. The spectrum shows no evidence of any low-lying a o r b i t a l below the i o n i z a -t i o n edge. The spectrum of (CH 3) 2NH (Figs. 8.3 and 8.4, Table 8.3) also shows t r a n s i t i o n s that can be assigned to the Rydberg l e v e l s . However, the i n t e n s i t y of the N Is •> 3s t r a n s i t i o n Is now larger than that of the N Is •*• 3p t r a n s i t i o n . However, the r e l a t i v e reduction i n the 3p i n t e n -s i t y i s consistent with that observed on going from NH3 to CH 3NH 2. The f i r s t feature i n the spectrum i s at 401.04 eV, which i s very close to the N Is -*• 11 feature of N 2 (401.10 eV). The p o s s i b i l i t y of t h i s feature a r i s i n g from an N 2 impurity, however, can be discounted since the valence spectrum was run and no trace was found of the very intense N 0 (X b 1 n ) feature at 12.93 eV [103]. The term values for the 3s and z u 3p l e v e l s are about ~0.3 eV larger than i n the valence spectra [198]. As with CH 3NH 2, the spectrum has a broad, intense feature that i s - 269 -assigned to a a (N-C) resonance. In t h i s case the feature i s centered just above the edge. Once again the spectrum shows no evidence for any low-lying a feature below the edge. The spectra of N(CH 3) 3 (Figs. 8.3 and 8.4, Table 8.3) shows very l i t t l e Rydberg character. There i s an i n d i c a t i o n of a weak feature, 1, with a term value of 3.0 eV, which i s assigned to a t r a n s i t i o n to the 3s Rydberg l e v e l . The term value compares with one of 3.03 eV from the valence s h e l l spectra. The broad feature 2 presumably encompasses t r a n s i t i o n s to the 3p Rydberg l e v e l s and higher. The lack of i n t e n s i t y for the 3s t r a n s i t i o n s consistent with the arguments made above i n the case of NH3. The spectrum i s t o t a l l y dominated by a a (N-C) shape resonance, centred j u s t above the edge. GENERAL DISCUSSION The spectra presented here a l l support the contention that there are no low-lying a v i r t u a l o r b i t a l s , and that the lowest l e v e l s are Rydberg i n nature. This agrees with the assignment of the valence s h e l l spectra (below the f i r s t IP) as t r a n s i t i o n s to l e v e l s of predominantly Rydberg character [12,200,201]. In the ISEELS spectra the a l e v e l s have been found to be at or only j u s t above the i o n i s a t i o n edges i n the amines (Tables 8.1 and 8.3) and seemingly absent i n NH3. For the valence spectra t r a n s i t i o n s to the a l e v e l s would be higher i n the continuum than for the core spectra since there would be no core hole to cause a large r e l a x a t i o n . The C Is spectra are very s i m i l a r and show an e s s e n t i a l l y constant 3s/3p i n t e n s i t y r a t i o . On the other hand, the - 270 -Rydberg c h a r a c t e r i s t i c s of the N Is spectra change dramatically and c e r t a i n l y do not p a r a l l e l the C Is spectra. Indeed, the C Is spectra could be s o l e l y interpreted i n terms of a l o c a l i z e d Rydberg structure based upon a CH3X model compound. This phenomenon has also been noted i n SKCHg)^ and P(CH 3) 3. A broad feature at or j u s t above the edge i s present i n a l l of * the methyl amine C Is and N Is spectra and i s associated with the a antibonding o r b i t a l formed by the C-N bond. In a multiple scattering picture i t i s also termed as a a shape-resonance [77], There has been much discussion recently (see Chapter 1, section F.III) on the r e l a t i o n -ship between bond length and the shape-resonance p o s i t i o n (E ) from the K edge (6 = E -IP). Hitchcock et a l . [100] have demonstrated that a K l i n e a r r e l a t i o n s h i p i s s u f f i c i e n t to describe the C-C bond length v a r i a -t i o n and shape-resonance p o s i t i o n i n a series of hydrocarbons. Sette et a l . [99] have made a systematic study on a larger v a r i e t y of systems. A study on some phosphorus compounds i n Chapter 6 also indicated that there i s a r e l a t i o n s h i p between bond length/type and resonance p o s i t i o n . Table 8.4 l i s t s the C-N bond lengths and resonance positions from the edge (6). I t i s seen that as the bond length decreases the r e l a t i v e p o s i t i o n of the resonance with respect to the edge moves to higher energy, as has been noted before [99,100]. The lack of data points and the error l i m i t s , however, does not allow the exact r e l a t i o n s h i p to be examined. F i n a l l y , i t i s of value to compare the spectra of NF 3, (CH 3) 3N and NH3 with t h e i r t h i r d row analogues PF 3, P(CH 3) 3 and PH 3. In - 271 -TABLE 8.4 Resonance Energy Positions (6) From Edge and C-N Bond Lengths (R) for the Methyl Amines Molecule R ( A ) ( a ) 6 ( e V ) ( b ) C Is N Is (CH 3) 3N 1.451 (3) 3.0 1.7 (CH 3) 2NH 1.462 (7) 1.8 0.9 (CH 3)NH 2 1.4714 (20) 0.2 -0.3 'From microwave spectroscopy. Landholt - BBrnstein (New Series) II/7 "Structure Data of Free Polyatomic Molecules," Springer-Verlag, B e r l i n (1976) (k)s = Resonance Energy - IP = -Term Value - 272 -previous chapters the e f f e c t s of the ligand on the P 2p spectra of some phosphorus compounds were compared to the e f f e c t s on the Si 2p spectra of s i l i c o n compounds. The ligand i d e n t i t y had a s i m i l a r e f f e c t i n both s e r i e s . B r i e f l y , the more highly electronegative ligands (e.g., F) enhance the p r o b a b i l i t y of t r a n s i t i o n s to the a v i r t u a l l e v e l s at the expense of t r a n s i t i o n s to the higher-lying Rydberg l e v e l s . The term values of the lowest a o r b i t a l s were much larger i n these compounds than those of compounds with less electronegative ligands such as H. The a l e v e l s i n the t h i r d row hydrides s t i l l precede the Rydberg l e v e l s . However, the spectrum i s much less dominated by the a l e v e l s and r e l a t i v e l y strong Rydberg t r a n s i t i o n s are observed. The e f f e c t of the electron donating CH 3 ligand leads to possible overlapping or mixed Rydberg-valence t r a n s i t i o n s . I t i s of i n t e r e s t to note, however, the complete absence of a 2p •*• 4s Rydberg t r a n s i t i o n i n the S i 2p spectrum of 81 (013 )^ , implying that t h i s spectrum consists e s s e n t i a l l y of t r a n s i -tions to valence o r b i t a l s . By contrast the C Is + 4s t r a n s i t i o n i s c l e a r l y seen i n the C Is ligand spectrum of Si(CH 3)^. Features i n the continuum spectra of S^CHg)^ and phosphorus compounds are also present * and can be i d e n t i f i e d i n part with d - l i k e a shape resonances. On comparing the ISEELS spectra of the phosphorus compounds with the spectra of the nitrogen compounds i t should be remembered that i n the former the 2p s p e c t r a l region i s being considered whereas the i s region i s being considered for the l a t t e r . However, important d i f f e r e n -ces (or s i m i l a r i t i e s ) can s t i l l be noted. The e f f e c t s of the ligands would appear to be very s i m i l a r i n both s e r i e s . The major differ e n c e - 273 -* a r i s e s with the p o s i t i o n of the a antibonding o r b i t a l s which are high l y i n g and i n the continuum for the N compounds except when i t i s asso-ciated with an electronegative ligand. Thus the N Is spectrum of NF 3 * has a very intense, low-lying N Is •+ o* t r a n s i t i o n and only weak t r a n s i -tions to the Rydberg l e v e l s . The NH3 spectrum can be mostly ascribed to Rydberg t r a n s i t i o n s with the high-lying and weak a (N-H) shape resonance * approximately 14 eV into continuum [99], whereas i n PH 3 the a t r a n s i -tions precede the r e l a t i v e l y intense Rydberg t r a n s i t i o n s . The N Is * spectrum of (CH 3) 3N i s dominated by a N Is + fl shape resonance occurring j u s t beyond the edge and shows l i t t l e Rydberg character. As i n S i ( C H 3 ) 4 and P(CH 3) 3, the ligand spectrum of (CH 3) 3N shows the t y p i c a l C Is •*• Rydberg t r a n s i t i o n s . Thus the N spectrum lends further support for the assignments previously given for P(CH 3) 3 and SiCCH^^, which indicate that the lowest Rydberg l e v e l s belong very much to the methyl group and have much C character. The assignment of the (CH 3) 3N valence s h e l l spectrum to Rydberg t r a n s i t i o n s [12,200,201] i s consistent with the above discussion since the a l e v e l i s i n the continuum (and at a higher energy since there would be no core hole). I t i s i n t e r e s t i n g to contrast the lack of i n t e n s i t y of t r a n s i t i o n s to Rydberg o r b i t a l s i n the N Is spectra of NF 3 and (CH 3) 3N. In the former molecule, NF 3, t h i s can be explained by the formation of a po t e n t i a l b a r r i e r r e s u l t i n g i n enhanced (inner-well) v i r t u a l valence occupied states, and depleted (outer-well) Rydberg states [73,77,92,93]. However, such a b a r r i e r should not e x i s t i n (CH 3) 3N, where the ligands are i n fact s l i g h t l y electron donating i n contrast to the highly electronegative F ligands i n - 274 -NF 3. Thus the lack of Rydberg t r a n s i t i o n i n (CH 3) 3N arises not from the formation of a b a r r i e r , but rather from the " l o c a l i z e d " nature of the lowest Rydberg members, which appear to be highly concentrated around the methyl groups. This i s consistent with the progressive reduction i n Rydberg i n t e n s i t y f or the N Is spectra progressing along the s e r i e s : NH3 -»• CH 3NH 2 •+• (CH 3) 2NH -*• (CH 3) 3N. No features a t t r i b u t a b l e to d - l i k e shape-resonances were noted i n the N spectra, and t h i s i s i n keeping with the fact that N i s a row 2 ( L i •*• Ne) atom. - 275 -CHAPTER 9 HIGH RESOLUTION C Is AND VALENCE SHELL ELECTRONIC EXCITATION SPECTRA OF ALLENE AND TRANS-1,3-BUTADIENE STUDIED BY ELECTRON ENERGY LOSS SPECTROSCOPY In the previous chapters, the e f f e c t s of electronegative ligands on the i n t e n s i t y d i s t r i b u t i o n s observed i n inner s h e l l electron e x c i t a t i o n spectra have been examined. The i n t e n s i t y d i s t r i b u t i o n s can be understood i n terms of shape-resonance which r e s u l t from the trapping of the photoelectron by some type of p o t e n t i a l b a r r i e r [73,77]. I n i t i a l l y a coulomb-repulsive type of b a r r i e r was advocated [73], however, observations of s i m i l a r phenomena i n molecules which do not possess electronegative ligands (eg. N 2 and CO[4,71]) led to the postulation of some form of c e n t r i f u g a l b a r r i e r caused by the anisotropic nature of the molecular f i e l d [74,75]. In the f i n a l study of t h i s work, the spectra of two molecules which f a l l into t h i s l a t t e r category are presented. Butadiene and allene are the simplest hydrocarbons possessing two carbon-carbon double bonds. Trans-1,3-butadiene (CH2=CH-CH=CH2), henceforth referred to as butadiene, i s the simplest conjugated system while allene (CH2=C=CH2) i s the simplest cumulene. Much attention has been focussed upon the lowest valence e l e c t r o n i c excited states of these two molecules, both experimentally and t h e o r e t i c a l l y . Valence s h e l l work on butadiene includes UV absorption studies [205 and references - 276 -with i n ] , variable angle electron energy loss spectroscopy (EELS) at low impact energies [206-211], multiphoton i o n i s a t i o n [212-214] and various t h e o r e t i c a l studies [215-218]. Likewise, the valence s h e l l spectrum of allene has also been studied extensively, including UV absorption spectra [219-222], electron impact spectroscopy [223] and t h e o r e t i c a l studies [215,224-226]. The only work on the carbon Is inner s h e l l electron excited states i s the low resol u t i o n (~leV FWHM) inner s h e l l electron energy loss spectrum (ISEELS) of butadiene reported recently by Hitchcock et a l . [100]. No inner s h e l l electron e x c i t a t i o n spectra of allene have so far been published. In t h i s chapter work high r e s o l u t i o n ISEELS spectra are presented for butadiene (0.07 - 0.21 eV FWHM) and allene (0.11 - 0.35 eV FWHM). Also shown are the extended valence s h e l l electron energy loss spectra (up to 25 eV) recorded at a resolution of ~0.03 eV FWHM. The previously reported valence s h e l l EELS spectra have only extended as far as ~11.5 eV [207,223]. There has been a great v a r i e t y of interpretations presented on the valence s h e l l spectra of these molecules, e s p e c i a l l y for allene. In the previous chapters, d i f f e r e n t relaxation e f f e c t s were noted on the Rydberg and v i r t u a l valence l e v e l s upon creation of a core hole as compared to a hole i n the valence s h e l l . Differences i n excess of 2 eV have been observed for term values between the two spectral regions for the v i r t u a l valence l e v e l s (see Chapters 3 and 7) whereas Rydberg term values are l a r g e l y unaffected by the l o c a t i o n of the hole. Increases i n term values on going to the inner s h e l l regions have been observed for the Rydberg l e v e l s , however, these Increases tend to be small (<0.5 eV) - 277 -- (see Chapter 8 and r e f . [72]). Thus a clear separation of t r a n s i t i o n s to the Rydberg and v i r t u a l valence l e v e l s i s often seen i n inner s h e l l spectra. Hence, from a consideration of the ISEELS spectra presented i n th i s work, i t should be possible to obtain an unambiguous assignment and to d i s t i n g u i s h c l e a r l y between the valence and Rydberg f i n a l states. Application of the term values thus obtained to the valence s h e l l spectra i s expected to give upper bounds to the term values for the various t r a n s i t i o n s and so help c l a r i f y the assignment of the valence s h e l l spectra. Experimental D e t a i l s A l l the spectra were obtained on the ISEELS spectrometer described i n Chapter 2. The inner s h e l l spectra were c a l i b r a t e d using the N 2 (N Is •*• % , v=l) t r a n s i t i o n at 401.10 eV while the valence s h e l l spectra were c a l i b r a t e d against the He I resonance l i n e at 21.218 eV [104]. Results and Discussion 1. Inner s h e l l Spectra (a) Butadiene (CH2=CH-CH=CH2) The carbon Is inner s h e l l spectra of butadiene at various energy resolutions are shown i n F i g . 9.1. The long range spectrum ( F i g . 9.1, lower section) agrees well with the previously published low r e s o l u t i o n spectrum by Hitchcock et a l . [100]. However, more features can now be observed due to the higher r e s o l u t i o n achieved i n the present work. The - 278 -a) A E = 0 . 0 7 e V 1 2 I i b ) A E = O.I4eV 1 2 l I 1,3 BUTADIENE CH 2=CH-CH=CH 2 CARBON K-SHELL -1s e d g e 3 A 5 I I I 284 285 286 284 286 288 290 c )AE=0 .2 leV i r X P S i R-1s f e d g e 1 2 11 A 6 7 8 I I I I 9 I 10 280 290 300 3I0 E N E R G Y L O S S (eV) F i g . 9.1: Inner s h e l l electron energy loss spectra of butadiene at (a) 0.07 eV, (b) 0.1A eV and (c) 0.21 eV spectral r e s o l u t i o n . The Is edge Is the estimate of Hitchcock et a l . [100]. The positio n s of the XPS s a t e l l i t e structure from the edge i s from Carlson et a l . [227]. - 279 -energies, estimated term values and proposed assignments of the features are summarised i n Table 9.1. There has been no reported value for the C Is IP's of butadiene and so the value (290.4(3) eV) estimated by Hitchcock et a l . [100] from a consideration of the IP's of rela t e d hydrocarbons has been used to predict the term values. Hitchcock et a l . [100] assumed that the IP's of the two d i f f e r e n t C Is environments would be very s i m i l a r and i n d i s t i n g u i s h a b l e . This i s supported by the (uncalibrated) XPS spectrum of butadiene [227] which shows a single peak i n the C Is region. Butadiene i s of symmetry and i t s two unoccupied n l e v e l s are of a and b symmetries. Transitions from the C Is le v e l s (which u g * have a^ and b u symmetries) to both of these Tt l e v e l s might be expected to be seen at the high resolution of the present ISEELS experiment. Features 1 and 2 ( F i g . 9.1a) are c l e a r l y a t t r i b u t a b l e to t r a n s i t i o n s to the it l e v e l s because of t h e i r very high term values (6.1 eV and 5.6 eV respectively) and thus they are assigned as the C ls(a^)-*-n ( a u ) and * * C ls(b U)-»-TC (bg) t r a n s i t i o n s accordingly. The term value f o r the n ( a u ) l e v e l (6.1 eV with respect to the estimated [100] IP) i s 3 eV greater than the corresponding term value obtained from the Tt(b )(H0M0)-»rt (a ) g u t r a n s i t i o n observed i n the valence s h e l l spectrum [207-211], The t r a n s i t i o n to the second TI l e v e l from the HOMO l e v e l ( i . e . n(b )-*-TI (b )) i s dipole forbidden and has been the subject of much - 280 -Table 9.1 E n e r g i e s , term v a l u e s and p o s s i b l e assignments f o r the C Is energy l o s s Spectrum of Butadiene Fe a t u r e Energy (eV) Term Va l u e ^ (eV) P o s s i b l e Assignment 1 2 284.32 284.83 T t 6.1 5.6 * C l s ( a ) •*• it (a ) 8 * u C l s ( b ) -» it (b ) u g 3s 3 287.1 3.2 4 287.56 2.8 3 P i 5 287.87 2.5 3 p 2 6 288.5 1.9 3d 7 289.52 0.9 4p e t c 8 292.4 Double e x c i t a t i o n 9 295.9 * a (C-C) shape resonance 10 -304 * a (C=C) shape resonance c a l c u l a t e d u s i n g an e s t i m a t e d IP of 290.4(3)eV [100]. c a l i b r a t e d f e a t u r e . E s t i m a t e d u n c e r t a i n t y ±0.08eV. - 281 -search [207-210,212,213]. I t has been recently i d e n t i f i e d using variable angle low energy EELS [209,210] and has been observed as a di f f u s e band underlying sharp Rydberg structure at ~7.4 eV i n the * valence s h e l l spectrum. The separation of the n states i n the ISEELS spectra ( F i g . 9.1a) i s 0.5 eV, whereas i t i s 1.5 eV i n the valence s h e l l spectrum [209,210]. This can be att r i b u t e d to several factors including the fact that a frozen o r b i t a l picture cannot be assumed, as has been noted by Hitchcock et a l . [100]. Adams [228] has estimated the r e l a x a -t i o n energies for some hydrocarbons including butadiene using CND0/2 cal c u l a t i o n s which indicate a d i f f e r e n t relaxation e f f e c t depending on which C atom the core hole resides. A further small difference may aris e i n that the a^ and b u C Is l e v e l s might be separated by as much as a few tenths of an eV, a separation which would not be seen i n most XPS work. The rest of the pre-edge features can be assigned to the various Rydberg t r a n s i t i o n s . The shoulder (feature 3), on the leading edge of peak 4, has a term value of ~3.2 eV and i s assigned to the C l s ( b )+3s(a ) Rydberg t r a n s i t i o n . Features 4 and 5 have approximate term values of 2.8 eV and 2.5 eV resp e c t i v e l y and are assigned to l s ( a )+3p 1 and l s ( a )-*3p2 t r a n s i t i o n s (note: the designations 3pj and 3p 2 are used [210] because the p Rydberg o r b i t a l s are of a u and b u sym-metries. Transitions to these l e v e l s are therefore dipole forbidden from the l s ( b ) l e v e l ) . Term values for features ascribed to the 3s, u 3p± and 3p 2 l e v e l s i n valence s h e l l spectra are 2.88 eV, 2.45 eV and 2.02 eV from EELS studies [210] and 2.88 eV, 2.43 eV and 2.06 eV from - 282 -multiphoton i o n i s a t i o n [213]. Thus i t i s seen that the ISEELS term values (3.2 eV, 2.8 eV and 2.5 eV respectively) are approximately 0. 4-0.5 eV larger than those for the corresponding valence s h e l l tran-s i t i o n s . As discussed e a r l i e r , this i s not unusual for Rydberg l e v e l s [Chapter 8 and r e f . [72]) and i s c e r t a i n l y much less than the diff e r e n c e for the 11 l e v e l s . Feature 6 i s ascribed to C Is + 3d t r a n s i t i o n s . The term value of ~1.9 eV i s 0.2 eV larger than that for the 3d l e v e l obtained from multiphoton i o n i s a t i o n of the valence s h e l l [213]. Feature 7 i s assigned to t r a n s i t i o n s to the 4p and higher Rydberg l e v e l s leading up to the edge. The spectrum shows considerable i n t e n s i t y beyond the C Is i o n i s a -t i o n edge. Features i n the continuum can a r i s e by several d i f f e r e n t means including (a) double excitations [50] (b) onsets of "shake-up continua" and (c) t r a n s i t i o n s to quasi-stationary states (shape-resonances). Hitchcock et a l . [100] attributed the broad continuum * feature centred at 295.9 eV and the structure at ~304 eV to a (C-C) and a (C=C) shape-resonances r e s p e c t i v e l y . There has been much discussionon the p o s s i b i l i t y of some c o r r e l a t i o n between shape-resonance p o s i t i o n and the distance between the ionised atom and i t s neighbour(s) (see Chapter 1, section F . I I I ) . The l i n e a r c o r r e l a t i o n obtained by Hitchcock et a l . [100] between features assigned to a shape-resonances and bond length i n a series of hydrocarbons, including butadiene, lend support to t h e i r * assignment. However while features 9 and 10 can be assigned to a shape-resonances, contributions from double e x c i t a t i o n and "shake-up" - 283 -continua cannot be discounted (see subsequent section on a l l e n e ) . Considerable "shake-up ( i . e . simultaneous e x c i t a t i o n of a valence electron upon core i o n i s a t i o n ) features can often be observed i n XPS spectra. The production of these states would be manifested i n ISEELS by onsets (adiabatic energies) of new continua. For example much of continuum structures i n the N Is ISEELS spectrum of NF 3 was assigned to "shake-up" continua by comparison with the corresponding XPS "shake-up" spectra. Carlson et a l . [227] have reported the "shake-up" spectrum of butadiene. The ( v e r t i c a l ) energies of the s a t e l l i t e peaks r e l a t i v e to the i o n i s a t i o n edge are indicated on the long range spec-trum i n F i g . 9.1. The f i r s t two (XPS) features are ascribed to u-nt e x c i t a t i o n s . They are close to the features assigned to the shape resonances i n the ISEELS spectrum [100]. However, the broad maximum (feature 9) i n the ISEELS spectrum i s s i m i l a r (though broader) to the continuum feature observed i n ethane [68] , which i s at t r i b u t e d to the a (C-C) resonance [100]. This feature i n ethane i s closer to the i o n i -sation edge, as would be expected from bond length considerations. Furthermore i t cannot be ascribed to "shake-up" since the f i r s t feature i n the XPS spectrum of ethane i s observed 14.7 eV away from the C Is peak [229]. Feature 8 ( F i g . 9.1) i s at t r i b u t e d to a double e x c i t a t i o n process. (b) Allene (CH2=C=CH2) The ISEELS spectra of allene are shown i n F i g . 9.2. The features as well as the possible assignments are summarised i n Table 9.2. The value for the C Is i o n i s a t i o n edge has been taken from X-ray - 284 -a ) A E = 0.l leV , b) AE=O.I8eV li ALLENE CH 2=C=CH 2 CARBON K - S H E L L i [ j 1 1 \ £edge i 1 2 1 1 J j I 2 "1'456 7 I I 1 H i I 9 i 1 1 i 285 286 1 284 1 1 ' 288 i 1 i 292 296 c ) A E = 0 . 3 5 e V [Sedge 46 8 I i I 91011 i i I 1— 12 I 13 280 290 300 3I0 32d E N E R G Y L O S S (eV) . 9.2 Inner s h e l l electron energy loss spectra of allene at (a) 0.11 eV, (b) 0.18 eV and (c) 0.35 eV spectral r e s o l u t i o n . The Is edge i s taken from XPS [230]. - 285 -Table 9.2 Energ ies , term values and poss ib le assignments for the C Is energy loss spectrum of a l l ene Feature Energy (eV) Term Value* (eV) Poss ib le Assignment** 1 2 8 5 . 4 0 t t f 5.05 C I s • « * ( a ) 2 286.01 4.44 C " l s •»• Tt 3 287.6 2.9 C l s ( b 2 ) + 3s 4 288.45 2.00 C I s * 3 P l ( b ) 5 288.8 1.7 C ' l s •> 3p 2 6 289.11 1.34 C " l s •»• 3p 7 289.57 0.88 C ' l s ->• 4p C Is edge* 290.45 8 290.5 i o n i s a t i o n edge 9 293.2 10 294.2 double e x c i t a t i o n / "shake-up" 11 295.5 12 -302.3 double e x c i t a t i o n / "shake-up" 13 -308.7 o(C=C) shape-resonance Estimated from Experimental XPS C Is IP [230]. t t C Is o r b i t a l s are of a^ symmetry for the c e n t r a l C atom and b 2 and combinations for the outermost C atomc C and C " . * * * C a l i b r a t e d feature , estimated uncerta ianty ± 0 . 0 8 e V . (a) * TC o r b i t a l s are a c c i d e n t a l l y degenerate and of e symmetry p o r b i t a l s are s p l i t i n t h i s symmetry as b 2 and e. Di f ference i s designated by p^ and p 2 « - 286 -photoelectron spectroscopy [230]. In spite of having two types of carbon environment, the allene XPS spectrum only shows one unstructured peak centred at 290.45 eV. However, t h i s feature i n the XPS spectrum has a FWHM of 1.47 eV [230] suggesting the presence of more than one c l o s e l y spaced l e v e l . The high r e s o l u t i o n ISEELS spectrum, ( F i g . 9.2a) shows two features (features 1 and 2) which can be assigned to the two expected C Is + i t t r a n s i t i o n s . Allene i s of D 2 (j symmetry and i t s two i t l e v e l s are degenerate [223] with e symmetry. Transitions from the C Is o r b i t a l s , which have symmetries of a^ for the ce n t r a l carbon and a^, b 2 combinations for the outer carbons [225,231], are a l l allowed. The separation of features 1 and 2 may be i n d i c a t i v e of the s p l i t t i n g of the Is l e v e l s due to the two d i f f e r e n t types of carbon environment and the d i f f e r e n t r elaxation e f f e c t s upon the creation of a core hole on the outer carbons as opposed to the c e n t r a l carbon. The rest of the pre-edge spectrum shows prominent Rydberg structure. The only dipole allowed t r a n s i t i o n to the s Rydberg l e v e l s i s from the C l s ( b 2 ) l e v e l . The weak structure at 287.6 eV (feature 3) with a term value of ~2.9 eV i s assigned to the C l s ( b 2 ) -*• 3s t r a n s i -t i o n . The rest of the Rydberg structure i s complex. As well as a possible separation between the C Is o r b i t a l s , the p Rydberg l e v e l s are s p l i t i n t h i s symmetry, having b 2 and e symmetries. Transitions to these l e v e l s are allowed from the C Is l e v e l s with the exception of the C l s ( b 2 ) •*• p(b 2) t r a n s i t i o n . Feature 4 i s assigned to a t r a n s i t i o n - 287 -to one of the p Rydberg series (3p^) with the shoulder on the high energy side (feature 5) assigned as a t r a n s i t i o n to the other p series ( 3 p 2 ) . The implied separation of the two series (0.3 eV) i s i n keeping with the separation of the 3p l e v e l s (0.12 eV) i n the valence s h e l l spectrum as assigned by Betts and McKoy [215]. However, as w i l l be seen i n the next section, the valence spectrum of allene i s complex and unlike butadiene i s subject to several very d i f f e r e n t i n t e r p r e t a t i o n s . Application of the Rydberg formula to feature 4 gives a quantum e f f e c t of 0.4, which implies a 4p term value of 1.0 eV. Feature 7 i s accordingly assigned to t r a n s i t i o n s to the 4p Rydberg l e v e l . From the term value of feature 3 and a p p l i c a t i o n of the formula a term value of 1.4 eV would be expected for the 4s Rydberg l e v e l . This would coincide with feature 6; however, the feature (3) assigned to the 3s l e v e l i s very weak and the higher s Rydberg l e v e l s , having even less i n t e n s i t y , should not be seen. Feature 6 i s therefore assigned to the t r a n s i t i o n to the 3p Rydberg l e v e l s from the other C Is l e v e l . I t should be noted that the separation of feature 6 from feature 4 i s close to the separa-tions of features 1 and 2. The term value of 1.34 eV Is very low for a 3p l e v e l , however, i t must be remembered that this i s estimated from the mean C Is edge and the true term value i s probably higher. While the term values cannot be established with absolute ce r t a i n t y to t h e i r respective edges i t can be concluded that the 3p term values for the inner s h e l l spectrum of allene are low and c e r t a i n l y not more than 2 eV. Structure leading up to feature 8 can be assigned to t r a n s i t i o n s to the higher Rydberg l e v e l s converging onto the edge. - 2 8 8 -Considerable structure can be seen i n the post-edge region. As i n the case of butadiene these can be assigned to double e x c i t a t i o n s , "shake-up" or a shape-resonances. The strong band encompassing f e a -tures 9 - 11 i s assigned to double excitation/"shake-up" continua. Strong % -*• n t r a n s i t i o n s can be expected i n allene but unfortunately no XPS "Bhake-up" spectrum has so far been reported. I t i s tempting to assign t h i s structure (9 - 11) as p a r t l y due to a shape-resonance. However, i n l i g h t of the discussion given i n the previous section and the proposed bond length/shape resonance re l a t i o n s h i p s [99,100] a shape-resonance only ~3 eV from the edge would imply a bond length t y p i c a l of a carbon-carbon single bond. Obviously t h i s i s not the case i n a l l e n e . Further continuum structures are seen at higher energies. These are * a t t r i b u t e d to double excitation/"shake-up" and the a (C=C) shape resonance features. From a l e a s t squares f i t on the resonance p o s i t i o n above the edge (6) against bond length (R) for a v a r i e t y of molecules with carbon-carbon bonds, Hitchcock et a l . [100] have obtained the following (empirical) r e l a t i o n s h i p : 6 = -54.8R + 84.8 Fourteen d i f f e r e n t carbon-carbon bonds were used and a c o r r e l a t i o n c o e f f i c i e n t of -0.977 was obtained. Substituting the approximate p o s i -tions of features 12 and 13 above the edge (11.8 eV and 18.2 eV) into t h i s equation r e s u l t s i n a prediction of ~1.33 A and 1.21 A for the - 289 -carbon-carbon bond length. The actual carbon-carbon bond length of allene found by other spectroscopic means i s 1.3084(3) A [162]. Thus on t h i s basis feature 12 rather than feature 13 should be assigned to the a (C=C) shape resonance. However, Sette et a l . [99] have noted that the s t r u c t u r a l l y related (valence i s o e l e c t r o n i c ) l i n e a r molecules C0 2, N 20 and COS give bond lengths predicted from the resonance positions that are systematically shorter than the actual bond length. Thus feature 13 * i s assigned to the a (C=C) shape-resonance. 2. Valence She l l Spectra (a) Butadiene The long range valence electron energy loss spectrum of butadiene i s shown i n F i g . 9.3 as well as a more detai l e d spectrum of the 5-10 eV region. The energies of the s p e c t r a l features are summarised i n Table 9.3. There has been considerable work done on the region leading up to the f i r s t i o n i s a t i o n edge at 9.09 eV [205-218]. The deta i l e d (dipole) spectrum ( F i g . 9.3c) presented here agrees well with the higher r e s o l u -t i o n UV absorption spectrum [205] and i s consistent with the e a r l i e r low energy EELS work [206-210]. There i s general agreement i n the assign-ment of the f i r s t members of the s and the two p Rydberg series o r i g i n a -t i n g from the HOMO ( l b ) o r b i t a l between the electron impact [209,210] and multiphoton i o n i s a t i o n [212,213] work. The positions of the (forbidden) 3s t r a n s i t i o n and (allowed) 3p± and 3p 2 t r a n s i t i o n s from the 1 b g o r b i t a l are indicated i n the spectrum (Fig. 9.3c). The o r i g i n of a - 2 9 0 -1,3 BUTADIENE CH2=CH-CH=CH2 VALENCE REGION 3rr~i6bu r n k7ar 2 o u 3p b) |1a u | 7 a g 6b u^ | 6 a g 23 4 56 769 IO 11 12 II I I I I I I I I I 3 s -1 tfau 2bg S I 10 15 3 s V T b u 3 s 2 b g r~n 23 4 56 789 10 11,12 II I I I I I I I I - | 5 b u 5a r io " T -20 15 I 25 3s 3Pi 1,3 BUTADIENE CH2=CH-CH=CH2 VALENCE REGION 5-10 eV DETAIL 3p- 1bG 0 I o b c d e f g h i j k l m n o p q r s t u v w x y z 1 i i i i i i i i i I I I ri i i i 11 11 i 11 i i i 2 3 I I 7 8 ENERGY LOSS (eV) io i g . 9.3 Valence s h e l l electron energy loss spectra of butadiene. The io n i s a t i o n edges are taken from PES [232]. Po s i t i o n of the tr a n s i t i o n s shown i n 3(a) are estimated from the term values (see Table 9.4). The features associated with the 3p l b •*• 3s and t r a n s i t i o n s are indicated i n 3(b). Table 9.3 Energies of the features in the valence shell electron energy loss spectrum of butadiene. Feature Energy Feature Energy Feature Energy Feature Energy (eV) (eV) (eV) (eV) 0 5.73 i 7.01 s 8.18 4 10.60 1 5.90* j 7.06 t 8.25 5 11.11 a 6.06 k 7.14 u 8.38 6 11.48 b 6.23 1 7.27 v 8.51 7 12.33 c 6.33 m 7.34 w 8.58 8 12.59 d 6.42 n 7.47 X 8.69 9 12.93 e 6.54 o 7.62 y 8.78 10 13.52 f 6.65 P 7.87 z 8.87 11 14.12 g 6.74 q 8.00 2 9.54 12 14.71 h 6.82 r 8.06 3 9.81 calibrated feature, estimated uncertainty ±0.05eV. - 292 -d Rydberg series has also been located by multiphoton i o n i s a t i o n [213]. A l l the sharp structure has been assigned to the various Rydberg series and t h e i r v i b r a t i o n s . There has been some disagreement, however, with e a r l i e r UV work [208] on the exact nature of the s e r i e s . These have been summarised by Mallard et a l . [214]. As stated e a r l i e r the two TE l e v e l s have a u and b^ symmetry. Only the u ( a u ) l e v e l i s accessible by dipole s e l e c t i o n rules from the l b l e v e l . This t r a n s i t i o n i s i d e n t i f i e d with the strong feature and g v i b r a t i o n a l structure centred at 5.90 eV i n agreement with e a r l i e r works * [206-210]. The t r a n s i t i o n to the TI (b ) l e v e l i s dipole forbidden but s has recently been i d e n t i f i e d by variable angle EELS spectroscopy [209, 210] to a broad structure under the sharp Rydberg t r a n s i t i o n s centred at 7.4 eV. It i s again of importance to note that the assignment of the presently obtained ISEELS spectrum i s consistent with the valence s h e l l * spectral assignments given e a r l i e r for the n and Rydberg l e v e l s . Transitions to a l l these l e v e l s are allowed and observed from the C Is l e v e l s (see F i g . 9.1 and Table 9.1). In spite of the separation of * 0.5 eV observed between the TI l e v e l s i n the ISEELS spectrum (Table 9.1) the assignment, as discussed e a r l i e r , i s s t i l l consistent with the valence s h e l l assignment as given i n the variable angle EELS work [209, 210]. Turning now to the long range spectrum shown i n F i g . 9.3, t r a n s i -tions from the more t i g h t l y bound valence o r b i t a l s can be seen. The - 293 -features at 9.54 eV and 1 1 . 1 1 eV have been noted previously [206,207] If the term values obtained for the t r a n s i t i o n s a r i s i n g from the lb l e v e l are assumed to be transferrable to the more t i g h t l y bound valence o r b i t a l , the t r a n s i t i o n energies from these l e v e l s can be estimated. Table 9.4 summarises the expected positions of these t r a n s i t i o n s . The estimate positions of the allowed t r a n s i t i o n s are also indicated i n F i g . 9.3. As noted by F l i c k e r et a l . [207], t h i s method should be more r e l i a b l e for the Rydberg t r a n s i t i o n s , although i t should be possible to t e n t a t i v e l y assign intravalence t r a n s i t i o n s i n a s i m i l a r fashion. This procedure has been used to help assign the valence s h e l l e x c i t a t i o n spectra of the phosphorus compounds i n Chapter 7 as well as for NF 3 (Chapter 3) 3nd 3 1 ( ^ 3 ) ^ (Chapter 4). From this procedure i t i s c l e a r l y seen that the features v-z ( F i g . 3a,c) can also contain contributions from the l a u " * 3s t r a n s i t i o n , a p o s s i b i l i t y not considered i n the e a r l i e r i n t e r p r e t a t i o n s . S i m i l a r l y feature 4 can be assigned to the 6b^ •*• 3s t r a n s i t i o n . The band encompassing features 2 and 3 can be assigned to * * 7a •*• 71 (a ) and l a + TE (b ) intravalence t r a n s i t i o n s . Likewise the g u' u g' band encompassing features 5 and 6 can be assigned to the 6a •*• i t (a ) * and 6b •*• TE (b ) t r a n s i t i o n s . However, t r a n s i t i o n s to the various u g Rydberg series would also contribute i n t e n s i t y , (b) Allene The 2.5 keV valence s h e l l electron energy loss spectrum of allene shown in F i g . 9.4 i s s i m i l a r to the EELS spectrum reported e a r l i e r by Mosher - 294 -Table 9.4 Estimated t r a n s i t i o n energies form the valence o r b i t a l s o f (a) butadiene assuming constant term values for the valence s h e l l O r i g i n a t i n g o r b i t a l T r a n s i t i o i I energies (b) and ( I P (eV))[232] «*(2.u) n*(2b ) g 3s 3 P l 3p 2 lb (9.09) g l a u (11.55) 5.9 (7.4) (6.21) 6.64 7.07 (8.5) 10.0 8.67 (9.11) (9.53) 7a (12.35) g 6b u (13.7) 9.3 (10.8) (9.47) 9.91 10.33 (10.6) 12.1 10.8 (11.3) (11.7) 6a (14.0) g 5a (15.3) g 5b (15.7) u 10.9 (12.4) (11.1) 11.6 12.0 12.2 (13.7) (12.4) 12.9 13.3 (12.6) 14.1 12.8 (13.3) (13.7) 4b (17.7) u (14.6) 16.1 14.8 (15.3) (15.7) Term Values : n1 - 3 . 1 eV, n2 - 1.6eV, s - 2.88 eV F i n a l o r b i t a l symmetries it^ - a u ' n2 ~ ^ s - a ; p - a ,b g u ' u Number i n parentheses s i g n i f y t r a n s i t i o n s which are d ipo le forb idden . - 295 ->-H CO UJ UJ > UJ ALLENE H2C = C=CH2 VALENCE REGION 2e 1e 3b2 67 8 9 l l l l I I 1 2 3 45 678910 11 III II llll I I "I I 10 15 ENERGY LOSS (eV) 20 25 F i g . 9.4 Valence s h e l l electron energy loss spectrum of al l e n e . The i o n i s a t i o n edges are taken from PES [232]. - 296 -et a l . [223] using 40 eV electron impact energy at zero degree s c a t t e r -ing angle and confirms the e a r l i e r work up to i t s l i m i t at 12 eV energy l o s s . Mosher et a l . [223] report that no features were observed between 12 eV and 16 eV; however, a broad structure centred at 13.85 eV can be c l e a r l y seen i n the presently reported spectrum ( F i g . 9.4). Higher r e s o l u t i o n UV absorption spectra have been reported by Rabelais et a l . [220] and Iverson et a l . [221], but only up to -10.2 eV. As i n the case of butadiene there has been much work on assign-ing the spectrum up to the f i r s t i o n i s a t i o n edge but for allene there i s considerable disagreement among the various i n t e r p r e t a t i o n s . The s i t u a t i o n i n allene i s further complicated by the existence of degene-rate states and hence the p o s s i b i l i t y of Jahn-Teller s p l i t t i n g . The f i r s t photoionisation band observed i n the photoelectron spectrum shows two components at 10.02 eV and 10.58 eV which have been assigned to Jahn-Teller s p l i t t i n g of the HOMO 2e o r b i t a l [233,234]. In view of the clear assignment which can be made i n the ISEELS spectrum between the TI and Rydberg l e v e l s , and given the fact that the term values of Rydberg le v e l s remain almost constant i r r e s p e c t i v e of the l o c a t i o n of the hole, i t i s thought that the presently reported ISEELS spectrum of allene can be used to c l a r i f y the valence s h e l l s p e c t r a l assignment. The most recent i n t e r p r e t a t i o n of the valence s h e l l electron e x c i t a t i o n spectrum i s that presented by Diamond and Segal [226] who have performed a large scale ab i n i t i o CI c a l c u l a t i o n of the excited s i n g l e t states of al l e n e . They have compared t h e i r r e s u l t s [226] with previous i n t e r p r e t a t i o n s of the o p t i c a l spectrum as well as that obtained using magnetic c i r c u l a r - 297 -dichroism [222]. The s p e c t r a l features below 10 eV i n the valence s h e l l spectrum of allene a r i s e from t r a n s i t i o n s from the n(2e) o r b i t a l to the * * % (3e) o r b i t a l and to the various Rydberg l e v e l s . No low l y i n g a l e v e l s are predicted i n keeping with the proposed assignments of the ISEELS spectrum (Table 9.2). Since allene Is of D^^ symmetry only f i n a l states of *B 2 and *E are accessible from the ground state. Table 9.5 summarises the energies and term values of the features, seen below the f i r s t IP i n F i g . 9.4. Also summarised are some of the dramatically d i f f e r e n t assignments given i n previous papers [215,220-224.226,235]. The most noteworthy disagreement comes with the recent high q u a l i t y (ab i n i t i o ) c a l c u l a t i o n of Diamond and Segal [226]. These authors assign [226] the intense band encompassing features 2 and 3 s o l e l y to t r a n s i t i o n s to the 3p Rydberg l e v e l s whereas most other e a r l i e r assignments a t t r i b u t e t h i s intense band as mainly due to the u -»• T t*( 1A 1 -> 1B 2) t r a n s i t i o n with either a 3s Rydberg [220,224] or 3s and 3p Rydberg [222,235] t r a n s i t i o n ( s ) also present. Diamond and Segal [226] then a t t r i b u t e the n -*• % ( • > lB2) t r a n s i t i o n to f e a t u r e d . Based upon t h e i r assignment, t h i s would imply a term value of 1.45 eV for the n l e v e l and term values of 2.82 eV and 2.60 eV for the 3p(e) and 3p(b 2) Rydberg l e v e l s . Comparing these with the term values obtained from the ISEELS spectrum implies that the term value for the it state decreases by at least 3.3 eV (an average value has been used for the it ISEELS values) while the 3p term value Increases by 0.8 eV. While the reduct-* ion i n term value for the TI o r b i t a l i s large (based upon t h i s assign-ment), i t i s nevertheless not unreasonable and nothing d e f i n i t e can be Table 9.5 Energies, Term Values and various assignments for the valence electron excitation spectrum of allene < , a' below the 1st I.P. Term Value* Assignments (b) Feature Energy u ; (eV) (eV) Rabelais Iverson Betts Robin Mosher Fuke et a l . Rauk et a l . Diamond This et a l . et a l . et a l . et a l . et a l . work ref[220] ref[221] ref[215] ref[235] ref[223] ref[222] ref[224] ref[2261 1 2 3 4 5 6 7 8 9 10 11 6.70 7.20 7.42 8.04 8.16 8.57 8.69 8.88 9.13 9.39 9.72 3.32 2.82 2.60(3.16) 1.98 1.86 1.45(2.01) 1.33 1.14 0.89 0.63 0.30 iE H w ( L B 2 ) 3p + vib 4s + vib 3d 4p etc diffuse diffuse 3 P diffuse 4s etc 3s 3p(b,) 3p(e) 4s etc 3s * V B 2 ) 3p(e), (3s)* 3p(b 2) 3 P ( b 2 ) * nVA,) or 3 P 4s 3s 3p,(3s)* T i V B p . 3 8 (3s)* 3p(b 2) 3p(e)( 1A 2) 4s,3p(e)( 1B 2) 3d(e) 3p(e) 3p(b 2) 3d, 4s n*( 1B 2) etc nVB 2) 3s 3p Calculated from the IP of the HOMO orbital (it(e)) at 10.02 eV except for the values in parentheses which are calculated from the higher Jahn-Teller component at 10.58 eV [233,234]. Upper Jahn-Teller component. ^ A l l e n e i s of D 2 symmetry. Only the l%2> * E 8 t a C e s a r e dipole accessible from the ground state, it* o r b i t a l i s of e ( ^ A l l transitions original from 11(e) HOMO level. Only f i n a l o r b i t a l and state ( i f needed) i s given. ^ T h e y assigned two Rydberg series with quantum defect of 0.92 and 0.372. These have been assumed to imply s and p Rydberg series respectively. ^^These interpretations involve conclusions based upon consideration of both valence and inner sh e l l spectra. See text for further discussion. - 299 -i n f e r r e d from this alone. However, the apparent increase i n 3p term value contradicts what would be expected and implies that these features cannot be assigned to the 3p t r a n s i t i o n . On t h i s basis the present work i s i n agreement with the e a r l i e r assignments [220,222-224,235] and the * 1 I major component of feature 2 i s attributed to the % •*• n ( A^ •*• AB 2) t r a n s i t i o n . I t should be noted that Diamond and Segal [226] acknowledge the d i f f i c u l t y i n describing t h i s ( 1B 2) state and c i t e the conclusions of Nascimento and Goddard [217,218] with regard to s i m i l a r problems i n the c a l c u l a t i o n of the ^ state of butadiene. They [226] suggest that more extensive basis functions and cal c u l a t i o n s at non-vertical geometries might be required to describe these excited states adequately. I t should also be noted that both the c a l c u l a t i o n s of Diamond and Segal [226] and also those by Rauk et a l . [224] indicate a strong mixing between the 3p(e) and Tt A B 2 states. With features 2 and * 3 assigned to the % t r a n s i t i o n features 4 and 5 are assigned to t r a n s i t i o n s to the 3p l e v e l s , i n agreement with the assignments given i n the e a r l i e r works (see Table 9.5), as opposed to the 3d/4s assignment suggested by Diamond and Segal [226]. B r i e f l y reviewing some of the other assignments, feature 1 has been attributed to e i t h e r a forbidden state of the i t •*• i t t r a n s i t i o n , the 3s Rydberg o r b i t a l or both [222-224,226,235]. Both of these assign-ments are reasonable; The term value (3.3 eV) i s larger than that for the weak feature assigned to the 3s Rydberg l e v e l i n the ISEELS spectrum (2.9 eV). However, i n the l a t t e r the feature, being very weak, i s hard to locate accurately, thus the term value may be i n erro r . Rabelais et - 300 -a l . [220] have assigned feature 1 as a t r a n s i t i o n to a *E f i n a l s t a t e . This was based upon i t s analogous r e l a t i o n to the forbidden 12 1 n t r a n s i t i o n i n systems. The lack of a low l y i n g a o r b i t a l rules out t h i s p o s s i b i l i t y of a valence-valence t r a n s i t i o n though i t should be remembered that the n •*• 3s t r a n s i t i o n leaves the molecule i n a *E f i n a l state. The assignment of a contribution from the TI •*• 3s t r a n s i t i o n (whether Jahn-Teller s p l i t t i n g occurs or not) to features 2 and 3 i s also reasonable. However, the assignment of a it ->• 3p contribution [222,235] as well i s unreasonable for the reasons stated above ( i . e . the term values for the 3p would not be expected to be larger for the valence s h e l l spectrum). Moving onto the spectrum above the f i r s t IP, Table 9.6 summarises the energies of the features and t h e i r term values form the various valence o r b i t a l s . These features, as i n butadiene, probably have a major contribution from intravalence-valence t r a n s i t i o n s . As Rydberg l e v e l term values are expected to remain constant, the term value of * 3.48 eV for feature 12 implies a t r a n s i t i o n to the it l e v e l and supports the above suggested assignments of the it l e v e l being below the 3p l e v e l . An increase, i n term value for a valence l e v e l i s not u n l i k e l y and i s consistent with a relaxation e f f e c t upon valence o r b i t a l s which may occur upon removing an inner electron. The lack of structure for the features rules out any further assignments. Contributions from valence-Rydberg t r a n s i t i o n s can also be expected. - 301 -Table 9.6 Energies and Term Values for the features above the 1st ionisation potential in the valence shell electron energy loss spectrum of allene Feature Energy (eV) Term Value (eV)* le (14.75eV) 3b 2 (15.5eV) 4a , (17.3eV) 2b 2 (22.0eV) 12 11.27 3.48 13 11.79 2.96 3.7 14 13.14 1.61 2.4 15 13.85 0.90 1.7 3.5 16 14.71 0.8 2.6 17 15.85 1.5 18 18.9 3.1 Estimated from the various ionisation potential obtained from photoelectron spectroscopy [232]. - 302 -Conclusions The ISEELS spectra of butadiene and allene have been presented and assigned. In both molecules, t r a n s i t i o n s to the TC l e v e l s were observed and seem to be well below the Rydberg structure. Transitions * * to both it l e v e l s were c l e a r l y i d e n t i f i e d i n butadiene. The Tt l e v e l s i n allene are degenerate, and therefore the two t r a n s i t i o n s which are seen r e f l e c t the d i f f e r e n t environments of the C Is o r b i t a l s . Both molecules show extensive structure above the i o n i s a t i o n edge. Some of t h i s i s * i d e n t i f i e d with a shape-resonances, but features a t t r i b u t a b l e to double excitations and "shake-up" are also seen and are e s p e c i a l l a y strong i n the case of a l l e n e . The valence s h e l l e x c i t a t i o n spectra have also been presented and are i n good agreement with previous UV and electron impact work. However, the spectral range has been considerably extended i n the present work. Previous assignments of the valence spectrum of butadiene are consistent with the findings of the ISEELS spectrum. As expected, the it term values are observed to decrease s u b s t a n t i a l l y , while the Rydberg term values decrease by only a small amount (<0.5 eV) on going from inner s h e l l e x c i t a t i o n to valence s h e l l e x c i t a t i o n . The observa-ti o n of the t r a n s i t i o n s to both of the Tt l e v e l s (separated by 0.5 eV) i n the ISEELS spectrum of butadiene supports the assignment of Doering and McDiarmid [209,210] of a forbidden it t r a n s i t i o n i n the low-energy, non-zero angle electron impact spectrum of the valence s h e l l of butadiene. Ap p l i c a t i o n of the 3s term value to the l a u valence o r b i t a l of butadiene indicates that some of the structure below the f i r s t IP - 3 0 3 -(lbg) i n the valence s h e l l spectrum might a r i s e form the l a u •* 3s t r a n s i t i o n . With the differences observed between the inner s h e l l and valence spectra of butadiene i n mind, the ISEELS spectrum of allene supports the assignment of the 7.20 eV feature i n the valence s h e l l spectrum as primarily due to a n •> n t r a n s i t i o n as predicted i n numerous e a r l i e r works and not as the TC -»• 3p t r a n s i t i o n suggested by the recent c a l c u -l a t i o n of Diamond and Segal [226]. It i s seen that as well as providing new spectral information, ISEELS can aid i n the i n t e r p r e t a t i o n of more complex valence s h e l l spectra. - 304 -CHAPTER 10 CONCLUDING REMARKS In t h i s work, the inner s h e l l electron energy loss spectra of several d i f f e r e n t series of compounds have been obtained. In almost a l l cases these spectra have not been previously reported. The s p e c t r a l assignments were aided by a consideration of the features i n a l l of the s e r i e s . It was seen that p a r t i c u l a r ligands had s i m i l a r e f f e c t s on the inner s h e l l spectra of the central atom i n the d i f f e r e n t s e r i e s . This was seen to be e s p e c i a l l y so for the P and Si L - s h e l l spectra, however, s i m i l a r e f f e c t s were also noted i n the N K-shell spectra. In addition to the inner s h e l l spectra, the valence s h e l l spectra of many of the compounds have been obtained and analysed with the aid of the corres-ponding ISEELS spectra. Knowledge of the inner s h e l l spectra have been shown to be very useful i n i n t e r p r e t i n g the valence s h e l l spectra. This arises from the r e l a t i v e ease i n the assignment of the inner s h e l l spectra i n conjunction with the d i f f e r i n g relaxation e f f e c t s of the Rydberg and v i r t u a l valence o r b i t a l s when an inner s h e l l vacancy i s created as opposed to a valence s h e l l vacancy. The Rydberg l e v e l s were seen to be less influenced than the v i r t u a l valence l e v e l s by the l o c a t i o n of the hole and hence have the same, or only s l i g h t l y higher (<0.5 eV) term values i n the inner s h e l l spectra as compared to the valence s h e l l spectra. Conversely, the term values of the features i n the discrete region of the spectra assigned to t r a n s i t i o n s to v i r t u a l - 305 -valence o r b i t a l s were seen to increase markedly i n the inner s h e l l spectra. In p a r t i c u l a r these factors allowed c l a r i f i c a t i o n of the complex and con t r o v e r s i a l valence s h e l l spectrum of a l l e n e . A l l of the inner s h e l l spectra presented i n t h i s work showed continuum structures which i n many cases could be reasonably ascribed to a shape-resonances. These can be thought of, i n a molecular o r b i t a l sense, i n terms of t r a n s i t i o n s to high l y i n g a v i r t u a l o r b i t a l s i n the continuum and hence can be assigned with a knowledge of the MO scheme. Continuum features a r i s i n g from many-electron processes such as "shake-up" can also occur. This was seen most c l e a r l y i n the ISEELS spectra of NF 3 where the t r a n s i t i o n s to the a l e v e l s appear i n the d i s c r e t e region of the spectra implying that the continuum structure arises from other f a c t o r s . Comparison of the N Is and F Is XPS s a t e l l i t e spectra c l e a r l y Indicated the presence of onsets of "shake-up" continua i n the ISEELS spectra. The proposed r e l a t i o n s h i p of shape-resonance p o s i t i o n and bond length was also examined i n the systems studied i n the present work. I t was seen that some form of c o r r e l a t i o n does e x i s t though care i s needed i n assigning the features. This factor provides a further possible * means of i d e n t i f y i n g features associated with a shape-resonances i n a series of related molecules. - 306 -REFERENCES 1. S. Trajmar, J.K. Rice and A. Kupperman, Adv. Chem. Phys. _18_ (Ed. I. Prigogine and S.A. Rice), Wiley Interscience, New York (1970) 15. 2. E.N. Lassettre and A. Skerbele, "Methods of Experimental Physics" 3B (Ed. D. Williams), Academic Press (1974) 868. 3. M.A. D i l l o n , "Creation and Detection of the Excited State", JJJ (Ed. A.A. Lamola), Marcel Dekker, New York (1971) 375. 4. G.R. Wight, C.E. Brion and M.J. van der Wiel, J. Electron Spectrosc. Relat. Phenom I (1972/73) 457. 5. G.R. Wight, Ph.D. Thesis, University of B r i t i s h Columbia (1974). 6. A.P. Hitchcock, Ph.D. Thesis, University of B r i t i s h Columbia (1978). 7. F.H. Read, J . Physique C o l l . Cl Supp 5, 39 (1978) 1. 8. C.E. Brion, S. Daviel, R. Sodhi and A.P. Hitchcock, AIP Conference Proceedings No. 94 "X-ray and Atomic Inner She l l Physics (Ed. B. Crasemann), American I n s t i t u t e of Physics (1982) 429. 9. C.E. Brion, "Physics of E l e c t r o n i c and Atomic C o l l i s i o n s " (Ed. S. Datz) North Holland (1982) 579. 10. A.P. Hitchcock, J. Electron Spectrosc. Relat. Phenom. 25_ (1982) 245. 11. J . Berkowitz, "Photoabsorption, Photoionisation and Photoelectron Spectroscopy", Academic Press, New York (1979). 12. M.B. Robin, "Higher Excited States of Polyatomic Molecules", Vol. 1, Academic Press, New York (1974). 13. K. W i t t e l and S.P. McGlynn, Chemical Reviews _7_7_ (1977) 745. 14. Ref. 15, p. 50. 15. J. Wayne Rabalais, " P r i n c i p l e s of U l t r a v i o l e t Spectroscopy", Wiley Interscience, New York (1977). 16. K. Siegbahn, C. Nordling, A. Fahlman, R. Nordberg, K. Hamrin, J . Hedman, G. Johansson, T. Bergmark, S.E. Karlson, I. Lindgren and B. Lindberg, "ESCA, Atomic, Molecular and So l i d State Structure studied by Means of Electron Spectroscopy", Nova Acta Regiae Soc. S c i . , Upsaliensis, Ser. 4, v o l . 20, Uppsala (1967). - 3 0 7 -17. K. Siegbahn, C. Nordling, G. Johansson, J . Redman, P.F. Heden, K. Hamrin, U. Gelius, T. Bergmark, L.O. Werme, R. Manne and Y. Baer, "ESCA Applied to Free Molecules", North Holland, Amsterdam (1969). 18. K. Siegbahn, Science 217_ (1982) 111. 19. U. Gelius, L. Asplund, E. B a s i l i e r , S. Hedman, K. Helenelund and K. Siegbahn, Nucl. Instr. Meth. Phys. Res. BJ^ (1984) 85. 20. R.J. Buenker and S.D. Peyerimhoff, Chem. Phys. Letters 36 (1975) 415. 21. R.L. Martin and D.A. Shi r l e y , J. Chem. Phys. 64 (1976) 3685. 22. R.L. Martin, B.E. M i l l s and D.A. Sh i r l e y , J. Chem. Phys. 64 (1976) 3690. 23. L.S. Cederbaum and W. Domcke, Adv. Chem. Phys. 36_ (Ed. I. Prigogine and S.A. Rice), Wiley Interscience, New York (1977) 205. 24. W. Domcke, L.S. Cederbaum, J . Schirmer, W. von Niessen, C E . Brion and K.H. Tan, Chem. Phys. 40 (1979) 171. 25. M.O. Krause, J . Phys. Chem. Ref. Data £ (1979) 307. 26. J . Nordgren, L. Selander, L. Pettersson, R. Bramner, M. BSckstrtfm, C. Nordling and H. Agren, Chem. Phys. JJ4 (1984) 333. 27. J . Nordgren, L. Pettersson, L. Selander, C. Nordling, K. Siegbahn and H. Agren, J . Phys. B _15 (1982) L153. 28. L. Pettersson, J . Nordgren, L. Selander, C. Nordling and K. Siegbahn, J . Electron Spectrosc. Relat. Phenom. 27_ (1982) 29. 29. K.D. Sevier, "Low Energy Electron Spectroscopy", Wiley Interscience, New York (1972). 30. P. Kelfve, B. Blomster, H. Siegbahn, K. Siegbahn, E. Sanhueza and 0. Goscinski, Phys. Scr. (1980) 75. 31. R. Sodhi and R.G. C a v e l l , J . Electron Spectrosc. Relat. Phenom. 32 (1983) 283. 32. R. Sodhi, M.Sc. Thesis, University of Alberta (1980). 33. E.J. Aitken, M.K. Bahl, K.D. Bomben, J.K. Gimzewski, G.S. Nolan and T.D. Thomas, J . Am. Chem. Soc. _102 (1980) 4873. - 308 -34. W.E. Moddeman, Ph.D. Thesis, Oak Ridge Nat. Lab., ORNL-TM-3012 (1970). 35. D.P. Chong, Chem. Phys. Letters J32 (1981) 511 and references within. 36. R.R. Rye, D.R. Jennison and J.E. Houston, J . Chem. Phys. 73 (1980) 4867 and references within. 37. M.O. Krause and J.H. O l i v e r , J . Phys. Chem. Ref. Data 8^ (1979) 329. 38. H. Bethe, Ann. Phys. (Leipzig) _5 (1930) 325. 39. M. Inokuti, Rev. Mod. Phys. 43 (1971) 297. 40. E.N. Lasettre, Rad. Research (Supp) 1_ (1959) 530. 41. E. McDaniel, " C o l l i s i o n Processes i n Ionised Gases", John Wiley, New York (1964). 42. M. Born, Z. Physik 38_ (1926) 803, or consult any standard physics quantum mechanics text, e.g. A. Messiah "Quantum Mechanics", J . Wiley and Son, New York. 43. E.N. Lassettre, A. Skerbele and M.A. D i l l o n , J. Chem. Phys. 50 (1969) 1829. 44. M.J. van der Wiel and G. Wiebes, Physica 53 (1971) 225. 45. C E . Brion and A. Hamnett. Adv. Chem. Phys. 4_5 (Ed. J. Wm. McGowan), Wiley Interscience, New York (1981) 1. 46. I.V. Hertel and K.J. Ross, J . Phys. B J_ (1968) 697. 47. I.V. Hertel and K.J. Ross, J . Phys. B 2 (1969) 285. 48. J.A.R. Samson, "Techniques of Vacuum U l t r a v i o l e t Spectroscopy", John Wiley, New York (1967). 49. F.C. Brown, "Synchrotron Radiation Research" (Ed. H. Winick and S. Doniach), Plenum Press (1980) 61. 50. D.A. Shaw, G.C. King, D. Cvejanovic and F.H. Read, J. Phys. B 17_ (1984) 2091. 51. G.C. King, F.H. Read and M. Tronc, Chem. Phys, Le t t e r s . 52 (1977) 50. 52. A.P. Hitchcock and C E . Brion, J . Electron Spectrosc. Relat. Phenom 18 (1980) 1. - 309 -53. S. Daviel, C E . Brion and A.P. Hitchcock, Rev. S c i . Instrum. Jj5 (1984) 182. 54. F.C. Brown, R.Z. Bachrach and A. Bianconi, Chem. Phys. Letters 5_4 (1978) 425. 55. P. GUrtler, V. Sa i l e and E.E. Koch, Chem. Phys. Letters 48 (1977) 245. 56. W. Eberhardt, R.-P. Haelbich, M. Iwan, E.E. Koch and C. Kunz, Chem. Phys. Letters 40 (1976) 180. 57. M.J. van der Wiel, Th. M. El-Sh e r b i n i and C E . Brion, Chem. Phys. Letters _7 (1970) 161. 58. M.J. van der Wiel and Th. E l - S h e r b i n i , Physica 59 (1972) 453. 59. R.B. Kay, Ph.E. van der Leeuw and M.J. van der Wiel, J. Phys. B (1977) 2521. 60. W.H.E. Schwarz, Chem. Phys. 9^ (1975) 157. 61. W.H.E. Schwarz, Chem. Phys. 1_1_ (1975) 217. 62. D.A. Shaw, G.C. King and F.H. Read, J. Phys. B 13_ (1980) L723. 63. A.P. Hitchcock and C E . Brion, J . Electron Spectrosc. Relat. Phenom. JL3 (1978) 193. 64. A.P. Hitchcock and C E . Brion, J. Electron Spectrosc. Relat. Phenom. _17 (1979) 139. 65. H. F r i e d r i c h , B. Sonntag, P. Rabe, W. Butscher and W.H.E. Schwarz, Chem. Phys. Letters 64 (1979) 360. 66. M. Tronc, G.C. King and F.H. Read, J . Phys. B 12_ (1979) 137. 67. A.P. Hitchcock, M. Pocock and C E . Brion, Chem. Phys. Letters 4£ (1977) 125. 68. A.P. Hitchcock and C E . Brion, J . Electron Spectrosc. Relat. Phenom. _10 (1977) 317. 69. A.P. Hitchcock and C E . Brion, Chem. Phys. 33 (1978) 55. 70. D. Blechschmidt, R. Haensel, E.E. Koch, U. Nielsen and T. Sagawa, Chem. Phys. Le t t . \A (1972) 33. 71. R.B. Kay, Ph.E. van der Leeuw and M.J. van der Wiel, J . Phys. B10 (1977) 2513. - 310 -72. G.R. Wight and C.E. Brion, J . Electron Spectrosc. Relat. Phenom. 4 (1974) 25. 73. J.L. Dehmer, J . Chem. Phys. 56_ (1972) 4496. 74. J.L. Dehmer and D. D i l l , Phys. Rev. Letters 35 (1975) 213. 75. J.L. Dehmer and D. D i l l , J . Chem. Phys. 65 (1976) 5327. 76. J.L. Dehmer and D. D i l l , "Electron-Molecule and Photon-Molecule C o l l i s i o n s " (Eds. T.N. Rescigno, V. McKoy and B. Schneider), Plenum, New York (1979) 225. 77. J.L. Dehmer, D. D i l l and A.C. Parr, "Photophysics and Photochemistry i n the Vacuum U l t r a v i o l e t " (Eds. S. McGlynn, G. Findley and R. Huebner). 78. T. Gustafsson, Atomic Physics 8_, Proceedings of the 8th International Conference, Goteburg, Sweden, Aug. 1982, 355. 79. G.J. Schulz, Rev. Mod. Phys. 45 (1973) 422. 80. M. Krauss and F.H. Mies, Phys. Rev. A J_ (1970> 1592. 81. J.L. Dehmer and D. D i l l , "Symposium on Electron-Molecule C o l l i s i o n s (Eds. I. Shimamura and M. Matsuzawa), Univ. Tokyo Press, Tokyo (1979) 95. 82. D. Loomba, S. Wallace, D. D i l l and J.L. Dehmer, J . Chem. Phys. 75 (1981) 4546. 83. P.W. Langhoff, "Electron-Molecule and Photon-Molecule C o l l i s i o n s " (Eds. T.N. Rescigno, V. McKoy and B. Schneider), Plenum, New York (1979) 183. 84. T.N. Rescigno and P.W. Langhoff, Chem. Phys. L e t t e r , 51. ( 1 9 7 7 ) 65. 85. T.N. Resigno, C.F. Bender, B.V. McKoy and P.W. Langhoff, J . Chem. Phys. 68 (1978) 970. 86. N. Pa d i a l , G. Csanak, B.V. McKoy and P.W. Langhoff, J. Chem. Phys. 69 (1978) 2992. 87. A.E. Orel, T.N. Resigno, B.V. McKoy and P.W. Langhoff, J . Chem. Phys. J72 (1980) 1265. 88. L.E. Machado, E.P. Le a l , G. Csanak, B.V. McKoy and P.W. Langhoff, J. Electron Spectrosc. Relat. Phenom. 25^ (1982) 1. - 311 -89. W.J. Hunt and W.A. Goddard, Chem. Phys. Letters _3 (1969) 414. 90. F.A. Gianturco, C. Guidotti and U. Lamanna, J . Chem. Phys. 57_ (1972) 840. 91. E. Ishiguro, S. Iwata, Y. Suzuki, A. Mikuri and T. Sasaki, J . Phys. B 15 (1982) 1841. 92. W.H.E. Schwarz, L. Mensching, K. Hallmeier and R. Szargan, Chem. Phys. 82 (1983) 57. 93. J.R. Swanson, D. D i l l and J.L. Dehmer, J . Chem. Phys. 15 (1981) 619. 94. J.L. Dehmer, J. Siegel and D. D i l l , J. Chem. Phys. 69 (1978) 5205. 95. C.R. N a t o l i , D.K. Misemer, S. Doniach and F.W. Kutzler, Phys. Rev. A 22 (1980) 1104. 96. T. Gustafsson and H.J. Levinson, Chem. Phys. Letters 78 (1981) 28. 97. C.R. N a t o l i , "EXAFS and Near Edge Structure" (Eds. A. Bianconi, L. Incoccia and S. S t i p c i c h ) , Springer-Verlag, B e r l i n (1983) 43. 98. A. Bianconi, M. D e l l 1 A r i c c i a , A. Gargano and C.R. N a t o l i , "EXAFS and Near Edge Structure" (Eds. A. Bianconi, L. Incoccia and S. S t i p c i c h ) , Springer-Verlag, B e r l i n (1983) 57. 99. F. Sette, J . StOhr and A.P. Hitchcock, J. Chem. Phys. (1984) submitted. 100. A.P. Hitchcock, S. Beaulieu, T. Steel , J. StOhr and F. Sette, J . Chem. Phys. 80 (1984) 3927. 101. Datron Instruments Limited, Norwich, U.K. 102. A.P. Hitchcock and C.E. Brion, J . Electron Spectrosc. Relat. Phenom. 22 (1981) 283. 103. a) E.N. Lassettre, A. Skerbele, M.A. D i l l o n and K. Ross, J . Chem. Phys. 48 (1968) 5066. b) J . Geiger and B. SchrOder, J . Chem. Phys. 50 (1969) 7. 104. C.E. Moore, Atomic Energy Levelss _1 U.S. Nat. Bureau Standards, C i r c u l a r L67 (1949). 105. C.E. Brion and S. Daviel, to be published. 106. D.A. Shaw, D. Cvajanovic", G.C. King and F.H. Read, J. Phys. B, - 312 -J7 (1984) 1173. 107. R.N. Lee, J . Electron Spectrosc. 28 (1982) 195. 108. G.C. King, M. Tronc, F.H. Read and R.C. Bradford, J . Phys. B, 1_0 (1977) 2479. 109. A.P. Hitchcock and C E . Brion, J . Phys. B, 13 (1980) 3269. 110. T.M. Zimkina and V.A. Fomichev, Soviet Phys. Doklady 11 (1967) 726. 111. D.M. Barrus, R.L. Blake, A.J. Burek, K. Chambers and A.L. Pregenzer, Phys. Rev. A, ^ 0 (1979) 1045. 112. R.E. La V i l l a , J. Chem. Phys. 57_ (1972) 899. 113. A.S. Vinogradov, T.M. Zimkina and V.A. Fomichev, J . Struct. Chem. 12 (1971) 823. 114. F. Wuilleumier, CR. Acad. S c i . , Paris B270 (1970) 825. 115. E.N. Laj3settre, F.M. Glaser, V.D. Meyer and A. Skerbele, J. Chem. Phys. 42 (1965) 3429. 116. CM. Lawrence, D.L. Mickey and K. Dressier, J. Chem. Phys. 4j$ (1968) 1989. 117. A.W. Potts, H.J. Lempka, D.C. Streets and W.C P r i c e , P h i l . Trans. Roy. Soc. London, A ^ 68 (1970) 59. 118. P.J. Bassett and D.R. Lloyd, Chem. Phys. L e t t . , 6_ (1970) 16; J.C.S. Dalton Trans., (1972) 248. 119. R.G. Cavell and R.N.S. Sodhi, to be published. 120. T.M. Zimkina and A.S. Vinogradov, B u l l . Acad. S c i . USSR Phys. Ser., 36 1972) 229 (Izv. Akad. Nauk. SSSR, Ser. F i z . , 36 (1972) 248). 121. R.L. B a r i n s k i i and I.M. Kulikova, B u l l . Acad. S c i . USSR Phys. Ser., 38 (1974) 16 (Izv. Akad. Nauk SSSR Ser. F i z . , 38 (3) (1974) 444). 122. B. C a d i o l i , U. P i n c e l l i , E. T o s a l t i , U. Fano and J.L. Dehmer, Chem. Phys. L e t t . , _17 (1972) 15. 123. CR. Wight and C E . Brion, J. Electron Spectrosc. Relat. Phenom., 4 (1974) 327. - 313 -124. F.C. Brown, R.Z. Bachrach and A. Bianconi, Chem. Phys. L e t t e r s , _54 (1978) 425. 125. A.S. Vinogradov, T.M. Zimkina, V.N. Akimov and B. Schlarbaum, Bullk. Acad. S c i . USSR Phys. Ser., _38 (3) (1974) 69 (Izv. Akad. Nauk. SSSR F i z . Ser., 38 (1974) 508). 126. S.R. La Paglla and A.B.F. Duncan, J. Chem. Phys., 34 (1961) 1003. 127. P. Finn, R.K. Pearson, J.M. Hollander and W.L. J o l l y , Inorg. Chem. , _10 (1971) 378. 128. D.W. Davis, Ph.D. Thesis, University of C a l i f o r n i a , Berkeley, Lawrence Berkeley Laboratory Report LBL-9900, May 1973. 129. M.L. Unland, J.M. Letcher and J.R. Van Wazer, J. Chem. Phys., 50 (1969) 3214. 130. B. K e l l e r e r , L.S. Cederbaum and G. Hohlneicher, J. Ele c t r o n Spectrosc. Relat. Phenom. _3 (1974) 107. 131. C. Chen, Int. J . Quant. Chem., 18 (1980) 199. 132. P.I. Mansell, C.J. Danby and I. Powis, J. Chem. Soc. Far. Trans. 2, TT_ (1981) 1449. 133. G. C i u l l o , A. Sgamellotti, F. T a r a n t e l l i , C. F a m i g l i e t t i , G. De A l t i and P. Decleva, Z. ftir Phys. Chem. Neue Folge, JL27 (1981) 1. 134. G. C i u l l o , A. Sgamellotti, F. T a r a n t e l l i , G. De A l t i and P. Decleva, Gazz. Chim. I t a l i a n o , _110 (1980) 305. 135. T.D. Thomas and R.W. Shaw, J . Electron Spectrosc. Relat. Phenom. _5 (1974) 1081. 136. W.C. Tarn and C.E. Brion, J . Electron Spectrosc. & Relat. Phenom., 2 (1973) 111. 137. M. Kobukovskii, private communication. 138. C.S. Fadley, Thesis, (Univ. of C a l i f o r n i a , 1970) UCRL 19535. 139. L. Asbrink, C. Fridh and E. Lindholm, Chem. Phys. L e t t . 52_ (1977) 63. 140. P. Langhoff, private communication. 141. L. Asbrink, C. Fridh and E. Lindholm, Chem. Phys. L e t t . 52 (1977) 72. 142. D.K. Creber, J.S. Tse and G.M. Bancroft, J . Chem. Phys., 72 (1980) - 314 -4291. 143. R. Roberge, C. Sandorfy, J . I . Matthews and O.P. Strausz, J . Chem. Phys. 69 (2978) 5105. 144. H. F r i e d r i c h , B. P i t t e l , P. Rabe, W.H.E. Schwartz and B. Sonntag, J. Phys. B: Atom Molec. Phys. _L3 (1980) 25. 145. T.M. Zimkina and A.C. Vinogradov, J. Phys. (Paris) (Suppl. 10) 32 (1971) C4-3. 146. W. Hayes and F.C. Brown, Phys. REv. A 6 (1972) 21. 147. V.A. Fomichev, T.M. Zimkina, A.S. Vinogradov and A.M. Evdokimov, Zhurnal Strukturnoi Khimii U_ (1970) 676. 148. S. Evans, J.C. Green, P.J. Joachim, A.F. Orchard, D.W. Turner and J.P. Maier, J . Chem. Soc. Faraday Trans. 2, 68 (1972) 905. 149. A.E. Jonas, G.K. Schweitzer, F.A. Grimm and T.A. Carlson; JH. Electron Spectrosc. J_ (1972/73) 29. 150. R. Boschi, M.F. Lappert, J.B. Pedley, W. Schmidt and B.T. Wilkins, J . Organometall. Chem. _50 (1973) 69. 151. W.B. Perry and W.L. J o l l y , Chem. Phys. Letters 17_ (1972) 611. 152. J.E. Drake, C. Riddle and L. Coatsworth, Can. J. Chem. 53_ (1975) 3602. 153. J.E. Drake, C. Riddle, B. Glavincevski, K. Gorzelska and H.E. Henderson, Inorg. Chem. JL7 (1978) 2333. 154. R.C. Gray, J.C. Carver and D.M. Hercules, J . Electron Spectrosc. *3 (1976) 343. 155. R.G. Cave l l and R.N.S. Sodhi, J . Electron Spectrosc. \5_ (1979) 145. 156. F.C. Brown i n So l i d State Physics (Ed. H. Ehrenreich et a l ) Volume 29 (1974), Academic Press, pages 1-73. 157. A. Venezia-Floriano and R.G. C a v e l l , private communication. 158. P.R. K e l l e r , J.W. Taylor, F.A. Grimm, P. Senn, T.A. Carlson and M.O. Krause, Chem. Physics, 74 (1983) 247. 159. A.A. Pavlychev, A.S. Vinogradov, T.M. Zimkina, D.E. Onopko and S.A. Titov, Opt. Spectrosc. (USSR) 47 (1979) 40. 160. V.I. Nefedov, Ya. V. Salyn, G. Leonhardt and R. Schreibe, J . - 315 -Electron. Spectrosc. Relat. Phenom. 10_ (1977) 121. 161. F. Bechstedt, Phys. Stat. Sol. (b) (1979) 167. 162. Landholdd-BBrnstein (New Series) II/7, "Structure Data of Free Polyatomic Molecules", Springer-Verlag, B e r l i n 1976. 163. Tables of Interatomic Distances, Chemical Society Special Publications (Ed. L.E. Sutton) _11 (1958). 164. W.B. Pery and W.L. J o l l y , J. Electron Spectros. Relat. Phenom., j4 (1974) 219. 165. J . J . Toman, A.A. Frost, S. To p l o l , S. Jacobson and M.A. Ratner, Theoret. Chim. Acta. 58 (1981) 285. 166. Z. Ber k o v i t c h - Y e l l i n , D.E. E l l i s and M.A. Ratner, Chem. Phys. 62 (1981) 21. 167. A.G. Alexander, O.P. Strausz, R. P o t t i e r and G.P. Semeluk, Chem. Phys. Lett. _13 (1972) 608. 168. Y. Harada,. J.N. Murrell and H.H. Skeena, Chem. Phys. Lett. , _1 (968) 595. 169. R.N.S. Sodhi, C E . Brion and R.G. C a v e l l , to be published. 170. M.B. Robin, Chem. Phys. L e t t . , 31^ (1975) 140. 171. L.N. Mazalov and A.V. Kondratenko, Proc. c6th Int. Conf. on VAc. UV Rad. Phys. 11-71 ( C h a r l o t t e s v i l l e , 1980). 172. I.A. Topol, A.V. Kondratenko, and L.N. Mazolov, Opt. Spectrosk. j>0 (1981) 494 [Engl, trans. Opt. Spectrosc. (USSR) _50 (1981) 267]. 173. I.A. Topol, A.V. Kondratenko, and L.N. Mazalov, Izv. Akad. Nauk (SSSR) Ser. F i z . 46 (1982) 776 [Engl. Trans: B u l l . Acad,. Sci (USSR) Phys. Ser. 46(4) (1982) 143]. 174. E.S. Ghuskin, L.N. Mazalov, A.A. Krasnoperova, V.A. Kochubei, S.I. Mishnev, A.N. S h r i n s k i i , E.M. Trahbtenberg, and CM. Tumaikin, Izv. Akad. Nauk. SSSR Ser. F i z . 40 (1976) 226 [Engl. Trans: B u l l . Acad. S c i . USSR Phys. Ser. 40(2) (1976) 1]. 175. R.N.S. Sodhi and R.G. C a v e l l , unpublished r e s u l t s . 176. S.-X. Xiao, W.C Trogler, D.E. E l l i s and Z. Berkov i t c h - Y e l l i n , J . Am. Chem. Soc. _105 (1983) 7033. 177. J . Norman, J r . , J. Chem. Phys. 61_ (1974) 4630. - 316 -178. A.D. Walsh and P.A. Warsop, "Advances i n Molecular Spectroscopy" (ed. A. Mangini) 2^ (1962) 582. 179. R.G. Ca v e l l , Inorg. Chem. 14 (1975) 2828. 180. W.B. Perry, T.F. Schaaf, and W.L. J o l l y , J. Am. Chem. Soc. 97 (1975) 4899 (corrected value from A.A. Bakke, H-W. Chen and W.L. J o l l y , J. Electron Spectrosc. Relat. Phenom. 20 (1980) 333. 181. A. B a r r i e r , I.W. Drummond, and Q.C. Herd, J . Electron Spectrosc. Relat. Phenom. _5 (1974) 217. 182. A.P. Hitchcock and C.E. Brion, J . Electron Spectrosc. Relat. Phenom. _14 (1978) 417. 183. R.W. Shaw, J r . , T.X. C a r r o l l and T.D. Thomas, J . Am. Chem. Soc. 95 (1973) 5870. 184. CM. Humphries, A.D. Walsh and P.A. Warsop, Discuss. Faraday Soc., 2 5 (1963) 148. 185. M. Halmann, J . Chem. Soc. (1963) 2853. 186. M.J. McAdams and B.R. Russ e l l , Chem. Phys. Letters ^8 (1973) 402. 187. I.H. H i l l i e r and V.R. Saunders, Trans. Faraday S o c , 66 (1970) 2401. 188. M. Xin, L-Y.C. Chiu, D. L i and Y-K. Pan, J . Electron Spectrosc. Relat. Phenom. 33 (1984) 93. 189. I.H. H i l l i e r and V.R. Saunders, J.C.S., Dalton Trans. (1972) 21. 190. T.H. Gan, J.B. Peel and G.D. W i l l e t t , Chem. Phys. L e t t e r s , 48 (1977) 483. 191. A. S t r i c h and A. V e i l l a r d , J. Am. Chem. Soc. 95 (1973) 5574. 192. J.P. Maier and D.W. Turner, J.C.S. Far. Trans. 2, 68 (1972) 711. 193. P.A. Cox, S. Evans, A.F. Orchard, N.V. Richardson and P.J. Roberts, Discuss. Faraday Soc. 5-\ (1972) 26. 194. D.W. Goodman, M.J.S. Dewar, J.R. Schweiger and A.H. Cowley, Chem. Phys. Letters J21 (1973) 474. 195. M.F. Lappert, J.B. Pedley, B.T. Wilkins, 0. Stelzer and E. Unger, J.C.S. Dalton Trans. (1975) 1207. 196. V.N. Akimov, A.S. Vinogradov and T.M. Zimkina, Opt. Spectrosk. 53 (1982) 476 [Engl. Trans. Opt. Spectrosc. (USSR) ^ 3 (3) (1982) - 317 -280]. 197. E. Tannenbaum, E.M. C o f f i n and A.J. Harrison, J . Chem. Phys. 21_ (1953) 311. 198. Ref. 12 + references within. 199. D.R. Salahub, Theor. Chim. Acta 22_ (1971) 325. 200. R. Rianda, R.P. Frueholz and W.A. Goddard I I I , Chem. Phys. _19 (1977) 131. 201. P. Avouris and A.R. Rossi, J. Phys. Chem. 85 (1981) 2340. 202. J.S. Jen and T.D. Thomas, J . Electron Spectrosc. Relat. Phenom. (1974) 43. 203. B.E. M i l l s , R.L. Martin and D.A. Shi r l e y , J. Am. Chem. S o c , 9fi (1976) 2380. 204. W.R. Harshbarger, J . Chem. Phys. 54 (1971) 2504. 205. R. McDiarmid, J . Chem. Phys. 64 (1976) 514. 206. O.A. Mosher, W.M. F l i c k e r and A Kuppermann, J. Chem. Phys., JS9_ (1973) 6502. 207. W.M. F l i c k e r , O.A. Mosher and A. Kuppermann, Chem. Phys. 30. (1978) 307. 208. J.P. Doering, J . Chem. Phys. 70 (1979) 3902. 209. J.P. Doering and R. McDiarmid, J. Chem. Phys. 73 (1980) 3617. 210. J.P. Doering and R. McDiarmid, J. Chem. Phys. 75_ (1981) 2477. 211. R. McDiarmid and J.P. Doering, Chem. Phys. L e t t . 88 (1982) 602. 212. P.M. Johnson, J . Chem. Phys. 64 (1976) 4638. 213. L.J. Rothberg, D.P. Gerr i t y and V. Vaida, J. Chem. Phys. 73 (1980) 5508. 214. W.G. Mallard, J.H. M i l l e r and K.C. Smyth, J . Chem. Phys. 79_ (1983) 5900. 215. T.C. Betts and V. McKoy, J . Chem. Phys. 60 (1974) 2947. 216. R.J. Buenker, S. Shih and S.D. Peyerimhoff, Chem. Phys. L e t t . 44 (1976) 385. - 318 -217. M.A.C. Nascimento and W.A. Goddard I I I , Chem. Phys. 36 (1979) 147. 218. M.A.C. Nascimento and W.A. Goddard I I I , Chem. Phys. 53 (1980) 251. 219. L.H. S u t c l i f f e and A.D. Walsh, J . Chem. Soc. (1952) 899. 220. J.W. Rabelais, J.M. McDonald, V. Scherr and S.P. McGlynn, Chem. Rev. 71 (1971) 73. 221. A.A. Iverson and B.R. Ru s s e l l , Spectrochim. Acta 28A (1972) 447. 222. K. Fuke and 0. Schnepp, Chem. Phys. 38 (1979) 211. 223. O.A. Mosher, W.M. F l i c k e r and A. Kupperman, J . Chem. Physs. 6>2_ (1975) 2600. 224. A. Rauk, A.F. Drake and S.F. Mason, J . Am. Chem. Soc. 101 (1979) 2284. 225. L.J. Schaad, L.A. Burnelle and K.P. Kressler, Theor. Chim. Acta. _15 (1969) 91. * 226. J . Diamond and G.A. Segal, J . Am. Chem. Soc. 106 (1984) 952. 227. T.A. Carlson, W.B. Dress, F.A. Grimm and J.S. Haggerty, J . Electron Spectrosc. Relat. Phenom. 10 (1977) 147. 228. D.B. Adams, J . Electron Spectrosc. Relat. Phenom. 13_ (1978) 123. 229. J . J . Pireaux. S. Svensson, E. B a s i l i e r , P.-A Malmquist, U. Gelius, R. Candano and K. Siegbahn, Phys. Rev. A _14 (1976) 2133. 230. R.G. C a v e l l , private communication. 231. Yu A. Kruglyak, H. Preuss and R. Janoschek, Teor.: Eksp. Khim. 7^ (1971) 291. 232. K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki and S. Iwata, "Handbook of Hel Photoelectron spectra of fundamental organic molecules", Japanese S c i e n t i f i c Societies Press, Tokyo, Halsted Press, New York (1981). 233. C. Baker and D.W. Turner, Chem. Commun. (1969) 480. 234. R.K. Thomas and H.T. Thompson, Proc. R. Soc. Lond. A 339 (1974) 29. 235. M.B. Robin, "Higher excited states of polyatomic molecules" ( V o l . 2), Academic Press, New York (1975).