Mir: «'4‘:

an.”

aw. m... ‘

1!-
:.

MFG}.

’ ‘60

“1'

i'ysgiyirfi?sé’ifi.rn

3'

WT." '

“if"

u-d“ V

 

 

 

‘HESiS

MICHIGANS

III IIIIIII IIIIIIII III III III I‘IIIIIIII

01413 7768

ll

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

The Study of Ion/Molecule Thermochemistry of
Chiorotitanium Ions Utilizing a Hybrid Mass Spectrometer

presented by

Kurtis Richard Kneen

has been accepted towards fulfillment
of the requirements for

Ph 4) . degree in Chem I. stny_

 

%/ i/e

Major professor

 

Date 1’4"“; 7‘

MSU i: an Affirmative Action/Eq ua! Opportunity Institution 0-12771

 

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thle checkout from your record.
70 AVOID F INES return on or bdore dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

I
I
i
i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Ie An Affirmative Action/Equal Opportunity lnetltulon
Wanna-9.1

THE STUDY OF
ION/MOLECULE THERMOCHEMISTRY
OF CHLOROTITANIUM IONS
UTILIZING A
HYBRID MASS SPECTROMETER

by

Kurtis Richard Kneen

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry
1996

ABSTRACT

THE STUDY OF
ION/MOLECULE THERMOCHEMISTRY
OF CHLOROTITANIUM IONS
UTILIZING A
HYBRID MASS SPECTROMETER

By
Kurtis R. Kneen

A hybrid mass spectrometer has been modified to study ion/molecule
thermochemistry. The principle components of the mass spectrometer are:
electron-impact (EI) ionization source, magnetic sector for mass selecting
the reagent ions, deceleration lens to reduce the reagent ion kinetic energy
from approximately 1000 eV to 0-50 eV, an octopole to act as a beam guide to
assure efficient collection of product ions, a collision cell in which the gas-
phase chemistry occurs, a quadrupole for mass selecting both product and
reagent ions exiting the octopole, and a daly-type detector with which
standard pulse-counting techniques were employed. The principle of
operation of each component is described here.

Experimental procedures are presented as well as the method for
data analysis. Experiments designed to evaluated both the accuracy of the
reagent ion kinetic energy and for the collection efficiency of the beam guide
are described.

The gas-phase chemistry of chlorotitanium ions was studied using
the mass spectrometer. Collision-induced dissociation (CID) experiments

performed on TiClm+ (n = 1-4) ions generated by 70 eV E1 on TiC14 provided a

measurement of the excess internal energy imparted to the ions during

ionization. This information was then used to evaluate chemistry of the

'I‘iCln+ ions with oxygen-containing compounds and isobutene.

This dissertation is dedicated to my grandfather, Bruce D. Caulkins.
His curiosity of the natural world around him, along with his great love of
teaching others, inspired me to fulfill my goal of attaining this advanced

degree. My only regret is not to have graduated before he left us.

iv

ACKNOWLEDGMENTS

Attaining my degree would not have been possible without the moral
and technical support, guidance, and encouragement from a number of
people.

I would like to thank those who made the adjustment to graduate
school during the first couple years a little more easier: Jeff and Kimberly
Gilbert, Jon and Karen Wahl, Jason and Sue Rouse, Paul O'Connor, Gary
Schultz, Dave Wagner, Mary Puzycki-Seeterlin, Dave Gale, Art Harms,
Laura Pence and, Mike Waldo. Not only did they help me in my studies but
also provided a release valve from the stress of school through activities
such as attending football, hockey, and basketball games together, ski trips
up north, Ceder Point each summer, parties at each other's
apartment/house, and of course, "going out" on the town to experience
some of the local taverns.

I also would like to thank those who continued to show support
throughout my career at MSU: Mark and Jan Sabo, Kris Kurtz and, Jim
Ridge. These were the ones who helped me make it through the tough times
by offering advice and moral support.

Ed and Maria Townsend are special for a number of reasons. On top
of the support at chemistry, I'll miss the early Tuesday and Thursday
morning basketball games and runs across the MSU campus and going to

MSU sporting events together. Ed and I seemed to always be the ones

V

planning and organizing and something tells me that it may continue if
and when everyone from MSU gets together in the future. Maria has the
distinction of introducing me to my wife. That is something for which I
have no way of expressing my gratitude. Thank you Maria.

Tracy and I met on a blind date while we both were attending MSU.
She has provided me with the love and support only a best friend can offer.
Her willingness to postpone her education and career while I attended
MSU is the kind of support for which I know I'll never truly be able to
return.

I am grateful that my advisor, John Allison, was able to offer the
kind of encouragement and guidance I needed to pursue my career at
MSU. The weekly project meetings always seemed to offer a little humor
which made the tougher times seem more bearable. JA, thank you for

being a challenging and fair advisor. »

TABLE OF CONTENTS

LIST OF TABLES x

LIST OF FIGURES xi

CHAPTER 1

INTRODUCTION 1
1. Significance of Gas-Phase Organometallic Chemistry 1

2. Mass Spectrometry: A Discipline for Identification, Quantification,

and Basic Research 3
3. Experimental Techniques for Studying Gas-Phase Ion/Molecule
Chemistry 5
A. General Description 5
B. ICR, FTICR Mass Spectrometry 7
C. Metastable/Collision-Induced Decomposition Experiments 10
D. Ion-Beam Experiments 10
CHAPTER 2
INSTRUMENTATION 17
1. Overview 17
2. Ion Source 19
3. Magnetic and Electric Sectors 25
4. Deceleration Optics in
5. Collision Region 31
A. Beam Guide 31
B. Collision Cell 37
6. Ion Optics 43
7. Quadrupole 43
8. Detector System 45
9. Computer System 47
CHAPTER 3
EXPERIMENTAL SECTION 49
1. Initial Mass Spectrum 49

vii

2. Initial Identification of Products 50
3. Optimizing Conditions 51
4. Performing the Energy-Resolved Experiment 53

A. Stopping-Potential Scan 53
B. Energy-Resolved Scan 54
C. Background Scan 57

5. Data Analysis 57
A . Determining Ion Beam Energy Spread and Point of Zero
Kinetic Energy in the Center-of-Mass Frame 57
B. Doppler Broadening 58
C. Incorporating Ion Beam Energy Spread 62
D. Experimental Cross Sections 63
E. Fitting the Data 64
CHAPTER 4
INSTRUMENT EVALUATION 67
1. Collision-Induced Dissociation of the Singly-Charged Manganese
Dimer Ion 67

2. Reaction of Singly-Charged Argon with Molecular Deuterium 76

CHAPTER 5
CHEMISTRY OF CHLOROTITANIUM IONS 84
1. CID Experiments on 70 eV EI Chlorotitanium Ions 84
2. Chemistry Between TiClx+ and Oxygen-Containing Compounds 100
A. Results 101
B. Discussion 118
3. Chemistry Between TiClx+ ions and Isobutene 124
A. Thermochemistry 127
B. Mechanisms 135
1. Tile and isobutene chemistry 135
2. TiCl3+ and isobutene chemistry 138
3. Tile and isobutene chemistry 140
4. TiCl+ and isobutene chemistry 142
4. Conclusion 143
APPENDIX A
ELECTRON ENERGY CALIBRATION 149

viii

APPENDIX B
DEFINITIONS 152

ix

Table 4.1.

Table 4.2.
Table 5.1.

Table 5.2.

Table 5.3.

Table 5.4.

LIST OF TABLES

Parameters of the Mn2+ CID Experiment.

Parameters of the Ar+ + D2 Experiment.

Heats of Formation for Ions Produced from TiCl4 [1].

Bond Dissociation Energies for Ionic and Neutral
Chlorotitanium Molecules.

List of products for the reaction TiClx+ (x: 1-4) +
acetylaldehyde.

Summary of Bond Dissociation Energies.

78

91

135

Figure 1. 1.
Figure 1.2.

Figure 1.3.

Figure 1.4.
Figure 2.1.
Figure 2.2.

Figure 2.3.
Figure 2.4.

Figure 2.5.

Figure 2.6.
Figure 2.7.

Figure 2.8.
Figure 2.9.

LIST OF FIGURES

Block diagram of a conventional mass spectrometer 6

Schematic representation of a cubic ICR cell located in a
magnetic field of strength B in the positive 2 direction. T
and T' are the trapping, E and E' are the excitation, and D
and D' are the analyzer plates. 8

Schematic diagram of a multisector mass spectrometer
consisting of an ion source, three mass analyzers (which
can be either magnetic or electrostatic), 2 collision cells,

and a detector. 11
Block diagram of the MSU ion beam instrument. 14
Ion Beam Instrument Schematic Diagram. 18

Schematic and Voltage Diagram of Intensitron EI
Source. 2)

Illustration of a vertical transition. 21

SIMION model of the Intensitron EI Source. Equipotential
lines are shown, separated by 1 V. 23

Dispersion of the ion beam by the magnetic sector. Fall is
the focal point for ions with mass 3 and velocity 1. F82 is
the focal point for ions with mass 3 and velocity 2. Fb1 and
sz are the focal points for ions with mass b and velocities
1 and 2. %

Schematic of Deceleration Lens System. I!)

Comparison of restoring forces for a quadrupole (Quad),
hexapole (Hexa), and an octopole (Oct).

Calculated radial ion kinetic energy in rf multipole [3]. 35

Schematic diagram of the oct0pole and collision cell. 39

xi

Figure 2.10.
Figure 2.11.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

Figure 4.6.

Figure 4.7.

Figure 5.1.

Figure 5.2.

Ionization gauge calibration curve for argon.

Schematic for transfer and extraction optics and the
detector system.

Potential energy diagram of the experiment.

A stopping potential curve (TiCl3+) with the quadrupole
operated in the rf-only mode.

A stopping potential curve (TiCl4+) with the quadrupole
operated in the rf-only mode.

The first derivative curve was calculated from the
stopping potential data plotted for TiCl4+ in fig. 3.3. The

Gaussion curve was then fit to the first derivative. The
FWHM is taken as the energy spread and the apex is
taken as the point of zero kinetic energy.

Comparison of three Mn2+ data curves. Eachcurve was
normalized to its maximum value.

Comparison of Mn2+ CID threshold regions.

Fit of Run 1 of the Mn2+ experiment. ET, n, and m are
defined in chapter 2.

Fit of Run 2 of the an+ experiment. ET, n, and m are
defined in chapter 2.

Fit of Run 3 of the Mn2+ experiment. ET, n, and m are
defined in chapter 2.

Illustration of product collection efficiency of the MSU ion
beam instrument.

Energy dependence of the cross-section for the reaction:
Ar+ 4» D2 --> ArD“ + D.

Energy resolved CID of TiCl4+. Argon was used as the
collision gas.

Energy resolved CID of TiCl3+. Argon was used as the
collision gas.

xii

it

8}

71

74

75

H)

81

87

88

Figure 5.3.

Figure 5.4.

Figure 5.5.

Figure 5.6.

Figure 5.7.

Figure 5.8.

Figure 5.9.

Figure 5.10.

Figure 5.11.

Figure 5.12.

Figure 5.13.

Energy resolved CID of Tile. Argon was used as the
collision gas. 89

Energy resolved CID of TiCl+. Argon was used as the
collision gas. 90

Bond dissociation energies of chlorine atoms from TiClx (x

= 1-4) where 1 represents the loss of the first chlorine
(TiCl3--Cl), 4 represents dissociation of the last chlorine
(Ti--C1), and 2 and 3 represent the intermediate cases. 92

Trends in successive bond dissociation energies (BDEs)
[8]. The first bar in each series represents the energy
required to remove the first atom, the second bar, the
second atom, etc. 94

Comparison between proposed geometry after EI
ionization and expected ground state geometry. %

A) Product mass scan for TiC14+ + CH3CHO with the
octopole DC offset set 4 Volts below Vacce]. B) Expanded
view of the spectrum. 102

A) Product mass scan for TiCl4+ + CH3CHO with the
octopole DC offset set 20 Volts below Vaccel- B) Expanded
view of the spectrum. 103

A) Energy dependent reaction cross-sections for the
reaction TiCl4+ + C2H4O (acetylaldehyde). B) Expanded
view of plot (A). 104

A) Energy dependent reaction cross-sections for the
reaction TiCl+ + CzH4O (acetylaldehyde). B) Expanded
view of plot (A). 106

A) Energy dependent reaction cross-sections for the
reaction TiC12+ + C2H4O (acetylaldehyde). B) Expanded
view of plot (A). 107

A) Energy dependent reaction cross-sections for the
reaction TiCl3+ + C2H4O (acetylaldehyde). B) Expanded
view of plot (A). 107

xiii

Figure 5.14.

Figure 5.15.

Figure 5.16.

Figure 5.17.

Figure 5.18.

Figure 5.19.

Figure 5.20.

Figure 5.21.

Figure 5.22.

Figure 5.23.

Figure 5.24.

A) Energy dependent reaction cross-sections for the
reaction TiCl+ + C3H60 (propanal). B) Expanded view of
plot (A). 110

A) Energy dependent reaction cross-sections for the
reaction TiC12+ + C3H60 (propanal). B) Expanded view of
plot (A). 111

A) Energy dependent reaction cross-sections for the
reaction TiCl3+ + C3H5O (propanal). B) Expanded view of
plot (A). 112

A) Energy dependent reaction cross-sections for the
reaction TiCl4+ + C3H60 (propanal). B) Expanded view of
plot (A). 113

A) Energy dependent reaction cross-sections for the
reaction TiCl“ + C3H60 (acetone). B) Expanded view of plot
(A). 114

A) Energy dependent reaction cross-sections for the

reaction TiC12+ + C3HGO (acetone). B) Expanded view of
plot (A). 115

A) Energy dependent reaction cross-sections for the
reaction TiCl3+ + C3H6O (acetone). B) Expanded view of
plot (A). 116

A) Energy dependent reaction cross-sections for the
reaction TiCl4+ + CgHGO (acetone). 117

A) Energy dependent reaction cross-sections for the
reaction TiCl” + C2H6O (dimethyl ether). B) Expanded
view of plot (A). 119

A) Energy dependent reaction cross-sections for the
reaction T1012+ + CzHeO (dimethyl ether). B) Expanded
view of plot (A). 120

A) Energy dependent reaction cross-sections for the
reaction TiCl3+ + CgHeO (dimethyl ether). B) Expanded
view of plot (A). 121

xiv

Figure 5.25.

Figure 5.26.

Figure 5.27.

Figure 5.28.

Figure 5.29.

Figure 5.30.

Figure 5.31.

Figure 5.32.

Figure 5.33.

Figure 5.34.

A) Energy dependent reaction cross-sections for the
reaction TiCl4+ + CZHGO (dimethyl ether). B) Expanded
view of plot (A). 12

Reaction scheme developed by AR2 [41] for the chemistry
between chlorotitanium ions and olefins. TiCl4+ was
found to be unreactive with olefins. 125

A) Energy dependence of reaction cross-sections for the
reaction TiCl4+ + C4H3 (isobutene). B) Expanded view of
plot (A). 128

A) Energy dependence of reaction cross-sections for the
reaction TiC13+ + C4H8 (isobutene). B) Expanded view of
plot (A). 129

A) Energy dependence of reaction cross-sections for the
reaction TiC12+ + C4H8 (isobutene). B) Expanded view of
plot (A). 130

A) Energy dependence of reaction cross-sections for the
reaction TiCl“ + C4H8 (isobutene). B) Expanded view of
plot (A). 131

Proposed mechanism for the reaction between TiCl4+ and
isobutene. 137

Proposed mechanism for the reaction between Tile and
isobutene. 139

Proposed mechanism for the reaction between TiC12+ and
isobutene. 141

Proposed mechanism for the reaction between TiCl+ and
isobutene. 144

XV

Chapter 1

Introduction

*#*************************Il!#1.!3H"!*4!*III*4!I”!!!IIIIIIIMHO!*************************

I . Significance of Gas-Phase Organometallic Chemistry

Methods of synthetic chemistry have exploded both in number and
complexity over the last 50 years. As a result, members of almost every
class of organic compound can be considered starting material for the
conversion to members of other classes or more complex materials such as
polymers, biomolecules, or organometallic compounds. An exception to this
generalization is the important class known as alkanes. Their lack of
reactivity is often attributed to the high alkane C-H bond energies (~90-100

kcal/mole) and the low acidities and basicities of alkanes [1,2].

The stability of alkanes represents a fundamental challenge to chemists
because the C-H and CC bonds represent the most prevalent forms of
chemical linkage in nature. Therefore, resolving the requirements needed
to break these bonds would greatly increase our understanding of chemical
reactivity. Also, alkanes are relatively inexpensive due to their high
concentrations in both natural gas and petroleum. They represent a

significant source of carbon potentially available for chemical synthesis.

2
One area of chemistry that holds promise in achieving CH and C-C bond

activation is that of organometallic chemistry.

Understanding of the gas-phase chemistry of organometallic systems has
undergone a rapid increase over the past decade. The motivation behind the
increase in interest is the realization that knowledge about the intrinsic
properties of the bare or ligated transition-metal ions can provide valuable
insight into both the mechanisms active in the condensed phase and
catalytic efficiencies in general. In fact, correlations have been made
between results from gas-phase transition-metal ion studies and the

reactivity of transition-metal-containing compounds in condensed phases

[3].

Gas-phase studies are uniquely suited for the characterization of isolated
molecules/ions as well as probing elementary reaction steps under well
defined conditions. The studies are not complicated by solvent effects such
as solvent-shell interactions, intermolecular processes that alter or destroy
reactive intermediates, and associations by ion pairing [4]. The advantage
of gas-phase studies stems from the spatial separation of the

molecules/ions characteristic during high vacuum conditions.

Many of the first studies on gas-phase organometallic chemistry were
concerned with alkane fuctionalization to make use of the chemical
potential of the natural gas and fuel feedstocks. Today, that interest
continues as we look to uncover critical steps [5-8] and identify and
characterize reaction intermediates [9-15] so that we may improve existing

catalysts. For example, low temperature and pressure polymerization of

3
olefins is accomplished through the use of organo-aluminum/titanium

compounds known as Ziegler-Natta catalysts [16].

Transition-metals have a number of low energy atomic orbitals that provide
bonding opportunities for organic compounds. This ability to provide a
favorable bonding environment may help to stabilize highly reactive
intermediates. However, the details behind the process of bond cleavage of
inactivated OH or C-C bonds of the metal-alkene derivatives are not yet

fully understood.

II. Mass Spectrometry: A Discipline for Identification, Quantification,
and Basic Research

Since its beginnings in 1913 as a method used for demonstrating the
existence of isotopic forms of stable elements by J. J. Thomson [17], mass
spectrometry has become one of the most widely applicable disciplines of
scientific inquiry. Mass spectrometry is being used to gather qualitative and
quantitative information in fields ranging from environmental trace
analysis through the biological and medicinal sciences to studying the
surface and atmospheric chemistry on other planets. Mass spectrometers
are routinely used to analyze samples where the mass of the analytes range
from one to near one hundred thousand Daltons, with an ever increasing
upper limit. Extremely low detection limits allow for precise measurements

when the analyte of interest may be present in sub-picomole levels.

Mass spectrometers are also utilized in fields of basic research. They are

used in nuclear and atomic physics as well as fundamental chemical

4
investigations. The low pressures under which the instruments are

generally operated make them ideal for studying gas-phase chemistry.
Mass spectrometers manipulate charged reactants and products through
applied magnetic and electric fields. This provides additional advantages
when studying ion/molecule or ion/ion chemistry because of the inherent
electrical properties of the ions. When mass spectrometers are operated
under proper conditions (e.g.. single-collision conditions), the experimental
environments allow a significant amount of chemical information to be
obtained. This dissertation describes the characteristics and performance
of a hybrid mass spectrometer (ion beam instrument) designed to study gas-

phase ion/molecule chemistry.

The MSU ion beam instrument enables the study of ion/neutral chemistry
under carefully controlled conditions. Cross-sections for ion/molecule
reactions and collision-induced dissociation (CID) may be evaluated as a
function of ion translational and internal energy. Information obtained in

this manner may then be related to bond energies and heats of formation.

Another technique being developed in this laboratory allows for the
structural investigation of isolated ions [18]. This effort involves the
deposition of mass-selected ions into a low temperature noble-gas matrix,
whereafter Fourier transform infrared spectroscopy (FTIR) is applied. It is
hoped that this technique will be developed to provide direct spectroscopic
analysis of transition-metal-containing ions. The MSU ion beam

instrument will provide complementary information.

 

5
The remaining section of this chapter describes other major techniques

used to study gas-phase ion chemistry with respect to this research project.
It also provides a general description of the MSU beam machine. The
succeeding chapters detail the design characteristics, instrument
evaluations, and experimental results of gas-phase chlorotitanium ion

chemistry.

I II. Experimental Techniques for Studying Gas-Phase Ion/Molecule
Chemistry

A . General Description

Before experimental mass spectrometric techniques can be discussed, a
general description of a mass spectrometer must be included. There is no
unique mass spectrometer design. The wide diversity of mass
spectrometers, however, all consist of four basic components: sample inlet
system, ion source, mass analyzer, and detector. The entire instrument is
operated under a vacuum; with the mass analyzer and detector pressures
maintained below 10'6 torr. The sample inlet system and ion source are
generally operated at slightly higher pressures (2 10‘5 torr). A general block

diagram of a conventional mass spectrometer is shown in Figure 1.1.

The sample is injected into the instrument, ionized, and extracted from the
ion source with electrostatic optics. The ions are focused into the mass
analyzer, separated according to their mass-to-charge ratio, and then
detected. There are a wide variety of methods to perform the function of

each basic component.

.8588?on $38 32835.58 a mo Eobwmmc #85 AA 9363

 

 

58¢qu

 

 

Al

 

553mg
mmog

 

 

 

mouoom
:3

 

 

acid
395%

 

B. ICR, FTICR Mass Spectrometry

Currently there are three dominant instrumental techniques in mass
spectrometry employed to study gas-phase ion/molecule chemistry. The
first technique to be discussed utilizes ion cyclotron resonance mass
spectrometers (ICR) [19]. When a Fourier transform is performed, the

technique is referred to as FTICR mass spectrometry or FTMS [20].

When this type of mass spectrometer is employed, ions can be stored in
crossed electric and magnetic fields generated in cells similar to the cell
shown in Figure 1.2. The magnetic field restricts the motion of the ions to a
circular path in the xy-plane. Static DC potentials placed on the trapping
plates T and T' create a potential well between the two plates, thereby

restricting the motion of the ions in the z-direction.

Ions of a given mass move in a circular motion (known as the cyclotron
. motion) and there is a characteristic frequency associated with this motion.
The frequency, f, is a function of the charge, q, and mass, m, of the ion and

the magnetic field strength, B, as expressed in the equation below:

f = qB/21tm (1)

The charge of the ion, q, is equal to the fundamental charge, 1.602x10’19 C
multiplied by the number of charges. The angular frequency, to = 21tf, is
known as the cyclotron frequency of the ion. The mass/charge value of the

ion is thus related to the frequency of the ion orbit. The ion is detected by

 

 

 

 

A

 

 

 

 

 

‘L

 

 

 

 

 

 

. /

Figure 1.2. Schematic representation of a cubic ICR cell located in
a magnetic field of strength B in the positive 2 direction. T and T'
are the trapping, E and E' are the excitation, and D and D' are the
analyzer plates.

9
applying a signal at the cyclotron frequency across or between the analyzer

plates. Ions orbiting at the applied frequency absorb power and experience
an increase in the radii of their cyclotron motion. The power absorption is
measured by the circuitry of the instrument with the absorption intensity

being related to the ion population within the cell.

FTMS is performed by applying a brief signal, referred to as a "chirp,"
which consists of a complex waveform usually containing all of the possible
ion cyclotron frequencies over a given mass range. Ions within the ICR cell

absorb power and begin to move in coherent ion "packets,' with each ion
packet corresponding to a particular m/z value. These packets, in turn,
induce a signal on the analyzer plates. This signal is a mix of all of the
frequencies corresponding to each ion m/z value present in the ICR cell. A
Fourier transform is applied to this signal which results in a plot of signal

amplitude vs. frequency, which is converted to a mass spectrum once the

relationship between frequency and m/z is determined (calibration).

There are two important advantages of using an ICR to study gas-phase
ion/molecule chemistry. Because of the instrument's ability to trap ions,
ions can be stored for relatively long periods of time (millisecond range).
This allows for long periods of interaction between ions and neutral reagent
species (also present in the cell) which result in product abundances that
are higher than those found when utilizing other mass spectrometers.
Double resonance techniques employed on ICR's allow the identification of
precursor ions that give rise to individual reaction products [21,22].
Therefore, primary, secondary, tertiary, etc. reactions can be determined

unambiguously.

10

C. Metastable/Collision-Induced Decomposition Experiments

The main difference between metastable/CID experiments and the other
two techniques described in this section is that organometallic complexes
are studied by investigating their decomposition pathways. The first
application of this technique was performed by Freas and Ridge in 1980 [23].
Multisector instruments are generally employed although, in principle,
every kind of tandem mass spectrometer could be used [24]. A schematic
representation is shown in Figure 1.3. Complexes are generated through
gas-phase chemistry in the ion source, extracted, and mass selected. In the
field-free region (collision cell) following the first mass analyzer, products
are generated either through unimolecular decomposition or CID. After
drifting through the field-free region, both products and reagent ions are
mass analyzed and detected. To obtain structural information on products,

additional collision cells and mass analyzers can be used.

An advantage of this technique is that the transition-metal-containing
complexes are formed, decomposed, and detected in separate sections of the
instrument. This decreases possible interferences such as additional gas-

phase chemistry of the complexes before they can be analyzed.

D. Ion-Beam Experiments

Ion-beam experiments are performed in guided beam instruments (mass

spectrometers) designed to measure product abundances as a function of

kinetic energy. Reactions studied are between mass selected gas-phase

11

 

€230

 

 

 

 

 

 

 

 

 

E
~3wa

mmmz

 

 

N :oO
pom—086D :oEEoO
EH
Al mouthing A.
$32
_ _
_. lllllllll I_

$9298 monsoon—Sm
8:33 .8.“ 3:330

.5835“. a new i=8
oommmzoo m .Aosfimouuoflo .8 38838 .856 on :8 533 mansions mmoE 8.23 .858
com no mo wcsmmmcoo 888950QO mama 583538 a mo 8.2.me osofiofim .mA whomrm

 

A

H
. Bahamas

mmmz

 

 

 

 

 

8.25m
:oH

 

 

12
reagent ions and stationary target neutral gases at ambient temperatures.

Because the intent of these experiments is to measure the amount of energy
required to "turn on" endothermic processes, conditions have to be set to
allow single collisions. If multiple collisions were to occur, uncertainty in
the measurements would arise from the inability to accurately measure the
energy of interaction during any subsequent collisions. Therefore, to ensure
single-collision conditions in the MSU ion-beam instrument (given the
collision-cell dimensions), pressures in the collision cell were maintained

at levels below 10’3 torr.

The neutral reagent/collision gas is characterized by an isotropic
Maxwellian velocity distribution. Also, the kinetic energy of the ions within
the ion beam can be described by a distribution. These two properties
combine to create an overall distribution of ion/neutral interaction energies.
Techniques used to compensate for this distribution are described in
Chapter 3. Theoretically, it is possible to increase the resolution of the
measured ion/neutral interaction energy by incorporating a supersonic
expansion neutral beam source in place of the static neutral collision gas

cell. However, because of the low number density within the neutral beam
and the small interaction volume (determined by the ion and neutral beam
physical dimensions) [25,26], no successful crossed beam-guided ion beam

instrument has been reported.

Early pioneers of this technique were Futrell and coworkers in 1965 [27].
Their instrument consisted of two double focusing (magnetic and electric)
mass spectrometers. Beauchamp and Armentrout were the first to apply

this technique to organometallic chemistry [28-31]. Their instrument used a

13
quadrupole instead of a magnet or double sector mass spectrometer as the

product mass analyzer. Armentrout further improved this method by
employing an octopole beam guide-collision cell to ensure efficient collection

of products [26,32].

The MSU ion beam instrument is of the guided beam design, incorporating
an octopole as an ion beam guide and a static gas ion/neutral interaction
region. A block diagram of the MSU instrument in shown in Figure 1.4.
Detailed descriptions of the instrumentation are provided in Chapter. 2,
including theory and operation. Chapter 3 covers experimental procedures
and methods of data analysis. Instrument performance for both exothermic
and endothermic chemical processes is given in Chapter 4. Chapter 5
concludes with experimental results of the gas-phase chemistry of

chlorotitanium ions.

14

 

 

583mm
33

€585.5me :83 :3 sz 23 mo Eamon. x85 .vA warm

 

 

 

~3th
mmmE
Eomawwsfiu

 

 

 

22830
35.:

 

 

 

 

I

 

 

 

sumo
nosmpflooom

 

 

uBoom
032%sz

 

 

 

 

 

8.26m
:8

 

10.
11.
12.
13.

14.

15.

LIST OF REFERENCES
Chapter 1

BA. Arndtsen, R.G. Bergman, T.A. Mobley, T.H. Peterson, Acc.
Chem. Res, 28, 154 (1995).

"Activation and Functionalization of Alkanes," C.L. Hill, Ed., John
Wiley and Sons, New York, (1989).

ED. Radecki, J. Allison, J. Am. Chem. Soc., 106, 946 (1984).
AD. Ryabov, Chem. Rev., 90, 403 (1990).

R.G. Bergman, P.F. Seidler, T.T. Wenzel, J. Am. Chem. Soc., 107,
4358 (1985).

M. Hackett, J.A. Ibers, G.M. Whitesides, J. Am. Chem. Soc., 110,
1436 (1988). .

T. Gregory, P. Harper, R.S. Shinomoto, M.A. Deming, T.C. Flood, J.
Am. Chem. Soc., 110, 7915 (1988).

W.A. Kiel, R.G. Ball, W.A.G. Graham, J. Organomet. Chem., 383,
481 (1990).

A.H. Janowicz, R.G. Bergman, J. Am. Chem. Soc., 105, 3929 (1983).
W.D. Jones, F.J. Feher, J. Am. Chem. Soc., 108, 4814 (1986).

M.V. Baker, L.D. Field, J. Am. Chem. Soc., 108, 7436 (1986).

G.L. Gould, D.M. Heinekey, J. Am. Chem. Soc., 111, 5502 (1989).

RM. Bullock, C.E.L. Headford, K.M. Hennessy, S.E. Kegley, J.R.
Norton, J. Am. Chem. Soc., 111, 3897 (1989).

A.A. Bengali, R.H. Schultz, C.B. Moore, BB. Bergman, J. Am.
Chem. Soc., 116, 9585 (1994).

R.H. Schultz, A.A. Bengali, M.J. Tauber, B.H. Weiller, E.P.
Wasserman, K.R. Kyle, C.B. Moore, RE. Bergman, J. Am. Chem.
Soc., 116, 7369 (1994).

16.
17.

18.

19.
20.
21.

26.
27.
28.

29.

30.

31.
32.

16

K. Ziegler, Adv. Organometal. Chem., 6, 1 (1968).

J .J. Thomson, "Rays of Positive Electricity," , Longmans, Green and
Co., London (1913).

T.M. Halasinski, J.T. Godbout, J. Allison, G.E. Leroi, J. Phys.
Chem., 98, 3930 (1994).

K.P. Wanczek, Int. J. Mass Spectrom. Ion Processes, 95, 1 (1989).
AG. Marshall, P.B. Grosshans, Anal. Chem., 63, 215A (1991).

MB. Comisarow, V. Grassi, G. Parisod, Chem. Phys. Lett., 57, 413
(1978).

RT. McIver, Jr., W.D. Bowers, "Tandem Mass Spectrometry," F.W.
McLafferty, Ed., Wiley-Interscience, New York (1983).

RB. Freas, D.P. Ridge, J. Am. Chem. Soc., 102, 7129 (1980).
F.W. McLafferty, Science, 214, 280 (1981).

S. Anderson, F. Houle, D. Gerlich, Y. Lee, J. Chem. Phys., 75, 2153
(1981).

K.M. Ervin, P.B. Armentrout, J. Chem. Phys., 83, 166 (1985).
J .H. Futrell, C.D. Miller, Rev. Sci. Instrum., 37, 1521 (1966).

PB. Armentrout, R.V. Hodges, J .L. Beauchamp, J. Chem. Phys., 66,
4683 (1977).

RB. Armentrout, R.V. Hodges, J.L. Beauchamp, J. Am. Chem.
Soc., 99, 3162 (1977).

R.V. Hodges, P.B. Armentrout, J.L. Beauchamp, Int. J. Mass
Spectrom. Ion Phys., 29, 375 (1979).

RB. Armentrout, J .L. Beauchamp, Chem. Phys., 48, 315 (1980).
J .L. Elkind, P.B. Armentrout, J. Am. Chem. Soc., 108, 2765 (1986).

Chapter 2

Instrumentation

************************************************************************

A schematic diagram of the MSU ion beam instrument is shown in
Figure 2.1. The primary ion beam is generated by using a saddle-field
electron-impact ionization (EI) source and a magnetic sector from a
VARIAN MAT CH5-DF double-focusing mass spectrometer. Ions are
generated within the ion volume, extracted from the source, accelerated,
and focused into the magnetic sector to be filtered according to their

momentum.

The mass-selected ions are then injected, with a desired kinetic energy, into
a radio-frequency (rf) only octopole. The ions drift into a collision cell,
through which the octopole passes, and collide with neutral reagent gas

molecules.

The product ions and transmitted reactant ions are then focused into a
quadrupole, mass filtered, and detected using a scintillation ion detector

(Daly detector) employing ion counting techniques.

17

Samoa osofionom Hannah—mom Soom :3 Ad 93mg

oouoom :3

rlu

is s is
uSoSmQ Shana. ”86:30 $2 b0

mafia
/E_.IPI____III\/___ L__||=_.H./_H_
Eel... J.LE. Q: 0%me

houooeom mango
ban usage—coon

 

 

 

 

 

 

 

 

 

l______\

 

19
II. Ion Source

A schematic of the source is shown in the top panel of Figure 2.2. Electrons
are boiled off a tungsten filament (0.007 in. in diameter) and accelerated to a
kinetic energy of 70 electron-volts (eV). The electrons enter the source
region via a small slit, pass through the source, exit through a slit on the
opposite side of the source, and impinge on a collector. To calibrate the
electron energy, appearance potential experiments using nitrogen as the

standard were performed. This is described in Appendix A.

When a neutral molecule interacts with a high energy electron, absorbs
energy, and is ionized, the process is believed to follow the Franck-Condon
principle. Since the nuclei are so much more massive than electrons,
electronic transitions occur on a much shorter time scale than the response
of the nuclei. This means that the electrons are promoted to higher energy

states while the nuclei maintain their position relative to one another.

Consider the potential energy curves in Figure 2.3. Quantum mechanics
allows us to calculate the energy level of each vibrational state. The form of
the vibrational wavefunction tells us the most probable distance between the
nuclei, Yprob» by the position of the maximum amplitude of the
wavefunction. The quantum mechanical version of the Franck-Condon
principle tells us that the most probable electronic transitions are those
transitions between vibrational states represented by wavefunctions that
have maximum amplitudes occurring at approximately the same Ypmb.
This means that electronic transitions are most likely to occur when the

nuclei are separated by Ypmb. This type of transition is represented by the

electron collector
pusher

W ionization box
/ filament M /
—cI—- - - -- Ie<

gas inlet W
ionization region I

x\

N\\ extraction plates///-——/——
—-fi shield plates /
—J |__\\ z 1-deflectiorw?

 

H9

 

II
II

lens plates

r-def'lection plates /—
m

z 2-def'lection plates

 

 

 

 

 

source slit

a) section in z-direction b) section in r-direction
U)
3
.‘3
o.
c:
O
s a:
‘3 J:
b s
5‘, Ionization Region 9.

1kV

 

990 V 990 V

 

Figure 2.2. Schematic and Voltage Diagram of Intensitron EI Source

21

electronic energy
level

 

 

1 - /

 

 

 

\
X

‘7 vibrational energy
_p‘ J7 levels

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A ‘ I
E electronic energy
level
I 1
1 I
j 7
j\ [I . .
t\
\\ I; 1‘ 11 inbrlational energy

Figure 2.3. Illustration of a vertical transition.

22
vertical line in Figure 2.3 and illustrates the origin of the term vertical

transition. Other transitions can and do occur, but to a lesser extent.

The consequence of vertical transitions is that the new electronic population
densities, as a result of the electron-neutral interaction, cannot be modeled
by a Maxwell-Boltzman distribution because the new population densities
are more of a function of the shape of the vibrational wavefunctions rather
than a simple distribution based strictly on the energetic separation of the
vibrational and electronic states. Without the ability to determine the
population densities of the ion after ionization we cannot predict the

amount of internal energy it possesses.

The lower panel of Figure 2.2 shows a voltage diagram of how the source is
operated. During an experiment, the acceleration voltage would be set to
approximately 1000 V. This means that 1000 V is applied to the ionization
box. An adjustable resistive divider network within the instrument
electronics (located in the BFD module of the CH5 instrument control panel)
would then apply voltages on the extraction and focusing lenses of the
source. Each lens is wired to a resistive potentiometer and can be
individually adjusted (within a specified range determined by the divider
network) to allow for optimal conditions. The pusher voltage is adjustable to
allow the operator to select any value within the range from approximately
20 V below the acceleration voltage up to the acceleration voltage. Setting
the pusher voltage to any value below the acceleration voltage would create
a potential surface within the source that resembles a saddle. This is
illustrated in Figure 2.4 when the acceleration voltage is set to 1000 V, the
pusher: 990 V, and the extraction plates: 990 V.

.> H .3 wSouoqom .8523 one 35—13583“:an
.ooBom Hm coufimooufi 2: mo ESE ZOHEw .vfi 9:63

 

24
By creating a saddle-field within the source, the operator decreases the

kinetic energy spread of the ion beam. Consider the voltage diagram in
Figure 2.2. The beam of electrons within the source has a finite width.
Neutral molecules that drift into this beam have an increased probability of
being ionized. If the voltages are set to create a field in the source such that
the maximum potential occurs within the electron beam and decreases as
the field approaches both the exit slit and the back of the source, then only
half of the ions created would be extracted from the source. This decreases
the signal but, according to the product literature, the source is designed to
produce ions with an initial kinetic energy spread of only approximately
0.3 eV. This has been verified experimentally (experiments to be presented
later), using retarding potential field measurements with the rf-only

octopole.

The control circuitry for the electron filament is housed in the BER module.
The nominal electron voltage is 70 volts. The emission of electrons is
monitored by the collector located on the opposite side of the ion source.
Electron currents are maintained by feedback circuits and can be adjusted
from < 1 uA up to 1 mA. The module also contains controls which allow the
operator to set the source to the electron energy mode. This mode allows the
electron energy to be adjusted while maintaining a constant electron
current of 15 [1A. The current is set to prevent space-charge burnout of the

filament at low electron energies.

25
III. Magnetic and Electric Sectors

The original MAT CH5-DF double-focusing mass spectrometer deployed a
magnetic and electric sector configured in a reverse geometry. Reverse
geometry means that the electric sector follows the magnetic sector. In this
arrangement, ions are accelerated out of the source and focused into the
magnetic sector. The magnetic field, of strength B, acts on the ions with a
force perpendicular to their trajectory, causing them to follow a circular

path with a radius, r, given by

r = mv/Bez

where m is the mass of the ion, v its velocity, e is the fundamental charge,
1.60219 x 10'19 C, and z is the number of charges. As seen from the
equation, the radius of the circular path is dependent on both the mass of
the ion and its velocity. Because the product of mass and velocity is
momentum, ions are said to be dispersed according to their momenta, not
just their m/z value. To a first approximation, the ion beam, as it exits the
magnetic sector, is spatially dispersed for any given velocity (due to
different mass ions) and diverging for any given mass (due to slight
variations in velocity). This is illustrated in Figure 2.5. The image curve
shown in Figure 2.5 is a curve along which ions with the same mass and

velocity are brought into focus.

The electric sector consists of two parallel curved plates between which an
electric field, E, is applied. The ions pass between these plates and follow a

curved trajectory with radius r, given by

Velocity Dispersion +|

Mass Dispersion +|

 

F F 1
szFbl 82 8
Image Curve
Ion beam

 

 

 

Figure 2.5. Dispersion of the ion beam by the magnetic sector.
Fall is the focal point for ions with mass a and velocity 1. F32 is

the focal point for ions with mass a and velocity 2. Fb1 and sz
are the focal points for ions with mass b and velocities 1 and 2.

27

r = 2(Kinetic Energy)/E

The electric sector performs two functions. It acts both as a focusing device
for diverging ion beams and as a dispersing device according to the kinetic
energy of the ions [1]. Ions coming from a point source with a small angle
of divergence enter the electric sector and are brought into focus and follow
the same approximate radius, r. The sector acts as an energy filter in that
ions with higher and lower kinetic energies are brought into focus at points
other than the exit slit, and thereby are not allowed to pass through the

device.

One of the goals of the experiments performed on this instrument is to
control the kinetic energy of the primary ion beam as precisely as possible.
A double focusing mass spectrometer is designed to focus isomass ions to
one point (usually the detector). Beyond that point, the beam is diverging
and can have a significant kinetic energy spread. The result is a non-
collimated beam of reagent ions that is inhomogeneous in energy. By
removing the electric sector, the kinetic energy spread of the beam is
decreased, the beam is more collimated and, as an added bonus, the ion
beam current is increased due to the removal of a device with a less than
perfect transmission coefficient. Experimental measurements of the kinetic
energy spread, to be explained in later sections, before and after the
removal of the electric sector have shown that the energy spread of the ion
beam has been decreased from 0.31 i 0.02 eV to 0.25 i 0.02 eV. Quantitative
measurements of the ion beam current did not show any marked

improvement. Differences in ion currents were attributed more to

B
generating different numbers of ions in the source rather than the

transmission through the electric sector.

Operation of the magnet is controlled by circuits housed in the BYA
module. This module allows the operator to manually scan the magnetic
field (required for optimizing ion current) or to hold the magnet to a specific
value during ion/molecule experiments. The BYD module allows the
operator to initiate an electronically controlled scan of the magnet, where
the rate at which the magnet is scanned is determined by the operator. The
magnetic field strength is measured using a Hall effect probe. Spectra are

recorded by a strip-chart recorder linked to the Hall probe.

IV. Deceleration Optics

The MSU ion beam instrument has been designed to study gas-phase
ion/molecule chemistry and collision induced dissociation cross-sections as
a function of translational energy. Because the translational energy must
be accurately determined, the ion deceleration optics comprise one of the

most critical components 0f the instrument.

In order to establish a well-defined interaction energy between the ion and
the neutral reagent particle, several criteria for the deceleration system
have to be satisfied. Most importantly, the deceleration system must
decelerate the ions from approximately 1000 eV to <2-100 eV in the

laboratory frame.

Z)
The system should provide good focusing (small ion beam diameter),

convert the beam symmetry from a rectangular shape (after the magnetic
sector) to a circular shape to match the axially symmetric octopole and
quadrupole sections of the instrument, and produce a highly-collimated
beam nearly parallel to the main axis of motion. Satisfaction of these
criteria help to minimize the effect rf-only multipole devices have on the
transverse energy of ions as the ions pass through the oscillating field. A
highly collimated beam also helps to decrease ion losses during collision
processes [2]. A small beam diameter ensures the highest transmission of

the primary ion beam through the multipole device.

When employing ion optics, aberrations in the focal properties of the lenses
can be the result of small imperfections in the construction and alignment
of the various components. It is, therefore, advantageous to limit the

number of elements in the deceleration assembly.

A schematic diagram of the ion deceleration optics is shown in Figure 2.6.
Details concerning construction and characterization can be found
elsewhere [3,4]. A brief summation of the conclusions from that work is

given below.

To convert the beam symmetry from a rectangular shape to a circular
profile, circular apertures were employed. Electrical potentials placed on
circular apertures generate electric fields that exhibit steep gradients
perpendicular to the central axis. These gradients result in over-focusing
(high angular divergence) of the ion beam and is more pronounced as the

initial beam diameter approaches the electrode aperture diameter.

Einzel Focusing Deceleration
Alignment Stage Stage Stage

r'I I-'—I I ' l

- I ”L.

 

 

 

  

Guard Guard
Cyl/inder Cylinder
I" I l |\ Injection
Element
Lens R L\ens Z
>

 

Ion trajectory

Figure 2.6. Schematic of Deceleration Lens System

31
Therefore, the most significant criterion in determining the performance of

a deceleration lens system is the ratio of the initial beam diameter to the
electrode aperture diameter. In order to obtain the best performance, the
beam must be constrained to the central axis of the lens assembly and
employ lenses with large apertures relative to the ion beam cross-sectional

diameter.

Monitoring and control of the applied voltages of the deceleration lens
system are implemented through an in-house built power supply mounted
on the CH5 electronic module frame. Each lens is independently controlled
to provide maximum optimization. A description of each lens element is

provided in Appendix B.

V . Collision Region

A. Beam Guide

As the reagent ions drift into the collision cell they undergo collisions with
neutral gas-phase atoms/molecules, resulting in the scattering of reactant
and product ions. For this reason, the beam guide must provide highly
efficient trapping to collect and focus the ions into the second mass
analyzer. Another criterion the beam guide must satisfy is that it must
have no or minimal mass discrimination effects. When the ions and
neutral atoms/molecules react, many different products may be generated,
representing a wide range of masses. To make quantitative measurements,
all of the resulting ionic species must be efficiently collected and detected.

And finally, in order to control the energy of interaction with precision and

32
accuracy, the trapping process generated by the beam guide must not

perturb the initial ion kinetic energy distribution. Szabo and Hagg
published a theoretical perspective on oscillatory electric field multipole
devices and their comparative study of quadrupole, hexapole, and octopole
ion traps concluded that the higher order multipole devices were superior

as rf-only beam guides [5-8].

Radio frequency potentials are applied to the rods of the multipole device
such that the potentials of adjacent rods are 180° out of phase. These
potentials generate a dynamic heterogeneous electric field that creates a

potential well in the center of the volume encircled by the rods. Ions

entering the device are then acted upon by a potential restoring force, Ueff,

towards the center. The relationship between Ueff and the radial distance

from the center, r, is shown below

nzqzvz r “‘2
U = __o _
9" [4mw2rfi ][ ro ]

where 2n is the number of poles, q is the charge of the ion, In is the mass of

the ion, and r0 is the inner radius of the device [9]. The rf potential applied

to the rods is Vocosz(wt).

From the above expression, the potential restoring force can be plotted
against the radial distance. The plot is shown in Figure 2.7. It can be seen
from the plot that the higher order multipole devices generate potential
wells that are broader near the minimum and have steeper slopes near the
extremes. This has an advantage in that as the ions enter the beam guide,

those ions that enter off-center experience a smaller restoring potential

_
a
0.
5

4.
A22 x .3 _a

1 7m

ESP— moans.—

 

1

h.

 

 

0.0

-0.20 0.00 0.20 0.40 0.60
Relative Radial Distance (r/ro)

p.40

-O.6O

Figure 2.7. Comparison of restoring forces for a quadrupole (Quad),

hexapole (Hexa), and an octopole (Oct).

34
force. This in turn means there is, on average, less transverse kinetic

energy imparted to the ions as they enter higher order multipole devices.

Another advantage of the higher order multipole devices is that they impart
less transverse energy to the ions as the ions drift through the beam guide.
The longitudinal kinetic energy is not affected. However, the rf field does
impose a time-dependent transverse energy component on the ions
perpendicular to the ions' trajectories causing the ions to oscillate as they
drift through the device. This additional energy must be taken into
consideration when measuring the total kinetic energy of interaction

between the ions and the neutral reagent gas.

Various factors that influence the transverse kinetic energy for both
quadrupole and octopole beam guides were studied by O'Connor [3]. Using
SIMION, an ion trajectory modeling program, the relative effects of rf
phase, initial position, and initial angle of entry were investigated. The
assigned variables for the study were: 20 ms residence time, ion m/z value
of 100, initial kinetic energy near thermal (0.02 eV), initial position of 0.2 r0,
and both beams guides operating at an rf amplitude of 300 volts with a
frequency of 1 MHz. The transverse kinetic energy was plotted as a function

of time shown in Figure 2.8.

From the results, O'Connor concluded that quadrupole beam guides would
introduce high levels of uncertainty in measured ion/neutral interaction
energies. Also, O'Connor found that even though the initial rf phase

relative to the point of ion injection was of little importance for the octopole

 

O
00‘0000‘...

 

 

 

 

.0...
. .....
..o- — A
V v _ A
--_..'
A “‘...
A- : ' fl - v
emu-XI: _ A - '
.‘ H
-‘C
A “v L -- --_ 3 i

 

Wan-z :

 

 

 

 

. - --_ _ ::-===--:-c.=:”..;: {
O 0.0...
eoeeeth- - '- - ::."_ __v “30‘ 2
“$1.112- 3 A - A ‘

 

 

ii
14

m‘$-.'.::: ;' A A - - i“: - - - 'A ~ A ‘

iI>

(I.

A““‘4

It

A A—
A

"‘1

......oo.oeooo ..

12

Time in Multtpoic (ms)

«“9"

.00....OO‘....
o

.l

10

A -A

5

)Il

it
Ib
‘9

“0:”
: “Que

4
8

Octopole KE.
.......-- Quadrupole ICE.

 

 

A-‘ w

Wo...oo.oooooeoo----
0......0.0.0....o.-—=::=____

l,

l

I
-A.A

mllz:: 'A _ ______ - A
v- v — v ' coo-3:3.....‘ @

 

 

 

 

 

 

 

 

 

 

-00.“..’

go
M
n
H
H
1
lb

 

jij'ijvj'VTTV‘1‘Vf'W"V‘VVTV'YV1'1'VT'17“

a a v N O. “2 ‘9. ‘1 N. O.
' ' .4 -i .- O _O O O 0
(Ac) 16.13113 onaum asxaasuen,

‘02::00000
_ --:_--- A--‘.;e
“1332........__‘ ' A':‘-"-::A VVVV v ‘
--_ ----_- _:...O.ooo....o.oo..-.o..oo.ooo
.‘z.
”rag-:5 — :3 - ----- _ "WA :A A A i
.e::.....:.. :3:-. -_-- —:-- v- 1
A _ A ._.‘. ‘
...... .
.e 1 N
.....¢oo-O."
‘Coe-e..-_: - ‘ ‘A ‘

C

Figure 2.8. Calculated radial ion kinetic energy in if multipole [3].

36
device, the rf phase had an extreme effect on the transverse kinetic energy

in the quadrupole beam guides.

The octopole beam guide is operated by using a WAVETEK HF Sweep
Generator (model # 144) to generate the primary rf signal. The resonant
frequency of the octopole system is approximately 2.5 MHz. The rf signal is
fed into an ENI Model A-150 power amplifier, amplified, and then fed into a
modified Extrel Model E High-Q head. The High-Q head acts as a
transformer supplying the rods of the beam guide with the amplified rf

signal.

The rod assembly of the octopole and the secondary coil of the High—Q head
act as an inductor (L, the coil) coupled in series to a capacitor (C, the rod
assembly). This arrangement is referred to as an LC tank circuit. If the
circuit is operated under the correct resonance frequency conditions, power
is alternately stored in the capacitor and the magnetic field of the inductor.
Due to practical finite resistance of the circuitry, resistive losses occur
which decrease the capacitor's ability to charge and discharge. This loss in
power is compensated for by applying additional power to the system.
Because an LC tank circuit operates within a specific frequency range, its
ability to absorb power is frequency dependent. If the rf power being
supplied to the circuit is not absorbed, the signal is reflected back to the

signal source.

To maximize the forward power and minimize the reflected power, a radio
antenna tuner has been inserted in the line between the power amplifier

and the primary coil of the High—Q head. The tuner (Heathkit Model

37
HFT-9A) consists of a variable inductor and two variable capacitors. By

adjusting each component, a balance between the amplifier and High-Q

head can be achieved.

The relationship between the forward and reflected power is quantitated as
the standing wave ratio (SWR). A perfect balance is reflected by a value of
unity. To monitor the forward power and the SWR, a wattmeter (Heathkit
Model HM-9 QRP) has been placed in the line between the power amplifier

and the antenna tuner.

The DC bias of the octopole (which determines the nominal interaction
energy in the laboratory frame) is applied through a center tap on the
secondary coil of the High-Q head. A floating power supply (referenced to
the acceleration voltage) is used to generate the offset potential. The range
of the power supply is O to -500 volts and it is programmed using a 0-5 volt
driver which acts to insulate the floating power supply from the operator's
computer. The potential of the driver is generated using a 12 bit digital-to-
analog converter (DAC). The 12 bit DAC allows for 4096 different voltage
settings. This corresponds to an increment of 0.1221 volts in the nominal
ion/molecule interaction energy. The uncertainty in the interaction energy

is therefore approximately i0.061 eV in the laboratory frame.

B. Collision Cell

To ensure efficient collection of products and transmitted reagent ions, the
collision cell was constructed from a 7.62 cm long, 5.1 cm ID stainless steel

cylinder placed over the central region of the beam guide. To limit the

38
conductance between the collision cell interior and the main chamber, an

extension tube with an ID of 2.22 cm and length of 6.35 cm placed at each
end of the collision cell. The ID of the extension tubes is slightly larger than
the OD of the octopole beam guide. A schematic representation of the

collision cell and beam guide is shown in Figure 2.9.

The most significant contributions to the uncertainty of measured absolute
reaction cross sections are the determination of the density of the neutral
reagent gas within the collision cell and of the effective collision cell
length [10]. Pressure within the collision cell was initially monitored by an

MKS Baratron 310 capacitance manometer.

The capacitance manometer has several advantages. First, a capacitance
manometer is a direct-reading gauge. It measures pressure by calculating
the force exerted on a diaphragm by the incident gas. This is a physical
measurement in that the gauge monitors the degree of deflection of the
diaphragm by observing the change in capacitance between the diaphragm
and a fixed counter electrode. The advantage is that the pressure
measurement does not rely on any individual property of the
atoms/molecules (e.g.. ionization energy) that comprise the gas. Therefore,
response is independent of the chemical nature of the gas. Another
advantage of this manometer is that it does not require any net throughput
of gas particles. The reference side of the gauge is sealed, thereby
terminating the tube leading from the collision cell to the manometer. Also,
the gauge does not consume any of the gas during the pressure

measurement, which would require a net flux of gas particles. This means

doc 5658 98 2338 23 .3 89%va oumfionum .md 98mg

925 mam manna. commcmaxm

   

mvom £3390

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_’ _
\ ‘7

acom :oU :omeoO $552 3535

 

(10:

bass

40
the system reaches a steady-state whereby the pressures in the collision

cell, the tube, and the manometer are equal.

The flow of gas through a tubular-shaped object is a function of the
pressure dr0p, AP, between the ends of the object as well as the item's
geometry. Division of the throughput, Q, by the pressure drop across the
object held at constant temperature yields a property known as the intrinsic

conductance, C, of the object (C = Q/AP) [11]. Rearrangement of the equation

gives the following relationship:

Q/C=AP

Because Q is zero and capacitance is zero only when the tube length
approaches infinity, the difference in pressure between the two volumes
must be zero. This means that the pressure measured at the gauge is equal

to the pressure in the collision cell.

The major disadvantage in using a manometer, however, is its lower
pressure limit of 0.1 mtorr. Furthermore, the uncertainty in the pressure
measurement near 0.1 mtorr may be greater than i25%, even with the best
capacitance manometer currently available. In order to obtain single
collision conditions during the experiments, the pressure within the

collision cell must be no greater than 0.1 mtorr.

In order to obtain more precise and accurate pressure measurements, a
hot cathode ionization gauge was used. The operation of the ion gauge is

based on the ionization of gas particles by electron impact and the

41
subsequent collection of the ions by an ion collector. The response of this

type of pressure gauge is dependent on the species of gas present because of
the ionization process. Pressures due to gas particles with higher
ionization energies would result in lower pressure readings, because less of

the gas population would be ionized.

To improve the accuracy of the ionization gauge, the capacitance
manometer was used to calibrate the ion gauge for each neutral reagent
gas. An example of a calibration curve for argon is shown in Figure 2.10.
Working curves, ionization gauge pressure readings verses capacitance
manometer pressure readings, were generated under molecular flow
conditions. Only during molecular flow conditions will the pressure drop
across the extension tubes be linear. Knudsen's number, Kn, is used to
determine the nature of a gas; it is a dimensionless ratio between the mean
free path of a particle and a characteristic dimension of the
instrumentation [11]. In this case, the characteristic dimension is the
diameter of the extension tubes of the collision cell. Gas flow is termed
molecular when Kn > 1.0. In order to maintain molecular flow conditions
in the experimental apparatus for argon, the pressure in the collision cell
must not exceed 2.36 mtorr. With working curves like Figure 2.10, the
pressure reading from the ionization gauge can be used to measure the
pressure within the collision cell. Using the ionization gauge with the
corresponding working curve allows pressures at or below 0.1 mtorr to be
monitored with greater precision and accuracy than with the capacitance

manometer.

downs “8 023 855:8 own—Wm 5:23:0— .o~ .N 8amE

E3 we: 888252 mogmoaaao

odm owfi 0.2 0.3 0.2 o.o_ o.» o6 o.v ON 06
_J W _ _ _ a n a _ d 0.9

 

x 0N

 

O.
V

E II
8.5 x

I
Q
0

I
o.
no

r o.-

l
O.
v
'—

segmmma
+ x Eggs u a

:3
S
(.110; [’01) 33mg) uoynzguol

l
o.
\O
o—

1 o.w~

 

r odN

43

VI. Ion Optics

Ions are extracted from the exit aperture of the octopole beam guide and
focused into an Extrel quadrupole mass filter for subsequent mass analysis.
The extraction and focusing of the ions are performed by transfer optics
consisting of a stack of 5 independently controlled electrodes. Voltages
applied to the optics are generated by a Spellman high voltage DC supply
(model # UHR10P100) and a resistive divider network. Figure 2.11 shows
the geometry of the lens system. SIMION modeling has shown that this
lens configuration can provide excellent ion transmission over a broad

energy range [3].

Extraction optics have been developed to extract ions from the quadrupole
and direct them to either of the 2 detectors. Voltages applied to this set of
optics are generated from the same voltage supply and divider network
employed for the transfer optics. These optics are a modified einzel lens

design and are also shown in Figure 2.11.

VII. Quadrupole

The quadrupole mass filter is a high transmission Extrel (model #4-324-9)
combined with a Model 15 High-Q head. The device has an upper m/z limit
of approximately 200 and the High-Q head has been modified to provide
electric isolation at DC-bias potentials up to 5 kV. Principles of quadrupole
mass filters are well documented in the literature [12]. The operating

manual provides an excellent resource for theory, electronics, and

839$ 38989 on» can mow—no ~833be can Emma—duh. mo oflmfionom Afim 93mg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

omega
568280 m
/ \ a. a
‘ m
5.222.530 _
I: ..I
I
l
I
l
Sam—550m '— — £3300
.. _ L
359 k.\ m ,5 anemomaG no a _uo mam."
swam 38m \ FA 3 o a 9

Ba. 8:3
cozfixdfioaoam 2°53be

4.5
operation of the mass filter. The quadrupole mass filter is an ideal

secondary mass analyzer for ion beam experiments because it performs

mass discrimination independent of the ion energy.

Although the DC offset voltage of the quadrupole is manually set, the rf
amplitude and DC potentials associated with mass filtering are controlled
through the quadrupole electronics. These control electronics are
programmable and a second 12 bit DAC is used to drive the quadrupole. The

setup allows the m/z value to be set to approximately :01 units.
VIII. DetectorSystem

Ions may be detected by either a continuous dynode electron multiplier or a
Daly detector. Both detectors are shown in Figure 2.11. The continuous
dynode detector consists of a Galileo 4800 series Channeltron detector with
the input horn placed coaxial with the ion optical axis. The detector is
normally operated in an analog current mode for measurement of
relatively high ion currents and is used primarily for instrument
diagnostics. When used with a picoameter, this detector can quantitate ion
currents from 10'17 to 10'10 amperes. This range corresponds to count rates

of approximately 100 ions/second to 109 ions/second.

To quantitate lower ions currents (those typically seen during threshold
measurements), ion counting techniques must be employed. The
Channeltron is not an ideal detector for these measurements because it
suffers from poor initial conversion efficiencies and mass discrimination.

To perform ion counting techniques, a high efficiency Daly detector has

46
been developed. Concepts, operation, and characterization of this detector

have been described elsewhere [13].

This design incorporates a conversion dynode constructed from a 1.25"
diameter stainless steel disk with a thickness of 0.25". The surface of the
dynode has been polished optically flat by using a 1 pm diamond polishing
compound on a polishing wheel. Voltages applied to the dynode are
approximately -27000 V. Ions exiting the quadrupole are accelerated by the
electric field generated by the negative potential, impact the conversion
dynode, and sputter secondary electrons off the surface. The electrons are
accelerated through the same field to a plastic scintillator disk that has
been coated by a 3 A thick aluminum film. The scintillator is optically
coupled to a Hamamatsu R-425 photomultiplier tube (PMT). These high
energy electrons stimulate photo emission in the scintillator which is
detected by the PMT. In response to a single ion event, the PMT generates a
negative pulse (<50 ns in duration). The signal from the PMT is amplified
by an MIT F-lOOT amplifier with an integral discriminator circuit. Because
of the high degree of polishing of the conversion dynode, the we were able to
set the discriminator to its lowest setting. The amplifier generates a TTL
pulse when the amplitude of the input signal exceeds the level set by the
discriminator. These TTL pulses are then detected by a six digit ORTEC 775
pulse counter. The counter is digitally interfaced to a computer which
controls the start and stops count states. Dark counts of this detection
system are dominated by the field emission of electrons from the dynode
and are typically <20 counts/sec when standard potentials of -27000 V to the
dynode and 1000 V to the PMT are applied.

47
IX. Computer System

The ion/molecule interaction energy, quadrupole mass filter setting, and
the detector signal are all controlled/monitored using a computer system
consisting of 10 MHz 80286 PC running MS DOS. Two IBM DACA interface
boards (DACs) are controlled through a custom program [14] written in the

ASYST 2.0 computer language.

3“

S°9°>39°Sn

11.

12.
13.
14.

LIST OF REFERENCES
Chapter 2

R.G. Cooks, J .H. Beynon, R.M. Caprioli, and GR. Lester, "Metastable
Ions." Elsevier, Amsterdam (1973).

JB. Hasted In Atomic and Molecular Processes; Vol. 13, DR. Bates,
Ed.; Academic Press: New York, p. 696-720 (1962).

P.J. O'Connor, Ph.D. Dissertation, Michigan State University (1991).

P.J. O'Connor, GE. Leroi, J. Allison, J. Am. Soc. Mass Spectrom., 2,
322 (1991).

I. Szabo, Int. J. Mass Spectrom. and Ion Proc., 73, 197 ( 1986).
C. Hagg, I. Szabo, Int. J. Mass Spectrom. and Ion Proc., 73, 237 (1986).
C. Hagg, I. Szabo, Int. J. Mass Spectrom. and Ian Proc., 73, 277 (1986).
C. Hagg, I. Szabo, Int. J. Mass Spectrom. and Ian Proc., 73, 295 (1986).
L. Landau, E. Lifschitz, "Mechanics." Permagon Press, Oxford (1960).
K.M. Ervin, P.B. Armentrout, J. Chem. Phys., 83, 166 (1985).

O'Hanlon, J .F., "A User's Guide to Vacuum Technology", 2nd ed.,
John Wiley & Sons, New York (1989).

P. Dawson, Mass Spec. Rev., 5, 1 (1986).
LA. Romero; M.S. Thesis, Michigan State Unversity (1989).

The majority of the ASYST program was written by Scott Kuiphoff, an
MSU senior majoring in computer science.

Chapter 3

Experimental Section

*****************************************5|!******************************

I . Initial Mass Spectrum

As mentioned earlier, the field generated in the magnetic sector, of
strength B, acts on an ion (of mass m and velocity v) with a force
perpendicular to the ion's trajectory, forcing the ion to follow a circular
path with a radius, r, given by:

r = mv/Bez

If we also consider the kinetic energy of the ion as the result of the
acceleration voltage, V, the standard expression for the separation of ions

by a magnetic sector is given as:
m/ze = r2B2/2V

From this expression it can be seen that a linear scan of B is proportional to
ml/z. Therefore, as the magnet is scanned, peaks representing adjacent
masses appear closer together and begin to overlap at higher masses. This

means that it is important to optimize the source and magnetic sector over

49

50
the desired mass range prior to performing energy-resolved experiments.

When the chemistry of an ion with a mass near the upper limit of the
quadrupole is studied, the resolution of the magnetic sector must be
increased to completely separate ions of adjacent masses. This decreases
the ambiguity of the source of a product ion mass which could be due to
isotopes of the reactant ion or to the presence of a hydrogen atom. When
studying the chemistry of an ion with a lower mass, the resolution can be
decreased to help increase the ion current while still completely separating

ions of consecutive m/z values.
I I . Initial Identification of Pmducts

After a chemical system has been chosen for study, a series of experiments
are performed to identify all products. When the beam instrument is used
to study gas-phase reactions, three types of products can be identified:
metastable, exothermic, and endothermic products. It is important to
identify the nature of each product in order to determine the optimum lens

[potentials required to monitor them.

The experiments to identify the products begin by setting the octopole 5 volts
below the acceleration voltage and scanning the quadrupole over its entire
mass range using 1 m/z steps. Because the products are not known and
therefore the instrumentation is not optimized for each product, the
resolution of the quadrupole is decreased in order to increase the product
ion current. After the spectrum is taken, the octopole is dropped another 5
volts and the quadrupole is again scanned. This is repeated until the

octopole voltage has been dropped a total of 50 volts.

51

The spectra are then studied and possible product peaks are identified. To
determine whether more experiments should be performed with the
octopole set to lower voltages, the most intense peak at the lowest voltage is
identified. If this is a collision-induced dissociation (CID) product
generated from the reactant ion, then this indicates the energetic limit of
the chemistry being studied. This will be explained in a later section.

For each individual product, the octopole potential is then set to the voltage
that generated the most intense peak for that product. The instrumentation
is optimized and the resolution of the quadrupole is set to its maximum
(approximately unit resolution). The quadrupole is then scanned over a
range of 10 mass units, centered over the m/z value where the peak first
appeared, using 0.1 m/z steps (the smallest possible step). Each quadrupole
scan is done in triplicate. After all spectra are collected, each product is
assigned an m/z value. It is these m/z values that are monitored during the

energy-resolved experiments.

III. Optimizing Conditions

To begin the experiment, the instrumentation is set to monitor either the
exothermic or endothermic products. Because optimization is very different
for the two types of products, reliable energy dependence curves can only be
generated one type at a time, not simultaneously. Optimization is different
because the products have maximum ion currents at very different kinetic

energies.

52
The appearance of the exothermic products coincides with the appearance

of the reactant ion without any offset as the voltage on the octopole is stepped
to lower voltages. The exothermic product signals then decrease rapidly as
the energy of the system is increased. This means that the largest product
ion currents occur at very low kinetic energies. Optimization must
therefore be performed very near zero kinetic energy in the center-of-mass
frame.

Endothermic products begin to appear later during the experiment and
therefore optimization for these products must be performed at voltages
reached at the end of the experiment. This is in agreement with

observations from other groups performing similar experiments [1].

There are three sets of lenses to be optimized: the lenses of the ionization
source, the deceleration lenses between the magnet and octopole, and the
transfer optics between the octopole and quadrupole. Voltages on the source
and deceleration lenses are used to optimize the ion current of the reactant
ion and the settings are independent of the kinetic energy range over which
the experiment is conducted. The deceleration lenses have been designed to
focus an impinging ion beam over a wide energy range, 0 - 100 Volts [2].

This is sufficient for all experiments performed on the instrument.

The transfer optics between the octopole and quadrupole are used to
optimize the ion currents of the products ions. Because these lenses are not
computer controlled, the voltages on the lenses may be optimized

simultaneously for product ions with masses within 20% of each other.

53
Product ions with a greater difference in mass must be monitored over two

separate experiments.
IV. Performing the Energy-Resolved Experiment

For each experiment, a set of three types of scans must be made: the
energy-dependent chemistry experiment and two supporting scans. The
first‘scan performed is a stopping-potential scan. This scan is used to
determine the point of zero kinetic energy and the kinetic energy spread of
the reactant ion beam. The second supporting scan is the background scan.
This scan is used to compensate for collisions that occur outside the
collision cell during the main experiment. Each will be explained in the

following section.
A. Stopping-Potential Scan

As mentioned earlier, a stopping-potential curve is generated to determine
the point of zero kinetic energy and the kinetic energy spread of the reactant
ion beam. Because the overall purpose of the experiments is to measure the
energy of interaction between the reactant ion and the neutral particle, it is
important to know the zero point from which all other kinetic energy values

are determined.

The experiment is performed by generating the ions of interest and
injecting them into the beam guide. Initially, the octopole DC-bias voltage is
set equal to the acceleration voltage. This prevents any ions from drifting

through the beam guide. As the experiment is conducted, the bias voltage

54
placed on the octopole is decreased. This is schematically shown in Figure

3.1. The signal (counts/second) due to the ion beam first appears, increases
as the oct0pole voltage continues to decrease, and finally levels off to a
constant value. The ion intensity is then plotted as a function of the
difference in electric potential between the ion source and the octopole DC-
bias (Vsource-Voctopole). No gas is introduced into the collision cell and the
quadrupole is operated in the rf-only mode. By using the octopole as a high-
efliciency retarding field analyzer, the interaction region and the retarding
region are physically the same. This has two advantages in that there are
no uncertainties introduced into the kinetic energy measurements due to
contact potentials or focusing aberrations. An example of a stopping curve

is shown in Figure 3.2.
B. Energy-Resolved Scan

After it has been determined which m/z values must be monitored, the
main experiment is performed. This experiment is equivalent to the
stopping-potential scan except now a gas is introduced into the collision cell
and the quadrupole is operated in the mass filter mode. The gas pressure
within the collision cell is maintained at pressures below 0.1 mtorr to
ensure single collision conditions. As the octopole DC-bias is stepped down,
the ion currents of the different ions are individually monitored. This
allows the onsets and reaction cross-sections for the different ionic species

to be determined.

.aqoatoaxo 23 mo 88»va 3.85 3593.5 Ad 3de

 

 

LBoSoQ
ran

 

Lea—E 334
£32,320

8:90 .30
8.2.95. commmzoo

    

:8
82:8

mam—

 

..mnbaé mam—z

Camacho—moon

0955

Hill

8.532

8:80.:5
augm—

mcoq mammaoch
um comaoeaxm

oocsom

/

l>xo

INC:

 

$318113 [etquaqod

o6

.288 3:3: 2: E @8826 2883:: 05 53, C32,: 3.50 3688 9:3on < .~.m EsmE

32 .3 3.8m 2.25.
ow oé o.m o.~ o; od
“H + w iiiliiT od

1 md

I
o.
F!

I
“n.
F-I

 

T
o.
N

I
V?
N

I
C.
M

1
m.
m

 

I oé

(,01 x) oas/smnoo

57
C. Background Scan

This scan is performed under the identical conditions under which the
main experiment was conducted except for the collision gas. For this scan,
the collision gas is diverted to an inlet which feeds into the chamber
containing the octopole and collision cell. The gas is no longer being leaked
directly to the collision cell. The pressure within the chamber is set equal to
the pressure of the chamber (as monitored by an ionization gauge) during
which the main experiment was performed. This scan is performed to
compensate for any collisions that may occur outside the collision cell
during the main experiment as well as any metastable products that may

appear.
V. Data Analysis

A. Determining Ion Beam Energy Spread and Point of Zero Kinetic

Energy in the Center-of-Mass Frame

The kinetic energy scale used to analyze the data is in the center-of-mass
(CM) frame. The data, however, are collected in the laboratory (lab) frame.
In order to relate the two frames of reference, the energy distribution of the

ion beam and the point of zero kinetic energy have to be determined.

From the stopping potential data, a first derivative curve with respect to

(Vsomce-Voctopole) is calculated. The resulting curve is Gaussian in shape

and, through a least-squares method, a Gaussian curve is fitted to it. An

example of a stopping potential curve and the Gaussian fit is shown in

58
Figures 3.3 and 3.4. The position of the apex of the Gaussian curve is then

taken as the point of zero kinetic energy. The full width at half-maximum
(fwhm) of the curve is taken as the kinetic energy spread of the ion beam.
Absolute uncertainties in the energy scale due to the source, deceleration

lens, and beam guide have been determined to be less than $0.1 eV lab [2].

B. Doppler Broadening

l

A larger effect than the absolute uncertainties described above, on the
energy scale arises from the thermal motion of the neutral reagent gas
(Doppler broadening). To address this problem, Chantry developed a three
dimensional treatment of the interaction energy [3]. He began by assuming
a monoenergetic ion beam interacting with neutral molecules having an

isotropic Maxwellian velocity distribution. The total energy, Etot, available

in the center-of-mass frame is defined as:

Etot = (1/2)[m / (M + m)]MV2 (1)

where V is the relative velocity of the incident ion and neutral reagent with
masses M and m, respectively. If the assumption is made that the neutral

molecule is a stationary target, the center-of-mass energy becomes:

ECM = [m / (M + m)]E]ab (2)

Chantry then presented an exact solution to the thermal distribution with

the form:

.ovoE bcof 2: 5 H3830 28233 05 53, +19 H .8 956 3:88; 9.3an .m.m SawE

Gd

1r—

-
‘

32 $3 3:5 2.25—

o6

H
A

 

o.m

o.~

0.“

od

 

0.0

md

04

n4

o.~

Wm

o.m

Wm

o6

(901 X) pnoaasmunoo

.335 unocc— 83 mo Eon 05 an :83 a 5%
2: can 38% 3.55 05 mm :85 fl 2:3”— BE. dzfizcov “a: 05 9 E 55 33 255 53330 2:. .m.m
.wE 5 +10; he case—q 8% 3888 wcaqoa 95 Eat new—=28 33 9:3 393:3. SE 2F in 8:9”—

 

32 53 Beam 3.25.
ll Noel
o6 oé o.m o.~ o;

TclolclfilIlhio o .p . L . iii . N
O 0 “All 0 \ o O o
.- No .
35cm £852. m.
2 o :5 A
on a . m N.
1. v o m
. _W
“E 3.50 5:330 LT ..V A >0 R n 212E m
0333qu E cam—330 o l c o W
w
#1 wd m

- . e o;

 

 

61
£12.80): (marl/21 exP[ - (elf2 - £0U2)2 1 - exp[ - (2V2 + 801/2)2 1 } de (3)

where 'y = M/(m + M), e = Etot/(kaTgas) and, so = ECM/(ykagas). In most
situations, when (IQtotEmvfll/2 > (1.15 ykagas). the contribution of the second

exponential is less than 0.01% that of the first and therefore can be

neglected. Thus, provided ECM is greater than kagas (4.1x10’21 J at 298 K),

the solution, to a very good approximation, becomes:

)

as, so) =(4neo)-1/2 exp[ - (.el/2 - 801/2)2 1 de (4)

This distribution peaks at e = so and has a full width at half-maximum

(FWHM) of:
FWHM = (11.1 ykagasECMfl/Z (5)

In general, if the product ion current is small compared to the incident ion
current, the ratio of secondary to primary ion currents will be proportional

to the effective cross section, Qem given by the following integral [4]:
_, 1/2
8
an = J(—) a(£)f(e,eo)de (6)

where the factor (9180)“ 2 is proportional to the effective path length, 0(a) is

the cross section of interest, and f(€, £0)de is the distribution function defined

by equation (4).

62
C. Incorporating Ion Beam Energy Spread

To account for the energy spread of the primary ion beam, Lifshitz et al.
expanded on Chantry's [3] work and developed an exact treatment for the
distribution [5]. They assumed the initial ion energy spread could be

represented by a Gaussian function. Given the nominal beam energy, Elab.

the probability of an ion having energy E" is given by:

P(E") = algexpl:-(§'&S£) ] (7)

where S is related to the energy resolution of the instrument and is
determined by the requirement that the full width at half-maximum be 0.3
eV (S = 0.2 eV). Inserting equation (7) into (6) gives the following double

integral for the effective cross section:

« ‘:;n.exp[-Ew(.- m

—S—
0 l2 0
(exp{ _a[En1/2 _E11/2J2}_exp{-a[En1/2 +E'U2 2})£%£dE"dEd/2 (8)

where E' = [(m + M)/m]Etot and a = m/(Mkagas).

By neglecting the second term [exp(--a(E"1/2 + E'1/2)2], when E131, is large,
and expanding E"1/2 of the first term in a Taylor series about E'Uz, equation

(8) reduces to the following equation:

 

_ 1 .. M W ‘(Eo'E')2 . .
Qm—nm {(82+4E'la) exp SZ+4E'/a 0(E)dE (9)

D. Experimental Cross Sections

—ln[fI—]
_. __°_ (10)

0’ 111

The total reaction cross section, 6, is calculated from the data using
equation (10), where n is the number density of the neutral gas, 1 is the
effective collision cell length, I is the measured intensity of the transmitted

ion beam, and 10 is the sum of the intensities of all the ions I + ZIj where

the subscript i refers to the a specific product signal.

a, = o :I—‘I— (11)

Cross sections for individual product ions, 6:, are calculated using equation
(11). Accuracy of the determined cross sections are limited primarily by our
ability to measure the neutral reagent gas density and to estimate the
effective collision cell length. Uncertainties are increased in the determined
cross section values near the threshold of an endothermic process due to
the relatively small ion currents. This is primarily attributed to random
counting noise (typically 5-10 counts/sec). Other sources of the background
signal are reactions occurring outside the collision cell. These
contributions are determined by subtracting the background scan from the
corresponding main experiment. The result of the subtraction yields the

signal of products generated within the collision cell. Random scatter in the

64
data results in some negative values for cross sections near the threshold.

These are truncated to zero.

As stated before, the diffusion of the collision gas cloud into the vacuum
chamber outside of the literal 7.62 cm by 5.1 cm ID stainless steel collision
cell results in uncertainties in the length of the effective collision cell.
Althovgh the background subtraction is used to help compensate for
reactions that occur outside of the effective collision cell, the exact
dimensions of the effective collision cell are still undetermined. Therefore,

relative cross sections (G/Gmax) were calculated as opposed to absolute cross

sections.

E. Fitting the Data

After the experimental curve (relative cross section vs. kinetic energy (CM))
has been determined (using equations (10) and (11)), data analysis
continues by solving equation (9) and comparing the theoretical curve to the
experimental curve. Adjustments to the parameters that define C(E') of
equation (9) are made through iterations involving least squares fitting
which causes the theoretical curve to converge to the experimental curve.
The best 6(E') function obtained using equation (9) is then used as the
starting function for equation (8), which is then further optimized. The

functional form used for 0(E') is:

(12)

65
where ET is the desired threshold energy, 11 and m are adjustable

parameters and A is a scaling factor. All four parameters, ET, n, m, and
A, are adjusted to obtain the best fit. For the systems studied during our
experiments, the best fits were obtained when the variable m was held
equal to 1 and n allowed to vary between 1 and 2. All fits where carried out
through the use of a software program "Crunch" given to our lab by

Professor Peter Armentrout and his group [6].

9‘99”.”

LIST OF REFERENCES
Chapter 3

LS. Sunderlin, personal communication.
P.J. O'Connor, Ph.D. Dissertation, Michigan State University, ( 1991).
P.J. Chantry, J. Chem. Phys., 55, 2746 (1971).

H. Pauly, J .P. Toennies, Methods Exptl. Phys., 7A, 283 (1968).

C. Lifshitz, R.L.C. Wu, T.O. Tiernan, D.T. Terwilliger, J. Chem.
Phys., 68, 247 (1978).

The "Crunch" program was written and continues to be modified by
the Armentrout group of the University of Utah.

Chapter 4

Instrument Evaluation

************************1"!III*****JIM!********III*****************************

Two experiments have been performed to evaluate the MSU ion beam
instrument. The first experiment involves the determination of the bond
dissociation energy of Mn2+. This study illustrates the energy accuracy and
precision of the instrument. The second experiment is a study of the energy
dependence of the gas-phase reaction between Ar“ and deuterium. This
investigation illustrates the product collection efficiency of the instrument.

This section discusses each experiment in further detail.

I . Collision-Induced Dissociation of the Singly-Charged Manganese
Dimer Ion

To demonstrate that the MSU ion beam instrument can be used to obtain
reproducible, accurate relative reaction cross-section measurements for
endothermic processes, the ion beam instrument was used to perform
collision-induced dissociation (CID) experiments to determine the metal-
metal bond energy of the singly-charged manganese dimer ion. These bond
energies have been previously reported and will be used to evaluate the

performance of the MSU ion beam instrument [1].

67

68

Mn2++ Ar --> Mn" + Mn + Ar AH=O.85iO.2eV (1)

The present CID experiment was conducted using argon as the collision
gas with the chemistry occurring as shown in reaction (1). A complete list
of the experimental parameters is documented in Table 4.1. The primary
reactant ions, Mn2+, were generated by electron impact ionization of
Mn2(CO)1o. An insertion probe was used to introduce the solid Mn2(CO)1o
directly into the ionization chamber. Because of the low vapor pressure of
the metal carbonyl at room temperature, the source diffusion pump had to
be throttled to allow the pressure within the source to be maintained at
approximately 8 x 10'6 torr. This generated a primary ion beam with a
current corresponding 1.0 x 106 counts/second. To limit the amount of
excess internal energy of the primary ion, 18.5 eV electronswere used to
generate the primary ion, rather than the normal 70 eV. As mentioned in
an earlier section, calibration of the electron energy is explained in

Appendix A.

With no gas introduced into the collision cell, the base pressure of the
instrument is approximately 2.1 x 10'7 torr. The pressure is monitored by
the ionization gauge mounted to the chamber containing the octopole and
collision cell. The gauge was calibrated for nitrogen. The stopping potential
experiment was performed by sweeping the octopole DC bias through the
acceleration potential while the quadrupole was operated in the rf-only
mode. The amplitude of the quadrupole rf signal was fixed to a setting that
would normally be used to pass an m/z value of 20. The value of 20 is chosen

because it corresponds to an m/z on the low end of the quadrupole range. If

.m £289? 5 3255 98 £888qu 2: mo 822.62“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82 a :3 33 88s 3 $3 on so 2 a 2 0?
258. 9v 8 28 2: E3 22 .=o> 283.
:3 so to 8 spews: no 8-2: 538.2. 5 S m S 85cm
:9 m-» a £8 828: m < cos; w: E a .NN 83%
23 E as on .36 as H .55; 8 E m 85am
am dc a sums: a H 55:. «.2. E m .23 85cm
9720 3 w £85 3. 8.5.0.5; > .t as E n .23 850m
956 E s 598:. and £2 5882:. cm do a ._N 8:3
8888s 3 o 3955 NE 9v 6%; _85 one 5 A ._N 88m
.9 94m sacs... 955 8.35m .88 m3 3 22.5
ma 3 4 .585 m2 8 a _88 82 E 5885
93 S m 923. 3.2 E S _85 82 3 Essex
N8 5 N Ems; 858% 9V 5 _85 so 2: .35 7.1:
80 E _ sag; S: in :3 0.0 w @289

 

 

 

 

 

 

.EoEtoaxm DU +82 2: mo £20883 ._.v 033.

 

70
the amplitude value set too high, the quadrupole would perform low mass

discrimination which may result in not detecting low mass products. The
primary ion beam kinetic energy spread was calculated to be 0.27 :t .02 eV,

as determined from three separate stopping potential curves.

For the CID experiments, high purity Ar (99.999% as stated by AGA Gas
Inc.) was used as the collision gas. Pressure within the collision cell was
maintained at 0.09 mtorr as measured by the ionization gauge calibrated
for argon (corresponds to a mean free path of 2.1 m/atom). Resolution of the
quadrupole was decreased such that transmission of the mass selected ions
was nearly 100%. The decrease in resolution was possible because of the

large difference in mass between the primary ion, Mn2+, and the product

ion, Mn+.

The background pressure of the chamber was measured to be 6.4 x 10'6
torr. The reactions were monitored from one to 21 eV in the laboratory
frame. After the zero kinetic energy point is determined and the energy
scale is converted to the center-of-mass frame, the effective measurement is
zero to 4.5 eV in the center-of—mass frame. A background scan was taken
for each CID experiment, with each pair of scans performed in triplicate.
Each background scan was subtracted from its corresponding CID
experiment. Because of this background subtraction, some of the relative
cross-section values are less than zero below the CID threshold. These
negative values have been set equal to zero. The results of the experiments

are shown in Figures 4.1-4.5.

71

.02? 53:88:. m: 8 33385: 33 023 scam .823 8% +82 09:: .«o near—3:50 A .v 2sz

A20 53 mayo—é 9:25—
mé oé m.m o.m Wm 9N m._ o._ m.o

 

a a A a a a _ “m? 0.0
‘
w _d

2;. h

N Sam I 4| 0.4
fl 53 o freq
4 no:
‘
b“.
1 0
Q" 4
i...
«no I
4
I 4 a“
on...

I Nd

I md

r v.9

I md

‘ od

.. 5.0

; wd

I ad

 

I o;

nouns-SSOJQ ”pupa

Relative C ross-Sections

Cross Section (10'1,6 cm2)

0.8 ~
MSU Ion Beam Instmment . R“
0.7 a I54 A.
F
0c
0.6 — IF! fin
I z
0.5 - -’~
I o
0.4 ~ - .
'I' ‘M I lst
0.3 " . .
A
0.2 ‘ Pie 1,. 0 2nd
. “ A 3rd
0.1 r I F"

 

O [#LLL%L1LLLIII‘1111+J
I 1

0.0 0.5 1.0 1.5 2.0 2.5 3.0
Kinetic Energy (eV, CM)

Energy (eV, Lab)

 

      
  
 

°~° 5- 0 mo
. D Y I I ' ‘ * f T *
' Armentrout et. al. results [1] . . (.
15b Mn2++AraMn++Mn+Ar .

1.0

AIAAPLILJJ

0.5

frj'l'UVVTV'V—Y"

l 1 A l A l l L 1

 

 

J l 1 j l l A I. l l J A A A 1

0.0 1.0 2.0 3.0

 

Energy (eV, CM)

Figure 4.2. Comparison of an+ CID threshold regions.

.N .5320 E wanton as E Ea .: .5 325393 +82 2: .8 fl and he an .mé 8:3”—

220 .>ov 38mm 0325.

 

 

    

 

m.~ o.~ 2 o._ md od
x . . _ + . . . . n 0.90.09.01.29
, fio
E £95 I.

3.50 Sam +32 0 I «d
83 n a ,2

who.“ u a

86 u h.m.
11o

0

go

lioness-$5013 M92193

74

.m .8320 5 33% 2a 8 can .: En.— .EoEtuaxo +32 05 mo N :3. mo HE .vé 2:3"—

E 555 In
330 Sam +82 0

96 53 98m 085

o.~

—4._

 

coo; H 8
3m; H a
36 N um

I

md

fl Yo

1

nd

I

ed

l
l‘.
o

 

Houses-8801;) analog

.m 8320 E cacao—u 03 E can .: .hm .EoEtuaxo +82 05 mo m =3. .8 gm .m.w oSmE

95 .3 3.8m 2.85M

Wm o.N _ m4 0;

5 £50 I
03:0 «:5 +32 0

75

  

83 n 8
mg; n a
85 n J

- fio

I
6!
o

M
:5
[109308-88ng ”99193

1 Yo

 

76
Precision of the instrument is illustrated when the results of the three

experiments are plotted together, as shown in Figure 4.1. An expanded plot
of the threshold region is shown in the top panel of Figure 4.2. The energy
accuracy of the instrument is illustrated by comparing the top panel of
Figure 4.2 with the lower panel of Figure 4.2. The bottom panel of Figure 4.2
shows the results of a previous study on Mn2+ by Armentrout and
coworkers [1]. Their results, plotted as absolute cross-section verses kinetic
energy (CM), were calculated from data generated from an+ ions formed
by ~18 eV electron ionization. Through curve-fitting and deconvolution,
Ervin et al. determined the an+ bond dissociation energy, D°(Mn2+), to be
0.85 :t 0.2 eV [1]. The fitting of our data is shown in Figures 4.3-4.5. From
these results D°(Mn2+) is determined to be 0.91 i 0.1 eV.

The Mng+ CID experiments demonstrate that the MSU ion beam
instrument can be used to produce reproducible, kinetic-energy-accurate
relative cross-section measurements for endothermic processes involving

ions generated through electron-impact ionization.

I I. Reaction of Singiy Charged Argon with Molecular Deuterium

The hydrogen transfer reaction of singly charged argon with molecular
hydrogen and its isotopic counterpart, deuterium, has been extensively
studied to the point that this system "represents one of the most thoroughly
investigated systems in the history of ion-molecule chemistry" [2]. The
reaction, shown below, has been found to be exothermic by 1.50 eV for

ground state reactants and products.

77
Ar+ + D2 --> ArD+ + D AH=-1.50i0.03eV (2)

The energy dependent reaction cross-sections of exothermic reactions are at
their maximum at or near the lowest translational energies. Also, due to
the release of the excess energy, products may possess high translational
energies. These two factors contribute to a high degree of scattering in the
collision cell of both the reactant and product ions. Because of this
scattering, there is a great deal of difficulty in performing these types of
experiments in a longitudinal ion beam instrument. Therefore, this system
represents an ideal candidate to aid in the critical evaluation of the

performance of the MSU ion beam instrument.

The complete list of the experimental parameters is shown in Table 4.2.
Argon ions were generated by 70 eV electron ionization at a source pressure
of approximately 1 x 10'5 torr. Primary ion currents of 2.5 x 106 ions/second
were easily obtainable. With the collision cell maintained at its base
pressure of 2.5 x 10'7 torr, as measured by the ionization gauge calibrated
for nitrogen, stopping potential analyses were performed. The quadrupole
was operated in the rf-only mode with the amplitude corresponding to m/z
20. The primary ion beam kinetic energy spread was calculated to be

0.29 :t .02 eV as determined from three separate stopping potential curves.

Technical grade deuterium was used for the energy resolved reaction
studies. The pressure within the collision cell was maintained at 0.09 mtorr
as measured by the ionization gauge calibrated for deuterium. The

background pressure of the chamber was measured to be 6.2 x 10‘6 torr. The

78

.m 525%? E 3255 08 fiflofinan 05 mo macaw—con—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82 cc 5: :5 8:2. 3 as on so- 2 .8 3 up
258. cc 8 38 :2 23m. 22 .=o> ._§<
:2 20 3 a 2222: no 235 .28: .t. a E m .3 82cm
:2 2... a sou 228: m < .32; 3 cc A .NN 85%
23 E as on .230 3 H :25: S E m 8.8m
5 E a .283. a a .23 ”2“ E x .23 028m
9220 co m .2222 E 322:5 >: as... S 2 .23 825m
2220 S s 3:25 SN 52 5:83.: :- 2: 9v a 2m 858
28.28% E o 52:22 E E :83 _85 82 E A 2N 85%
2.: S m .2822 955 cc _ogm _85 :3 8 222m
5 9v 2. .222; 8. _- cc 2 s25 82 3 5822m—
Eo E m 5.25:2 2 .2 E «N _85 82 $12.22.:
as co N .222... 282.8% E _N .85 2o 2: .35 3:
as E _ .2822 :2 __sm :2 m-» _ A332:

 

 

.2585qu NO + +54. 2: «0 280823 NV 035.

 

79
reactions were monitored from 1 to 9.5 eV in the laboratory frame which

resulted in a range from 0 to 0.63 eV in the center-of-mass frame. To
achieve the highest ion beam current, the quadrupole was operated with
the lowest resolution that completely separated the primary reactant ions

from the product ions.

Results of the experiments are shown in Figures 4.6 and 4.7. Stopping
potential experiments were performed by monitoring the primary argon ion
current as a function of kinetic energy with no gas introduced into the
collision cell. Deuterium was then introduced into the collision cell. The
reactant ion, Ar+, and product ion, ArD+, were then monitored as a
function of kinetic energy. A background scan was then taken, as described
in an earlier section. Each scan was performed in triplicate. As shown in
Figure 4.6, the sum of the transmitted ion beam current and the product
ion current is almost equal to the primary ion intensity when no gas was
present in the collision cell. These results indicate that the dynamic
trapping process of the octopole is effective at collecting highly scattered

reactant and product ions even at low interaction energies.

The data were collected and converted to the center-of-mass frame as
described in an earlier section. The energy-dependent reaction cross-
sections are shown in Figure 4.7. The curve is an average of three separate
experiments. As expected, the cross-section is at a maximum at the lowest
interaction energy and decreases exponentially as the energy is increased.
This is consistent with Langevin—Gioumousis-Stevenson rate theories based
on ion/neutral captures driven by ion/dipole and ion-induced/dipole

attractive forces [3].

2:2:ng Enos :2 DmE 05 mo .8565: 5:00:00 8385 mo 2.2852: 6.... BamE

32 .>2 2.522 2322

 

 

 

 

 

- We
:8 222:8 2 6 .EE 2. +2 2 saw I...
:8 222:8 2 Na .+:< +

:8 222:8 2 d was. IT - 3
:8 82:8 2 5 oz .1... IT

1 m2

- o.~

2 3

 

(,oI) pauses/saunas

81

.D + b? A: NO + .2. ”22828 2: 8.2 2.288-280 2: mo 85988: $225 .5“ 2:me

Nmo

 

2o 0o 2o m.o No 2.0 0.0
_ O 0% w L. + m +
o oooooooooooooooooooooooooo o
.00 09909990009090... 9999909000 11

O.

00

0.

 

r

o.o
o.N
oé
0.0
o.w
0.3
92

0
fl'
‘I-t

0.3

(can "_01) uoyoas-ssoJQ

82

The monitoring of the reaction between singly-charged argon ions and
molecular deuterium has demonstrated the high trapping and
transmission efficiencies for exothermic ion-molecule reactions attainable

with the MSU ion beam instrument.

LIST OF REFERENCES
Chapter 4

KM. Ervin, S.K. Loh, N. Aristov, P.B. Armentrout, J. Phys. Chem.,
87, 3593 (1983).

K.M. Ervin, P.B. Armentrout, J. Chem. Phys., 83, 166 (1985).
G. Gioumousis, D. Stevenson, J. Chem. Phys., 29, 294 (1958).

Chapter 5

Chemistry of Chlorotitanium Ions

***************************ItIIIIII*5!It*********1!*t**************************
I

I . CID Experiments on 70 eV EI Chlorotitanium Ions

In 1968, Kiser, Dillard, and Dugger (KDD) performed ionization and
appearance potential experiments to determine the heats of formation for
all of the positive chlorotitanium ions generated from titanium
tetrachloride [1]. Data obtained were plotted as ionization efficiency curves
by the semi-logarithmic plot technique described by Lossing, Tickner, and
Bryce [2]. Potentials were then extrapolated from the data using a
procedure outlined by Warren [3]. A brief description of the procedure is

given below.

The instrument on which the experiments were performed was a time-of-
flight mass spectrometer employing electron-impact ionization. Two gases
were leaked into the instrument simultaneously; one was the compound
under investigation, while the other, a standard (nitrogen), for which the
ionization energy is already accurately known. Measurements of the ion
current were made as a function of electron energy for both the unknown

and standard.

85

Partial pressures of the gases were adjusted to give signals of
approximately the same intensity for each gas when the electron energy is
set to 50 eV. The peak heights were then plotted on a logarithmic scale as a
percentage of the peak height at 50 eV as a function of electron energy.
Because the curves were parallel within experimental error at the one
percent level of the 50 eV peak or below, voltage differences were measured
at the one percent level. The difference between the voltages required to
generate a peak one percent of the height of the 50 eV peak for each
compound was taken as the difference between the ionization and
appearance energies of the two compounds. Results of the experiments are

shown in Table 5.1.

Table 5.1. Heats of Formation for Ions Produced from TiCl4 [1].

 

 

 

 

 

 

Ion AHf (ion) BDEl
(kcal/mole) (kcal/mole)
TiCl3*--C1 87 38
TiClz+--Cl 96 78
Tier--01 145 90
Tit-Cl 206 102
Ti“ 279

 

 

 

 

 

1Calculated from information provided in the paper.

KDD calculated the AHf for each ion using the general reaction below:

TiCl4 —. T101,+ + (4-x)Clo (1)

86
By including -181.6 kcal/mole for the AHf of TiCl4, 29.08 kcal/mole for the

heat of formation of Cl- , and measuring the appearance energy of each

'I‘iCl,‘+ ion (which is equal to the Aern) the authors were able to generate

values for the heats of formation for the four chlorotitanium ions.

Bond dissociation energies (BDEs) were calculated using the general

equation below for x = 0-3:

AHf (1301;) + AHf(C1-) - AHf(TiCl(x+1)+) = BDE(TiClx*--Cl) (2)

We have performed collision-induced dissociation experiments on
chlorotitanium ions generated by 70 eV EI. Experimental procedures and
data analysis were described in Chapter 3. Plots of our experimental

results and curves of best fit are shown in Figures 5.1-5.4.

Table 5.2 lists the results of our experiments, the BDEs calculated from the
heats of formation determined by KDD, and the BDEs of the neutral
chlorotitanium molecules calculated from the heats of formation reported

in reference [4].

.me 222:8 05 mm :8: $3 coma 30:. mo 90 3282 Exam Ad 8:me

:5 $2 22.2 2.25.
E S 2 23 we to No 0

A

 
 

 

 

We

 
 

coco; H E
35m; n 2
>036 u 2m

I
°°.
o

 

uopoas-ssmg) M93193

7P

.mmw 222:8 2: 3 cow: was» coma 30E. mo 90 3282 .35an .~.m onE

340 $3 .333.— 9:25—
m.m m we v Wm m Wm N m._ 2 We

._
—<r—

 

. 88.2.1.2
EB“:
.. . >s 521m

 

mo

5. ‘0. 'n. “I "l
O O O O O
comes-smug axpnpu

°°.
9

ad

.mmw 222:8 2: mm :82 83 28.3. SUP mo 90 8388 Exam .m.m 22mm”:

:5 $2 22.2 2.222
8.2 8.2 cos 8.: com cod cod cod co:

— : a _ _ :

 

LDC. o o 00 o COCOA H E

.. . 22: u a

oo o. oo >uwh.~flhm
o o. o

 

l

5.0

r ad

 

uoyaas-ssmg analog

.mmm 222:8 2: m2 .88 22>» 282 «5:. mo 90 8282 $82M in 22me

:20 $2 .882: 2.25.
0.2 8 8 S. 8 8
. . . . _ . 3...; 8
. S

 

 

O
O
O

O
T

I

No

l

m:

I

2.:

l

m:

l

0.0
coco; H E .
om_w.— H 2 b o
. . >o 3N u 2m r w.o

l

 

I

O O O Q.o

 

names-smug GAQBPH

91

Table 5.2. Bond Dissociation Energies for Ionic and Neutral
Chlorotitanium Molecules.

 

 

 

 

 

 

Ions Neutral
BDEl 13m:2 BDE3
(kcal/mole) (kcal/mole) (kcal/mole)
TiC13--Cl 8.3 38 82.5
, Ti012--Cl 64.7 78 101.2
TiCl--C1 41.0 90 122.6
Ti--Cl 63.8 102 105.1

 

 

 

 

 

 

1Results from our experiments.
2BDEs calculated from heats of formations determined by KDD.
3BDEs calculated from heats of formation reported in reference [4].

A graphical comparison of the bond dissociation energies is shown in
Figure 5.5. The techniques used by the two studies to generate the BDEs for
the ions are very different. The approach used during our study involved
generating chlorotitanium ions and injecting a single TiClx+ species into
the collision cell. This enabled us to measure the amount of energy
required to remove one chlorine atom from the ion. This approach is
limited by the unknown amount of internal energy imparted to the ion
during the ionization process. The procedure employed by KDD generated
ions through the successive loss of atoms following ionization. All the

chlorotitanium ions were believed to be derived directly from TiCl4. The

probability is low that each chlorine atom will depart with zero momentum

and leave all the energy available for further dissociation of the TiClx.1+

fragment. Energy carried away from the system by the exiting atom must

be compensated for and requires higher electron energies. The approach

.880 o§3§§£
05 882%: m can m 98 .cuib 8:25 .8. 05 .6 cocaoommE aaomoae v .A_U--M_U_.C onto—go 82c 05 .8 $2 2:
$5338 _ 823 clubs +19 ._. Ea clans JOE. Eat mesa 2:820 .8 flamenco 223885 25m .n.n 959;

553982.. 88a 05..ch E 85.5%

 

v m N 2 o
_ a _ _ o
I on
m
#53 £5 P
T ow .
/ .
./. ,./ .I 8 m.
8:8 COM M
-e m
, m
/ - 82 w
./ m
8:8 .830: m
1 ofi (

 

[ OE

93
followed by KDD is also limited by the ability to accurately determine the

electron energy and the efficiency of energy transfer from the electron to
the molecule during the ionization process. A more complete explanation
is provided in Appendix A. Therefore, only upper limit values can be

derived from appearance potential experiments.

The consecutive BDEs determined by our study follow an alternating
"weaker-stronger" pattern. This differs from the pattern of increased
BDEs determined by KDD. Below is a plot of trends in the central ion-atom
BDEs (Figure 5.6) which shows there is only one known series (MoF3+)
where the successive removal of each halogen atom requires a greater

amount of energy.

Titanium tetrachloride is known to be tetrahedral [5-7] which would
suggest an sp3 hybridization. Ionization of TiCl4 most likely involves
removal of an electron from one of the unshared pairs of electrons on one
of the chlorine atoms, because these electrons are the least tightly bound to
the molecule. Other valence electrons participate in the bonds between the
metal center and the chlorine atoms. The ion is believed to retain the
tetrahedral shape of the neutral molecule. This is shown in Figure 5.7.
The ionization energy of a chlorine atom (12.967 eV [8]) is greater than that
of titanium (6.82 eV [8]) . This could result in the transfer of an electron
from the titanium atom to the chlorine atom with the partial charge now
residing on the metal center. The loss of the electron decreases the electron
density about the metal center which increases the repulsion between the
titanium nucleus and the chlorine nuclei, thereby decreasing the bond

strength. The average bond order would decrease to 7/8 (0.875). This may

 

 

 

 

 

 

 

 

 

   
  

 

 

  

 

 

1606

1401

120 A

(atom/mu) 308

 

the energy

series represents

m each

(BDE's) [8]. The first bar

ton energies

t

' bond dissoci

1n SUCCESSIVC

Figure 5.6. Trends

required to remove the first atom, the second bar, the second atom, etc.

Configuration
After EI Ionization

\
TiCI4
/\

. l'e'
\ .o“
TiCl4+

 

 

 

 

A
l a) tetrahedral
\/+\ TiC13+
i c) trigonal pyramid
'+
lex Ti012+
C1 C1
l e) bent
'I‘i+
\ T'Cl+
C1 1
g)

Proposed
Ground State
Configuration

 

\
\ “go
A

b) tetrahedral

)a

d) planar

C1 — Ti+— Cl

f) linear

Ti+- Cl
h)

Figure 5.7. Comparison between proposed geometry after EI
ionization and expected ground state geometry.

96
explain why the BDE for the removal of the first chlorine atom determined

by KDD and our experiments is lower than that of the neutral species.

Our determined BDE is 29.7 kcal/mole lower in energy than that
determined by KDD for the dissociation of the first chlorine atom. The
nominal electron energy used during our experiments was 70 eV as

compared to approximately 13 eV electrons used by KDD. This means that

the energy of interaction between the electron and TiCl4 is much higher
during our experiments and because TiCl4 has many degrees of freedom
which can store energy, the resulting ion may be excited. An excited ion
would have a lower apparent BDE. The BDE determined by KDD may also
be slightly higher than the "true" BDE as a result of having to compensate

for the departing chlorine atom's momentum.

Titanium trichloride (TiCl3) is known to be planar [7] which implies an sz

hybridization. The third p-orbital, which does not participate in the
hybridization, would serve as an empty orbital into which the chlorine
atoms could donate electrons through ligand 1: donation. TiCl4 has no
opportunity for back bonding, which could explain the increase in BDE for
dissociation of the second chlorine atom from the neutral molecule.
Influence of the unpaired electron in one of the d-orbitals would not be
significant because influence of d electrons decrease as the both the ligand

atom size and electronegativity increase [6].

The titanium atom in TiC13+ is in the +4 oxidation state. Performing CID

on TiCl3+ results in a change in the formal oxidation state of the metal

from IV to III. This is similar to the neutral case when a chlorine atom is

97
removed from TiCl4 and, as seen from Figure 5.5, the BDEs of the Tile

are approximately equal to the BDE of the neutral TiCl4 case.

Both studies of the Tile ion, ours and that of KDD, found an increase in

BDE when the second chlorine was dissociated from the metal center. The
discrepancy between the BDE's, 13.3 kcal/mole, may again be attributed to

the difference in interaction energy between the ionizing electron and

TiCl4. KDD attempted to measure the onset at which TiCl3+ appeared. The

result of their experiments may have been a TiCl3+ ion with a planar

geometry as shown in Figure 5.7d. The experiments performed here at
Michigan State may have resulted in an ion with a trigonal pyramid

geometry, Figure 5.7c. This geometry would be similar to the geometry of a

vibrationally excited TiCl3+ ion. However, TiCl3+ has fewer degrees of

freedom with which to store energy than TiCl4+. This may contribute to the

decrease in the difference between the two studies (13.3 verses 29.7

kcal/mole).

The oxidation state of the titanium in Ti012 is II and the bonds may be

essentially ionic [7]. Two anions ionically bound to a (+2) cation would

explain the higher BDE of the third chlorine (third Cl atom removed from

TiCl4) from the metal center. The titanium atom in TiClz+ has an

oxidation state of III. Removing a chlorine atom from this ion is similar to

cleaving the titanium-chlorine bond in TiC13 in that the metal center is

reduced from an oxidation state of III to II.

The difference in the determined BDEs for the ion, TiClz”, between our

results and those of KDD may stem from the difference in the experiments.

frm

sel

98
KDD generated the ions directly from TiCl4, immediately extracted them

from the source, and used a time-of-f'light mass spectrometer to mass
select and detect the ions. Their experiments were essentially performed
in one step. During our experiments, we first generated the precursor ion

and then performed CID experiments on the ion.

When TiC12+, a doublet, is generated from TiCl4, the resulting ion may
initially resemble a tetrahedron without two ligands, i.e., would be bent,
Figure 5.7e. If the ionization and dissociation processes occur before the
remaining atoms can adjust, the resulting ion may initially have a highly
strained geometry as shown. This geometry would be equivalent to a
vibrationally excited linear ion. Performing CID experiments on this

excited ion would result in low BDE's.

'I‘iCl has a lower BDE than TiClg because the ionic bond for TiCl is between
two singly charged ions. As stated above, the bonds in TiC12 are between a
(+2) ion and two (-1) ions. The higher charged state would produce

stronger bonds.

For the ion, TiCl”, breaking the bond would be similar to the neutral case
when a chlorine atom is dissociated from TiClg in that the titanium atom
would be reduced from an oxidation state of II to an oxidation state of I.
However, in the case of the ion, KDD would be breaking covalent bonds of
TiCl4, not ionic bonds and therefore would require less energy. During our
experiments, TiCl+ is first generated and then CID experiments are
performed on the ion. After the ion is generated, the geometry may

resemble a vibrationally excited state. But, because TiCl+ has fewer

99
degrees of freedom, less energy is internally stored. Having less internal

energy would require higher CID energies to break the bond. Also, the
increase in BDE (from 'I‘iC12+ to TiCl‘”) may be explained by examining the

nature of the bond in TiCl“.

The energy required to generate Ti+2 from Ti+ is 313.9 kcal/mole [29]. The
energy released when a gas-phase chlorine accepts an electron is
83 keel/mole [calculated as the AH for the reaction Cl(g) --> Cl'(g)] [29]. The
difference between the ionization energy of Ti+ and the electron affinity of
Cl is 231 kcal/mole. When we use Coulomb's Law to evaluate the potential
energy between the two ions (Ti‘”2 and Cl'), we find a distance of <2.9 A
between the two ions is required to generate a potential energy equal to the
231 kcal/mole energy barrier. This would lead us to believe that the bond is
an ionic bond between Ti”2 and C1" instead of a covalent bond between 'I‘i+
and C1, the energy required to separate two oppositely ions is greater that

the energy required to separate a neutral atom from an ion.

As mentioned above, the ions studied during our experiments were
generated by performing E1 on TiCl4. An attempt to mathematically
correct for the BDEs determined during the present study would not be
possible since the ionization process is believed to be dominated by vertical
processes that do not allow the state populations to be characterized by a
Maxwell-Boltzman distribution. It is also noted that the curves in Figures
5.2 and 5.4 are not well fit in the threshold region. The shape of the
experimental plots resembles plots generated by Ervin, Loh, Aristov, and
Armentrout [9] during their study on the effect electron energy has on

determined thresholds. They concluded that higher electron energies

100
decrease the apparent onset. Reasons why data in Figures 5.1 and 5.3 do

not possess the same characteristic are unknown.

Therefore, considering that our determined BDEs may be those
characteristic of excited ions and that the determined heat of formation for
Ti+ by KDD agrees very well with others [9, 10], we conclude that the
difference between our determined BDEs and those determined by KDD
represent an estimate of the minimum amounts of internal energy the
ions possess. Our results are consistent in that the experimentally
determined BDEs for the excited chlorotitanium ions never exceed the BDE
values determined by KDD. With these approximations, we are then able to

gain more insight into the chemistry of chlorotitanium ions generated by

70 eV EI.

I I . Chemistry Between 'l‘iClx“ and Oxygen-Containing Commands

Titanium compounds containing oxygen have dominated the
organometallic chemistry of this metal in the past [10,11]. The first organic
derivative of titanium was an alkoxytitanium compound prepared in 1875
[12]. Oxygen-containing titanium compounds have been used in paints, in

the treatment of textiles and paper, and as polymerization catalysts [10,11].

Titanium tetrachloride (TiCl4) has been used as a Lewis acid to catalyze
condensation reactions involving enol acetates, acetals, carbonyl
compounds, and other oxygen-containing organic molecules [13-16].

Studies of the gas-phase chemistry between chlorotitanium ions and

101
oxygen-containing compounds at room temperature have also been

conducted [17].

More recently, the organometallic chemistry of titanium metal has
focused on the chemistry of the bare-metal ion in the gas phase [18-40].
Here we present an attempt to investigate the endothermic and
exothermic, gas-phase chemistry of chlorotitanium ions with oxygen-
containing compounds. These studies were carried out in the hope that

any discovered processes may help us to more fully understand reactions

characteristic of condensed phase titanium compounds.

A. Results

Exothermic products are more abundant at lower kinetic energies while
products generated through endothermic processes are more abundant at
higher kinetic energies. For this reason, in order to identify all products
for a given system, several product ion spectra are collected over a range of

kinetic energies.

Figures 5.8 and 5.9 show the product spectrum of TiCl4+ + CH3CHO

(acetaldehyde) at two different kinetic energies. These two spectra are used
to determine the mass values monitored during the energy-resolved
experiment. The results of the energy-resolved experiment are shown in
Figure 5.10. As seen from Figures 5.8-5.10, there are two exothermic

products, TiCl302H4O+ (III/Z 197) and TiC12C2H3O+ (m/z 161). Only the

product appearing at m/z 197 was previously reported by Allison and Ridge
(ARl) [17]. The endothermic product is TiCl3+ (m/z 153). The appearance of

5.E+6 —

4.E+6 .

L»

m

3‘
A

Counts/sec

N
m
3;

41

1.E+6 «

 

102

(A)

 

all.“
h l l

 

O.E+0 I

7.E+4 -

6.E+4 4

3'"
m
E

4.E+4 ~

Counts/sec
ii;
i

.N
m
:

LEM -

20 40 6O

 

O.E+O ‘

20 40 6O

80 100 120 140 160 180 200

m/z

(B)

 

 

 

 

1

 

80 100 120 140 160 180 200

m/z

Figure 5.8. A) Product mass scan for Tile + CH3CHO with the octopole
DC offset set 4 Volts below V“. B) Expanded view of the spectrum.

4.E+6 ~

3.E+6 4

Counts/sec
'm
3

1.E+6 -

 

 

 

(A)

 

0.E+0

4.E+4 '1

3.E+4 4

Counts/sec
.N
m
K

1.E+4 -

 

 

100 120 140 160 180 200

(B)

 

 

0.E+0 -

Lt

100 120 140 160 180 200

 

Figure 5.9. A) Product mass scan for riot.“ + cnscno with the octopole
DC offset set 20 Volts below v.00... B) Expanded view of the spectrum.

104

 

 

1.2E+04 -,
(A)
1.01904
"s
U
3 8.0E+03
S
‘5' 601,303 1 A m/z 197,TiC1_,C2H.,o+
f3 , x m/z 153, TiCl3+
I § . ‘
U '
0.0E+00 ; ~ --------
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
Kinetic Energy (eV, CM)
4.0E+03 1.-
5 (B)
3.6E+03 ] 5
3.2E+03

2.8E+03
2.4E+03
2.0E+03
1.6E+03
l .2E+03
8.0E+02

Cross-Section (10“ cm’)

4.0E+02

 

0.0E+00

 

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50

Kinetic Energy (eV, CM)

Figure 5.10. A) Energy dependent reaction cross—sections for the reaction
TiCi,+ + c.1440 (acetylaldehyde). B) Expanded view of plot (A).

105
the collision-induced dissociation (CID) product, TiCl3+, signals the limit

of the amount of energy that can be injected into the system through
kinetic energy. As the kinetic energy is increased, the CID product
abundance increases until its signal reaches a maximum and then

decreases to zero. At higher kinetic energies, there are no product signals.

Figures 5.11-5.13 show the results of the energy-resolved experiments for
the reactions 'I‘iClx” (x = 1-3) + acetaldehyde. In each case, two exothermic
products were observed with a CID product appearing at higher kinetic
energies. For each experiment, the exothermic product with the higher
reaction cross-section is the product observed by ARl. Allison and Ridge

did not report the observation of the second exothermic product.-

The experiments performed by ARI employed an ion cyclotron resonance
(ICR) mass spectrometer which did not allow them to precisely inject
additional energy beyond that of the ionization process. Motion within the
mass spectrometer is governed by thermal energy and the magnetic fields
generated by the magnet. The products they observed would have to be the
result of either exothermic or endothermic processes with barriers small
enough to be easily overcome by the energy imparted to the system during
ionization. The exponentially decaying curves generated from our
experimental data suggest these products to be the result of exothermic
processes. We conclude there are no endothermic processes within this

system available to the injection of kinetic energy.

The next system we studied involved a slightly larger aldehyde: propanal.

Results of the energy-resolved experiments from the TiClx+ (x = 1-4) +

106

 

 

1.6E+05
(A)
1.4E+05

F

a 1.2E+05

U

2

b 1.0E+05

_

v

8 8.0E+04

0?

r1.) 6.0E+04 . +

, a o A m/Z 48, Ti

0 . .

5 4-0E+04 x m/z 64, “no+
2.0E+04 1 O m/z 99. TiClO+
0.0E+00

0.00 2.00 4.00 6.00 8.00 10.00
Kinetic Energy (eV, CM)
1.6E+03 1
(B)
X A
1.4E+03 ~ 4“
X

A x ‘i?

"E 1.2E+03 ~ x. mu

0

2 2"

b 1.0E+03 . ",

—1 x5: 5*

v X l“

8 8.0E+02 - “2 x i

i»: ' ‘2. . 5‘“

v.2 6.0E+02 ~ xii: “

n “will“

a

8 40E+02 . ‘W
. “ A

U °. 5.: * xw
2.013+02 ~ ‘9 558'; ,

£ . . .9. 9 o t’sit
5“ o.oo~.. °~’° 0:“ of.
0.0E+00
0.00 2.00 4.00 6.00 3.00 10.00
Kinetic Energy (eV, CM)

Figure 5.11. A) Energy dependent reaction cross-sections for the reaction TiCl+
+ C2H40 (acetylaldehyde). B) Expanded view of plot (A).

107

3.5E+05

(A)

3.0E+05
2.5E+05
2.013405

"5905 A m/z 126, Tic1c2H3o+

1.0E+05 X m/Z 99. TiClOJr
. m/z 83, TiCl+

Cross-Section (10.15 cmz)

5 .0E+04

0.0E+00 _—1

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Kinetic Energy (eV, CM)

3.5E+03 -

(B)

3.0E+03— ,,
2.5E+03 - e

2.0E+03 a

 

1.5E+03

1.0E+03

Cross-Section (10“ cm’)

5 .0E+02

 

0.0E+00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Kinetic Energy (eV, CM)

Figure 5.12. A) Energy dependent reaction cross-sections for the reaction
TiCiz+ + c.1140 (acetylaldehyde). B) Expanded view of plot (A).

108

2.5E+05 '1
(A)
i)
A 2.0E+05 ~
E
U
3; 0
3 1.5305]
2:
.2
fig .
a.) 1.0E+05 i
5 J A m/z 118, Ticlz+
r . +
U 5.03.04 . x m/z 126, T1C1C2H3O+
' O m/z 161, T1C12C2H30

 

O
0
0.0E+00 w

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00

Kinetic Energy (eV, CM)

5.0E+03 1 0

4.5E+03 1 (B)
4.0E+03‘1 .

3.5E+03
3.0E+03
2.5E+03
2.0E+03
1.5E+03
1.0E+03

Cross-Section (10“ cm’)

5.0B+02

 

0.0E+00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00

Kinetic Energy (eV, CM)

Figure 5.13. A) Energy dependent reaction cross-sections for the reaction
Tie],+ + C2H40 (acetylaldehyde). B) Expanded view of plot (A).

109
propanal reactions are shown in Figures 5.14-5.17. As with the

acetaldehyde system, we observed two exothermic products as the result of
the reactions involving TiClx+ (x: 1-3). And, as before, the products with
the larger reaction cross-sections were the same products reported by ARI.
A11 curves generated from the data of the propanal system exhibit an
exponentially decaying signal which identifies them as being exothermic
products. Again, we conclude there are no endothermic processes with the

propanal system available to the injection of kinetic energy.

The reaction TiCl4+ + propanal was carried out with only a single

exothermic product being observed at m/z 175. ARI performed the reaction
and reported the displacement of a Cl atom from the chlorotitanium ion

resulting in the product TiCl3C3H60+ which has an m/z value of 211. This

m/z value is beyond the range of our quadrupole.

The next compound to be studied was acetone. This step in the neutral
reagent series maintained a carbon-oxygen double bond, but placed the
oxygen in the middle of the carbon backbone rather than at the end of the
chain. Figures 5.18-5.21 show the experimental results for the reaction
TiClx” (x = 1-4) + acetone. The TiCl4+ + acetone reaction generated results
similar to those of the propanal system. As with the aldehydes, the other
three chlorotitanium ions generated two exothermic products. Again, only
the product with the higher reaction cross-section for each reaction was

reported by ARI.

In our attempt to uncover endothermic processes other than CID, we

shifted our search from molecules containing a carbon-oxygen double bond

110

25E+5
l (A)
A 2.01951
E
U
2
a 1.5E+5
:
e
'5
8 10E+5
a.) .
g ’. A m/z 48, “n"
h . . +
U 5.0E+4 , x m/z 64, T10
.0 . mlz 99, T1C10+

 

0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

2.5E+3 n

 

 

A
"a 2.0E+3 "
U
.‘S
3 1.5E+3 «
v
’5
3 1.0E+3 «
‘9
In
(I:
5 5.01=.+2 ~
00E+o
0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

Figure 5.14. A) Energy dependent reaction cross-sections for the reaction
m“ + C3H50 (propanal). B) Expanded view of plot (A).

111

5.0E+5

4.SE+5 1 (A)
4.0E+5 '

3.5E+5

3.0E+5

2.5E+5

2.0E+5

A m/z 140, T1C1C3H50+
x m/z 99,TiC10+
LOB” o m/z 83, TiCi+

5.0E+4

1.5E+5

Cross-Section (10'“5 cm’)

 

0.0E+0
0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

(3)

Cross-Section 00'" cm’)

 

 

Kinetic Energy (eV, CM)

Figure 5.15. A) Energy dependent reaction cross-sections for the reaction
TiCiz+ + c.1160 (propanal). B) Expanded view of plot (A).

112

LOE+5

(A)

8.0E+4

6.0E+4

4.0E+4
A m/z 118, TiClz+

x m/z 140, TiClC3H50+
O m/z 175, TiC12C3H50+

Cross-Section (10“ cm’)

2.0E+4

 

0.0E+0
0.0 1.0 2.0 3.0 4.0 5.0

Kinetic Energy (eV, CM)

2.0E+3 '1
1.8E+3 i (B)
L6E+31
1.4E+3 ~
1.2E+3 — :,
LOB+33K
8.0E+2 1. °
60E+2
4.0E+2

 

Cross-Section (10“ cm’)
s

2.0E+2

 

0.0E+0
0.0 1.0 2.0 3.0 4.0 5.0

Kinetic Energy (eV, CM)

Figure 5.16. A) Energy dependent reaction cross-sections for the reaction
TiCl3+ + c.1160 (propanal). B) Expanded view of plot (A).

113

 

 

3.0E+4
1 (A)
2.5E+4 .
"s
3 2.0E+4 n
'c
3
.5 1.5E+4 «
8
"3 y,
g "OE” A m/z118,'I'iC12C3H50+
U x m/z 140, TiCl3+
5.0E+3 1+
0...... M
0.0 1.0 2.0 3.0 4.0 5.0
Kinetic Energy (eV, CM)
5.0E+3 -
(B)
h
4“ 4.0E 3 .
E +
3 ”x: :51”; x:
a 3.0E+3 «A xii-xix" "8 5"“ ’51?"
= 8 X X x
3 A gnaw
8 2.0E+3 . A
"a f
3 X41.
2 31‘ [A
U 1.0E+3 32‘: A?

 

 

0.0 1.0 2.0 3.0 4.0 5.0

Kinetic Energy (eV, CM)

Figure 5.17. A) Energy dependent reaction cross—sections for the reaction
'ric1,+ + €311.30 (propanal). B) Expanded view of plot (A).

114

4.0E+5 }

(A)
3.5E+5 4
3.0E+S .
2.5E+5 ~

2.0E+5 .

v

1.5E+5 4

Cross-Section (10“ cm’)

A m/z 48, Ti+
1.0E+5 4 . +
t x m/z 64, T10

D

5.0E+4 o m/z 99, TiClO+
0.0E+0 _

0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

 

5.0E+3 ‘1
4.5E+3 « (B)
4.0E+3 ~
3.5E+3 ~
)1
3.0E+3 d
2.5E+3 -
2.0E+3 .11
X
1.5F.+3 «1;,

1.0E+3 i ”1“,...” - M

5.0E+2 .

Cross-Section (10“ cm’)

  
 
 

 

0.0E+O
0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

Figure 5.18. A) Energy dependent reaction cross-sections for the reaction
TiCl+ + €311.50 (acetone). B) Expanded view of plot (A).

115

1.5E+4 n
1.4E+4 a, (A)
1.2E-t-4 1
1.1E+4 3,
9.0E+3 .
7.SE+3 .

6-0E+3 I A m/z 83, TiCl+
4,513+3 a x m/z 99, TiClO+
O m/z 140, TiClC3H50+

.evvv

Cross-Section (10" cm’)

3.0E+3 d

s M
1.5E+3 1 .~ x
0.0E+0
0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

0“.

 

2.0E+3 1 ‘

1.8E+3 4 A“

1.6F.+3 . ° 5413414 A“
A ‘ ‘

1.4E+3 . A}

1.2E+3 . ° N
1.0E+3 '1 o
8.0E+2 . H

 

Cross-Section (10“ cm’)

 

0.0 2.0 4.0 6.0 8.0 10.0

Kinetic Energy (eV, CM)

Figure 5.19. A) Energy dependent reaction cross-sections for the reaction
Tim,“ + c.1160 (acetone). B) Expanded view of plot (A).

116

 

 

3.01~:+5 -.
(A)
2.5E+5 ]
E
a 2.0E+5 *
'c
fl
V
g 1.5E+5 «
H
8
0') 10E 5 1
an . + a» .
8 g A m/z 118, TiClz+
i-
U . X m/z 140, TIC]C3I‘I{,()+
.0 .
5 E” , o m/z 175, T1C12C3H50+
O
O
0.0E+0
0.0 1.0 2.0 3.0 4.0 5.0
Kinetic Energy (eV, CM)
3.013+3 . .
(B)
A 2.5E+3 «
N
E a
3; 2.0E+3 ~ ,
S
V o
S 1.5E+3 ~" .
’5 .
0 X o
‘2 1.0E+3 - 2..
g x “'0
U 50E+2 x O"
a o
. ”at“? .‘. fiM
ODE-+0 W

Kinetic Energy (eV, CM)

Figure 5.20. A) Energy dependent reaction cross-sections for the reaction
TiCl3+ + c.1160 (acetone). B) Expanded view of plot (A).

117

6:988 056 + +302. 5:88 05 no.2 8288-820 .6288 Eugene—u 328m .86 uSmE

A20 $3 3.5:”..— 9385—
me o.v mgm o.m Wm od m._ o._ md o.o

 

 

 

4 .

Jami?! « D
$8.2. m
{83.3%fo m w
. w
20:822.: x a $85 m
b.8388: as < m...
- $8.” a...
mz

a. $8.2

 

Alma.—

118
to one having two carbon-oxygen single bonds: dimethyl ether. Figures

5.22-5.25 show the experimental results for the reaction 'I‘iClx+ (x = I-4) +
dimethyl ether. As seen from the plots, the results are similar to those of

the ketones and aldehydes. Only exothermic products where observed.

When the results of the dimethyl ether chemistry are compared to those of
ARI, there appear to be two differences. Previously, when our results for
the chlorotitanium ion chemistry are compared to those of ARI, we found

that we observed one additional exothermic product for each reaction. For

the reaction involving dimethyl ether, this is also true for the Ti012+ and

TiCl4+ reactions. The reactions involving TiCl” and 'I‘iC13+ however, yield

the same number of products in both studies. We did not observe any

additional products from those reported by ARI.

B. Discussion

The two studies are in general agreement. For most of the individual
reactions, we have observed one additional product to those reported by
ARI. This may appear to be additional chemistry, previously not observed.
However, when the results of the two studies are compared as series of
chlorotitanium ions reacting with each compound, no new chemical

species are discovered.

In every case, the product with the larger reaction cross-section is the
same product observed by ARI. The product with the smaller cross-section
is the same product observed by ARI during the next reaction in the series.

This is shown in Table 5.2 where the products of the chlorotitanium

L2E+5

1.0E+5

8.0E+4

-I‘cm2)

6.0E+4

Ion (10

4.0E+4

Cross-Sect

2.0E+4 "

0.0E+0
0.0

2.0E+03 ~
1.8E+03 *
'5‘ 1.6E+03 «

X

m

3

w
1

1.2E+03 A
1.0E+O3 4
8.0E+02 -
6.0E+02 *

ross-Section (10“‘cm

C
.e
E:
s

 

.0.

2.0E+02 1
0.0E+00 %

119

(A)

X m/z 114, TiClOCH3+
O m/z 48, Ti+

 

1.0 2.0 3.0 4.0 5.0

Kinetic Energy (eV, CM)

6.0 7.0

x O.

X .00

x
5% up”

‘0
0....

«W it 3'} I“ ’0‘ x
”.3... J 1 i! l x X&VM‘M¥¢£¢&*¥ 1

 

0.00

1.00 2.00 3.00 4.00 5.00

Kinetic Energy (eV, CM)

6.00 7.00

Figure 5.22. A) Energy dependent reaction cross-sections for the reaction
Tic:+ + CZHGO (dimethyl ether). B) Expanded view of plot (A).

2.0E+5 —
1.8E+5 1’ (A)
«renews 1
U
2 1.4E+5j
31.2E+5 -
81.01435 4:

fl
38.0E+4 -
(I)

$601544 d A m/Z 83, T1Cl+
54 0E+4 3x x m/z 149,TiC120CH3+
20W 9 m/z 114, TiCIOCH;

w
0.0E+0

0.0 1.0 2.0 3.0 4.0

 

5.0 6.0 7.0
Kinetic Energy (eV, CM)

2.0E+3 -
1.8E+3 e (B)

.1"
E 1.6E+3
u

1.4E+3 e

2
31.2193 — "
V

E x
l.OE+3 fl

8 x m“

88.01342 -

V.) ‘ ‘5“

36.0192 ~ “at

O R x A

54.0E+2

U x “no“.

2.0E+2 « °°..

 

 

Kinetic Energy (eV, CM)

Figure 5.23. A) Energy dependent reaction cross-sections for the reaction
Tie]; + C2H60 (dimethyl ether). B) Expanded view of plot (A).

121

4.0E+4
(A)
15E+4

10E+4
2.5E+4
2.0E+4

L5E+4

A m/z 163,TiC12CH20CH3+
x m/z 149, TiC120CH3”
o m/2118,TiC12+

LOE+4

Cross-Section (10“ cm’)

50E+3

 

QOE+0
00 05 ID 15 20 25 3D 35

Kinetic Energy (eV, CM)

2.0E+3 1|
1.8E+3 ,, (B)
1.6E+3

1.4E+3 "

1.2E+3 "

LOE+3 .

8.0E+2 ,.

6.0E+2 ”k

4.0E+2 ‘, ”a x . . MW“:

X
2.0E+2 “ . ...°o. xx °’

0.0E+0 O.” l‘h’e‘efifiéw

00 05 ID 15 20 25 30 35

Kinetic Energy (eV, CM)

Cross-Section (104‘ cm’)

 

Figure 5.24. A) Energy dependent reaction cross-sections for the reaction
TiCl3* + C2H60 (dimethyl ether). B) Expanded view of plot (A).

 

 

 

 

 

 

2.5E+4 -
4) (A)
p 2.0E+4 ~
E
U
:2. ,
2 1.5E+4 a
=
O
'5
.J
(I)? 1.0E+4
g x m/z 153, TiCl3+
65 5,013,, _ e m/z 149,TiC120CH3+
W
0.0E+0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Kinetic Energy (eV, CM)
5.0E+3 —
(B)
.1" 4.0E+3 J
E
U
2
3 3.0E+3 - “XFXMW
v
g sag-9w
.3 l) x W
8 2.0E+3 . ’9'”?
m ’5! "
a'a at,“
E . x
U 1.0E+3 a "‘
:x
D
0.0E+0 & - 7w" '. W fr“? T2“ -, w
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Kinetic Energy (eV, CM)

Figure 5.25. A) Energy dependent reaction cross-sections for the reaction
Tia,” + (221160 (dimethyl ether). B) Expanded view of plot (A).

123
reaction series with acetylaldehyde are listed. Branching ratios for the

reactions were calculated using the following formula:

ratio value A = (peak intensityA/Z product peak intensities) x 100%

Table 5.2. List of products for the reaction 'I‘iClx+ (x = 1-4) + acetylaldehyde.

 

 

Ion ARI Results Our Results Branching

(m/z) (m/z) ratios (%)
TiCl+ TiClO” TiClO“ 63.6
("V2 83) Ti0+ 36.4
Ti012+ TiCngH3O+ 'I‘iClC2H3O+ 99.4
(m/z 118) TiClO+ 0.6
Tile TiClgC2H3O+ TiClngH3O+ 98.3
(m/z 153) 'I‘iClC2H3O+ 1.7
Tile 'I‘iCl3CZH4O+ TiC13C2H4O+ 80.6
(m/z 188) 'I‘iC1202H3O+ 19.4

 

 

 

 

 

 

The difference between the two studies may be explained by our use of a
more sensitive instrument. If the instrumentation used by ARI were more

sensitive, they may have also observed a second reaction product.

Our results indicate that there does not appear to be any endothermic
processes available to increased reactant ion kinetic energy for the systems
studied. Because the bond between titanium and oxygen is so strong, the
interaction between the metal and oxygen releases a large amount of
energy. This excess energy is then transferred throughout the newly

formed ion. With relatively small compounds, the number of modes into

124
which this excess energy can be stored is limited. This results in the

breaking of weaker bonds, to form both neutral and ionic fragments with a
wide range of kinetic energies. It is of interest to note that the onset
associated with the chlorotitanium CID product in every system studied
occurred within the range 0.5-1.0 eV below the onset observed during the

CID experiments with argon as the target gas.

If we had been able to work with larger compounds, we may have been able

to observe endothermic processes other than the CID of the reactant ion.

I I I . Chemistry between 'I‘iClx+ ions and isobutene

In the previous section, CID experiments performed on chlorotitanium
ions generated by 70 eV EI were described. BDE's generated from the
results of those experiments differed from BDE values previously
determined Kiser, Dillard, and Duggar (KDD) [I]. The differences allow us
to estimate the amount of internal energy present in the ions. In this
context, we investigated the chemistry between TiCln+ ions and isobutene.
This work refines previous conclusions and provides an important context
for understanding the influence of internal energy on ion/molecule

reactions of metal-containing ions generated by 70 eV EI.

Gas-phase chemistry between positive chlorotitanium ions and olefins has
previously been investigated by Allison and Ridge (AR2) [41]. The general

reaction scheme developed from the results is shown in Figure 5.26.

deco? 5m? 2,3283: 3 3 case 33 “Hop 85.33 28 meow
828359820 5253 humane”? 2.3 .8.“ SE mg .3 39233 08658 :oBowom find 3ng

ammao + anaemmadoioa. AIIJ

 

5m + mammaoaeaa. I ammao + mam.

 

mm + Jamaammaojofit All.

126
Chlorotitanium ions were found to react with small olefms through the

elimination of molecular hydrogen and hydrogen chloride. Larger alkenes
also reacted through the elimination of smaller olefins. TiCl4+ was found
to be unreactive. Reactions (1-12) illustrate the observed chemistry for
experiments performed at Michigan State. For comparison, previously
observed chemistry is indicated by an asterisk (*). Also included with the
reactions are the branching ratios. Reactions (6-8) contain two sets of
branching ratios. The first set of ratios corresponds to ratios occurring at
K.E.CM = 0. The second set corresponds to ion kinetic energy values where
the endothermic product has a maximum intensity. Curves generated
from our energy-dependence experiments are shown in Figures 5.27-5.30.

Experimental procedures and data analysis are described in Chapter 3.

 

Branching
ratios
TiClzC4H7+ + Cl- + HCl 92.1 (1)
TiCl4++ 04H8 {
TiClZ+ C4H7+ + C1- + HCl 7.9 (2)
Ti012C4H7+ + H0] 94.0 * (3)
TiCl3++ C4H8 TiClC4H6+ + 2HCl 3.1 (4)
TiCl2+ C4H7+ + HCl 2.9 (5)
TiClZC 4H8+ 20.3 8.0 (6)
TiC12++ C4H8 TiClC4H7+ + H0] 79.6 60.2 * (7)
TiCl + 04H?+ + HCl 0.1 31.8 (8)
TiClC4H6+ + H2 78.8 * (9)
LTiClC3H4“ + CH4 18.7 (10)
TiCl" + 04H8 -
'> TiC4H5+ + HCl + H2 0.9 (12)

A. Thermochemistry

Considering the results of the previous study performed by AR2, several

conclusions could be drawn about the thermochemical information.

60000 a (A)

50000 .1
t.‘
5 40000 —
'5 I!
§ 30000 a
282
a
a 20000 . -°-m/z 55
U + m/z 153

 

 

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Kinetic Energy (eV, CM)

60000 .

(B)

50000 d

40000 -

 

30000 a

20000 ~

Cross-section (10“ cm‘)

1 0000

 

 

 

 

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Kinetic Energy (eV, CM)

Figure 5.27 . A) Energy dependence of reaction cross-sections for the reaction
Tich,+ + c411,. (isobutene). B) Expanded view of plot (A).

25000

20000

g—+%~Lu_aun-—J

15000 1‘

10000 ~

Cross-section (10" cm’)

5000 ~

 

2000 1

1600 -

‘5
U
’5, 1200 A
E’
.2 ’
¥
'5
E
o

800 1,

 

(A)

0 m/255
. m/z118
-+-m/z 137
x m/z 173

 

 

1.0 2.0 3.0 4.0 5.0
Kinetic Energy (eV, CM)
* (B)
x
o: O .0
o 00. .
x9... .0. 0..
0‘ . o .
o. 00. O
: O .000 .
O O .9
in ° ° 0

 

 

1.0 2.0 3.0 4.0 5.0
Kinetic Energy (eV, CM)

Figure 5.28. A) Energy dependence of reaction cross-sections for the reaction
Tic13+ + C4118 (isobutene). B) Expanded view of plot (A).

130

 

 

 

9000
l (A)
(It
8000 3,
X
7000 -
6‘ X
.5, 6000 .
=5 3:
c. 5000 -
g x 3: m/z 138
'3 4000 ~ +an 174
g, x —Cnmch Fit (83)
§ 3000 . x -—Crunch Fit (55)
U x o m/z 55
2°00 ' "x + m/z 83
X
1000 (
0 w ~ I ~~W'a~¢;.‘a..ux....- .A .
0.0 2.0 4.0 6.0 8.0 10.0
Kinetic Energy (eV, CM)
1000 ~
2 (B)
900 — x
I!
X
300 ‘ x 12.05) = 1.73 eV ET(33) =2.98ev
It _ =
5; 700 4 n - 1.0371 R 1.3841
9 x m = 1.0000 m = 1.0000
3 600 J "‘2,
G x
g 500 4 ,, ’9
3 400 - E'us.‘ firing ++
a +4. + ++ ++++
é’ - ++
Q 300 r I
200 -
, ,
100 - *
0 A...“ .. M
0.0 2.0 4.0 6.0 8.0 10.0
Kinetic Energy (eV, CM)

Figure 5.29. A) Energy dependence of reaction cross-sections for the reaction
Tile + C4113 (isobutene). B) Expanded view of plot (A).

131

S i

 

Cross-section (10" cm’)
0)

 

 

(A)

 

+m/z 101
A m/z 110
2000 - +m/z 123
+m/z 137
x m/z 48
1°00 --Crunch En
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Kinetic Energy (eV,CM)
120 - ‘ ‘
. (B)
h 11
100 J 2
. ‘ ET = 1.97 eV
80 g ‘ n = 1.7326
+ A m - 1.0000
1 r

 

 

Cross-section (10“ cm’)
8

 

 

 

0.0 1.0 2.0 3.0 4.0 5.0

6.0 7.0 8.0 9.0 10.0

Kinetic Energy (eV, CM)

Figure 5.30. A) Energy dependence of reaction cross-sections for the reaction
TiCl+ + Cal-13 (isobutene). B) Expanded view of plot (A).

132
Reactions (1315) show three ways to generate TiCIZC4H7+ by reacting the

various chlorotitanium ions with isobutene. Shown with each reaction

equation is the limit of the BDE(TiClz+--C4H7) required for the chemistry to

occur as described by the equation.

BDE(TiClz+--C4H7)

(kcal/mole)
TiCl4+ + C4H8 --> TiClzC4H-7+ 4» HO] + Cl' <97.8 (13)
Tile 4» C4H3 --> TiClzC4H7+ + HCl >59.9 (14)
'file + C4H8 --> TiClgC4H7+ + H' <85.0 (15)

However, reaction (14) is the only chemistry observed by AR2 that
generated the ion. Based on available thermochemical information and the
fact that the reaction must be at least thermoneutral, a lower limit of 59.9
kcal/mole can be placed on the BDE; Reactions (13) and (15) were not
observed, possibly due to an energy barrier to those reaction pathways.
This allows an upper limit of 85.0 to be placed on the BDE. Therefore, the
BDE can be bracketed:

59.9 kcal/mole < BDE(TiClz+--C4H7) < 85.0 kcal/mole

In a similar manner, TiC1C4H6+ could be generated a number of ways

which are shown by reaction equations (16-19).
BDE(TiCl+--C4H6)

(kcal/mole)
TiCl4+ + C4H8 --> TiClC4H6+ + 2HC1 + Cl' <146.6 (16)
'I‘iCl3+ + C4H3 --> 'I‘iClC4H6+ + 2HCl <108.6 (17)
'I‘i012+ + C4H3 --> 'I‘iClC4H6+ + HCl + H' <I33.7 (18)

TiCl+ + C4H8 --> TlClC4H6+ + H2 >428 (19)

133

Reaction (19) was the only chemistry observed that generated the ion.

Again, because this reaction is required to be thermoneutral/exothermic, a

lower limit of 42.8 kcal/mole can be placed on the BDE(TiCl+--C4H6).

Chemistry described by reactions (16-18) was not observed which places an

upper limit on the BDE of 108.6 kcal/mole. Therefore:

42.8 kcal/mole < BDE(TiCl+--C4H6) < 108.6 kcal/mole

Proposing different neutral products for reaction (18), such as H2 and Cl°,
would raise the upper limit of the BDE to 134.7 kcal/mole for the chemistry

described by that equation.

We have found the Tile ion is reactive with isobutene as shown by

reaction (1). This does not agree with the results generated by AR2, who

found TiC14+ to unreactive with olefins. Our data indicate the product ion

curve has the general form consistent with exothermic reactions, in that

the curve peaks very close to K.E.CM = 0 and undergoes an exponential

decrease. The results are shown in Figure 5.27. The peak of the curve,
however, does not occur at K.E.CM = 0 which would indicate that a small
energy barrier (possibly due to steric hindrances) does exist to the
chemistry. The reaction pathway may be more accessible to increased
K.E., rather than internal energy. This would explain why the product
was not observed by AR2, because they had no means by which to inject
additional energy into the system beyond that available during the

ionization process.

134
Therefore, if reaction (1) is exothermic, the thermochemistry tells us that

97.8 kcal/mole are required for the reaction to occur. By approximating the

internal energy of the Tile ion (shown in the previous section) to be 30

kcal/mole, the energy requirement is reduced to 67.8 kcal/mole. This

means that the value for the BDE(TiC12+--C4H7) would have a lower limit of

67.8 kcal/mole. Therefore, bracketing of the BDE is adjusted to:

67.8 kcaJ/mole < BDE(TiC12+--C4H7)< 85.0 kcal/mole

In a similar manner, our experimental results tell us that the TiCl3+ ion

reacts with isobutene to eliminate I HCl (reaction (14)) and 2 HCls
(reaction (17)). AR2 indicated only the elimination of a single HCl. The
product ion curve generated from our data indicates an exothermic
process. However, because the product ion current was low, the

instrumentation employed by AR2 may not have been sensitive enough to

detect the 'I‘iClC4H6+ ion.

,In order for the exothermic chemistry to occur as described by reaction
(17), 108.6 kcal/mole of energy are required. By including the 13 kcal/mole
excess internal energy estimated from the CID experiments, the lower

limit of the BDE(TiCl+--C4H6) would be 95.6 kcal/mole. Reaction (18) would

then determine the upper limit. Therefore, bracketing of the BDE(TiCl+--
C4H6) would be adjusted to:

95.6 kcal/mole < BDE('I‘iCl+--C4H6) < 133.7 kcal/mole

135
Table 5.3 summarizes the thermochemical information derived from the

AR2 results and how our results make adjustments to those conclusions.

Table 5.3. Summary of Bond Dissociation Energies.

 

 

 

 

 

 

 

 

Ion Complex Lower BDE Limit Adjusted Lower BDE
(kcal/mole)a Limit (kcal/mole)b
'I‘iClz+--C4H3 2 O 2 O
'I‘iClz+--C4H7 59.9 67.91t
TiCl+--C4H8 72.0 23.0
TiCl+--C4H6 42.8 95.61
'I‘iCl+--C3H4 31.8 2 0
TiCl+--C2H3 95.4 56.4
'I‘i+--C4H5 24 2 0

 

 

 

 

 

a) determined by assuming chemistry to be exothermic
b) calculated by subtracting the approximate internal energy due
to the ionization process.

i) discussed in text

B. Mechanisms

1. Tile and isobutene chemistry

As stated before, AR2 found TiCl4+ to be unreactive with unsaturated

hydrocarbons, whereas the present study found isobutene to react through

an apparent exothermic pathway resulting in two product ions (the initial

complex, TiCl4C4H3+, may not have been observed due to the mass

136
limitations of the quadrupole). A proposed mechanism to help explain the

observed chemistry is shown in Figure 5.31.

The appearance of the products can be explained by an initial ligand
exchange reaction which would suggest the TiCl3+--C1 bond to be a weak
bond. Ligand exchange was observed during both our study and the study
performed by AR2 of the chemistry between oxygen-containing compounds

and Tile-. However, when other chlorotitanium ions (TiCl3,2,1+) were

used as reactants, ligand exchange was not observed. This chemistry

supports the findings of the KDD study, which determined the TiCl3+--Cl

bond to be the weakest chlorotitanium bond.

Following the ligand exchange, the resulting complex may be in an excited
state. Excess internal energy (as a result of the ionization process and/or
interaction between the titanium atom and the double bond of the olefin)
may then result in electron rearrangement. As a consequence of this
rearrangement, HCl is eliminated, generating a complex (represented by
the peak at m/z 173) consisting of a dichlorotitanium ion bonded to a
hydrocarbon through a four electron, three center bond. If this final
complex still contains excess energy, it may further decompose. The
charge of the resulting products is governed by the ionization energies

(I.E.) of the neutrals. The fragment with the lower ionization energy is the

fragment most likely to be charged. As shown in Figure 5.31, I.E.(TiC12) =
9.3 eV and I.E.(C4H7) = 7.9 eV which explains why C4H7+ (m/z 55) and not
'I‘iC12+ (m/z 118) is detected.

137

C1\+
/Ti +<4:) >—C3H + HCl

M I..E(C4H7)=7.9eV
I..E(T1012 )=9.3eV

TiCl2
12:2._&7

Figure 5.31. Proposed mechanism for the reaction between
TiCl: and isobutene.

138

2. TiCl3+ and isobutene chemistry

AR2 found TiCl3+ to react with isobutene through the elimination of HCl.

During our investigation we found evidence of a loss of a second HCl as

well. A proposed mechanism is shown in Figure 5.32.

As with the chemistry of Tile, the initial complex, TiCl3C4H8+, was not

observed, possibly due to the mass limitations of the quadrupole. Without
the steric hindrance of a fourth chlorine, the titanium is more exposed and
therefore able to react more readily. This is evidenced in the fact that AR2
did observe chemistry and reported an ion at m/z 173. Although they were
able to detect an ion at m/z 173, it is of interest that they did not report an
ion at m/z 209 representing TiCl3C4H8+, the assumed precursor. One
explanation is that the initial complex is unstable and decomposes very
quickly. The lifetime of an ion must exceed the period of one cycle to be
detected in an ICR instrument. Because ICRs generally operate in the 100-

200 kilohertz range, the lifetime of an ion must exceed 5-10 microseconds.

During our investigation, we also observed an ion at m/z 137. This can be
explained by the elimination of HCl from the precursor represented by the
peak at m/z 173. A possible mechanism is illustrated in Figure 5.32. One
driving force behind this further decomposition may be the stability of the
product ion. Finseth, Sourisseau, and Miller (FSM) [42] obtained complete
vibrational spectra for tricarbonyl(trimethylenemethane)iron. They found
that the terminal carbons of the ligand should be described as having

appreciable sp2 character and that trimethylenemethane behaves as a

139

01
+ '4. .....
T1013 + —> Cl—Ti'j
c':1\>
H
H01
H01
C A
1>Ti+(\4’>—CH3 H2
Cl v
m/z 173*
04H; + T1012 + HCl

 

Figure 5.32. Proposed mechanism for the reaction between
1301; and isobutene.

140
strong ligand by making substantial bonding contributions through 1:

donations. According to the authors, the bonding description of the iron
complex in terms of a single bond between the metal and the central
carbon of the ligand is a misleading oversimplification. It must also be
pointed out that FSM found that less than 0.3% of the molecules were in
the ground state at 300 K. The majority of our experiments were carried
out at temperatures ranging from 298-300 K. Further decomposition was
not observed possibly because of the already low ion current of the

chlorotitanium-trimethylenemethane complex.

The ion appearing at m/z 55 (C4H7+) is thought to be the result of the

decompostion of the ion at m/z 173. This follows the same reasoning

provided in the previous section concerning the chemistry between Tile

and isobutene.

3. TiClz” and isobutene chemistry

AR2 found TiClz+ to react with isobutene through the elimination of HCl.

During our investigation, we also observed the HCl elimination. And,
because the mass of the initial complex, TiClzC4H8+, is within the mass
range of our quadrupole, we were able to observe a peak at m/z 174. A

proposed mechanism is shown in Figure 5.33.

As before, the product represented by the peak at m/z 138 can be explained
by the formation of a four center, four electron bond resulting in the
elimination of HCl. Because of the relatively high product ion current,

further decomposition was observed. As shown in Figure 5.33, the

 

CID i
01 —Ti+<’4‘>—0H3 + H01
\vl

m/z 138*

 

I.E.(TiCl) = 7.0 eV

  

I.E.(C4H7) = 7.9 eV

1301+
m/z 33
+
C4H7
m/z 55

Figure 5.33. Proposed mechanism for the reaction between
TiCl; and isobutene.

142

ionization energies favored the formation of the TiCl” ion. Observation of

this ion was complicated by the fact that 'I‘iCl” is also the CID product of
the reactant ion TiC12+.

We were, however, able to generate our first measurable onset (other than

CID) by observing the formation of C4H7+ from the reactants Tile and
C4H8. This is shown in Figure 5.29. The proposed mechanism for this
reaction predicts C4H7+ is the result of the dissociation of the ion at m/z
138. Ionization energies of TiCl and C4H7 favor the formation of TiCl+ and
suggest that this reaction pathway is not exothermic. Therefore, the

assumed reaction is:

'1‘1012+ + C4H8 —' H01 + TiCl + 04H7+ (20)

Using available thermochemical information, reaction (20) is endothermic
by 2.93 eV. Our experimental measurement generated a value of 1.73 eV.
The low experimental value is consistent with performing experiments on

"hot" reactant ions.

4. TiCl+ and isobutene chemistry

AR2 found TiCl” to react with isobutene through the elimination of H2.

During our investigation, we also observed a more rich chemistry through
the elimination of H2, CH4, C2H5, and HCl. It is noted that even though the
mass of the initial complex, TiClC4,H3+ at m/z 139, falls within the mass
range of the quadruple, the ion was not observed. Repeated attempts to

increase the resolution of the quadruple to separate the possible peak from

143
that appearing at m/z 137 were unsuccessful. An added complication may

be that the peak is very small due to a short lifetime of the ion. A proposed

mechanism is shown in Figure 5.34.

Because of the relatively high product ion currents of the ions represented
by the peaks at m/z 137 and 123, it is believed that there are two competing
reaction pathways rather than a sequential mechanism. This competition
may contribute to the small signal of the initial complex. Although there
were no measurable onsets (other than by CID) it is believed that the
chemistry represented by the mechanism that generates the ion at m/z 123
has some small energy barrier. This is supported by the fact that AR2 did
not report the observation of this ion, even though our experiments suggest

a relatively high reaction cross-section.

The reaction between TiCl+ and isobutene generated a higher ion current
of the proposed chlorotitanium-trimethylenemethane complex than any
other reaction between a chlorotitanium ion and isobutene. Because of this
higher current it is believed that observation of the decomposition of this
complex is now possible. Product ions believed to be the result of this
decomposition are represented by the peaks at m/z 110 and 101, however, a

reasonable mechanism has yet to be proposed.

IV. Conclusion

It has been demonstrated that the MSU ion beam instrument can be used
to generate thermochemical information from gas-phase chemistry

studies. The utility of the instrument would be increased greatly by the

144

+,*
C1—Ti(\4> + CH4

\
>
I
V

m/z 123

 

Figure 5.34. Proposed mechanism for the reaction between
TiCl+ and isobutene.

145
replacement of the quadrupole with a higher performance mass analyzer.

Also, the installment of a drift cell between the source and the magnet to
facilitate collisional relaxation of excited states of the reactant ion, would
enable the investigation of ground state chemistry. To increase the
transmission efficiency of the intrument for product ions which differ
greatly in mass, programable power sources could be employed to apply
potentials to the ion optics. Coupling these power sources to the quadrupole
would allow for the monitoring of each ion under ideal transmission

conditions.

10.
11.

12.
13.

146

LIST OF REFERENCES
Chapter 5

R.W. Kiser, J.G. Dillard, D.R. Dugger, "Mass Spectrometry in
Inorganic Chemistry”, Adv. Chem. Ser., No. 72, 153 (1968).

RF. Lossing, A.W. Tickner, W.A. Bryce, J. Chem. Phys., 19, 1254
(1951).

J .W. Warren, Nature, 165, 810 (1950).

O. Kubaschewski, O. Kubaschewski-von Goldbeck, P. Rogl, H.F.
Franzen, "Atomic Energy Review-Special Issue No. 9; Titanium:
Physico-Chemical Properties of its Compounds and Alloys," ed. K.L.
Kamarek, International Atomic Energy Agency, Vienna, Austria
(1983).

G. Skinner, H.L. Johnston, C. Beckett, "Titanium and Its
Compounds," ed. H.L. Johnston, Herrick L. Johnston Enterprises,
Columbus, OH, p. 43 (1954).

U. Muller, "Inorganic Structural Chemistry," 2nd ed., John Wiley &
Sons, New York, NY, p. 70 (1990).

FA. Cotton, G. Wilkinson, "Advanced Inorganic Chemistry," 5th ed.,
John Wiley & Sons, New York, NY, p. 653 (1988).

SO. Lias, J .E. Bartmess, J .F. Liebman, J .L. Holmes, R.D. Levin, and
WC. Mallard, J. Phys. Chem. Refi Data, 17, Suppl. 1 (1988).

K. Ervin, S.K. Loh, N. Aristov, P.B. Armentrout, J. Chem. Phys., 87 ,
3593 (1983).

R.S.P. Coutts, P.C. Wailes, Adv. Organomet. Chem., 9, 136 (1970).

R. Feld, "The Organic Chemistry of Titanium", Butterworths,
Washington, DC, (1965).

C.R. Demarcay, C. R. Acad. Sci., 80, 51 (1875).
K. Saigo, M. Osaki, T. Mukaiyama, Chem. Lett., 989 (1975).

14.
15.
16.
17.
18.

19.

20.
21.

22.

23.

25.

26.
27.

30.
31.

32.

147
T. Mukaiyama, Chem. Lett., 1189 (1974).

T. Mukaiyama, T. Izawa, K. Saigo, Chem. Lett., 323 ( 1974).
T. Mukaiyama, M. Hayashi, Chem. Lett., 15 ( 1974).
J. Allison, D.P. Ridge, J. Am. Chem. Soc., 78, 163 (1977).

R. Kinser, J. Allison, T.C. Dietz, M. deAngelis, D.P. Ridge, J. Am.
Chem. Soc., 100, 2706 (1978).

J. Allison, R.B. Freus, D.P. Ridge, J. Am. Chem. Soc., 101, 1332
(1979).

J.S. Uppal, R.H. Staley, J. Am. Chem. Soc., 102, 4144 (1980).

GD. Byrd, R.C. Burnier, B.S. Freiser, J. Am. Chem. Soc., 104, 3565
(1982).

RC. Burnier, G.D. Byrd, T.J. Carlin, M.B. Wiese, R.B. Cody, B.S.
Freiser In Ion Cyclotron Resonance Spectrometry II; Lecture Notes in
Chemistry Vol. 31; H. Hartmann, K.-P. Wanczek, Eds.,Springer-
Verlag, Berlin, p. 98 (1982).

M.A. Tolbert, J .L. Beauchamp, J. Am. Chem. Soc., 108, 7509 (1986).

T.C. Jackson, T.J. Carlin, B.S. Freiser, J. Am. Chem. Soc., 108, 1120
(1986).

H. Kang, D.B. Jacobson, S.K. Shin, J.L. Beauchamp, M.T. Bowers, J.
Am. Chem. Soc., 108, 5668 (1986).

M.A. Tolbert, J .L. Beauchamp, J. Phys. Chem., 90, 5015, (1986).

J .L. Elkind, P.B. Armentrout, Int. J. Mass Spect. Ion Processes, 83,
259 (1988).

R. Tonkyn, M. Ronan, J .C. Weisshaar, J. Phys. Chem., 92, 92 (1988).

LS. Sunderlin, P.B. Armentrout, Int. J. Mass Spectrom. Ion
Processes, 94, 149 (1989).

K.K. Irikua, J .L. Beauchamp, J. Am. Chem. Soc., 111,75 (1989).

H. Higashide, T. Oka, K. Kasatani, H. Shinohara, H. Sabo, Chem.
Phys. Lett., 163, 485 (1989).

T.J. Carlin, M.B. Wise, B.S. Freiser, Inorg. Chem., 20, 2743 (1989).

33.

36.

37.

39.
40.
41.
42.

148
P. Braunstein, D. Nobel, Chem. Rev., 89, 1927 (1989).

K. Eller, W. Zummack, H. Schwarz, J. Am. Chem. Soc., 112, 621
(1990).

RB. Armentrout, Ann. Rev. Phys. Chem., 41, 313 (1990).

DE. Clemmer, L.S. Sunderlin, P.B. Armentrout, J. Phys. Chem., 94,
3008(1990).

P.B. Armentrout, Int. Rev. Phys. Chem., 9, 115 ( 1990).

K. Eller, H. Schwarz, Chem. Rev., 91, 1121 (1991).

BC. Guo, A.W. Castleman, Jr., J. Am. Chem. Soc., 114, 6152 (1992).
G.M. Daly, M.S. El-Shall, J. Phys. Chem., 98, 696 ( 1994).

J. Allison, D.P. Ridge, J. Am. Chem. Soc., 99, 35 (1977).

D.H. Finseth, C. Sourisseau, F.A. Miller, J. Phys. Chem., 80, 1248
(1976).

APPENDICES

Appendix A

Electron Energy Calibration

The electron energy calibration was performed under identical source

conditions to those under which the Mng+ CID experiments were

performed. Nitrogen was used as the calibration compound. Ion currents
(counts/sec) of N2+ (m/z 28) were recorded for as a function of nominal
electron energy. The plot of the experimental results is shown in Figure
A.1. The extrapolated onset for ion production is approximately 16.3 eV.
The accepted literature value for the ionization energy of nitrogen is 15.57
eV. Therefore, a correction of -0.7 eV must be applied to the measured

nominal electron energy.

The nominal electron energy is determined by the potential difference
between the filament and the ionization box. As the electrons boil off the
filament wire, they experience an electric field that accelerates them
toward the ionization box. As the electrons enter the ionization region they
experience an electric field generated by the potentials applied to the
pusher, ionization box, and the extraction lens. The design of the source is
such that these applied potentials generate a potential energy surface

resembling a saddle (discussed in chapter 2).

The potential energy surface (following the trajectory of an electron from
the filament to the electron collector) has a minimum value near the center

of the source with the field reaching a maximum at the sides of the

149

Counts/sec

1.6E+6 a

1.4E+6 ~

1.2E+6 4

1.0E+6 J

8.0E+5 —

6.0E+5 *

4.0E+5 J ,

2.0E+5 4 .

O

0.0E+0 11.1.1.1.1.Wl+4144
0 2 4 6 810121416182022

 

 

 

-1

Nominal Electron Energy (eV)

Figure A1 Electron energy calibration performed by using nitrogen as the
calibration compound. Mass monitored was that associated with m/z 28.

151
ionization box. The purpose of this field is to force positively charged

particles toward the center of the source, where they are more efficiently
extracted. The field experienced by the electrons, however, has an opposite
effect due to the negative charge of the electron. Instead of the potential
surface having a minimum at the center of the source (along the electron
trajectory), it would have a maximum. Stopping potential experiments have
shown that positively charged ions have been generated at potentials as
much as 5 volts below the acceleration potential (the voltage applied to the
ionization box). Electrons experiencing the same electric field would have to

overcome a 5 volt barrier.

As the electrons approach the potential energy surface apex, their kinetic
energy decreases. Electrons interacting with neutral molecules near this
point would be at some energy less than the potential difference between the
filament wire and the ionization box. When appearance potential
experiments are performed, the measured electron energy thus would be

greater than the "true" electron energy, creating false high reading.

Adding to this high reading would be the energy transfer efficiency between
the electron and the molecule/atom being ionized. A 13 eV electron most
likely would not transfer 13 eV, but some amount less than 13 eV. This
would also result in a measured electron energy higher than the "true"

electron energy.

AppendixB

Definitions

P(source) Gas pressure within the source housing as measured by
a Penning gauge.

I(e-) Amps. Measured electron current (Amperes) at the electron
collector. Used for regulating the filament electron emission. See
Figure 2.2.

Repeller(V) Voltage applied to the repeller. The repeller is also
referred to as the pusher. Used to create a saddle-field-shaped
potential energy surface within the source. See Figure 2.2.

Extractor (V) Voltage applied to the extraction plate. Used to create
a field that extracts ions out of the source. See Figure 2.2.

Shield (V) Voltage applied to the shield plate. Used to focus the ion
beam into the magnetic sector. See Figure 2.2.

Source Zl,L(V) Voltage applied to one of the two 2 l-deflection
plates. Used to focus the ion beam into the magnetic sector. See
Figure 2.2.

Source Z], R (V) Voltage applied to the second of two 2 l-deflection
plates. Used to focus the ion beam into the magnetic sector. See
Figure 2.2.

Source Lens, L (V) Voltage applied to one of two lens plates. Used to
focus the ion beam into the magnetic sector. See Figure 2.2.

Source Lens, R (V) Voltage applied to the second of two lens plates.
Used to focus the ion beam into the magnetic sector. See Figure 2.2.

152

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

153
SourceR(V) Voltage applied to the r-deflection plate. Used to focus

the ion beam into the magnetic sector. See Figure 2.2.
SourceZZ,L(V) Voltage applied to one of the two 2 2-deflection
plates. Used to focus the ion beam into the magnetic sector. See
Figure 2.2.

Source 22,11 (V) Voltage applied to the second of two 2 2-deflection
plates. Used to focus the ion beam into the magnetic sector. See
Figure 2.2.

Accel. Volt. Voltage applied to the ionization box. This defines the
nominal potential energy of the ions. See Figure 2.2.

TIC Total ion current as measured at the exit slit of the source.
The current is given as the meter reading with the corresponding
meter scale. The meter is on the BDG module of the CH5.

Hall Hall probe value of the magnetic sector. This provides a
measure of the magnetic field strength.

DecelZl (V) Voltage applied to one of two z-deflection plates of the
deceleration lens system. Used to focus the ion beam into the
deceleration stage. See Figure 2.6.

DecelZZ (V) Voltage applied to the second of two z-deflection plates
of the deceleration lens system. Used to focus the ion beam into the
deceleration stage. See Figure 2.6.

DecelR(V) Voltage applied to the r-deflection plate of the
deceleration lens system. Used to focus the ion beam into the
deceleration stage. See Figure 2.6.

Decel Einzel(V) Voltage applied to the middle plate of the Einzel
focusing stage. Used to focus the ion beam into the deceleration stage.

See Figure 2.6.

20.

21.

22.

26.

27.

29.

154
Decel Inject.(V) Voltage applied to the injection element of the

deceleration stage of the deceleration lens system. See Figure 2.6.
ranequencyMHz Frequency of the rf signal applied to the
octopole. Reported in megahertz.

rf V (ms-amp) Value of the readout on the rf power amplifier.
This is the root-mean—square amplitude of the rf signal applied to the
octopole.

Tuner,T Setting for the first capacitor of the radio antenna tuner
used to create a balance between the amplifier and High-Q head of
the octopole. See chapter 2.

Tuner, I Setting for the inductor of the radio antenna tuner used to
create a balance between the amplifier and High-Q head of the
octopole. See chapter 2.

Tuner,A Setting for the second capacitor of the radio antenna
tuner used to create a balance between the amplifier and High-Q
head of the octopole. See chapter 2.

rf Power HM-9 A measure of the forward power from the amplifier
to the octopole High-Q head. Readout provided by the Heathkit Model
HM-9 QRP Wattmeter. See chapter 2.

SWR Standing wave ratio providing a measure of the relationship
between the forward and reflected power between the amplifier and
the octopole High-Q head. See chapter 2.

Oct. DC bias (V) DC bias applied to the octopole. Computer
controlled. See chapter 2.

Transfer] (V) Voltage applied to the transfer lens 1. This lens
helps to extract ions from the beam guide and focus them into the

quadrupole. See Figure 2.11.

30.

31.

32.

33.

35.

36.

37.

38.

39.

40.

155
Transfer2(V) Voltage applied to the transfer lens 2. This lens

helps to focus the ion beam into the quadrupole. See Figure 2.11.
’I‘ransfer3(V) Voltage applied to the transfer lens 3. This lens
helps to focus the ion beam into the quadrupole. See Figure 2.11.
’I‘ransfer4(V) Voltage applied to the transfer lens 4. This lens
helps to focus the ion beam into the quadrupole. See Figure 2.11.
Transfer5(V) Voltage applied to the transfer lens 5. This lens
helps to focus the ion beam into the quadrupole. See Figure 2.11.
'l‘ransfer6 (V) This lens was found to be unnecessary and was
removed. Transfer lenses 7-9 where never renamed.

Transfer7 (V) Voltage applied to the transfer lens 7. This lens is
used to help focus the ion beam into the detector system. See Figure
2.11.

'l‘ransferS (V) Voltage applied to the transfer plate 8. This plate is
used to help focus the ion beam into the detector system. See Figure
2.11.

Transfer9 (V) Voltage applied to the transfer plate 9. This plate is
used to help focus the ion beam into the detector system. See Figure
2.11.

Quad. DCbias (V) DC bias applied to the quadrupole. This is
manually set using a Hewlett-Packard DC Power Supply model
6110A.

P(Collision Cell) Gas pressure within the collision cell as
measured by a calibrated ion gauge. See chapter 2.

P(chamber B) Gas pressure of the instrument chamber housing the
beam guide and collision cell. The pressure is monitored by the same

ion gauge used to monitor P(Collision Cell).

41.

42.

43.

44.
45.

156
Daly CD(V) Voltage applied to the conversion dynode of the Daly

detector. See chapter 2.

Daly PMT (V) Voltage applied to the photomultiplier tube of the Daly
detector. See chapter 2.

Er Energy of onset. This value is generated during the curve fitting
routine. See chapter 3.

n Curve fitting parameter. See chapter 3.

m Curve fitting parameter. See chapter 3.