Photoelectron Spectroscopy of Organornetallic Compounds

by

Jingcun Wu

Department of Chemistiy

Subrnitted in partial fulfilment

of the requirements for the degree of

Master of Science

Faculty of Graduate Studies

The University of Western Ontario

London, Ontario

May, 1997

8 Jingcun Wu 1997

National Library )*I of Canada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395, rue Wellington Ottawa ON K I A ON4 Ottawa ON K 1 A ON4 Canada Canada

Your hb Votre relerence

Our file Notre réfdrence

The author has granted a non- L'auteur a accordé une licence non exclusive licence dlowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sel1 reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la fome de microfiche/film, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or othenvise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

ABSTRACT

High resolution gas phase photoelectron spectra are reported for a series of

trimethylphosphme substiMed twigsten carbonyls and cyclopentadienyl derivatives of indium

(CpIn) and thallium (CpTI). The advantages of using monochromatic synchrotron radiation

PES for studying the electronic structure of organometallic molecules are demonstrated.

For esch of the substituted tungsten complexes, al1 electronic levels fiom valence to

inner-valence end core levels can be studied in one spectrurn with high resolution. The inner

valence and core level spectm can be interpreted based on cornparison with published results.

Better resolution has been achieved in our newly obtained He 1 spectra of the valence level

and W 5d regions of these complexes. Spin-orbit splittings, ligand field splittings, and

vibrational structures are observed in the spectra of both W Sd and W 4f regions. Ligand

field splittings on both the W 5d and W 4f levels increase in the order of W(COI6 = foc-

W(CO),(PM%), 4 W(CO),PMe, s cis-W(CO),(PM%), < Irm-W(CO),(PMeJ,. Because

phosphine is a stronger o donor and weaker K acceptor than CO, al1 the metal and ligand

orbitals shift with different degrees to lower energies when CO is substituted successively by

phosphine. Linear binding energy shifl trends are found in both core and valence levels of

tungsten and phosphorus ionizations, which confirm the ligand additivity predictions for these

complexes. The corevalence ionhion correlation principle can be illustrated by comparing

the binding energy SM data between core and valence levels. The phosphorus 2p spinsrbit

components of the phosphine substituted complexes have been resolved for the first tirne.

To resolve the major controversial among theoretical treatrnents on the electronic

structure and bonding of CpIn and CpTl, a photoelectron spectroscopie study with variable

photon energies has been carried out. The experimental results, especially the variations in

band intetlsity as a function of photon energy, confirm our assignments of the spectra which

are in good agreement with the results of both previous PES studies and Xa-SW and SCF

caîculations. Due to the high resolution of our spectra, the vibrational structure in the lowest

ionizations has bem resolved, and the broadening on the metal d levels caused by ligand field

spiittings has been observed. In addition, the shake-up structures of metal d levels and core

level TI 4f have been studied for the first tirne with synchrotron radiation.

ACKNOWLEDGEMENTS

1 w d d üke to take this oppomuiity to thank my supervisors, Dr. G.M. Bancroft and

Dr. RJ. Puddephatt, for th& guidance, encouragement, patience and fnendship, which have

made my study and research here possible.

1 am very pitefiil to Xiaorong Li and YongFeng Hu for their continued assistance in

my research, to Doug Hainine for his technicd assistance with the ESCA photoelectron

spectrometer, and to Kim Tan for his support with the synchrotron radiation experiments.

1 would also ike to express my Sncere thanks to my fnends and CO-workers for their

support: DrXping Zhang, Dr. Sam Choi, Dr. Hilary Jenkins, Dr. Lou Rendina, Dr. Mike

Scaini, Mike Irwin, Greg Spivak, Joshi Kuncheria, Daniel Legrand, Marina Fuller, Greg

Canning, Geoff Hi& Kim PoUard, Mary-Anne MacDonald, Wei Hong, Jayasree Sankar, Cliff

Baar, and Michael Janzen. It is these fiiends, CO-workers and the supervison who have

aeated and maintained the acadernic and fnendly atmosphere which have made my work in

Western rewarding and enjoyable.

1 thank Dr. P.A.W. Dean, Dr. M.J. Stillman, Dr. N.C. Payne, Dr. N.S. McIntyre,

Dr. RJ. Puddephatt, Dr. W.N. Lennarâ, Dr. O.L. Warren, Dr. K. Grifnths, Dr. J.G. Shapter,

Dr. A. Bassi, Dr. M.S. Workentin, and Dr. D. McConville for their instructions. 1 thank

Dr. T.K. Sham, Dr. R.K. Chan, Dr. M.C. Usselman, Dr. R.G. Kidd, Dr. J.D.Talman, and

Dr. P.A.W. Dean for their advice.

1 am gratefiil for the h c i a l support provided by the University of Westem Ontario

over the term of my studies.

Finaüy, 1 wouid like to tbank rny parents and my younger sister for their understanding

and support, with special thanks to my wife Chunning Li for her love and encouragement.

This thesis is dedicated to my dear father Yuguo Wu and my late mother Qinxiou Li.

TABLE OF CONTENTS Page

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CERTIFICATE OF EXAMINATION ii

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ABSTRACT iii

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ACKNOW LEDGEMENTS iv

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TABLE OF CONTENTS v

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LIST OF TABLES vii

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LIST OF FIGURES vüi

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LIST OF ABBREVIATIONS x

CHAPTER 1 Introduction

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Introduction 1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Gas Phase Photoelectron Spectroscopy 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1. Basic PMciple 2

. . . . . . . . . . . . . . . . . 1.2.2. Studies of Organometallic Compounds by Gas Phase PES 6

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ionkation (or Binding) Energy Trends 6

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Splitting Effects and Fine Stnicture 8

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shake-up and ûther Effects 12

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Band Intensities 13

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3. Conclusions 18

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3. Outline of the Thesis 18

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4. References 20

CaAPTER 2 Experimental

. . . . . . . . . . . . . . . . . . . 2.1. Preparation, Rirification, and Introduction of Sarnples 23

. . . . . . . . . . . . 2.2. Recording the Photoelectron Spectra with Helium Light Source 25

. . . . . . . . . . 2.3. Recordhg the Photoelectron Spectra with Synchrotron Radiation 28

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. References 30

CHAPTER 3 Photoclectron Sptctra of Trimethylphosphine Substituted Tungsten

Carbonyls

3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.2. Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.3. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.3.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.3.2. ValenceLevelWSdandCoreLevelW4f . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.3.3. Higher Energy Spectra and Phosphorus 2p Bands . . . . . . . . . . . . . . . . . . . . . . 55

. . . . . . . . . . . . . . . . . 3.3.4. High Resolution Photoelectron Spectra of W(CO), NBD 60

3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 61

3.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

CHAPTER 4 Photdectmn Spectra of Cydoptntadicnyl Derivatives of indium(0 and

Tha l l ium~

4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.2. Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.3. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.3.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.3.2. Variable Energy Photoelectron Spectra of Cpln and CpTl . . . . . . . . . . . . . . . . 72

4.3.3. The Shake-up Satellites of the Metal d Levels and Tl 4f Bands . . . . . . . . . . . . 88

4.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

4.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CaAPTER 5 Conclusions 104

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . APPENDIX 106

VITA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

List of Tables

Table 2-1

Table 2-2

Table 3 - 1

Table 3-2

Table 3-3

Table 3-4

Table 4- 1

Table 4-2

Table 4-3

Table 4-4

Table 4-5

The sublimation temperatures melting points, and references of

the organometaflic compounds studied in this work . . . . . . . . . . . . . . . 25

Worbg parameters for recording the PE spectra of the studied

compounds with helium light source . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Binding energies &), widths (Wd, and assignments of the

. . . inner-valence and core level spectra of W(CO),(PMe& (n = 1 - 3) 4 1

Band positions (eV), widths (eV), assignments, spin-orbit coupling

constants (0, ligand field splittings (A = b, - e or ba - e,), average

binding energies (eV), and their shifts (eV) relative to W(CO),

. . . . . . . . . . . . . . . . . . . . . . . . in W Sd spectra of the listed complexes. 46

. . . . . . . . . . Fitting parameters of W 4f spectra of the listed complexes 50

Phosphorus 'lone pair' or o(W-P) and phosphorus 2p ionizations

in W(CO),(PMeJn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

Binding energies (4) and relative intensities (Ir) of the peaks in CpIn

by He 1, He II, and SR (at 80 eV) PES and calculated binding

energies (4) and eigenvalues (q) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Binding energies (&) and relative intensities (Ir) of the peaks for CpTl

by He 1, He II, and SR (at 80 eV) PES . . . . . . . . . . . . . . . . . . . . . . . . . 77

Binding energies (4) and widths (FWtPul) of metd nd bands

for TU(, TlCp, and InCp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Derived crystai field parameters (eV) for TlX, WI, TICp, and InCp . . . 93

Shake-up energies (A) and widths of CpIn and CpTl . . . . . . . . . . . . . . . 95

List of Figures

Figure 1-1

Figure 1-2

Figure 3 - 1

Figure 3-2

Figure 3-3

Figure 3-4

Figure 3-5

Figure 3-6

Figure 3-7

Figure 3-8

Block diagram showing the arrangement of the principal parts of

a photoelectron spectrometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Photoelectron spectra of W(C%, (a) broad-scan spectrum with

synchrotron radiation source; (b) high resolution spectrum of

W 5d region 4 t h He 1 source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Broad-scan PE spectra of (a) W(COl6, @) W(CO),PMq,

(4 cis-W(C0)4(PM4%)2, (4 tr~-W(C0)4(PM%),,

(e)/ac-W(CO),(PMe&, and (f) W(CO),NBD . . . . . . . . . . . . . . . . . . . 40

He I valence level spectra of (a) W(CO),PM%,

@) trans-W(CO)4(PM&, cis-W(C0)4(PM&

and (d) W(CO),NBD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

He 1 spectra of W 5d levels in (a) W(CO),PM%,

(b) C~.S-W(CO)~(PM~J, (c) ~~L~zs-W(CO),(PM~),, (d) W(CO),NBD . . 45

High resolution W 4f core level spectra of (a) W(CO),,

@) W O ) , + W(CO),PM%, (4 cis-W(CO)4(PM%)*,

(dl ~~s-W(COX(PMs) , (e)fac-W(COX(PMe,)3,

and ( f ) W(CO)4NBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

A diagram showing the correlation between the ligand field

. . . . . . . . . . . . . . splitting ofW 5d bands and the width ofW 4f bands. S i

Shift c o m p ~ s o n diagram for tungsten and phosphorus binding

energy shifts: W 5d (valence), W 4f (core),

. . . . . . . . . . . . . . . . . . . . . . . . . P 'lone pair' (valence), and P 2p (core) 52

Core - valence shifi correlation for tungsten and phosphorus

. . . . . . . . . . . . . . . . ionizations: W 5d - W 4 î and P 'lone pair' - P 2p 53

High resolution PE spectra: (a) broad-scan of c~s-W(CO)~(PM~&

at 1 00 eV, (b) phosphorus 2p bands in cis-W(CO),(PMe&

at 152 eV, (c) W 4f bands and the second order bands of

Figure 4- 1

Figure 4-2

Figure 4-3

Figure 4-4

Figure 4-5

Figure 4-6

Figure 4-7

Figure 4-8

Figure 4-9

Figure 4-1 0

Figure 4-1 1

Figure 4- 12

Figure 4- 13

Figure 4- 14

phosphorus 2p at 101 eV, and (d) W 4f bands and the second

order bands of phosphorus 2p at 102 eV . . . . . . . . . . . . . . . . . . . . . . . . 59

High resolution broad-scan photoelectron spectrum of InCp at 80 eV . . 73

High resolution broad-scan photoelectron spectrum of TlCp at 80 eV . . 74

Valence level PE spectra of (a) InCp (He I), @) InCp (He II),

(c) TlCp (He I), and (d) TlCp (He II) . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Variable energy photoelectron spectra of InCp at (a) 80 eV,

(b) 130 eV, (c) 140 eV, (d) 150 eV, (e) 160 eV, ( f ) 180 eV . . . . . . . . . 78

Variable energy photoelectron spectra of TlCp at (a) 80 eV,

@) 130 eV, (c) 140 eV, and (d) 160 eV . . . . . . . . . . . . . . . . . . . . . . . . 79

Photoionization cross section for atornic C 2s, C 2p, In Ss, In 5p,

In 4d, Tl 6s, Tl 6p, and Tl Sd subshells . . . . . . . . . . . . . . . . . . . . . . . . . 80

Variation in relative intensity of band A and band B as

a function of photon energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Variation in relative intensity of band C and band D as

a fùnction of photon energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

Variation in relative intensity of band C, D and E as

a fùnction of photon energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

PE spectra of metal d region in (a) InCp (He II),

@) InCp at 70 eV (SR), (c) TlCp (He II), and

(d) TlCp at 80 eV (SR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Variation in relative intensity of the bands as

a function of photon energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Shake-up bands of In 4d and Tl 5d regions of (a) InCp and (b) TlCp . . . 94

Variation of the intensity ratio of metal d bands (F+F') with their

shake-up bands (H,+H,') as a function of photon energy . . . . . . . . . . . . 96

Photoelectron spectrum of Tl 4f region in TlCp obtained at 280 eV . . . 97

List of Abbrcviations

PE

PES

SR

UPS

XPS

eV

FWHM MO

HOMO

A 0

BE (or 4,) E

BR

W C )

SCF

Xa-SW

PSD

CP

CpIn (or InCp)

CpTl (orTlCp)

Me

PM% NBD

photoelectron

photoelectron spectroscopy

synchrotron radiation

ultraviolet photoelectron spectroscopy

X-ray photoelectron spectroscopy

electron volt (8065.73 cm -')

full width at baif the maximum intensity

molecular orbital

highest occupied molecular orbital

atornic orbital

binding energy

ionization energy

branching ratio

spin-orbit coupling (splitting)

self-consistent field

Xa scattered-wave

position sensitive detector

q5-cyclo pentadienyl

cy clopentadieny lindiumo

cyclopentadienylthallium(I)

met hyl

trirnethylphosphine

norbomadiene

Chapter 1

Introduction

1.1. Introduction

The chemistry of organometallic compounds is fundamental to many organic and

inorganic syntheses, cataiysis and surface reactions, and bioinorganic processes (or biological

systems). In order to understand this chemistry in ternis of the basic electronic structure

mon thaî control electron distribution, bonding, geometry, and reactivity or stability of the

cornpoundq many experimmtal and theoretical studies by photoelectron spectroswpy (PES)

have been camed out since the pioneering works in 1960.' in the past, photoelectron

spectroscopy (PES) mainly relied on laboratory light sources and was classified into two

areas based on the sources used: (1) molecular photoelectron spectroscopy, or more

cornmoniy cailed, ultraviolet photoelectron spectroscopy (LPS), and (2) X-ray photoelectron

spectroscopy (XPS). UPS is used to study the valence bands. The most useful and

cornrnonly utilized UV light sources are He I (2 1.2 eV) and He Il (40.8 eV) resonance lines.

One of the advantages of using üPS is its high resolution. The variations of band shape and

fine structure due to various spiitting effècts and vibrational couplings can be observed in the

valence spectra of many molecules, * S U C ~ as the W Sd spectra of W(CO),(PMq), in this

study. XPS is usuaily used to study the core level binding energies of atorns in molecules.

The lllnited resolution of X-ray sources bas restricted its application for valence band studies.

The most versatile photon energy source is synchrotron radiation (SR), which has the

2

advantages of high htensity, high resolution (using a monochrornator), continuously tunable

radiation, and wide spectral range covering fiom the vacuum W to soft X-ray regions. In

the last ten years, monochromatic synchrotron radiation (SR) has been used in combination

with He I / He II photon sources for gas phase photoelectron spectroscopic studies of

inorganic and organometallic compounds, in which information about both energy and

intensity, as well as fine structure can be obtained. Such studies have greatly increased the

power of PES. Photoelectron spectroscopy has now become one of the most direct

experimentd methods for probing the energy levels of the electrons in substances, and

characterizing their electronic str~cture.'~ ' This thesis will focus on the gas phase

photoelectron spectroscopic studies of organornetallic compounds.

There have been a lot of PES studies on transition metal carbonyls and their

denvatives, possibly due to their typical synergic bonding characters and high volatility.'* ' The Group lII cyclopentadieny1 derivatives MC& M = In or Tl, are of interest because they

are rare examples of 'halfsandwich' organometallic rnole~ules.~ In this work, the electronic

structure of a series of phosphine substituted tungsten carbonyls and the cyclopentadienides

of indium (I) and thallium (1) are studied with a combination of synchrotron radiation and

He 1 / He II photoelectron spectroscopy. Special attention and interest are focused on

studying the effects of ligand replacement on the PE spectra of the phosphine substituted

tungsten carbonyls.

1.2. Gis Phase Photoelectron Spectmscopy

1.2.1. Basic Principle

When photons of sutncient energy interact with a molecule, ionization can occur with

ejection of electrons (so-ded photoelectron).

M+hu- .M++e ' (1)

The photon energy, hu, can be transferreû to an electron, enabling it to overcome the electron

binding energy, E, (the force which binds the electron within a molecule), and giving it kinetic

energy, E, . Based on Einstein's photoeiectric effect:

& = h u e & (2)

Since the photon energy (hu) is known in a certain photoelectron experiment, and the

kinetic energy (&) is measured accurately by an electron energy analyzer (see Figure L I ) ,

the electron bhding energy (4) can be detennineâ by experiment (if the ho is large enough,

al1 electrons fiom valence, i~er-valence and core levels of a molecule can be ejected, and

their 4,'s can be detennined by the photoelectron experirnent). Photoelectron spectroscopy

(PES) is an expenmentai method for studying these photoelectrons, their energy levels and

their relationship with the electronic structure of a molecule.

A photoelectron (PE) spectrum consists of a plot of the 4, or E, versus intensity in

electron counts per second (or the number of photoelectron with a certain energy), which

represents the energy distribution of photoelectrons in a molecule. According to convention,

the spectra are ploned with E, increasing fiom lefi to right, and E, increasing fiom right to

left (see Figure 1-2). A PE spectrum can be relatd to the molenilar orbital energy pichire

of a molecule by K W P ~ Q I ~ S ' %wem ' which States that the ionization energy (E) is equal

to the negative of the self-consistent field (SCF) orbital energy (-ej ).

IE = -ej (3)

C O M P U T E R

RECORDER r l P H O T O N T A R G E T bI E L E C T R O N E N E R G Y 1 SOURCE C H A M B E R A N A L Y Z E R

F-1 DETECTOR

Figiisc I - 1. Block diagrani sliowiiig tlir arrangement ol'tlir priiicipal parts of r7 photoelectron spectronieter.

Figiire 1-2. Photoelectron spectra of W(CO),. (a) broad-scan spectriirn witli synchrotron radiation source: ( b ) Iiigli resoliition sprctsiini of W 5d regioti witli He 1 soiirce.

5

Although various approximations are involved in this theorem and quantitative predictions

ofIE are not accurate, this theorem has been used widely in the interpretation of PE spectra

(such as in predicting the number, type and energy of primary PE spectral bands of a

moleaile). The most signaficant deduction from the correlation of Kmprn~ l l~ ' Theorem and

mo lda r orbital energy diagram of a molecule is that there is a one-to-one conespondence

between the primary band features of a PE spectrum and the occupied molecular orbitals of

a closed sheU moleaile. In its sirnplest form, one moldar or atomic orbital gives rise to one

spectral band, which provides us the best expenmental method of obtaining molecular and

atomic orbital binding energies. For example, in the PE spectnim of W(CO), " ( Figure

1-Za), band 1 ( at - 8.5 eV) aises from the ejection of electrons fiom the valence W 5d

orbital (i.e. the t, orbital, see the qualitative molecular orbital diagram for W(CO), in

Appendix A). Band 2 is due to ionizations of electrons fiorn mainly CO 5 0 and 1 n orbitals.

Band 3 is from ionization of CO 40 orbitals. Bands C, and C, are resulted from core level

W 4f electrons. The bands from S to D are rnainly ligand-based inner valence orbitals, and

they are ofien ditticult to interpret accurately for organometallic compounds because of the

orbit overlapping and shake-up interactions in t his region.

When analyzing the PE spectrum of a molecule, dl band features such as energy,

width, shape, resolved fine structure and relative intensity, as well as their changes with

different photon energies and chemical variations, should be taken into account.

Helium light sources can provide enough energy for ionization of the majority of

valence electrons. The high resolution of He 1 PES is critical for the studies of the valence

electrons in organometdic compounds.* Observations of resolved fine structure due to

6

vatious spiitting effects and vibrational stretchings have not only increased Our understanding

of moleailar electronic structure, but also greatly assisted Our assignrnent of PE spectra for

the studieà molecules (see Figure 1-Pb" and Chapter 3). However, in order to study core

level electrons and obtain both energy and intensity information about the ionization processes

of a molecule, the use of continuously tunable Synchrotron Radiation (SR) is ofien

necessary, and has proven to be very powerful for studies of the electronic structure of

organometdîic compounds in the last ten years (see the following section and references

cited).

1.2.2, Studies of Organometallic Compounds by Gas Phaae PES

As mentioned previously, al1 the characteristic features of a PE spectral band

(binding energy, width, band shape, resolved fine structure, and relative intensity) ought to

be considered when studying a PE spectrum, because al1 the features are sensitive to the

electronic structure of the m d d e . The ionization energies (IE 's) or binding energies (4,'s)

are most signifiant in terms of trends between related molecules. The other band features

can directly or indirectly reveal information about the electron localization on the metal and

ligands, the interactions between metal and Ligands (bonding, or nonbonding nature of the

orbitalq splittings and vibrations), and the different variations in photoionization cross section

between metal and ligand orbitais.

Ionization (or Binding) Energy Trends. P@s the most important use of PES

is to obtain IE's or l$'s of atornic and molecular orbitals based on equation 2 (6 = hu - Q.

Although the IE's or 4 ' s are characteristic of the electrons in a molecule which are

independent of the photon energy, they do Vary over a certain range depending on the

7

chernical variations on the molecule. The shift in an IE or E, between electronically or

chemically related molecules is an especially revealing and usefiil feature of the electronic

structure. In sorne cases. the correlation of IE 's or 4 ' s between related molecules cm assist

in the assignrnent of spectra, such as in the comparison of the IE's or 4 ' s of published

W(C% spectra with those of phosphine substituted tungsten complexes, as illustrated in

chapter 3. In other cases, the electronic perturbations caused by chernical group substitutions

produce identifiable IE or E, shifis which r evd the localized or delocalized character of the

electronic states and the 0uiuidity of charge in the system. In these cases, the IE 's or Eb7s cm

serve as a good maure of the substituent effects and can assist in monitoring the effects of

chernical variation on a molecule. For exarnple, the IE or E, shifls in both valence and core

levels of organometallic compounds can be very useful in studying the a-donor and rr-

acceptor properties of ligands, and the additivity of the o and x effects of ligands on a metal

tenter.' The ligand additivity rnodel states that valence metal orbital IE's or 4 ' s are shified

in a linear way as one kind of ligand in a moleaile is substituted successively by another kind

of ligand.' Although this rnodel was proposed initidly for valence rnetal ionizations, it has

proved to be valid for both valence and core level ionizations by several studies (see chapter

3 and cited references). For instance, the series of M(CO),(PMe&, complexes (M = Mo and

W, and n = O - 3) have been studied in detail (see chapter 3), in which both valence and core

metal IE's or &'s are shifteâ linearly to lower energies with each step of ligand replacement

because the total donor ability (O-donation minus n-acceptance) of PM% is greater than CO

(i.e. PM% is a stronger a-donor but weaker n-acceptor than CO). A third application of

ionization energy trends is baEed on the correlation of core and valence IE or E, shifts, which

8

allows the differentiation of the influences of bonding / overlap and charge potential

contributions to 1 . 5 or &S. The principle of core-valence ionization correlation states that,

when comparing huo related molecules, the binding energy shift of a nonbonding valence

orbital ofa certain atom is approximately eight tenths of the core binding energy shift

for that atom between the two molecules, i.e. = 0.8 &m. A value of'

aEMdma) /A&-) > 0.8 indicates the contribution of bonding /overlap interactions to the

valence SM. The reason why the core level E,, shift is larger than the valence E, shift is t hat,

according to ~ollf or Lichtenberger, the core &'s are mainly determined by the charge

distributions and relaxation energies, whereas the valence 4's are affected by these factors

as weii as aD aspects of chernical bonding (Le. valence &'s are more sensitive to the bonding

effects than core G's) Although organometallic compounds seldorn have the strictly non-

bonding valence orbital, the correlation values obtained from the above M(CO),(PMq),

complexes have been shown to be in reasonable agreement with the principle's predictions

of 0.8 i 0.1.

The examples of carbonyl and phosphine complexes mentioned above demonstrate

how PES can provide detailed and sometimes quantitative information on each individual

interaction of a iigand with a metal center. The interactions are observeci directly in terrns of

stabhtion or destabilization of the orbital IE 's. The interaction of the ligand a orbital, or

r orbital is a separate effect." The binding energy shüt (4) is dependent on the overall

charge potentiai on a orbital. We can apply these pnnciples to the studies of various metal-

ligand interactions in organometallic compounds. " Splitting Eff'ts and Fine Structure. PE spectral bands often exhibit fine

9

structure resulted fiom various splitting and vibrational effects which are associateci with the

molecuiar ion states. Even though such structure is often not well resolved, it can affect the

band shape. Such fine stnicture not ody provides detailed information about the molecular

electronic structure but also offers considetable help in the interpretation and assignrnent of

the PE spectra. However, in order to study these fine structure features, high resolution PES

is ofien needed.

Spin-Orbit Spliffing. We often see from the PE spectra that one band is split into

two. The major reason for this splitîing is due to the coupling of spin angular momentum (S)

with orbital angular momentum (L) (so-called spin-orbit coupling) based on J = L k S.

Removal of an electron fiom a filled p, d, and f orbital, leaving the orbital with an

unpaired electron, always gives a doublet in the PE spectrum. For example, the P 2p and W

4f doublets in phosphine substitutd tungsten complexes (sec chapter 3), and the In 4d, TI 5d

and Tl 4f doublets in InC5H5 and TlC5H5 (see chapter 4) have been resolved in this work

(because S = W 2 ; L = 1, 2, and 3, respectively for p, d, and f orbitais). The magnitude of

spin-orbit splitting is approximately proportional to the square of the atomic number of the

atom for the valence shells of a many-electron system;1° a larger splitting is expected for a

second or third row transition metal valence d orbital, and this is somethes called the heavy

metal effea.21' It is important to notice that for core levels, the spin-orbit splitting is not

chemicdly sensitive compared with that for valence levels. The splitting for a given atom

increases 60rn valence to a r e level " (e.g. see chapter 3 for the spin-orbit spütting values of

W 5d and W 4f).

The spborbit coupling theory was first used by Hal to interprete PE spectra of

10

transition metal systems, which 14 to a definitive assignment for the spectra of XRe(CO),

species? This theory has been shown to be very useful for the assignment of metal d orbitals

of the second or third row transition metal complexes.* " Ligmd Field Splitting. Ligand field effects on the metai centet can dso lead to

spiittings in metal orbitais, which are often shown as split or broadened spectral bands in the

PE spectra. Crystal field and ligand field theories have been used to account for the effects

on the PE spectra. l2

Previous studies have shown that the ligand o-donor and x-acceptor effects on the

valence d levels can be sepamed. The biiding energy shifis of the met al orbit als wit h ligand

substitutions depend on the total donor ability (a-donation minus x-acceptance) of the

substituted ligand relative to that of parent ligand, while the ligand field splittings of the metal

orbitals depend only on the relative x-acceptor ability of the substituted ligand and the parent

ligand. The magnitude of the spütting is proportional to the difference in x-acceptor abilities

of the two different kinds of ligands. This has been confirmeci by the study of a series of

LM(CO), complexes (M = Cr, Mo and W; L = PEt,, PM%, P(NMeJ,, P(OEt),, P(OMe),,

PF,)? The ability of the ligands to split the orbitals of the parent hexacarbonyl into the

e and 4 components decreases in the above order (fiorn left to right), showing that the

n-acceptor ability increases in the same order. This ligand field splitting effect will dso be

discussed in chapter 3 for compounds of W(CO),(PMq), where n = O - 3.

Ligand field splittings on core levels of main group compounds and metai surfaces

from photoelectron spectra have been reviewed.12 On core d levels of non-cubic compounds

such as (CH3)$d and TU[ (X = Cl, Br, and I), the d, level splits into two and the d, level

11

spüts into three. Simiiarly, for the metal 4f orbitais in complexes of Os(CO),L (L = CO, and

PMeJ," lSaD Re(CO)&(X = Re(CO), CI, Br, and I),% l5 and W(CO),(PM%)n (chapter 3),

the f,, level splits into three and the f , level splits into four. This latter effect results in

brodening of the spectra of metal 4f levels in these complexes. The crystal field Hamiltonian

for the ligands interacting with the core d or f levels is :

H = Cp [3L: - L(L+l)] +C:[35L: - 30L(L+ 1)L: -25L:- 6L(L+1)+ ~ L * ( L + I ) ~ ] (4)

where, the nonaibic C: term dominates over the cubic C," term. It is possible to diagonize

the Hamiltonian math and obtain five equations for the four unknown E, C:, Ct and the

spin-orbit splitting A.& This splitting effect wil1 be examined in more detail when discussing

the In 4d and Tl 5d spectra of InCp and TlCp compounds (see chapter 4).

fibrutionai Splitfiing. The principle of vibrational splitting in the valence PE spectra

has been described pre~iously.~ 'O Different electron transitions fiom the ground state of a

m o l d e (v = O for most molecules) to a series of vibrationai energy levels of the molecular

ion state, govemed by Franck-Condon rule, usually lead to the vibrational structure in both

valence and core level PE bands. For organometallic compounds, observation of well-

resolved vibrationai fine structure is not often possible due to the small metal-ligand

vibrational fkquencies, or because severai difEerent vibrational progressions are excited by

a single ioni~ation.~ Therefore, high resolution PES is required for studying these fine

structure featuns. For valence bands, vibrationai progressions are often related to ionizations

from bondhg or antibonding MO'S.' Even for unresolved bands, the band shape can indicate

to some extent the bonding nature of the vacated orbital, e.g. narrow bands are associated

with nonbonding orbitals and broad bands with bonding (or antibonding) orbitals. How ever,

the observation of core level vibrational splittings cm be best interpreted by considering the

core quivalent model, which states that when a core electron is removed fiom an atom or

molde , the valence electrons relax as if the nuclear charge of the atom had increased by one

unit. l3 According to this model," the properties of a molecular ion with a core hole are

approximated by the moldar with the Z+1 atom. For example, the core-equivalent species

for core ionized W(COI6 is Re(CO),'. ' In this work, aii the spütting e f f i s and fine structure features mentioned above have

been observed in the study of phosphine substituted tungsten carbonyls (see chapter 3).

Jahn- Teller SpIitting. In addition to the above splitting effects, sometimes Jahn-

Teller qlifting cm be important in yielding extra peaks in the valence band. The Juhn- Te ller

horem states that a non-linear molecule in a degenerate electronic state is unstable towards

distonions which remove the degeneracyZh 'O Jahn-Teller spiitting often fùrther splits a

spin-orbit split state. These splittings have been found in the PE spectra of metal d orbitals

of Fe(CO)514 and Os(CO),.'

Shabup and Othtr Eff'cfa Extra bands are sometimes seen on the low kinetic

energy (hi@ &) side of a core l t d or vaience ievei band, such as in the spectra of W(CO)6.'5

These bands, so-calleû shake-up satellites, are oflen broader than the main bands '' and can

be illustrateci by the foliowing processed6 As the result of a vacancy formed in a given

orbital due to ionisations, the electrons in the sarne orbital or other orbitals see a change in

the effkctive nucleat charge due to an alteration in the electron screening. This change in

effective nuclear charge can give rise to an excited state in which an electron may undergo

a transition to a discrete state - shake-up which is shown as low A?$ satellites or it may go to

the continuum (shake-on) state.

The shake-up peaks of metal d orbitals in InC,H, and TIC& have been seen using

high resolution PES with SR and wiU be discusd in chapter 4. The intensity of the shake-up

band S observed previously l5 in W(COI6 decreases wit h each CO being replaced by PM%,

which further confimis that this shake-up results fiom a CO valence band between

13 - 1 5 eV (see chapter 3).

The extra peaks created by other eff'ects, such as the "self-ionization" peaks (He*) and

the peaks excited by He II P(48.37 eV) or He Ii y (SI .O1 eV) satellite lines of He Il emission

are found in the He II spectra of InC& and TiC& (chapter 4). Most of these peaks overlap

seriously with the main bands of the samples, which make it difficult to assign the spectra

accurately. To overcome these problems and obtain the red spectra of sarnples, high

resolution synchrotron radiation PES is used in this study (which can get rid of the excitation

fiom He II satellite lines). The second-order ionization bands of P 2p spinsrbit components

are also observed in the phosphine complexes (chapter 3).

Band Intensities. In addition to the information mentioned above, a PE spectrum

cm also provide information about band intensities. Aithough IE's or 6 ' s are characteristic

of the shidied moleaile and are independent of the photon energy, the band intensities depend

on the angle of observation of the photoelectron with respect to the photon beam, the

polarization of the photons, and the photon energy. If the electrons are observed at an angle

8 to the direction of the electric vector of a plane polarized light beam, the intensity I(8) is

given by eq~a t ion :~~

I(0) = a/4n[l + (P/2)(3cos2 0 - 1)] (5)

14

where o is the total cross section integrated over al1 angles, and P, known as the anisotropy

parameter, is the only parameter required to descnbe the angular distribution of the

photoelectrons. However, for most studies on inorganic and organometallic compounds, the

measurements of band intensities are perfonned at a "magic angle" (8, = 0, = 8, = 54.7')

where the band intensity is directly proportional to the cross section ( the probabitity of

photoionization to an ion state is nonnally designated as photoionization or photoelectron

cross ~ection).~ The photon energy dependence of a band's intensity or cross section is

characteristic of the nature of the ionized (or vacated) molecular orbital (MO). This nature

depends on the types and the compositions of atornic orbitals (AO's) which make up the

MO." Thus, intensity or cross section studies cm provide more ~seful information about

molecular electronic structure and give a firm basis for PE band assignrnent.

The d y studies on band intensity were based on the cornparison of He 1 and He II

spectra of the studied molecules. It had been shown that the intensity of the valence bands

fiom ligand orbitals (e.g. the s and p orbitals of C, N, O, P, and S. especially C 2p orbital for

organometaiiic compounds) d e c f w with photon energy increasing fiom 2 1.2 eV (He 1) to

40.8 eV (He II), while the band intensity of valence d orbitals of most transition metals

inmeases drarnatically with photon energy changing within the same region. This empirical

relationship of relative He 1 and He II ionhation band intensities had been used to interpret

the valence band spectra of many organometaüic compounds." However, this method was

lllnited to onS, two separate photon energies, and in this energy region ( around 40 eV) there

exist other variation effects on band intensities, such as shape resonances and interchannel

cou pling which may lead to rnistakes in the assignrnent of spectra.' Therefore, in order to

15

obtain diable and definite assignments of PE spectra, the variable photon energy PES with

SR should be applied, which provides the intense tunable source of photons required to study

the continuous variation of atomic and molecuiar photoionization cross sections within a wide

range of photon energies.

The applications of variable energy PES in the interpretations of cross section

variations and in the PE band assignmnts rely on the Gelius model," which assumes that the

cross section of a molecular orbital (MO) is mainly determined by the atomic orbital

components of that molecular orbital,

q ZP,, oy0 (6)

where, qMo is the cross section of the jth MO, oAt0 are the atomic cross sections for al1

orbitals of atoms Ai in an MO, and PA, is the "probability" of finding in the jth MO an

electron belonging to the atomic Ai ohitai. This means, for organometallic compounds, the

cross section variation of non-bonding metal d and ligand carbon 2p orbitals behave like their

atornic countqarts, and those of the bonding MO'S behave in an intermediate manner. This

mode1 has been used successfully in many recent variable energy PES studies of

organometallic wmpounds. 'D ' Several characteristic features of MO'S can be obtained Born studying cross section

or intensity variations with photon energies. Most of these qualitative features can be

understood in terms of the interactions between the ionized (vacated) MO and the outgoing

electron wave.'

A Generd Trend in the Phtoionidon C m Section. A cross section is generally

highest near the ionkation threshold, and afler that, it decreases with the increase of photon

16

energy, such as the ligand C 2p orbital cross section. When photon energy Uicreases, the

kinetic energy of the electron increases and its wavelength decreases, the electron wave

b m e s more oscillatory, the positive and negative parts of the dipole math element with

the ionized orbital tend to canal one another which lead to the rapid decay in cross section."

However, for some metal d and f orbitals, there are other features superimposeci on the

generai decay of the cross section, such as delayed maxima, Cooper minima, resonance effects

and so on. These features are characteristic of these metal d and f orbitals.

Cwper Minima For orbitais whose radial wavefùnctions have a node (the number

of nodes = n 4- 1), there is a minimum in their cross sections- so-called Cooper minimum,

such as for the 4d and 5f transition metd orbitals. In contrast, other orbitals without a radial

node (Is, Zp, 3 4 and 4f) do not show Cooper minima, such as C 2p and Ni 3d cross sections.

This Cooper minimum can be illustrated by a change in the phase of the initial state wave

function which results in cancellation of the electron dipole transition moment to the final

state wavefùnction at some energies. In previous studies, Li has used the Cooper minimum

in the 4d ionization of Pd to assign the PE spectnim of Pd(q-Ca,), and show that the ion-

state ordenng of this cornplex is different fiom that of ~ i (q -C~H~) , . l ~

Delayed Mmima. Orbitals with high angular momentum, such as the d and f

orbitals of transition metals, of€en show maxima in their cross sections some way above

threshold. This is in contrast to s and p orbitals whose maxima (if they have one)

tend to be near their threshold. The delayed maxima in the cross sections of d and f orbitais

can be explained by the larger centrifiigal barrier effects of the high angular momentum

electrons?

17

Resonunce Eflects. For nd (or nt) orbitals of transition metai complexes, great

cross d o n or band intensity changes may be obsaved at the corresponding i ~ e r np (or nd,

for nd-nfresonance) threshold. Such great variations in cross section or band intensity result

fiom indirect ionizations or resonant excitations, which can be illustratecl by a two-stage

process:

n p 6 n d x - n p S d n l - n p 6 n d x - ) + e - or .

dionfx ,nd9nf*' -&'Onf'-1 + e-

First, an inner np (or nd) electron is excited to one of the empty nd (or nf) orbital.

Subsequently, an electron Ws back into the np (or nd) hole and an outer nd (or nf ) electron

is ionid - the d e d super Coster Kronig (SCK) transition. Therefore, an enhancement

in the nd (or nf) orbital cross section may be observed in the region of the np (or nd)

absorption.

The large metd np - nd resonant effects observed in some organometallic systems

have assisted the definitive assignments of metal d-based ionkations in the valence spectra.)

For example, these resonance have aided the assignments of valence d bands in the complexes

of CpM(CO), (M = Mn and Re)" and Os(CO),PM%.

The theories and appiications of these cross section features in variable energy

photoeltxtron spectroscopie studies of transition metal systems have been discussed in detail

in recent review articles.' In surnmary, the cross sections or intensities of d and f PE bands

of gas-phase molecules show a number of characteristic features, including maxima at

relatively low photon energies due to centrifûgal banier effkcts, resonant effects (Le.

pronounced maxima and minima) at photon energies corresponding to resonant absorption

18

by inner shell eiectrons, Cooper minima at high photon energies for ionization from orbitals

with radial nodes, and shape resonances. These features lead to highly different photon

energy dependences between the cross section or intensity of rnetal d (or f) bands and that

of ligand-baseû bands. Studying these difEerent features between metal and ligand bands can

not only lead to firm experimentaiiy based band assignments, but also can increase Our

understanding of the interactions between metal and ligand orbitals within a molecule. The

achievements in this area have been dernonstrated by the recent studies of inorganic and

organometallic compounds with variable energy PES? ' 1.2.3. Conclusions

Studying organometallic compounds by gas phase PES is not only desirable but is also

possible since volatile compounds cm be obtained either cornmercially or by synthesis. %y

synthesizing a group of compounds (such as by keeping the same metal center and changing

the nurnber or type of the ligands, or by keeping ligand the same but varying the metals), the

ligand substitution effects on the metal center (chapter 3) or the periodic trend of rnetal

ionizations 4b can be studied. Wth the combined use of helium and synchrotron radiation

sources, PES has become one of the most important and direct experimental methods for

studying the electronic structure of organometaîüc compounds.

1.3. Outline of the Thmis

This thesis is composed of 5 chapters. This first chapter offers a general introduction

of gas phase photoelectron spectroscopy (PES), including the basic princi ple and some

important applications of gas phase PES in the study of organometallic compounds. Chapter

2 descr i i the experirnental methods and working conditions used in this study of work. In

19

chapter 3, high resolution gas phase photoelectron spectra are presented for a series of

trimethylphosphine substituted tungsten carbonyls. Spinsrbit splittings, ligand field effects

and vibrational stmctures are observed in the spectra of both W 5d and W 4f regions. The

linear binding energy difi in both valence W 5d and core W 4f levels confirms the validity of

ligand additivity principle. The wre-valence ionization correlation can be illustrated by

comparing the binding energy shift data between valence and core ionizations. Chapter 4

discusses the study of InC,H, and TICsH5 by PES. Our He 1 and He II spectra show better

resolution than the previously reported. With synchrotron radiation PES, the shake-up

satellites of metal d orbital are observed which are included in the broad-scan spectra of

I n C a and nC&. The variation in the relative intensities of the spectral bands can be seen

in the variable energy PE spectra. For the first time, the Tl 4f spin orbit components are

resolved in this study. Chapter 5 provides some of the conclusions obtained from this study.

Referenca

(a) Lloyd, D. R.; Schlag, E. W . Inorg. Chem. 1969, 8, 2544. @) Evans, S.;

Green, J. C.; Green, M. L. H. : Orchard, A. F.; Turner, D. W . Dismss. Fwarioy Soc.

1969, 47, 112.

(a) Lichtenberger, D. L.; Kellogg, G. E. Acc. C h . Res. 1987, 20, 379. @)

Eland, J. H. D. "Photoeiechon Spctroscopy, An Inrrodirclion to Ultr~viofet

Photoeiectron Spctroscopy in the Gus Phase ", Buthenvorth & Co (Publishers)

Ltd., Second Edition, 1984.

(a) Green, J. C . Acc. Chem. Res. 1994, 27, 13 1. O>) Green, J. C . Encyciopedia of

Inorg. Chem. 1994, 6, 3257. (c) Bancroft, G. M.; Hu, Y. F (Review, to be

published).

(a) Hu, Y. F. Ph. D. Dissertation, The University of Western Ontario, London,

Ontario, Canada, 1996. (b) Li, X . Ph. D. Dissertation, The University of Western

Ontario, London, Ontario, Canada, 1995. (c) Yang, D. S . Ph. D. Dissertation, The

University of Western Ontario, London, Ontario, Canada, 1989. (d) Dignard, L. M .

Ph D. Dissertation, The University of Western Ontario, London, Ontario, Canada,

1986.

(a) Yarborough, L. W.; Hall, M. B. I m g . C'Che. 1978,17,2269. @) Bursten, B. E.;

Darensbourg, D. J.; Kellogg, G. E.; Lichtenberger, D. L. lnorg. Chem. 1984, 23,

4361. (c) Banaoft, G. M.; Dignard-Bailey, L.; Puddephatt, R J. Inorg. Chem. 1984,

23, 2369. (d) Bancroft, G. M.; Dignard-Bailey, L.; Puddephatt, R J. Inorg. Chem.

1986. 25. 367. (e) Pudde~hatt. R J.: Dinnard-Bailev. L.: Bancroft. G. M . Inorp.

21

Chim. Ac& 1985,96, L91. (f) Lichtenberger, D. L.; Kellogg, G. E.; Landis, G. H.

J. Chem. Phys. 1985, 83, 2759. (g) Lichtenberger, D. L.; Kellogg, G. E. Acc.

Chem. Res. 1987,20,379. (h) Hall, M. B. J . Am. Chem. Soc., 1975,97,2057.

(a) Craddock, S.; Duncan, W.; J. C h . Soc., F w a h y Trmts, 2, 1978, 74, 194.

@) Egdell, R. G. ; Fragala, 1.; Orchard, A. F. J . Electron S'ctrosc. Relat. Pherom.

1978, 14, 467.

Koopmans, T. Physica, 1934, 1, 104.

Bursten, B. E. J. Am. Chem. Sm. 1982,104, 1299.

(a) Jolly, W. L. AccChem. Res. 1983, 16, 370. @) Beach, D. B.; Jolly, W. L.

Inorg. Chem. 1986, 25, 875. (c) Jolly, W. L. C h . Phys. Lett. 1983,100, 546.

(d) Jolly, W. L.; Eyermann, C. J. J. Phys.Chenr. 1982,86,4834. (e) Jolly, W. L.

J. Phys. Chem. 1983,87,26.

Eland, J. H. D. " Pbtmlectron Spctroscopy"; John Wiley & Sons: Toronto, 1974,

p. 132.

B6hq M. C.; Gleiter, R. Angew. Ch., Int. M. Engl. 1983, 22, 329.

Bancrofl, G. M.; Tse, J . S . Comments Inorg. C'hem. 1986,5, 89.

(a) Jolly, W. L.; Hendrickson, D. N. J . Am. Chem. Soc. 1980, 92, 1863. (b)

Agren, H.; Selander, L.; Nordgren, J.; Nordling, C.; Siegbahn, K.; Mulier, J.

Chem. Phys. 1979,3 7, 16 1.

Hubbard, J. L.; Lichterberger, D. L. J. Chem. P h p . 1981, 75, 2560.

Hu, Y. F.; Bancroft, G.M.; Liu, 2.; Tan, K . H. Inorg. C h . 1995,34, 37 16.

Carlson, T. A " Photoelectrm and Auger Spctroscopy ", Plenum Press, New York,

22

1975.

(a) Gelius, U. "Electm Spectroscopy "; Shirley, O. A., Ed.; North Holland:

Amsterdam, 1972; pp3 1 1 . (b) Gelius, W.; Siegbahn, K. Furauby Disnrîs. Chem. Soc.

1972, 54, 257. (c) Bancroft, G. M.; Malmquist, P.-A.; Svensson, S.; Basilier, E.;

Gelius U.; Siegbahn, K. Inorg. C h 1978, 17, 1595.

(a) Yeh, J. I.; Lindau, 1. At. Data Nucl. Data Tables, 1985, 32, 1 . @) Green, J. C.

Stmct. BBonng (Berlin), 1981,43, 37. (c) Cowley, A. H. Prog. horg. Chem. 1979,

26, 46.

(a) Li, X. R.; Bancroft, G. M.; Puddephatt, R. I.; Hu, Y. F.; Liu, 2.; Sutherland,

D. G. J.; Tan. K. H. J. Chem. Soc., Chem. Commzin. 1993, 67. (b) Li, X. R.;

Bancroft, G. M.; Puddephatt, R. J.; Hu, Y. F.; Liu, 2.; Tan. K. H . Inorg. Chem.

1992,31,5 162. (c) Li, X. R; Bancroft, G. M.; Puddephatt, R. J.; Liu, 2.; Hu, Y. F.;

Tan, K. H. J. Am. Chem. Soc. 1994,116,9543.

Hu, Y. F.; Bancroft, G. M.; Davis, H. B.; Male, J. 1.; Pomeroy, R. K.; Tse, J. S.;

Tan, K. H. Organometallics, lW6,iS, 4493.

Chapter 2

Expet-imental

2.1. Pnpantion, Purification and Introduction of Samples

The compounds W(CO),PM%, W(CO),NBD (NBD = norbomadiene), cis-

W(CO),(PMe& fim-W(CO),(PMe& and ~oc-W(CO),(PM%)~ were prepared and purified

by methods in the literature,' with some modifications. For example, a column separation

method was used for the purification of cis-W(CO),(PMeJ, rather than the sublimation

method, ' because small arnounts of cis-W(CO),(PMe& could be converted to the tram-

isomer dwing the sublimation process. For the same reason, the temperature should be kept

as low as possible in the process of introducing the sarnple to the gas ceIl for evaporation in

the photoelectron spectrometers. A culumn separation process was also used to purif) ~rans-

W(CO),(PMe& afler the cis- to tram- isomerization reaction was complete. Two eluents

were used in order to separate the tram-isomer from the remaining cis-isomer and decalin

(the heating solvent). The latter was difncult to remove by evaporation ' because of the high

boiling point (> 1 60°C).

InC,H, and TIC5H5 were obtained cornrnercially fiom Strem Chernicals and were

purified by vacuum sublimation.' The extremely air sensitive and moisture sensitive

compound InC5H5 was handled in a dry nitrogen atmosphere (a dry box, schlenk tube, and

a vacuum line were required). The relatively stable T l C a could be handled in the air for a

short period of tirne, but it should be kept cold for future use.'

The pur@ of sarnples was confirmed by known methods (melting points, IR,

24

MS and Chromatography).'* ' Special precautions were taken for transportation of the

sensitive samples to the remote synchrotron radiationcentre. It could be carried safely on a

long journey by pachg the sarnple tubes Ued with dry ice together in a Dewar covered with

cotton wool. The needle valves of the tubes were not allowed to touch the dry ice since that

might break the seal of the "0" ring.

AU the sarnples were introduced into the gas ce11 of the spectrometer directly via the

heatable probe, except for InC5H5 which was introduced under dry nitrogen (the operating

methods for air sensitive samples were describeci in detail elsewhere. '* ' The less volatile

samples required heating in order to generate enough vapor pressure. The pressure in the

sample chamber was controlled to be - IO*' Torr, and the pressure in the gas ceIl was

around 5 x 10" Torr.

It has been show- by our scpniments that a good sample, for gas phase photoelectron

spectroscopie stuclies, should be volatile and stable (to heat and light of the light source) or

easy to volatilize by heating without any decomposition. In other words, the sarnple should

have low sublimation temperature and high melting point or high decomposition temperature

in order to obtain intense and reliable photoelectron spectra. In addition, it has been found

haî, the lmger the dfereence between the sublimation tempercture d the meiting point or

the &composition temperature of the sample , the better the sumple is for gas phase

photuelecrm pc trawpic &es. For example, W(CO), is an excellent compound for this

study since it can be evaponited without heating under vacuum condition. TIC& is another

very good sample because it can be volatilized easily by heating without any sample

decomposition (very high m.p. 300 O C ) . inC5H, is volatile under vacuum even at room

25

temperature, but special skills and equipments are required to handle this air and rnoisture

sensitive compound. The sublimation temperature of jac-W(CO),(PMq), is very high

(- 180 OC), therefore, it is difncult to record high quality spectra for this kind of compounds.

Compounds W(CO),PM%, irans and cis isomer of W(CO),(PMe&, and W(CO),NBD al1

show small differences between their sublimation temperatures and melting points. To get

reüable spectra of these compounds the sarnple heating temperatures must be controiîed

strictly below their melting points, because when the temperatures reach their melting points,

big fluctuations in the sample pressure &en occur which can lead to deformation of the

spectra and ( in the worst case) even shutdown of the pumps or whole instruments. The

sublimation temperatures and melting points of the organometallic compounds studied in this

work are summarized in Table 2- 1.

Table 2- 1. The sublimation temperatures, melting points, and references of the

organometallic compounds studied in this work

Compound Sublimation temperature Melting point Reference

w~5G&) 40 k 10 OC 2

Tl(%H5) 90 * 10 OC 300 "C 2

w(Co), 30 10 OC 170 OC 12

w(cO)phle, 45 10 "C 56 "C 1

fr41tsœw(c0)4(PM% )2 70 IO "C 82 O C 1

cis-W(CO),(PM% ), 90* 10°C 108 OC I

W(CO),NBD 80* 10°C 90 OC 1

fww(co)3(PM% )3 190A 10°C 300 "C 1

2.2 Recordiag the Pbotoclcetron Speetra with Heliurn Light Source

26

In this work, the He 1 and He il photoelectron spectra of the studied organornetallic

compounds were obtained by using a modified McPherson ESCA-36 photoelectron

spectrometer. This spectrometer has been described in detail previously.' Briefiy speaking,

it has 6 major components: (1) a helium hoUow cathode discharge lamp which can generate

He 1 (21.22 eV) and He II (40.81 eV) resonance lines; (2) a vacuum (sample) charnber with

a Edwards Senes 100 diffusion pumping system and a gas cell; (3) a hemispherical sector

electrostatic analyzer with a 36 cm radius and a 10 cm gap between the spheres. The

photoelectrons with difEerent kinetic energies can be separated and analyzed when passing

through this analyzer. It is pumped by a turbomolecular pumping system; (4) a channeltron

detector which rneasures the intensity of the photoelectrons with certain (or selected) kinetic

energy; (5) a Zenith 2-158 PC microcornputer, which collects the electron signals Born

detector and controls the kinetic energy scanning of the electron energy analyzer; and (6) a

heatable sample probe through which solid sample can be evaporated and introduced to the

gas cell.

Table 2-2. Working parameters for recording the PE spectra of the studied

compounds with helium light source - --

Name Parameter - -- - --

Base pressure of the chamber (Ph) s 2 x 1 0 " t o ~

Pressure with helium gas (P, + P,) 3.5 x 104 tom

Total pressure with sarnple (Ph + P, + P-) - 3 x 20 torr

Working voltage of the channeltron 2450 V

Working cument of the He lamp 300 mA

The Ar 2p,, line at a binding energy (BE) of 15.759 eV was used as the interna1

caliiration for the spectni of shidied compounds dunng data acquisition. The typical working

parameters for recording the photoelectron spectra with the helium light source are listed in

Table 2-2. Under these conditions, the resolution (de6neû as the FWHM of Ar 2p, line) was

better than 25 meV for the He 1 spectra and about 36 meV for the He ïi spectra, and the

intensity of Ar 2p, line was about 15,000 counts/second for He 1 spectra and about 300

countskand for He U spectra. The PE spectra were fitted by using a linear combination of

Gaussian-Lorentzian line shapes with an iterative procedure described previ~usly.~

The main purposes of recording helium PE spectra of the studied compounds are as

foliows: (i) The srnail natural width of the helium beam allows us to obtain high resolution

He I spectra, in which not only the levels due to different electronic states can be separated,

but afso the vibrational structure can be observed. For example, the splittings due to spin-

orbit coupling and ligand field effect, as well as the vibrational structures due to CO and M-

CO stretchings have been resolved for W 5d spectra of W(CO), and its denvatives in this

study. The observation of these fine structures is helpfùl for the correct assignment of the

spectra of metal d orbitals and for the study of the interactions (or bindings) between metal

and ligands. (ii) The optimized working conditions can be obtained by recording the He

spectra of each compound in our laboratory. Because these conditions are the sarne as those

required in the synchrotron radiation centre, this work can save a lot of precious beamtime

in the synchrotron radiation centre. (iii) The He 1 spectra of the studied compounds are

caiibrated i n t d y by the Ar Zp, line; therefore, these spectra can be used as the references

for the studies by synchrotron radiation.

28

2.3. Recording the Photoelcctron Spectra with Synchrotron Radiation

The spectra at higher photon energies were recorded on a similar PE spectrometer

with the Grasshopper bearnline ' at the Canadian Synchrotron Radiation Facility (CSRF)

which is located at the Aladdin storage ring, University of Wisconsin-Madison.' The

synchrotron radiation was monochromatized by a Mark IV Grasshopper monochromat or

which provides light with energy ranging from 22 to 500 eV. The 600 groovehm grating

and the 1800 groovdmm grating were usd in this monochromator, respectively, to offer

photons with energies ranging fiom 20 to 75 eV and photons of energy from 70 to 200 eV.

Inside the Grasshopper b d i n e , the light is first focused by a mirror and then passes the

entrance slit. Mer that it strikes on the grating, which rnonochromatizes the light. Finally the

light bearn passes the exit slit and enters the gas chamber. The wavelength of the

monochromatized üght can be chosen by changing the position of the mirror and grating. The

photon resolution (4 of the rnonochromatized light depends on three parameters, the

spacing of the ruling on the grating (associated with x parameter), the adjustable widths of

the dits (w,pm), and the selected photon energy (E, eV), based on

AE = E2 aÂ/12398 (2.1)

where al = nu and x = 0.008 for the 600 groove/mm grating, and 0.0027 for the 1800

groove/mrn grat

A Quantar (Mode1 3395 A) position sensitive detector (PSD) has been used since

199 1 together with the ESCA 36 photoelectron spectrometer to enhance the intensity of the

signal and minimize the experimental time.l0 The operation of the CSRF spectrometer was

similar to that of our laboratory spectrometer with the helium light source. The working

29

conditions in CSRF for the studied samples were the same as those in our lab with helium

light .

The spectra were caiibrated using Xe gas and the calibrated He 1 spectra of the

samples. The recent reported spectra, such as the spectra of W(CO), which has been

calibrated, can also be used as the references for spectral calibration. Al1 spectra were

deconvoluted with a Gaussian-hrentzian iine shape using a noniinear least-squares procedure

described e1sewheree6 The peak (or band) areas were used to calculate the experimental

branching ratios (BR,) or relative intensities for each peak (or band), based on the simple

formula, BR, = A, /CA, where A, is the individual peak area. The cornparison of the

experimental BR, with the thearetical BR, calculatecl by Xa rnethod or Gelius mode1 can assist

the spectral a~signrnents.~ l2

Since the photon energies can be changed continuously over a wide range with the

synchrotron radiation source, variable energy photoelectron spectra can be obtained with

great convenience In this work, rnany high resolution broad s«ui and narrow valence spectra

of the studied samples were recorded under different photon energies, which demonstrated

the great power of synchrotron radiation in photoelectron spectroscopy.

30

2.4. Refcrencts

(1) (a) Strohmeier, W. Angew. C h , lit. Ed Engl. 1964, 3, 730-737. (b)

Darensbourg, M. Y.; Conder, H. L.; Darensbourg, D. L.; H d a y , C. J. AmChem.

Soc. 1973,95, 5919. (c) Mathieu, R. ; Lenzi, M.; Poilblanc, R. Inorg. C'Che. 1970,

9, 203 0. (d) King, R. B.; Fronzagiia, A. Inorg. Chem. l966,S, 1 83 7- 1 846. (e)

King, R. B.; Raghu Veer, K. S. Inorg. Chem. 1984,23, 2482. ( f ) Jenkins, I. M.;

Verkaâe, J. G. h r g . Chem. 1967,6,2250. (g) Jenkins J. M.; Moss, J. R.; Shaw,

B. L. J. Chem. Soc. (A), 1969, 2796. (h) Bancroft, G. M.; Dignard-Bailey, L.;

Puddephutt, R. J. Inorg. C h . 1984, 23, 2369.

(2) (a) Fischer, E. O. Angew. Chem. 1957,69, 207. (b) Meister, H. Angew. Chem.

1957,69, 533. (c) Cotton, F. A.; Reynolds, L. T. J. Am. Chem. Soc. 1958,80,269.

(d) Nielson, A. J.; Rickard, C. E. F.; Smith, J. M. Inorg. S ' t h . 1986, 24, 97. (e)

Fischer, E.O.; Hofinam, H.P. Angew. Chem., 1957, 69,639. (0 Poland, J. S.; Tuck,

D.G. J Olgmmet. Chem., 1972,12,307. (g) Peppe, C.; Tuck, D. G.; Victoriano,L.

J. Chem. Soc., Dalion, Truns. 1981, 2592. (h) Lalancette, J. M.; Lachance, A.

Gan J. Chm. 1971,19,2996. (i) Shibata, S.; Bartell, L. S.; Gaviq Jr. R. M. J

Chem. Phys. 1964, 41, 717. (j) Egdell, R G.; Fragaia, 1.; Orchard, A. F. 1

Electton Spctrosc. relut. phenom. 19711,14,467.

(3) Li, X. "Photoelectron Spctroscopy of Organometaliic Compoundr", Ph.D.

Dissertation, The University of Western Ontario, London, Ontario, Canada, 1995.

(4) Shriver, D. F.; Drezdzon, M. A "me Mmîpla~ion of Air-Semîtive Compounds ",

Second Edition, A Wrley-Interscience Publication, by John Wiley & Sons, hc. 1986.

3 1

( 5 ) (a) Coatsworth, L. L.; Bancroft, G. M.; Creber, D. K.; Lazier, R. J. D.; Jacobs, P.

W. P. J Elect. S ' c t . Reiat. Phenom. 1978, 13, 395. (b) Dignard, L. M. Ph.D.

Dissertation, The University of Western Ontario, London, Ontario, Canada, 1986.

(c)Yang, D. S. Ph D. Dissertation, The University of Westem Ontario, London,

Ontario, Canada, 1989.

(6) Bancroft, G. M.; Adams, 1.; Coatsworth, L. L.; Bemewitz, C. D.; Brown, J. D.;

Westwood, W. D. Aml. Chem. 1975,47, 586.

(7) (a) Bozek, J. D.; Cutler, J. N.; Bancroft, G. M.; Coatsworth, L. L.; Tan, K. H.;

Yang, D. S.; Cavell, R. G. C'hem. Phys. Lett. 1990, 165, 1, @) Cutler, J. N.;

Bancroft, G. M.; Bozek, J. D.; Tan, K. H.; Schrobilgen, G. .I. Am. Chem. Suc. 1991,

113, 9125. (c) Cutler, J. N.; Bancroft, G. M.; Tan, K. H. Chem. Phys. 1994, 181,

46 1. (d) Sutherland, D. G. J.; Bancrofl, G. M.; Tan, K. H. J. Chem. Phys. 1992, 97,

7918.

(8) Tan,K.H.; Bancrofk,G.M.; Coatsworth,L.L.; Yates,B.W. C'and Phys. 1982,

60, 131.

(9) Cutler, J . N. PhD. Dissertation, The University of Western Ontario, London,

Ontario, Canada, 1992.

(10) Liu, 2. F.; Bancroft, G. M.; Coatsworth, L. L.; Tan, K. H. Chem. Phys. htf. 1993,

203,337.

(1 1) (a) Hu, Y. F.; Bancroft, G. M; Bozk, J. D.; Liu, 2. F.; Sutherland, D. G. J.; Tan,

K. H. J. Chem. Suc, Chem. Commun. 1992,12764278. (b) Hu, Y. F.; Bancroft.

G. M.; Liu, 2.; Tan, K. H. Inorg. Chem. 1995,34,37 16-3723.

32

(1 2) Hu, Y. F. Ph D. Dissewon, The University of Western Ontario, London, Ontario,

Canada, 1996.

Chapter 3

Photoelectron Spectra of Trimethylphosphine Substituted Tungsten

Carbonyls

3.1. Introduction

Photoelectron specîroscopy (PES) has proven to be a valuable tool for the study of

the electronic structure of transition-metal complexes since the first studies of Ni(CO),,

Fe(CO),, and Mn(CO)5X (X = halogen, etc.) in 1969.' Possibly, because of their synergic

bonding properties, many PES studies have been focused on transition-metal carbonyls and

their denvatives, with the aim of studying their electronic stnictures. In particular, the

M(CO)& complexes (M = Cr, Mo, and W; L = substituted ligand, such as phosphine, etc.,

and n = 0, 1,2, 3) have amacted considerable attention2 Two principles (or models) related

to the ligand electronic effects on a transition-rnetal center have been found usefiil in the

photoelectron spectroscopic studies of these complexes. The first one, the ligand additivity

model proposed by Bursten in 1982,' States that valence metal orbital ionizations are

systernatidy and repramicibly shiAed by an amount directly proportional to the nurnber and

type of ligand substitutions on the metal center. This model was proved to work well in

Bursten's electrochemical studies of valence metal d orbitals in M(CO),(CNR),, complexes,'

and in valence photoelectron spectroscopic studies of MO(CO),(PR,),~" and

W(CO),(PR,),~. A similar principle was found to be true in Feltharn and Brant's X-ray

photoelectron spectroscopic WS) studies of core ionkition shifts of solid ~omplexes.~ The

34

second principle States that, when compating related molecules, the binding energy shift of

a nonbonding valence orbital localized on a particular atom of these molecules should be

eight-tenths of that particular atom's core binding energy shift between two molecules, i.e.

di -, = 0.8 AE (,, . This me-valence ionization correlation pnnciple was descnbed by

Jolly and was applied to understand the valence spectra of main-group molecules.' This

principle was also found valid for Fe(CO),L complexes by Joliy and for the

Mo(CO),(PM%), series by Lichtenberger and co-workers. *

Previous studies * have show that the binding energy shifis in metal orbitals with

ligand substitutions depend on the total donor ability (a-donation - n-acceptance) or donor -

acceptor ratio of the substituted ligand relative to that of parent ligand; while the ligand field

splittings in the metal orbitals of low spin d6 octahedral complexes depend only on the

relative x-accepting ability of the substituted ligand and the parent ligand. For example,

ligand field spüttuigs of the $ orbitals and binding energy shifts toward lower energy for the

metal d orbitals have been observed in the spectra of MQ(CO),(PR,),~ - " %* and

W(CO)~(PR&~"-* when CO is replaced by PR, which is a stronger a donor but a weaker

n acceptor than CO.

The spinsrbit coupling theory was first used by Hall to interpret photoelectron

spectra of transition metai systerns. In particular, the large spin-orbit splitting in the 5d

orbitals of the third row transition metal complexes has been usefùl to obtain a definitive

assignrnent for the spectm of XRe(CO), species? Based on the sarne t heory, the spin-orbit

coupling constants (0 and ligand field splittings (e-bJ were obtained in the studies of

M(CO)&= (M = Cr, Mo, and W, L = PM% PEt, P(NMeh P(OMe),, P(OEt),, and PF,) and

w(co),(PR3)n.h

Recently, the high resolution broad-scan photoelectron spectmrn of W(CO), was

obtained in our group using synchrotron Tadiation. For the first time, this spectrum covered

the valence, h e r valence, and core level regions with high resolution. The vibrational

structures in W 4f core level spectrum were aiso observed.' In this chapter, high resolution

broad-scan gas phase photoelectron spectra (which cover valence, imer valence, and core

levels) are reported for a series of trimethylphosphine substitut4 tungsten carbonyls. The

inner valence and core levd spectra can be interpreted and assigned based on cornparison with

published results. Better resolution has been achieved in the newly obtained He 1 spectra of

the valence level and W 5d regions of these complexes (compareci with the previous spectra).

Spin-orbit splittings, ligand field splittings, and vibrational structures are observed in the

spectra of both W 5d and W 4f regions. The phosphorus 2p spin-orbit components of the

phosphine substituted complexes have been resolved for the first tirne. With ligand

substitution, al1 the metal and ligand orbitals shifi with different degrees to lower energy,

because phosphine is a stronger a donor and weaker ir acceptor than CO. Linear binding

energy shift trends are found in both core and valence levels of metal and phosphorus

ionizations, which confirm the ligand additivity predictions for these complexes. The core-

valence ionization correlation principle can be illustrated by comparing the binding energy

shift data between cote and valence levels.

3.2. Experimentai Section

The compounds W(CO)5PM%, W(CO),NBD, ch-W(CO),(PM+), &ms-

W(CO),(PMe& and fac-W(CO),(PMeJ, were prepared b y met hods in the literature, ' wit h

36

some modifications. For example, a colurnn separation method was used for the purification

ofcis-W(CO),(PMe& rather than the sublimation method,' because small amounts of

cjs-W(CO),(PMe& muid be converted to the hm-isomer during the sublimation process.

For the same reeson, the temperature should be kept as low as possible in the process of

introducuig the sarnple to the gas ceIl for evaporation in the photoelectron spectrometers. A

column separation process was also used to puri@ frans-W(CO),(PM%), &er the cis- to

tram- isomhtion reaction was complete. Two eluents were used in order to separate the

banoisorner &om the remaining cis-isomer and decalin (the heating solvent). The latter was

dificult to remove by evaporation8 because of the high boiling point (>160°C).

AU the sarnples were introduced into the gas ce11 of the spectrometer directly via the

heatable probe. The less volatile sarnples required heating in order to generate enough vapor

pressure. Temperatures for vaporization of these samples were as follows: W(CO), (35

* 5 T), W(CO)pMe, (50 + 5 OC), cis-W(CO),(PMe,), (90 5 OC), pans-W(CO),(PMe&

(65 5 OC), fac-W(CO)3(PMq)3 (180 5 O C ) , and W(CO),NBD (75 + 5 OC). The pressure

in the sample chamber was cuntrolled to be around 4 x 10" Torr, and the pressure in the gas

ceIl was around 5 x lO" Torr.

The photoelectron spectra were obtained by using two different photoelectron

spectrometers. He 1 spectra of sarnples were recorded using a McPherson ESCA-36

photoelectron spe~trometer.~ The Ar 2p, üne at a binding energy (BE) of 1 5.759 eV was

used as intemal celibration dunng data acquisition. The spectra at higher photon energies

were recorded on the modified ESCA-36 spectrometer with the Grasshopper beamlllie 'O at

the Canadian Synchrotron Radiation Facility (CSRF) which is located at the Aladdin storage

37

ring, University of Wisconsin-Madison." A Quantar Mode1 3395 A position sensitive

detector (PSD) was used together with the ESCA 36 photoelectron spectrometer to enhance

the imensity of the signal and m h h h the acpnimentai tirne.'* The spectra et higher photon

energies were calibrated using the Xe 5s line at a BE of 23.397 eV and the calibrated He 1

spectra of the samples. The spectra of W(COl6 which have been calibrated and pubüshed'

can also be used as the intenial caiibration data. Spectra were deconvoluted wit h a Gaussian-

Lorentzian line shape using a nonlinear least-squares procedure described elsewhere."

3.3. Resulb and Discussion

3,3.1. Central Features

The high resolution broad-scan photoelectron spectra of W(CO),, W(CO),PM%, cis-

W(CO),(PMe& trm-W(CO)4(PMe&,, jac-W(CO),(PMeJ,, and W(CO)4NBD at 80 eV

photon energy are presented in Figure 3-1. One of the important points of recording these

broad-scan spectra is that, for each cornplex, ail features can be seen immediately in one

specûum: the relatively intense and narrow valence bands with Eb less than 20 eV; the weak,

broad inner valence bands with Eb between 20 - 40 eV (from S to D); and both very narrow

core levels (C, and CJ and a weak, broad core level (C,) with Eb around 40 eV. Another

important point for obtaining these broad-scan spema is that the difference and sirnilarity of

these complexes in the whole spectral range can be observed clearly by cornparison, which,

in tum, can assist us to interpret and assign these spectra based on our recently reported

results of W(C0): (the spectnim was recordeci again and is shown in Figure 3 - 1 (a)). Al1 the

phosphine substituted tungsten wbnyls have similar featues in their spectra, except for their

different shifis in E, and the bigger spîitting of W Sd and phosphorus 'lone pair' bonding

38

orbital in the tram-isomer. However, they are dinerent h m the starting material W(CO&

in that they have an extra band P around E, 10 eV which has been assigned previously as the

phosphorus 'lone pair' bonding orbital or ~ ( w - P ) ~ ~ " . Another difference is observed in the

inner valence level: the intensity of band S decreases with the increase of PM% ligmd

substitution (or with the decrease of the number of CO); however, the intensity of band T

increases with the increase of ligand substitution. This trend clearly ülustrates that band S

is related to the CO ligand and band T to the PM% ligand. In fact, band T can be assigned

to an orbital containing mainly C 2s character of the substituted groups according to the

published results.14 A sirnilar trend was also found in the spectra of Os(CO), and

Os(CO),PM%, but with a different assignment.lS We think that our new assipunent and

interpretation for the i ~ e r valence levels of phosphine complexes are more reasonable than

the previous one" because they are based on the systematic study of the ligand substitutions.

The intense and very narrow peak between band S and band T, only observed in thefac-

complex of this study, possibly results from the sarnple decomposition at bigh temperature

(185 O C ) . Since di other bands are similar to those of W(CO),, they can be interpreted and

assigned according to w published resuits of w(C0);. Table 3-1 lists the positions, widths

and assignments of bands in the i ~ e r valence and core level regions. For the valence level,

band 3 arises from three MO'S (6t,,4eP and 7a,, orbitals) of mainly CO 40 character. ûther

valence bands are shown in Figure 3-2. Previous studies of these complexes onfy reported

the results of the outer valence or W 5d region," - " and the inner valence and core level

bands of the phosphine substituted complexes have not been observed before. In addition,

Our newly obtained valence level He 1 spectra for W(CO),PMq, cis-W(CO),(PMe&, and

39

I~~-W(CO),(PM& (Figure 3-2) have better resolution than the previous ones." In these

spectra, the band assignments are show on the figure. There are two band regions which are

worthy of special attention. First, the bands in the region around 10 eV (shown as P in

Figure 3- 1) are due to ionizations âom the preûominantely o(W-P) orbitals. The cas-isomer

shows only one ionization band in this region, which is broader and more intense than the

similar band in the monophosphine cornplex. The intensity of this band in the cis-isomer

indicaies that the ionizations of the two o(W-P) orbitals are essentially degenerate. A large

splitting (1.46 eV) can be observed in the ionkations of the two o(W-P) orbitals from the

ttm-isomer, and the two bands are well resolved. In the fac-cornplex, these ionkations are

again close (only one band can be seen, see Figure 3-1). These experimental results are

very similar to those obtained by Lichtenberger and co-workers in their study of

Mo(CO),(PMe3, complexes,* and therefore, the large splitting of the a(W-P) band in the

truns-complex can be arplained similady as due to the energy differences between the d, and

pz stabilizations.* Second, four components can be resolved from the W 5d band in almost

al1 these spectra (especiaily for tram-isomer), which are due to spin-orbit coupling, ligand

field spütting and CO vibrational splitting. These wül be dimssed in detail in the next section

using our newly obtained high resolution W Sd spectra. The bands in the region with E,

greater than 11 eV in Figure 3-2 derive prllnarily from the Su and lx orbitals of the CO

ligands. Also observed in this region are bands due to o(C-H) and a(P-C) for phosphine

complexes. This region is & d t to interpret in detail for organometallic complexes because

of the large number of overlapping ionization bands," and we do not attempt to give a

detaiied interpretation.

4 (b) W(CO),PMe, 1

Figure 3- l . Broad-scan PE spectra of (a) W(CO),, (b) W(CO)5PM%, (c) cis-W(CO),(PM% h , ( d ) r i ~ r i ~ s - W(CO),(PMe3)2, (e) fac- W(CO)3(PMel))o and ( f ) W(CO),NBD.

Table 3-1. Binding energies &), widths (WJ, and assignments of the inner-valence and core level spectra

of W(CO),(PMe& (n = 1 - 3)

A 24.0 1 2.12 23.60 2.13 23.18 2.13 23 .O7 2.14 satellite 1

B 26.92 2.3 1 26.14 2.33 25.96 2.30 25.55 2.23 satellite 2

C 30.3 1 2.74 29.2 1 2.75 29.08 2.75 28.44 2.80 satellite 3

3.3.2. Valence Ltvel W Sd and Corc Levtl W 4f

A High resolution close-up of W Sd spectra. Our previous photoelectron stud?

on the tungsten 5d orbitais in W(CO),(PR,), complexes has confirmed experimentally the

general vaiidity of the ligand additivity principle in that (i) the experimental ratio of the W

b splitting (or ligand field splitthg) for W(CO)SR, cis-W(CO),(PR,), trans-W(CO),(PR,),,

and fac-W(CO&(PR& is in qualitative agreement with the theoretical predictions (1 :- l:2:0);

(i) a plot of the first ionization potential (iP) or E, for the W 5d levels vs. n (number of ligand

substitution) shows a good linear correlation and the first IP's of cis and isomers are

very sVnilar as prediaed. In figure 1 of the previous studp (which showed the spectra of W

5d region for some of the substituted W(CO), species), ody a doublet of intensity - 2: 1 due

to the spin-orbit splitting of the ta orbital was seen for W(CO), and fac-W(CO),(PMeJ,;

however, three peaks were observed for cis and truns complexes due to the ligand field

splitting of the ta level into bt and e, (trats) or b, and e (cis) and the spin-orbit splitting of

the e (or eJ MO. Because PR,(such as PMQ is a poorer sr acceptor than CO, the b, (or bS

MO has a larger IP than the e (or eJ MO in W(CO)QR, and ~~~s-W(CO),(PR,)~, but with

the O pposit e order in cis-W(CO),(PR&.

Our newly obtained spectra of W 5d region for W(CO)5PM%, cis-W(CO),(PMq),

and &m-W(CO),(PMeJ2 (see Figure 3 -3) show better resolution t han the previous ones .&

In addition to the components of spin-orbit splitting (the splitting between band 1 and

band 2) and ligand field splitting (the splitting between band 3 and the average of band 1

+ banà 2), another component (the high energy shoulder) which is due to the vibrational fine

structure of CO is aiso observed very clearly in these new speara, such as 2' and 3' bands.

44

Additionai vibrational structure fiom W-C vibration ( -50 meV) has been resolved in the

spectrum of W(CO)6.74 l6 This vibrational structure broadens the spectra in Figure 3-3, but

c m not be fitted readily. The C-O vibrational splitting has been reported for the valence

spectra of w(cO),,"'~ CpM(CO), (M = Re and Mn)17 and Mo(CO),(PR,)/'

AU the bands are fitted with the same width for a h spectrum. Because of the larger

ligand field splittings in the spectra of iras-W(CO),(PMq), and W(CO),NBD, another CO

vibrational band has to be fitted in order to obtain a reasonable fit for both of them. It is

noteworthy that our new spectrum of cis-W(CO),(PMq), is different from the old one,"

which contained a minor component due to trm-isomer (the product was purified by

sublimation at high temperature in the previous wo*" which evidently causes a small amount

of cis-isomer to be converted to the trm-isomer). The fitting parameters of W 5d spectra

are listed in Table 3-2. The spin-orbit coupling constants (c) and ligand field splinings (A )

are obtained based on spin-orbit coupling the or^,^ which are in rather good agreement with

out previous resuitsa and theoretical predictions for this series of phosphine complexes (see

Appendk B.2) The ligand field splitting (A) in the W Sd region increases in the order of

w(co)6=f~-w((coX(PM~ < W(CO)QMe, < cis-W(C0),(PMeJ2 < tram-W(CO),(PMe,),

< W(CO)JWD, which lads to the increase in width of the whole W Sd spectral envelope in

the sarne order. However, the tungsten spin-orbit coupling parameters are almost constant

for W(CO), and the phosphine substituted complexes. A diagram showing the synergic

bonding in W(CO),(PM& is given in AppendDr B. 1. A cornparison of ligand field splittings

in the isomers of W(CO),(PMq), is included in Appendk B.2.

The spectrum of W(CO),NBD wiU be disaisseci in a separate section.

Counts

Table 3-2. Band positions (eV), widths (eV), assignments, spin-orbit coupling constants (O, ligand field splittings (A = 4 - e or b,, - e,), average binding energies (eV), and their shifts (eV) relative to W(CO), in W 5d spectra of the listed complexes.

W(cOk

1

2

2'

W(~OhPM%

1

2

3

3'

nans- W(CO),(F'Mc,h

I

2

2'

3

3'

cis- W(CO),(PMC,~

3

1

2

2'

foc-W(COh(PMG

1

2

W(COpJE3D

3

3'

1

2

2'

47

IL High resolution spectra of corn level W 4J In the past, core level spectra of

inorganic and organornetauic complexes were recorded by using XPS with low resolution and

these spectra could only be useâ to study chernical shift effkcts. Recently, however, with high

resolution synchrotron radiation, it has been possible to resolve vibrationai and ligand field

sphttings on the core p and d levels of inorganic m~lecules'~~ l8 and on the metal 4f levels of

organometallic complexes.'p '' Studying high resolution core level spectra of a group of

phosphine substituted tungsten carbonyls is important because we want ed to know not on1 y

the effect of chernical shiAs but also the infiuence on the width of the core level spectra by

ligand replacement. Figure 3-4 shows the high resolution spectra of W 4f levels for W(CO),

and its phosphine substituted complexes. Spectnun (a) has been published recently by our

group.' This spectrum shows mainly two spin-orbit components W 4 f , and W 4fjn with

binding energy at 37.98 eV and 40.16 eV, respectively. The CO vibrational structure has

been resolved for the first time in the hi& energy shoulder of the bands.' Spectnim (b) in

Figure 3-4 is obtained by mixing a smdl amount of starting material W(CO), with the

W(CO)QMe, sample and recording the spectra at different temperatures. Since the binding

magies of the two bands fiom W(COI6 are known (spectrum (a)), the binding energy shift

of the core level W 4f bands caused by ligand substitution can be seen directly and

immediately f?om spectrum @) (in which band 1 and 3 belong to W(CO),PMe, and band 2

and 4 to W(CO),J. In addition, the two bands of W(CO), can be used as intemal calibration

of the bands that belong to its monophosphine derivative.

In Figure 3-4@-e), the high resolution W 4f core level spectra of a senes of

phosphine substitut& tungsten carbonyls are reported for the first the. AU these spectra

48

show botb the two strong bands of W 4f spin orbi components and the smaU shoulders of CO

vibraiionai fine structure. In order to compare the influence of ligand replacement on the W

4f spectra without involving the effect of variation in photon energies, spectra @) - (f) in

Figure 3-4 are ail recorded at the same photon energy (102 eV). The experimental results and

fitting parameters for the binding energies (E,), shifts, and widths of W 4f bands of the

studied complexes are iisted in Table 3-3. A greater binding energy shiR can be seen

obviously from these results with ligand substitutions and an almost linear correlation is

established between the core binding energy shift and the number of ligand substitutions. In

addition, the width of W 4f bands increases slightly from W(CO), to nrms-W(CO),(PMe,J,

following the order: W(CO), (0.30) = fac-W(CO),(PM& 4 W(CO),PMe, (0.3 1) + cis-

W(CO),(PMe& (0.32) 4 rrmt-W(CO),(PMeJ, (0.33) + W(CO),NBD (0.35). The trend of

increase in the width of W 4f bands is similar to the trend of ligand field splitting observed in

the W Sd spectra and therefore can be explained as due to ligand field effects on the core 4f

orbitals. A diagram showing the correlation betwm the ligand field splitting of W 5d spectra

and the width of W 4f bands is given in Figure 3-5 (which has a regression coefficient of

? = 0.9065). An even larger ligand field broadening has been seen recently on the Os 4f

levels of complexes Os(CO),L (L = CO and PM%). " "

Figure 3-4. High resolution W -If core level sprctra of (a ) W(CO),. (b) W(CO), +

\\.'(CO),PiCle,. (c ) cis-W(CO),(Ph.lrj h . (d ) rram- W(CO1 (PM% h . (e) foc-W(COh (PM% h . and ( t) W( CO),NBD.

50

Table 3-3. Fitting parameters' of W 4f spectra of the listed complexes

binding energy (eV) spin - orbit shift width cornplex 4f, 4% average splitting (eV) (eV) (eV)

W(cO)6 37.94 40.10 39.02 2.165 0.00 0.30

'error for binding energy: 0.02 eV; for spin - orbit splitting: i 0.005 eV; and for width: 0.01 eV.

0.29 0.30 0.31 0.33 0.33 0.34 0.35 0.36

Half Widtli (eV) of W -Cf

Figure 3-5. A diagram showing the correlation between the ligand field

splittin; of'W 5d bands and the width of W 4f bands.

ci W 5 d

W-I f

A P lone pair

V P 2p

O 1 2 3 4

Number of Pliosphine Ligands (n )

Figure 3-6. Shi ft comparison diagrmi for tungsten and p liosplioriis binding energy

shifis. W 5d (valence), W I f (core). P lone pair (valence). and P I p (core).

-3 -2 - 1 O

Core Level Binding Energy SliiR (eV)

Figure 3-7. Core - valence shifi correlation for tungsten and phosphorus

ionizations. W 5d - W 4f+, and P lone pair - P 2p.

54

C Sliajt compm'sons and core-valence ionizaîion correlaîionr. Bursten' s ligand

additivity model was onginally designed for valence metd ionizations. However, it can also

be used directiy for core level ionizations as shown in our above expenmental data and other

published r e~u l t s .~ '~~ Compared with valence ionizations, the overlap and hyperconjugative

interactions between metal m e orbitals and ligand valence orbitals are much smaller than that

between metal valence orbitals and ligand valence orbitals? a Therefore, a greater ionization

(or binding energy) shifi in core level than in valence level is expected when the change in

charge potential on the metal center (due to ligand replacement) occurs. According to Jolly's

core-valence ionization correlation model,' when comparing the spectra of two related

molecules, the binding energy shift of a valence orbital localued on a particular atom of the

molecules should be eight-tenths the core binding energy shi . of that particular atom between

the two moldes, i.e. AE-, = 0.8 Our experimental data on binding energies and

their shifts relative to the starting matenal, W(CO),, in both W 5d and W 4f levels of these

phosphine complexes are listed in Table 3-2 and Table 3-3, respectively. A graphical

presentation of the core and valence data is shown in the shifl cornparison diagram of

Figure 3-6. The abscissa of the diagram is the number of phosphines in the cornplex and the

ordinate is the shifl in electron volts(eV) relative to the starting material, W(CO),. The

valence metal shifts shown in Figure 3-6 and Table 3-2 are obtained by comparing the average

W 5d binding energy values of these complexes with that of W(CO),. The core shifts for both

W 4 f , and W 4f, components are the sarne, and therefore, either their average binding

energy values or the binding energies of any one of the spin orbit components can be used to

compare the shifts. As illustrated in Figure 3-6 and Table 3-2 - Table 3-3, the binding

55

energies (or ionization potentials) for both valence W 5d and core W 4f levels are shified

almost linearly toward lower energy regions with each successive ligand substitut ion. The

shidt per phosphine substitution is - 0.66 k 0.03 eV for the W 5d ionization (using the average

value of the binding energies of al1 fined W Sd peaks except for the peak of vibrational

shoulder), and - 0.76 0.03 eV for W 4f ionkations ( for both W 4 f , and W 4fd . These

data confirm the validity of ligand additivity predictions for both valence and core level shifts

in these complexes. The core-valence ionization correlation can be seen immediately for

these complexes when the W 5d shifi is plotted against the W 4f shi f i (Figure 3-7). The ratio

of the valence metal d level shifts to the core metal shifts is 0.86 * 0.03 (Le. &(,, /AE(-,

= 0.86 * 0.03). 3.3.3. Eigher Energy Spectn and Phosphorus 2p Bands

The most important advantage of using synchrotron radiation is that the photon

energy of the light source can be changed continuously within a wide range. Thus, relative

partial photoionization cross sections as a function of photon energy can be exarnined in

detail. This has proven to be an invaluable assignment tool in photoelectron spectroscopy and

has provided crucial information about fundamental photoionization processes because

photoionkation cross sections for different atomic and molecular orbitals Vary greatly with

photon energy (due, for exarnple, to shape resonances, pronounced maxima and minima,

Cooper minima effect s and etc.). 19* For M(C% (where M = Cr, Mo, and W) complexes,

the energy dependence of cross sections has been studied by Green and CO-workêrs.'"

Similar variation trends in cross sections of valence W 5d and core W 4f bands are also found

in Our studies for phosphine substituted tungsten complexes. For example, with photon

56

energy increase from 80 eV to about 100 eV, the relative intensities of valence W 5d bands

decrease slowly and that of core W 4f bands increase greatly to almost their maxima. These

variations can be seen obviously by cornpuhg the broad-scan spectra obtained at 80 eV

(Figure 3-1) and 100 eV (Figure 3-8 (a)). In addition to these features, new bands on the low

energy side of the W 4f bands (shown as 2p in Figure 3-8 (a), for example) are observed in

ail 100 eV spectra of the phosphine substituted complexes (we oniy show the spectra of cis-

W(CO),(PM& ). Baseû on our experimental data, it is clear that these bands are related to

the phosphine ligands and the photon energy of 100 eV, because (i) they are found only in the

spectra of phosphine complexes, not in those of W(CO), and W(CO),NBD; (ii) they are

found only in the spectra recorded at about 100 eV for phosphine complexes, not in the

spectra at 80 eV and 90 eV; (i) the purity of samples has been proved by the data from other

techniques, such as NMR, melting point measurement and chromatography. By carefùlly

studying the position of these bands and comparing them with the reference binding

energy values for phosphorus 2p, these bands can be assigned to the second order ionizations

(at - 200 eV) of phosphorus 2p, and 2p, orbitals. This assignrnent is confirmed by carefùlly

calculating the kinetic energies of these two bands, then their binding energies for second

order ionization, and by comparing these values with that obtained directly fiom higher

energy first order spectra of the 2p level, such as the spectrum of cis-W(CO),(PMeJ,

obtained at 152 eV photon energy in Figure 3-8 (b). Since binding energy of electrons in a

certain orbitai does not change with photon energy, only kinetic energy changes with photon

energy. When the photon energy changes from 101 eV to 102 eV, the positions (or binding

energies) of W 4f bands for cis-W(CO),(PMe,), (Figures 3 -8(c) and 3-8(d)) do not change;

57

but the positions of the P 2p bands shifi relatively by 1 eV toward lower E, (higher Ed,

because the second order photon energy increases by 2 eV (fiom 202 eV to 204 eV). This

experimental evidence further confhns our assignrnent of the two bands. The widths of the

phosphorus 2p bands are broadened by vibrational splittings as for Our previously published

Si 2p spectra of si(CH,),.'" Table 3-4 gives the binding energies (E,) and widths of the

phosphorus 2p bands in W(CO),(PMe&, together with the mean binding energy of

the o(W-P) orbital. These data show that there is an initial large increase in phosphine

bhding energies when the Gsst phosphine bonds to the metal because the metal accepts some

of the electrons of the ligand. The P 'lone pair' is stablized by 1.5 1 (1 0.09 - 8.587 eV, and

the P 2p , orbital by 0.48 (136.58 - 136.107 eV. The additional stabilization of the P 'lone

pair' is mainly due to the strong bonding interaction with the metal center. With fiirther

substitutions d e r the monophosphine cornplex, the phosphine levels show a destabilization

trend which is additive like that in the metal levels with shifts of - 0.46 k 0.02 eV for the

phosphorus valence 'lone pair' or o(W-P), and - 0.62 0.06 eV for core level phosphorous

2p,and 2 p , Also like that in the metal levels, another evidence of additivity is that cis and

tram isomers of W(CO),(PM& have aimost identical phosphorus ionizations (although the

'lone pair' binding energies for the cis and trms isomers of W(CO),(PMe& are quite

different, the averages of the two 'lone pair' binding energies for each cornplex are very

similar). The shift cornparison of phosphorus 'lone pair' and 2p ionizations of the studied

phosphine complexes is iliustrated graphicaliy in Figure 3-6, and their correlation is shown

in Figure 3-7. The ratio of the 'lone pair' shifis to the 2p shifts is 0.73 0.04. SUnilar additive

shifts were also found in the study of MO(CO),(PM~&,.~

Table 3-4. Phosphorus 'lone pair' or o(W-P) and phosphorus 2p ionizations in

W(CO),(PM%)" -- - -

~(w-p) phosphorus 2p,, phosphorus 2p,

cornplex mean E, (eV) E, (eV) width (eV) E, (eV) width (eV)

W(CO)$'M% 10.09 137.43(5) 0.47(3) 136.58(5) 0.48(3)

c~s-W(CO)~(PM& 9.62 136.76(8) OSO(3) 135.90(8) OSO(3)

fim-w(co),(PM%)* 9.64 136.78(5) 0.5 l(3) 135.93(5) 0.52(3)

fac-w(cw3(PM%)3 9.18 136.19(5) 0.49(3) 135.34(5) OSO(3)

60

3.3.4. Higb Rcsolution Photoelectron Spcetn of W(CO),NBD

The valence level UPS spectra of norbornadiene (NBD) and similar organic

compounds were published many years ago by Heilbromer and co-~orkers,~' the

photoelectron spectroswpic study of Me,PtNBD cornplex was &ed out previously by our

group.= In this chapter, the high resolution spectra of W(CO),NBD are reported for the first

time and are shown in Figure 3-l(f), 3-2(d), 3-3(d), and 3-4(f), where Figure 3-l(f) is the

broad-scan spectnim at 80 eV photon energy; Figure 3-2(d) is He 1 valence level spectnirn;

Figure 3-3(d) is W 5d region close-up spectrum; and Figure 340 is W 4f level spectrum

recurded at 102 eV. These spectra differ fiom those of other tungsten complexes as follows:

(i) in the spectra of W(CO),NBD, the band at around 10 eV (shown as x band in Figure 3 - I(f)) results from the two R bonding orbitals of norbornadiene ligand. In the valence

spectnirn of norb~rnadiene,~' t hese two x orbitals gave two peaks at 8.69 eV and 9.55 eV,

and SQ they shift about 1 eV to high energy when coordinated to tungsten. (ii) Band 2

contains not only Sa and lx components of CO ligands, but also o(C-H) and a(C-C)

components of norbomadiene. (iii) The spectra in both W 5d and W 4f regions show much

larger ligand field splittings than other complexes reported in this paper. The very weak

n-acceptor ability of norbomadiene is largely responsible for the large splitting. In addition,

the bhding energy shift in the spectra of W(CO),NBD is smaller than that of other complexes

because the total donor abiiity (odonation - x-acceptance) of norbomadiene is siightly larger

than CO but smalla than phosphines. Thedore, the a-donor ability of M3D is weaker than

that of CO, and much weaker than that of PM%. Since other bands are sirnilar to those of

phosphine substinients, they can be assigned siilarly .

61

3.4. Conclusions

High resolution photoelectron spectra of W(CO)& W(CO)QMe, cis-W(CO),(PMe&,

hm-W(CO),(PMqX, fac-W(CO),(PMe&, and W(CO),NBD have been reponed. The

advantages of monochromatized synchrotron radiation (SR) for studying the electronic

structure of organometallic complexes have been demonstrated further in this paper: we can

study ail the levels from valence to imer-valence and core levels with high resolution in one

specûum for each of these complexes. The high resolution and high intensity of SR is critical

for the study of the inner-valence and core level spectra. The inner-vdence spectra of the

substituted tungsten complexes are similar to that of W(CO), which is dominated by the

contribution fiom CO. However, noticeable differences are seen in the relative intendties of

bands S and T: the contribution fiom mainly C 2s of the substituted ligands should be

considered for phosphine and norbodene complexes. For the first time in these phosphine

complexes, the spin orbit components of phosphorus 2p have been resolved and their second

order ionizations been observed. Spin-orbit splittings, ligand field efFects and vibrational

stmctures are observed in the spectra of both W 5d and W 4f regions.

As the CO ligands are systematicaliy replaced by phosphines on the metal center, al1

the metal and l i g d orbitals SM?. The changes in charge potential cause t hese shifts t owards

lower energy because phosphine is a stronger o donor and weaker x acceptor than CO.

Meanwhüe, the difference in x acceptor ability of the ligands lads to ligand field splittings

of the metal t, orbitals (since phosphine is a weaker x acceptor than CO). Ligand field

splittings on both the W 5d and W 4f levels increase in the order of W(CO), = fut-

W(CO),(PMt& < W(CO),PMe, s cis-W(CO),(PM%), < frans-W(CO),(PMe& *

62

W(CO),NBD. The linear shi€t of binding energy in both metal valence and core levels, and

in phosphorus valence and core levels prove the validity of ligand additivity p ~ c i p l e for these

complexes. The core-valence ionization correlation cm be estabiished behveen valence and

mre levels in both metal tungsten and phosphoms ionizations. This expenmental evidence

represents an extension of these principles to many well defined systems of transition metal

complexes where strong x back bonding ligands are replaced by weak n acids in either

direction. The smdest binding energy shifi and the largest ligand field splittings obsened

in the spectra of W(CO),NBD indicate that norbomadiene (NBD) ligand is a very weak

x-acceptor and a weak o-donor.

63

References

(a) Lloyd, D. R; Schlag, E. W . Inorg. Chem. 1969,8,2544. (b) Evans, S.; Green, J.

C.; Green, M. L. H.; Orcharci, A F.; Turner, D. W. Discuss. Fm* Soc. 1969,47,

112.

(a) Yarbrough, L. W.; Hall, M. B. Inorg. Chem. 1978,17,2269. (b) Bursten, B. E.;

Darensbourg, D. J.; Kellogg, G. E.; Lichtenberger, D. L. Inorg. Chem. 1984, 23,

4361. (c) Bancroft, G. M.; Dignard-Bailey, L.; Puddephatt, R J. Inorg. Chem. 1984,

23,2369. ( d ) Bancroft, G. M.; Dignard-Bailey, L.; Puddephatt, R. J. Inorg. Chem.

1986, 25, 3675. (e) Puddephatt, R. J.; Dignard-Bailey, L.; Bancroft, G. M. Inorg.

Chim. Acta. 1985,96, L91. ( f ) Lichtenberger, D. L.; Kellogg, G. E.; Landis, G. H.

J. C h . PPhys 1985,83,2759. (g) Lichtenberger, D. L.; Kellogg, G. E. Acc. Chern.

Res. 1987,20,379. (h) Hall, M. B. J Am. Chem. Soc., 1975, 97,2057.

Bursten, B. E. J. Am. Chem. Soc. 1982,104, 1299.

Feltham, R. D.; Brant, P. J . Am. Chem. Soc. 1982, 101,645.

(a) Joily, W. L. Acc.Chem. Res. 19113,16,370. @) Beach, D. B.; Jolly, W. L. Inorg.

Chem. 1986,25,875. (c) JoUy, W. L. Chm Phys. ka. 1983,100, 546. (d) JoUy, W.

L.; Eyermann, C. J . J . Phys. Chem. 1982,864834. (e) Jolly, W. L. J . Phys. Chem.

1983, 87,26.

(a) Beach, D. B.; SMt, S. P.; Jolly, W. L. Orgonometalfics, l984,3, 556. (b) Beach,

D. B.; Bertoncello, R; Granoz5, G.; JoUy, W. L. O r g ~ e t a I f i c s , 1985,4, 3 1 1. (c)

Beach, D. B.; Jolly, W. L. Inorg. C'Che. 1983,22, 2 137.

(a) Hu, Y. F.; Bancroft, G. M; Bozek, J. D.; Liu, 2. F.; Sutherland, D. G. J.; Tan, K.

64

H. J. C h . SocSoc, Chem. Commun, 1992,1276. (b) Hu, Y. F.; Bancroft. G. M.; Liu,

2.; Tan, K. H. Inorg. Chem. 1995,34,3716.

(8) (a) Strohrneier, W. Angou. Chem., Int. Eil Engl. 1964,3, 730. (b) Darensbourg, M.

Y.; Conder, H. L.; Darensbowg, D. L.; Hasday, C. J. AmChem. Soc. 1973,95,5919.

(c) Mathieu, R; Lenq M.; Poilblanc, R I m g . Chem. 1970,9,2030. (d) King, R. B.;

Fromigüa, A I k r g . C k m 1966,5,,1837. (e) King, R. B.; Raghu Veer, K. S. Inorg.

Chem. 19û4,23,2482. (0 Jenkins, J. M.; Verkade, J. G. Inorg. Chem. 1967,6,2250.

(9) Coatsworth, L. L.; Bancroft, G. M.; Creber, D. K.; Lazier, R. 1. D.; Jacobs, P. W. P.

J. Elect. Spct. Relut. Phenom. 1978, 13, 395.

(10) (a) Bo* J. D.; Cutler, J. N.; Bancrofl, G. M.; Coatsworth, L. L.; Tan, K. H.; Yang,

D. S.; Caveli, R. G. Chem. Phys. &!tell. 1990,165, 1 . (b) Cutler, J. N.; Bancroft, G.

M.; Bozek, J. D.; Tan, K. H.; Schrobilgen, G. J. Am. Chent. Soc. 1991,113, 9125.

(c) Cutler, J. N.; Bancroft, G. M.; Tan, K. H. C h Phys.1994, 181, 461. (d)

Sutherland, D. G. J. ; Bancroft, G. M.; Tan, K. H. J Chem. Phys. 1992, 97, 79 1 8.

(1 1) Tan, K. H.; Bancroft, G. M.; Coatsworth, L. L.; Yates, B. W. Can. J . Phys. 1982, 60,

131.

(12) Liu, 2. F.; Bancroft, G. M.; Coatsworth, L. L.; Tan, K. H. Chem. Phys. Lett. 1993,

203, 337.

(13) Bancroft, G. M.; Adams, 1.; Coatsworth, L. L.; Bennewitz, C. D.; Brown, J. D.;

Westwood, W. D. Anal. C h . 1975,47, 586.

(1 4) (a) Egdell, R G. ; Fragala, 1.; Orchard, A. F. J . Elect. Spect. Relat. Phenom. 1978,

11, 467. (b) Kimura, K.; Katsumata, S.; Achiba, Y.; Yamazaki, T.; Iwata, 1.

65

H a ojHe ~Photm~ecfron Spctta of Fundamental Organic MoIecules, Japan

Scientific Societies Press, 198 1 , pp47.

(15) HyY.F.;Bancroft,G.M.;Davis,H.B.;Male,J.I.;Pomeroy,R.K.;Tse,J. S.;Tan,

H. K. OrganwmefaIItcs, 1996, 15,4493.

(16) Hubbard, J. L.; Lichtenberger, D. L. J. Am. Chem. Soc. 1982, 104, 2132.

(17) (a) Calaùro, D. C.; Hubbarâ, J. L.; BleWis, C. H., II; Campbell, A. C.; Lichtenberger,

D. L. J . Am. Chem. Soc. 1981,103,6839. (b) Lichtenberger, D. L.; Fenske, R. F.

J . Am. Chern. Soc. 1976,98,50.

(18) (a) Cutler, J. N.;Bancroft, G. M.;Tan, K. H.J. Chem Phys. 1992, 97, 7932. @)

Sutherland, D. G. J.; Bancroft, G.M.; Liu, 2. F. Nucl. h t . Meth& B 1994,

87,183. (c) Liu, 2. F.; Bancroft, G. M.; Cu*, J. N.; Sutherland, D. G. J.; Tan, K. H.

Phys. Rev. A 1992,46, 1688. (d) Svensson, S.; Ausrnees, A.; Osborne, S. J.; Bray,

G.; Gel'mukhanov, F.; &yen, R; Naves, de Bnto, A.; Sairanen, O.-P.; Kivimaki, A.;

N6mmiste, E.; Aksela, H.; Aksela, S. Phys. Rev. Lett. 1994, 72, 302 1.

(19) (a) Green, J. C. Acc. Chem. Res. 1994,27, 13 1. @) Cooper, G.; Green, J. C.; Payne,

M. P.; Dobson, P. R; mer , 1. H. 1 Am. Chem. $oc. 1987, 109, 3836. (c) Breman,

1. G.;Green. J. G.; Redfem, C. M. J. Am. Chem. Sm. 1989, 111, 1989. (d)

Brennan, J. G.; Cooper, G.; Green. J. G.; Kaltsoyannis, N.; MacDonald, M. A.;

Payne, M. P.; Redfem, C. M.; Sze, K. H. Chem. Phys. 1992,64, 27 1.

(20) (a) Yang, D. S.; Bancroft, G. M.; Puddephatt, R. I.; Tan, K. H.; Cutler, J. N.; Bozek,

J. D. Ihng. C h . 11990,29,4956. @) Li, X.; Bancroft, G. M.; Puddephatt, R J.;

Hu, Y. F.; Liu, 2. F.; Tan. K. H. Inorg. Chem. l992,31, 5162. (c) Li, X.; Bancroft,

66

G. M.; Puddephatt, R. J.; Hu, Y. F.; Liu, 2. F.; Sutherland, D. G. J.; Tan. K. H. J.

Chem. k., C h . Commun. 1993,67. (d) Li, X.; Bancroft, G. M.; Puddephatt, R.

J.; Liu, 2. F.; Hu. Y. F.; Tan, K. H. J. Am. Chem. Soc. 1994, 116, 9543. (e) Li, X.;

Bancroft, G. M.; Puddephatt, R J.; Hu, Y. F.; Tan, K. H. Organometallics, 1996,I5,

2890.

(a) Bischof, P.; Hashrnall, J. A.; Heilbromer, E.; Homung, V. Helv. Chim. Acta,

1%9,52, 1745. (b) Heilbromer, E.; Ma* H. D. Helv. Chint. Acta, 1972, 55, 1490.

Yang, D. S. PhD. Di.wrta?i'on, The University of Western Ontario, London, Canada,

1989.

Chapter 4

Photoelectron Spectra of Cyclopentadienyl Derivatives of Indium(1) and

Thallium(1)

4.1. Introduction

Cyclopentadienylindium (CpIn) and cyclopentadienylthallium (CpTI) were first

synthesized by Fischer and HofmaM about 40 years ago,' and since then, the two compounds

and their derivatives have attracted great interest of chemists for at least two main reasons.

The fkst reason is that CpIn (including its MeCp analogue) and CpTl (including its analogue,

the bonnates) are the ody readily available, stable and soluble M(1) organometallic

compounds (CpIn is usually stable in the solid state, being sensitive to air but unaffected by

water and is soluble in common organic solvents; CpTl is moderately soluble in polar

solvents and is stable in air or water)? and as such they are usehl starting materials for

synthesizing other important metal complexes. For example, CpTi has found wide application

as a cyclopentadieny1 group donor not ody to metal complexes (especially those of transition

metals), but aiso to several organic moiet ie~.~ " The second reason, which is more

important and apposite to this chapter, is that the discussion about the electronic structure and

the bonding of these molecules has been very controversial.'

In the solid state, CpIn and CpTl have been shown by X-ray structural studies to

contain a s i i a r zig-zag polymeric chah structure, in which the MCp distances are equal and

considerably longer than those in the monomer."' The monomer structure of CpIn and CpTl

68

in the gas phase has been estabtisheâ by electron diffraction studies and microwave

spectroscopie studies, 3b which show that both molecules have the half-sandwich structure

"th precise C, symmetry (as originally suggested by Fischer and Hohandb). It was found

that the cyclopentadienyl hydrogens in Cpln were bent away from the metal atom by

approximately Similar structures were also reported for (CsMes)Sn+ ' and (C,Me,)1n3j

in which the methg groups were bent away fiom the metal. INtially, the bonding in CpIn was

considered to be rnainiy ~ovalent,'~ but ionic bonding was also proposed for CpIn and

CpTl, respedvely.*" The fact that the M-C distance is decidedly smaller than the sum of

the ionic radius of M+ and the van der Waals radius of C 34 3b suggests strongly that the ionic

structural argument is incorrect. In addition, several latter experimental and theoretical

studies 3* have shown that the bonding in CpIn and CpTl in the gas phase is essentiaily

covalent, but the ionic charaaer of these compounds increases in the solid state and in

s~lution.~

The He I and He II photoelectron spectra have been reported wk for CpIn and CpTl.

Several theoreticai calculations have a h been performed in order to probe the electronic

structure and molecular orbital energy levels of both molecules, and to interpret the results

of PES. Generdy, a syrnbiotic relationship (or a one to one correspondence) exists between

molecular orbital calculations and photoelectron spectroscopy (K00pmatt.s ' theorem '). The

results Gom the recent Xa-SW "and pseudo potential SCF a dailations on these molecules

are indeed in very good agreement with the assignment s of PES ." " However, the previous

NOCOR 3h and Cm0 " calculations do not agree with the PES data (especially for the

energy levels of orbitals 3e, and 4aJ; and unfortunately, the results from the more recent

69

ab initio calculations on CpIn and (C$fe,)In also disagree with the PES results for CpIn,

although this shidy does show the same &kas of methyl groups on the bonding of (C$de,)In

(Le. the orbital energies of the pemethylated compound are lowered in wmparison with

those of the non-methylated one) as indicated by the previous PES studies for a variety of

methylated cyclopentadienyl complexes.' The major disagreement between these calculations

on CpIn and CpTi, or between some of the calculations with the results of PES experiments,

is that the order of the low energy orbitals (i.e. the mainly Cp n orbitals, e, and the orbital

with mainly metal ns character, a,) is reversed (our assignments are listed in Table 4-1 and

Table 4-2).

This series of controversial results has show that a clear understanding of the

bonding and the electronic structure of these relatively simple molecules has not yet been

adiieved, and fllrther experhental and theoretical studies are still necessary. In this chapter,

the photoelectron spectroscopic studies of Cph and CpTl are reported with the combination

use of helium light sources and synchrotron radiation (SR). The information obtained on the

variations of band intensity with photon energy confirms the vaiidity of our assignments on

the PE spectra. b r new He 1 and higher energy PE spectra for Cph and CpTl compounds

show better resolution than those reported p r e v i o ~ s l y . ~ ~ The broadening on the metal d

levels cause. by iigand field splitting eEects cm be observed clearly in our spectra, and even

the asymmetrical spiitting can be seen on the Tl 5d bands. The ligand field splittings on the

metal d orbitals provide a new experimental evidence for the covalent bonding in these

m o l d e s . In addition, the interfig h d s (in He 1 and He II spectra ) coming from helium

'self-ionization' and excitations of He II p and He II y satellite lines can be overcome and

70

new spectral Monnafion such as the shake-up bands and Tl 4f bands can be detected by using

SR The results from our Xa - SW calculations on CpIn and CpTI are very close to those of

the recent Xa - SW "and pseudo potential SCF % calculations on these molecules and will

be published elsewhere.

4.2. Experimental

CpIn and C p l were obtained comrnercially from Strem Chemicals and were purified

by vacuum sublimation. ' The extremely air sensitive and moisture sensitive compound C pIn

was handled in a dry Ntrogen atmosphere (a dry box, Schlenk tube, and a vacuum line were

required). The relatively stable CpTl could be handled in the air for a short period of tirne,

but it should be kept cold for future use.'

CpTl was introduced into the gas ceIl of the spectrometer directly via the heatable

probe, whiîe Cph was introduced under dry nitrogen (the operating methods for air sensitive

sarnples were desctibed in deîad elsewhere '4. In order to generate enough vapor pressure,

the vaporization temperatures were maintained at 90 * 5 O C for CpTl and 40 * 5 O C for CpIn,

respectively. The pressure in the sample chamber was controlled to be - 10" Torr, and the

pressure in the gas ce1 was around 5 x 10" Torr.

The photoelectron spectra were obtained by using two dflerent photoelectron

spectrometers. He 1 spectra of sarnples were recorded on our laboratory McPherson ESCA-

36 photoelectron spectrometer." The Ar Zp, line at a binding energy (BE) of 15.759 eV

was used as intemal calibration during data acquisition. The spectra at higher photon energies

were recorded on the modified ESCA-36 spectrometer with the Grasshopper bearnline '* at

the Canadian Synchrotron Radiation Facility (CSRF) which is located at the Aladdin storage

71

ring, University of Wisconsin-Madison." A Quantar Model 3395 A position sensitive

detector (PSD) was used together with the ESCA 36 photoelectron spectrometer to enhance

the intensity of the signal and Mnimize the Bcperimental time." The spectrum of Tl 4f region

was calibrated by using the Xe 4p, line at a binding energy (BE) of 145.5 1 eV,'' other

spectra at higher photon energies were calibrated using the Xe 5s line at a BE of 23.397 eV

and the calibrated He 1 spectra of the sarnples. Spectra were deconvoluted with a Gaussian-

Lorentzian line shape using a nonlinear least-squares procedure described elsewhere.16

4.3. Results and Discussion

4.3.1. General Features

The high resolution broad-scan photoelectron spectra of CpIn and CpTl at 80 eV

photon energy are presented in Figure 4-1 and Figure 4-2. The valence level PE spectra with

He 1 and He II light sources are illustrated in Figure 4-3. Peak (or band) positions and

assignrnents are given in Table 4-1 and Table 4-2 for CpIn and CpTl, respectively. Both our

He I spectra and the higher photon energy spectra show some improvements compared

with the published s p e ~ t r a . ~ " However, the assignments of our PE spectra are similar to

those of the previously reported PES studies, and are in good agreement with the results

of the recent Xa-SW %nd pseudo potential SCF calculations. Band A at the lowest E,

of the spectra is assigned to the highest occupied orbital (HOMO of CpM), i.e. the pair of

degenerate el orbitals, which originates m d y from the bonding interaction of the Cp e,(n)

molecular orbitals (MOs) with the valence p, and p, orbitals of in and TI. Band B a n be

assigned to the ionization Erom the a, MO having a majority of metal ns character (or metd

'lone pair'). Band C is very broad and includes ionizations fiom mainly ligand C-C and C-H

72

o bonding orbitals and metal-Cp Nig x bonding molecular orbitals, a,@). Bands D, E, and

G (G is only observed in o u specûa for CpIn) cari be assigned to ionizat ions from MOs with

substantid C 2s chanider. The bands F and F , which are very intense compared with other

bands at 80 eV (Figure 4- l), correspond to the 3, and 3, multiplet States caused by the

spin-orbit splitting on the metal d orbitals. The bands H, and Hl' , only observed in our

spectra with use of SR source, can be assigned to the shake-up satellites induced by the

ionizations of the metal d orbitals. The interpretation of our assignments for the spectra will

be discussed in detail in the following sections.

4.3.2. Variable Encrgy Photoelectron Spectn of CpIn and CpTl

Gas phase PES is one of the most usehl and direct experimental techniques available

for probing the energy levels of the MOs and studying the electronic stnictures of inorganic

and organometallic compounds, because it can provide not only the ionization energy (IE or

Eb ciata and the fine stmctud information, but the intensity information as welI.17 Variable

energy PES has proven to be the most powerful tool for solving the conflicting problems on

m o l d a r electronic structure, especially with the application of synchrotron radiation. 17~189P

To interpret Our assignments of the PE spectra for CpIn and CpTl and to resolve the major

disagreement between difl'erent calculations on these molecules, the variable energy PE

spectra of CpIn and CpTi have been recorded (Figure 4-4 and Figure 4-5). The low

energy spectra at 2 1.2 eV (He I) and at 40.8 eV (He II) are shown in Figure 4-3. The

ionization cross-sections ( h m theoretical data for atoms l9 ) are iiiustrated in Figure 4-6.

The w e s for variation of relative band intensity as a ninction of photon energy are shown

in Figure 4-7 to Figure 4-9 for band A and B, and for band C and D (or E), respectively.

Counts

Counts

Coun t s

Counts Counts

Table 4-1. Binding energies (4)' and relative intensities (lry of the peaks in CpIn by He 1, He II, and SR (at 80 eV) PES and calculated binding energies (Qb.

experimental results calculated band

&(eW IXHeI) Iflen) I X W &(eV)S &(eV)" MOb assignment

A

B

C

X

D

Y E

F F

G

H, Hl'

In 5s ('lone pair') x(1n-ring) + o(C-H) + o(C-C) (mainly c 2 ~ ) -m (He II Y)

shake-up satellites

Errors are 0.02 in E, and 1, fiom reference (39, ' fiom reference (3e).

(a) InCp rit 80 eV F

H,' H, G . E D C A

A L 6 gfq

(c ) InCp at 1 -IO e V ' F 1

(t') InCp at 160 eV L 5c

F' - - - .. -Ci

- .- - - - E D C

d . - - - -

i.- 4 -

-

1 (d) InCp at 130 e V

Figure 4-1. Variable energy photoelectron spectra of InCp at (a) 80 eV, (b) 130 eV, (c) 110 eV, (d) 150 eV, ( e ) 160 eV, and (0 180 eV.

Cross Section (Mb)

Cross Section (Mb)

Relative intensity of band A and band B

Rclativt. intensity of band .4 anci band B

Relative intensity of band C and band D

Rclativc intcnsity of band C and band D

Cpln

-O- Band C 1 C+D+E 13 - Band D / C+D+E -A- Band E 1 C+D+E

60 80 100 120 140 160 180 200

Photon E n e r g . (eV)

Figure 4-9. Variation in relative intensity of band C. D. and E

as a Function of photon energy.

&nd A and B

The main disagreement between different calculations or between some of the

calculations and the previous PES studies on CpIn and CpTl is focused on the energy

ordering of the low energy orbitds el and a,. Which one is the HOMO of CpM and how

should the spectrai bands A and B be assigned ? Based on the intensity information obtained

from our variable energy PE spectra, band A can be assigned to a pair of degenerate 3el

orbitals (HOM& of CpM) which are the bonding combination of the Cp el@) orbitals with

the pq, orbitals of In and Tl; band £3 is assigned to the ionization from the 4a, MO (or the

metal 'lone pair') which contains mainly metal ns character. These assignrnents are the sarne

as those of the reported PES studies,' and can be verified by comparing the expenmental

band intensity data with the theoretical. cross d o n vahies of atoms (see Figure 4-3 to Figure

4-7). Fint, the relative intensity of band B compared to band A increases greatly when the

photon energy changes from 21.2 eV (He i ) to 40.8 eV (He II) (Figure 4-3), which

corresponds to the increase in cross sections of In 5s and TI 6s and the decrease in cross

sections of C 2p and metal p orbitals (Figure 4-6). Second, the relative intensity of band B

increases gradually compared with band A with energy vaqing from 80 eV to 160 eV

(Figure 4-4 to 4-5, and Figure 4-7), because the cross section of metal ns does not decline

as fiist as those of C 2p and metal p orbitals (Figure 4-6). In addition, the shifi of band B to

higher E, on going tom CpIn to CpTl agrees weU with the metal ns orbitals' E, data obtained

fiom the atomic speara of In and The computed charge densities and MO compositions

on both species also indicate that the major contributions to the 4a, MO corne from the rnetal

ns atomic orbital (AO).'~ *

85

It is noteworthy that ou spectra show better resolution than those reported earlier for

CpIn and CpTl " since at least two peaks can be resolved from band A., which indicate the

existence of vibrational structure or Jahn-Teller splitting as discussed in detail recently by

Green2' for rnagnesocene and osmocene compoundq e.g. M(qC,D,), and theu undeuterated

analogues (M = Mg and Os). The existence of the vibrational splitting for the Cp ring e, n

orbitals and the lack of this vibrationai splitting for metal ns orbital are consistent with Our

assignments for band A and B @and B is much narrower than band A).

Band C, D, E, G, x undy

There is less or no disagreement regarding the assignments and interpretations of

band C, D, and E. Actuaiiy, the PE spectxaî bands and their interpretations in these E, ranges

are very sirnilu for al1 the metal cyclopentadienyl derivatives because these bands mainly

belong to iiganâ Cp M0s.'~ 'W Band C can be assigned to ionizations h m mainiy Cp C-C

and C-H a orbitals with a greater contribution from the C 2p orbitals t han C 2s. Band C also

contains the metal-ring x bondiig OMS, a,@); and in the He II spectra, to some extent the

metal d components excited by He II f l and He II y satellite lines from the He II emission.

Band D is probably the overlap of signals fiom metal-ring a bonding MO, a,@) and one of

the ~g o bonding MOs, e(o) with a great C 2s contribution. Band E can be assigned to the

remaining e(o) MO which contains substantial C 2s character. The band E in CpTl is,

however, obscureci and ovedapped by the band F from ionization of Tl 5d electrons. The

x and y bands in He II spectra (Figure 4-3, @) and (d)) corne from the excitation of metal nd

orbitals by He II P (48.37 eV) and He II y (5 1 .O1 eV) satellite lines of the He II

emission which can be avoided by using the SR source.

86

Our variable energy PE spectra of both species (Figure 4-3 to Figure 4-9, when

cornpared with the theoretical cross section data (Figure 4-6), can provide further evidence

for these assignments (Figure 4-8 to Figure 4-9). First , when the photon energy increases

âom 21.2 eV (He 1) to 40.8 eV (He II), the relative intensity of band C falls quickly due to

the decrease in cross section ofC 2p orbitals, while the relative intensity of band D increases

greatly because of the increase in cross section of C 2s in this energy range. Second, when

the photon energy changes fiom 80 eV to 160 eV, since the cross section of C 2p decreases

much faster than that of C 2s in this energy range, the intensity of band D continues to

increase gradudly relative to that of band C, and finally the intensity of both bands become

closer and closer in the higher photon energy levels (Figure 4-8). Band E shows the similar

variation trend in relative intensity to the band D (Figure 4-9), and therefore can be

interpreted similady. Band G is only resolved in our spectra of CpIn, which can be assigned

tentatively to a MO related to the ionkation 6om C 2s according to the previous PE

studies and calculations on M(~'-C,H,), (M = Ni, Pd, and ~t ) . ' "

Bond F and F'

Band F and F obviously correspond to the two spii-orbit components (3, and Q&

ofthe metal d orbitals. The spacing between band F and band F' for both CpIn and CpTl is

close to that of the hafides MX observed in the PE spectra. Coincidently, the strong and

sharp helium "~e~ionization" peak He' overlaps ~mpletely with the band F in the He II

spectrum of C p h (Figure 4-10, (a)), which makes it impossible to obtain a pure In 4d

spectrum with He II radiation. This He' peak has also been observed previously in the core

d level spectra of Me& and Et,Pb. '' In order to overcome this problem, the spectra with

SR have been recurded (Figure 4- 10, @)).

The relationship between photon energies and relative band intensities uui be found

cleady fiom the variable energy PE spectra (Figure 4-4 to 4-5). With the Uicrease of photon

energy from 80 eV to 180 eV, the intbties of band F and F for In 4d compared with other

bands first decrease greatly, then reach the Cooper minimum *' at around 140 eV and then

increarie slowly (Figure 4-1 1). However, the intensities of band F and F for Tl 5d decrease

continuously relative to ot her bands, wit hout showing any minimum. This relationshi p

observed by the experiment (Figure 4-1 1) is in accord with the theoretical cross section data

for In and Tl atoms (Figure 4-6).

Both the In 4d and Ti 5d spectra are broadened greatly from the expected - 0.15 eV

linewidth? For example, in Figure 4- 10, the linewidt hs of the In 4d, or In 4d, and Tl 5d,

or TI Sd, bands are around 0.4 eV. In addition, the TI Sd, band shows a distinct shoulder,

and the TI Sd, band shows a distinct asymrnetry. This shoulder, asymrnetry or broadening

is due to unresolved ligand field spütting similar to that observed on the TI Sd levels in the Tl

halides. 26 Ligand field splittings are due mainly to the asymmetric electronic field set up

by an unequal distribution of valence p electrons in the 5p (In) and 6p (TI) levels. For

example, ifthe bonding is ionic in CpTI, no ligand field splitting would be expected because

Tl' has no electron in the 6p orbitals. In the Tl halides, the halogens donate some electron

density to the 6p, orbital, resulting in a measurable ligand field splitting (Table 4-3) and a

negative CZ (the asyrnmetric crystal field tenn (Table 4-4)). For the TI halides, as the

electronegativity increass fiorn I to Br and to Ci, the 6pz electron density decreases, resulting

in a smaller Ct and a larger Sd binding energy ( E d .

88

In C, symmetry, the 5d orbitais are again split into three sets (dz; d, Q; and &-2,

Q. Along with the spin-orbit splitting, which removes the degeneracy, five levels are

expected. The Cph and CpTl spectra have been fitted to five peaks of closely sirnilar width

and shape (Table 4-3), and these values have been put into the standard five equations "

to obtain Ct, C,9 1 and E,. The results are given in Table 4-4. Because of the asyrnmetry

of the Ti Sd, band, it is apparent that C; for CpTl is positive. TNS indicates' ,bat the 6p, and

6p,, electron density (due to the TlCp bonding in the 3e, MO) is larger than the 6p, electron

density (due to the TlCp bonding in the 4a, MO). The large value for CT shows that the

donation of electron density fiom the Cp to the 'Iï is strong. The very srnall E, value also

shows that the Cp donates very substantid electron density to the TI. These results are

consistent with the latest SCF pseudopotential caîculations. CpIn gives a slightly smaller

C," value as expe*ed from previous results for analogous molecules. " 4.3.3. The Shakeup Satellitts of Metnl d Levels and Ti 4f Bands

The core level shake-up satefites in organometallic compounds are usually measured

by using low resolution laboratoty X-ray sources.27 However, the much narrower 4f shake-

up bands in W(COI6 have been obtained recently by using high resolution PES with SR

source." Simiiariy, the high resolution PES has been employed in this work to examine the

effêcts of high resolution, variable photon energy studies on the iine width of the nomally

broad shake-up bands and the relative intensity (relative to the main Iine) of the shake-up

bands as a hction of photon energy. The shake-up spectra of CpIn and CpTl at 120 eV are

show in Figure 4-12. The resuhs indicate that both the main metal d bands and their shake-

up bands are narrowed (Table 4-5) with the decrease of photon energy (i.e. the increase of

89

experirnental resolution, see equation 2.1 of chapter two). Because the relative intensities

of metal d bands (F and F ) vary greatly fiom 80 eV to 180 eV (Figure 4- 1 l), the intensity

ratios of the shake-up bands (H,+H,') to the metal d bands (F+F ) change drarnatically

(Figure 4-13). For CpTl, the Uitensity ratio increases alrnost linearly with the increase of

energy (80 - 160 eV), since the intdties of Tl 5d bands decrease continuously in this energy

range. However, the intensity ratio for CpIn first increases with the increase of photon

energy, then reach a maximum at about 140 eV, and then decreases gradually with the

increase of energies due to the opposite and dominant variation trend in the intensities of In

4d bands. This study demonstrates that good experimental resolution can improve the

shake-up widths considerably, and that the relative shake-up intensities do not vaty in the

same ways as those of In 4d and Tl 5d bands over the studied photon energy range. In

addition, this variation trend in shake-up intensities appears to be a completely novel

observation and is still not understood.

The two spiinsrbit components of TI 4f level of CpTl are observeci for the first time

by gas phase PES with SR at 280 eV (Figure 4-14). The binding energy for TI 4 f , is

124.3 l(3) eV, the spin-orbit splitting of T14f bands is 4.468(5) eV and the width of the bands

is 1.5 5(2) eV. The brdening of the bands is probably caused by one or aU of the following

reasons: (a) the relatively lower experimentai resolution at high photon energy; @) vibrational

broadening and (c) ligand field splitting.

C o u n t s

6 N (Zr O O

O O O O O O O

Relative intensity of bands

Relative intensity of bands

Table 4-3. Binding energies (E,) and widths (FWHM) of metal nd bands for TIX/ TlCp and CpIn.

binding energy (eV) width (eV)

- - -- - -

TIC1 20.986 21.199 21.283 23.286 23.505 0.112 0.070 0.085 0.150 0.103 TlBr 20.849 2 1 .O56 2 1.161 23 .O88 23.245 0.118 0.120 0.149 0.101 0.154 Tl1 20.550 20.74 1 20.848 22.809 23 .O27 O. 1 13 0.084 0.150 0.183 0.191

TlCp 19.761 19.580 19.445 22.144 21.862 0.263 0.210 0.217 0.273 0.267 InCp 23.942 23.770 23.6 13 24.71 4 24.547 0.228 0.187 0.185 0.194 0.187

-- -

' Reference (23a).

Table 4-4. Derived crystal field parameters (eV) for TIX,' TlCp and InCp

SA12 compound End (2'

exptl free atomC

TICI 22.05 -0.02 1 O -0.002 1 2.28 2.25 TlBr 2 1.86 -0.0233 -0.00 1 8 2.14 2.25 TI1 21.58 -0.0258 -0.00 t 6 2.20 2.25

TlCp 20.56 0.0360 0.0004 2.28 2.25 InCl 25.75 - - 0.90 0.85 InBr 25.70 - - 0.88 0.85 I d 25.50 - - 0.89 0.85

InCp 24.12 0.0280 0.0001 0.87 0.85 ' Reference (23a). reference (23c), and ' teference (20).

2500 f (a) InCp ai 120 eV

Binding Energy (eV)

(b) TlCp at 12OeV

F ' 1

Binding Energy ( e ~ )

Figure 4- 12. Shake-up bands of in I d and Tl jd regions of (a) InCp and (b) TiCp.

Table 4-5. Shake-up energies (a) and widths of CpIn and CpTl '

a binding enagy (eV) width (eV)

Cpin CpTl CpIn C P ~

band 80eV 120eV 140eV 80 eV 120cV 140cV 80cV 120eV 140cV 80cV 120cV f40eV hgnmcnt

Errors are 0.1 in the a binding energies and are 0.05 in band widths.

60 80 100 120 140 160 180 200

Photon Energy (eV)

60 80 100 120 140 160 180

Photon Energy (eV)

Figure 4- 1 3 . Variation of the iiitensity ratio of nietal d baiids (F + F') witli tlieir slinke-iip bsiids ( H t - H as a

fiiiict ion o f photon energy.

TlCp at 280 eV

Fisure 4-11. Photoelectron spectruiii of TI I f region in TlCp obtained at 280 eV.

98

4.4, Conclusions

High resolution photoelectron spectra of CpIn and CpTl have been reporteci with the

use of a combination of helium light and synchrotron radiation (SR). The advantages of

using variable energy PES with SR for studying the electronic structure of organometallic

compounds have been fùrther demonstrated in this chapter; Our assignments of the PE

spectni on CpIn and CpTl have been confinned by the variable energy PES, which are in good

agreement with the results ofboth the previous PES studies " and the Xa-SW 3f and pseudo

potential SCF * calculations. The broadening on the metal d levels caused by ligand field

splittings has been observed in Our high resolution spectra which indicate the covaient

b o n h g character ofthese moldes in the gas phase. We have also shown that the shake-up

structures of metal Ymer' d levels can be studied by synchrotron radiation, and enhanced

resolution can be achieved by improving the experimental conditions. In addition, the

unwanted bands fiom the excitation of He Il satellite lines c m be overcome and the core TI

4f bands can be measured by synchrotron radiation PES.

99

4.5. Referencts

(1) (a) Fischer, E. O. Angew. Chem. 1957, 69, 207. (b) Fischer, E. O.; Hofmann, H. P.

Angew. Chem. 1957,69,639.

(2) (a) Gmelin H m d b o k of Inorgm>ic Chemisiry, Orgrnoindium compnds , Part rt 1;

edited by WoIfgang Petz, Springer-Verlag: Berlin, 1991, p372. @) Schmidbaur, H.

Angew. Chem. 1985, 97, 893; Angew. Chem. , Inf. Ed Engl. 1985, 24, 893. (c)

Jutzi, P. A&. Organomet. Chem. 1986, 26,2 17. (d) Paver, M. A.; Russell, C. A.;

Wright, D. S. in CumprehenFiw Orgunometallic Chemistry II , Volume 1 ( I I .

Gallium, Indium md Thailium denvatives); edited by Catherine E. HousecroA,

Elsevier Science Ltd., 1995. (e) Marko, 1. E.; Leung, C. W. in Comprehensive

Organumetallic Chemistry II. Volume 11 (1 0. i%aIIium); edited by Alexander

Mckillop, Elsevier Science Ltd., 1995. ( f ) Tuck, D.G.; Kurosawa, H. in

Comprekensnte Organome fdIic Chemistry, Volume 1 (7. Gallium and Indium; and

8. Thaliium); edited by Sir GeoEey Wilkinson FRS, Pergamon Press Ltd., 1982.

(3) (a)Shibata,S.;BarieU,L.S.;Gavin,R.M.,Jr.J.Chem.ms.1964,41,717. (ô)

Tyler, J. K.; Coq A P.; Sheridan, J. N e e (London) 1959, 183, 1 182. (c) Cotton,

F. A.; Reynolds, L. T. J Am. Chem. Soc. 1958,80,269. (d) Cradock, S.; Duncan,

W . J. Chem. Sm, Farad@ Tram 2 1978, 74, 194. (e) Edgell, R. G.; Fragala, 1.;

Orchard, A. F. J . EIeIctron. Spcîrosc. Relut. Phenom. 1978, 14,467. (f) Lattman,

M.; Cowley, A. H. I'g. Chern. 1984, 23,24 1 . (g) Canadell, E.; Eisenstein, O.;

Rubio, J . Orgm~metallics 1984, 3, 759. (h) Ewig, C. S.; Osman, R.; Van Wazer,

J. R J. Am. C h . .Soc. 1978,100,5017. O Lin, C. S.; Tuck, D. G. Cm. J. Chem.

1 O0

1982, 60, 699. (j) Beachley, O. T., Jr.; Blom, R; Churchill, M. R.; faegri, K., Jr.;

Fettinger, J. C.; Pazik, J. C.; Vicotnano, L. OrgmmetdIics 1989,8, 346.

(4) (a) Frasson, E.; Menegus, F.; Panattoni, C. Nature (London) 1963, 199, 1087. @)

Berar, J. F.; Calvarin, G.; Pommier, C.; Weigel, D. J. Appl. Crystaffogr. 1975, 8,

386. (c) Freeman, M. B.; Sneddon, L. G.; Huffman, J. C . J. Am. Chem. Soc. 1977,

99, 5194. (d) Coates, G. E.; Green, M. L. H.; Wade, K. 'Organornetallic

Cm-, Methuen, London, 1968, Vol. 1, p.363. (e) Beachley, O. T.; Pazik, J.

C., Jr.; Giassman, T. E.; Churchill, M. R; Fettiriger, J. C. ; Blom, R. Organometallics

1988, 7, 1051.

( 5 ) (a) Jutzi, P.; Kchl, F.; Krüger, C . Angov. C h . , Int. Ei En& 1979, 18, 59. @)

Kohi, F.; JutD, P. Chem. Ber. 1981,114, 488. (c) Jutzi, P.; Kohl, F.; Hohann, P.;

Krüger, C.; Tsay, Y,-H. Chem. Ber. 1980, 113, 757.

(6) (a)Fri~H.P.;Kohler,F.H. J.Organomet.Chern.1971,30,177. @)Koridze,A.

A.; Gubin, S. P.; Ogorodnikova, N. A. J Organomet. Chem. 1974, 74, C37. (c)

Fritz, H. P. A h . Organomet. Chem. 1964,1,240.

(7) Koopmans, T. Physica 1934,1, 104.

(8) (a) Evens, S.; Green, M. L. H.; Jewitt, B.; Orchard, A. F.; Pygall, C. F. J . Chem.

Soc., Fm&y Trms. 2, 1972,68, 1847. @) Evens, S.; Green, M. L. H.; Jewitt, B.;

King, G. H.; Orchard, A. F. A Chem. Sa:. Fw@ T r m . 2, 1974, 70, 356. (c)

Green, J . C . Strrrct. Bonding (Berlin) 1981, 43,37. (d) Caiabro, D. C.; Hubbard,

J. L.; BleWis, C. H., II; Campbell, A. C.; Lichtenberger, D. L. J Am. Chem. Soc.

1981, 103, 6839. (e) Baxter, S. G.; Cowley, A H.; Lasch, J. G.; Lattman, M.;

101

Shanim, W. P.; Stewart, C. A J. Am. Chem. Soc. 1982, 104, 4064. ( f )

Lichtenberger, D. L.; Kellogg, G. E. Acc. Chem. Res. 1987, 20,379.

(9) (a) Burroughs, P.; Evans, S.; Hamnett, A.; Orchard, A. F.; Richardson, N. V. J.

Chem. Soc., C h Commun. 1974, 921. (b) Egdell, R. G.; Orchard, A. F.;

Richardson, N. V.; Lloyd, D. R. J. Electron Spctrosc. Relat. Phenom. 1977, 12,

41 5 .

(10) Shriver, D. F.; Drezdzon, M. A. " n e Muni'iation of Air-Semitive Compounds",

Second Edition, A Wdey-Merscience Publication, by John WiIey & Sons, Inc. 1986.

( I I ) Coatsworth,L.L.;Bmcrofi,G.M.;Creber,D. K.;Lazier,R. J.D.; Jacobs,P. W.P.

J. Elect. Spct. Relat. Phenom. 1978, 13, 395.

(1 2) (a) Bozek, J. D.; Cutler, J. N.; Bancroft, G. M.; Coatsworth, L. L.; Tan, K. H.; Yang,

D. S.; CaveU, R. G. Chem. Phys. Lett. 1990,165, 1 . @) Cutler, J. N.; Bancroft, G.

M.; Bozek, J. D.; Tan, K. H.; Schrobilgen, G. J. Am. Chem. Soc. 1991,113, 9 125.

(c) Cutler, J. N.; Bancroft, G. M.; Tan, K. H. C h Phys. 1994, 181, 46 1 . (d)

Sutherland, D. G. J.; Bancroft, G. M.; Tan, K. H. J. Chem. Phys. 1992, 97, 7918.

(13) Tan, K. H.; Bancrofk, G. M.; Coatsworth, L. L.; Yates, B. W. ch. J. Phys. 1982, 60,

13 1 .

(14) Liu, 2. F.; Bancroft, G. M.; Coatsworth, L. L.; Tan, K . H . Chem. Phys. Lett. 1993,

203, 337.

(1 5) Bancroft, G.M.; Malmquist, ~ . k ; Svetlsson, S .; Basilier, E.; Gelius, U.; Siegbahn, K .

Inorg. Chem. 1978,17, 1595.

(16) Bancroft, G. M.; Adams, L; Coatsworth, L. L.; Bennewitz, C. D.; Brown, J. D.;

102

Westwood, W. D. And. Chem. 1975.47, 586.

( 1 7) (a) Green, J. C. Acc. Chem Res. 1994, 2 7, 1 3 1. @) Green, J. C. Encycîopedia of

Inorg. Chem. 1994, 6, 3257. (c) Bancroft, G. M.; Hu, Y. F. (Review, to be

published).

(1 8) (a) Li, X PhD. D ~ ~ a t i o t t , The University of Westem Ontario, London, Canada,

1995. @) Hu, Y. F. PhD. Dissemon, The University of Western Ontario, London,

Canada, 1996. (c) Yang, D. S. Ph.D. Dissertation, The University of Westem

Ontario, London, Cana@ 1989. (d) Li, X.; Bancroft, G. M.; Puddephatt, R. J.; Liu,

2.; Hu, Y. F.; Tan, K. H. A Am. Chem. Soc. 1994, 116, 9543. (e) Yang, D. S.;

Bancroft, G. M.; Puddephatt, R J.; Tan, K. H.; Cutler, J. N.; Bozek, J. D. Inorg.

C h . 1990,29,4956.

(19) Yeh, J. J.;Landau,I. At. Nucl.DatuTables, 1985,32, 1.

(20) Moore, C. E. Atomic Energv Lovels, circ. no. 467, Natl. Bur. Stand., Washington,

D. C., 1952, Vol. 2; 1958, Vol. 3.

(2 1) Bahr, A; Cooper, G.; Green, J. C.; Longiey, K. A; LoveIl-Smith, M.; McGrady, G.S.

Chem. Phys. 1996,203,223.

(22) (a)Brennan,J.;Cooper,G.;Green,J.C.;Payne,M.P.;Redfeni,C.M. J. EIectron

Spech.oscc. Relat. Phenom. 1993, 66, 10 1. (b) Cooper, G.; Green, J. C.; Payne, M.

P. Mol. P h 1988, 63, 103 1.

(23) (a) Bancroft, G. M.; Bristow, D. J. C a J. C h m 19û3, 61,2669. (b) Egdeiî, R. G.;

Orchard, A F. J. Chem Soc., Faraday Trans. 2, 1978, 74, 1179. (c) Potts, A. W.;

Lyus, M. L. J. Electrm Specttosc. Relat. Phenom. 1978, 13,327.

1 O3

(24) Bancroft, G.M.; Coatsworth, L.L.; Creber, D.K.; Tse, J. Physica Scripta. 1977, 16,

220

(25) Cooper, J. W. Phys. Rev. Lat . 1964,13, 762.

(26) Bancroft, G. M.; Tse, J. S. Comments Inorg. C h . 1986,5,89.

(27) (a) Bancroft, G. M.; Boyd, B. D.; Creber, D. K. inorg. C h . 1978, 17, 1008. (b)

Brisk, M. A; Baker, A. D. J. Electron Spctrosc. ReIat. Phenom. 1975, 7, 197. (c)

vernon, G. A.; Stucky, G.; Carlson, T. A. Inorg. Chem. 1975,iS, 278.

(28) Hu, Y. F.; Bancroft, G. M.; Liu, 2.; Tan, K. H. Inorg. Chem. 1995, 31,3716.

Chapter 5

Conclusions

In the last ten years, monochromatic synchrotron radiation has been used in

combination with He 1 I He II photon sources for gas phase photoelectron spectroscopic

studies of inorganic and organometallic compounds, in which information about both energy

and intensity, as well as fine structure can be obtained. Such studies have greatly increased

the power of photoelectron spectroscopy, which has now become one of the most direct

experimental rnethods for probing the energy levels of the electrons in substances, and

characterizing their electronic structure. The recent advance in gas phase photoelectron

spectroscopic studies of organometailic compounds have been reviewed briefly in the first

chapter. The advantages of monochromatic synchrotron radiation have been demonstrated

fùrther (chapter 3 and chapter 4):

(a) For each compound, al levels of the electrons from valence to imer valence and core

levels can be studied readily in one spectrum with high resolution by synchrotron

radiation.

@) Binding energies and their shifis in both the valence and the core levels can be

measured and compared, and therefore ligand additivity effect s and core-valence

ionization correlations can be studied for compounds with successive ligand

substitutions.

(c) Information on not only ionization (binding) energy but also band intensity cm be

obtained with tunable synchrotron radiation source, which provides a firm

experimental buis for PE spectral band assignment.

1 OS

(d) The high remlution and high intensity of synchrotron radiation are critical for studies

of the fine structures in both valence and core levels caused by different splitting

effects, of the weak and broad imer-valence spectra, and of the shake-up structures

of the inner or wre level electrons.

(e) The undesirable signals due to the 'self-ionization' of helium or / and the excitations

of He II satellite lines can be examined and overcome by synchrotron radiation.

Appendix A

A Qualitative Molecular Orbital (MO) Diagram for M(CO),, M = W

tl u* - (a) a- interaction -

8-1 .' alg* : 8" 8-8,'1,

0' $4

tl u ," 8'

, ' '. '6

metai AOs ligand field complex MOs ligand

a-complex complex ligand

Appendix B.1

Synergic Bonding in the Meta1 - Carbonyl Complexes

(a) A Diagram Showing the Synergic Bonding in M(CO),, M = W.

WEAKER t BOND

(b) Pi ( x ) - bonding in the metal - phosphine bond.

Appendix B. 2

Spin-orbit and ligand field spüttings in W 5d band of W(CO),L,

Two fàctors contribute to the splitting of the W 5d band, spborbit coupling and ligand field

effect s (without considering the vibrational structure).

(A) Ligand field splitting effécts in the isomen o f W(CO),Lm, L = PMe, and n+3

The ligand field splittings in the series of W(CO),(PMq), are mainly depended on

the x acceptor ability of the substiaited ligand ( such as phosphine ligand) compared with the

parent ligand (CO). The phosphine ligands are known to be stronger o donors than CO

ligands, however theu x bonding (acceptance) ability is considered to be quite weaker

compared with CO.

The ligand x orbitals interact directly with each of the three d orbitais (4, d, Q.

In W(CO),, each of the d orbitals has four CO ligands around it, and therefore they are

degenerate (Q.

In W(CO)& &=PM@, where the z axis is de fM dong the W-L bond, the 4, orbital

is mounded by four CO ligands, however both 4, and a orbitals are surrounded by three

COS and one L. The effect of the weaker rr acceptor L is to destabilize the d, and a compared to d, by an amount A, which is defined as the ligand field splitting.

In the cis- W(CO)& isomer, whae the two L ligands are defined in the xy plane, the

a orbital is surrounded by two CO and two L ligands and therefore it is predicted to be

destabiiized by a &or of two. The d, and a orbitals are surrounded by three CO and one

L, therefore it is destabilized by a factor of one. The splitting between these two levels is

predicted to be equal to that of the mono substituted isomer, but the ordering of the levels has

reversed.

In the trans- W(CO)& isomer, where the two L ligands are defined dong the z axis,

the 4 and orbitais are surrounded by two CO and two L ligands and are destabilized by

a factor of two compared to the 4, orbital (which is surrounded by four CO ligands). The

spiitting in the tram-isomer is predicted to be twice as large as that in the mono substituted

isomer.

In the fuc-W(CO),L, isomer, aU the three 4 orbitals are destabiiized by a factor of

two since each one is surrounded by two CO ligands and two L ligands. Therefore no

observable ligand field spiitting is expected in these orbitals.

In summary, the ratio of the ligand field splitting for mono : cis : trams : foc is

predicted to be 1 : -1 : 2 : O. The result fYom our PES studies is close to this ratio (see

chapter three).

(B) Calculations of spin-orbit coupling constants (C) and ligand field splittings (A)

The method for the calculations of spinsrbit coupling constants (0 and ligand field

splittings (A) has been described in the Ph.D thesis by ~ignard', which is based on the spin-

orbit coupling theory proposed by Hall.'

The vaiues for q, a+, E. are obtained from the photoelectron spectra as defined in the

figure below, where the ionization energy (IE) follows the order of a+ < 6, < E.. Examples

for the evaiuation of the ligand field splittings and spin-orbit coupling constants are outlined

below for the isomers of W(CO),JPM%), (n = O -3) and W(CO),NBD.

I l l

IE (eV)

7.20

7.30

7.40 --

(1) Dignard, L.M. Ph.D. Thesis, The University of Western Ontario, London, 1986.

(2) Haii,M.B. J.Am.Cheni.Soc.1975,97,2057.

IMAGE EVALUATION frC*T T h - - - -

l C 3 I intwti @i\-sj

APPLIED & IMAGE. lnc 1653 East Main Street - -. - Rochester. NY 14609 USA

I -- - - Phone: 716/482-0300 -- -- - - FU: 71 6/288-5989

O 1993. Appiied Image, Inc.. AI1 Righîs Reserved