 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE ll

—? i F

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU is An Afﬁrmative Action/Equal Opportunity Institution
CWMMJ

THE MODIFICATION OF RESIDUAL GAS ANALYZERS
TO PRODUCE MASS-SELECTED ION BEAMS

by

J effrey Ross Gilbert

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1990

L
s.

'3 .

a45— '75

ABSTRACT

THE MODIFICATION OF RESIDUAL GAS ANALYZERS
TO PRODUCE MASS-SELECTED ION BEAMS

by

Jeffrey Ross Gilbert

We have constructed an instrument designed to trap mass-selected
ions at low temperatures within a solid inert gas matrix for spectrosc0pic
analysis. Our goal was to construct a ﬂexible instrument that would
permit the study of a wide variety of mass-selected positive ions, and which
could also be used to investigate the role that counterions play in the
effective trapping of ionic species in inert cryogenic hosts. This type of
experiment requires the production of two mass-selected ion beams, a
beam of mass-selected cations that are deposited simultaneously with a
beam of anions to reduce space-charging within the matrix. The
instrument was designed to utilize both laser-induced ﬂuorescence (LIF)
and Fourier transform infrared (FTIR) spectroscopies to identify and
investigate the structure of the trapped species. The sources employed in
this experiment must produce high current ion beams for extended
periods to allow the accumulation of a signiﬁcant number of absorbers in
the optical beam for FTIR investigation. Residual gas analyzers (RGAs)
were selected as the basis for the mass-selected ion sources for this
instrument. An RGA is a specialized high vacuum pressure gauge that
consists of a high efﬁciency electron impact ionization source, injection

optics, a miniature quadrupole mass ﬁlter, and a detector.

This dissertation focuses on the modiﬁcation of two RGAs to produce
controlled beams of mass-selected positive and negative ions that can be
directed onto a remote surface for matrix isolation experiments. The
discussion includes descriptions of the modiﬁcations made to the RGA ion
sources and to a commercially available chemical ionization source to
produce ions by surface emission, chemical ionization, and negative
surface ionization. The mass-selected beams produced by the RGA
quadrupoles were focused and deﬂected using a series of electrostatic
optics. The design of these elements was optimized using computer
modeling and ion beam visualization techniques. The modiﬁcations to the
ion sources, quadrupole mass ﬁlters, and electrostatic Optics of these
systems have allowed these RGAs to produce mass-selected ion beams that
have been effectively used in the isolation of mass-selected ions within solid

inert gas matrices.

Copyright by
JEFFREY ROSS GILBERT

1990

iv

In dedication to my parents
whose lives continue to provide inspiration in all that I do
and in memory of
Carbon
the best friend I ever had

ACKNOWLEDGMENTS

This project has been a conglomeration of the efforts of many
individuals, all of whom have been invaluable in its completion. First of
all I would like to thank Mark Sabo who has made my graduate experience
more than just a bunch of lab work (even if at times it was more odiferous
than I might have desired). I think that the all night experiments
watching “The Big Brawl” movies will always be my favorite memory of lab
‘work’. I started this project with little or no experience in the hardware
side of mass spectrometry, and I’d like to thank Paul O’Connor for all his
help in running these experiments. Finally I’d like to thank John Allison
and George Leroi for their guidance and supervision. In this project I was
working toward a goal of their conception, and I hope that my efforts have
aided in the completion of that goal.

I’d also like to thank all the support staff for the hundreds of hours
that they have spent trying to interpret my somewhat confused needs.
Russ Geyer and Dick Menke made many of my convoluted designs into
something that actually worked. Their help with my work and in the
overall construction of the matrix isolation instrument was truly
invaluable. Marty Rabb and Ron Haas put up with uncounted hours of
pestering when I had incorrectly designed or misconnected various
electrical components, and I’d like to thank them for both their help and
tolerance.

Last, but far from least, I’d like to thank all the people here at State
that made graduate school an experience that I will always look back on
fondly. I’d like to thank Berly for putting up with me these last few
months while I’ve been under a lot of stress, and for all the support/stress

relief eﬂ’orts she has made. I will always miss ‘the gang’ here at MSU.
vi

Kris, Sabo, Jason, D.G., Laura, Karen, Ed and Maria, Judy, Jim D., Gary,
Tom, Nurt, and the rest of the gang have made grad school fun. I’ll
always remember the rowboat party my ﬁrst year, Mac’s Bar fall of my
second year, poker parties in cigar smoke ﬁlled rooms, keg parties in our
old apartment, wild friday nights of pre, during, and post-hockey partying,
and other fun events too numerous (or embarrassing) to mention here. I
don’t think that I would have survived without you guys ..... thanks!!!

Table of Contents

Page

List of Tables .......................................................................... xi
List of Figures ........................................................................ xii
Chapter 1. Introduction ........................................................... 1
Chapter 2. Theory and Operation of RGA components ................. 13
I. Ion Sources .................................................................... 13

A. Electron Impact Ionization ......................................... 13

B. Chemical Ionization ................................................... 22

C. Surface Ionization ...................................................... 28

II. The Quadrupole Mass Filter ............................................ 36
A. Introduction .............................................................. 36

B. Theory of Quadrupole Operation ................................... 39

C. Practical Operation of a Quadrupole ............................. 49

D. Basic Concepts in Radio Frequency Circuitry ................. 52

E. Quadrupole rf and dc generation .................................. 58

III. Ion Optical System Design ............................................. 61

A. Basic Concepts in Ion Optics ........................................ 61

B. Practical Ion Optical Design Tools ................................ 68
Chapter 3. Anavac-2 Ion Source Modiﬁcations ............................ 75
I. The design and performance of the Anavac-2 RGA .............. 75

II. Modiﬁcation of the Anavac-Z EI source ............................. 86

A. SIMION modeling of design modiﬁcations .................... 86

B. Design and performance evaluation of modiﬁed source 94

III. Modiﬁcation of the quadrupole wiring ............................. 98
IV. Design and construction of complete source mount/ﬂange . 102

V. Post-quadrupole focusing and deﬂection optics ................... 105

viii

A. Beam visualization and divergence measurements ....... 105
B. SIMION modeling and construction of post-quadrupole

optical systems ........................................................... 114
VI. Performance evaluation of the completed ion source .......... 124
A. Faraday plate experiments in the matrix chamber ......... 124
B. CEMA visualization at cryostat window ......................... 131
VII. Modiﬁcations to produce mass-selected ions using
alternate sources ......................................................... 133
A. The modiﬁcation of a Nermag EI/CI source ................... 133
B. Surface Ionization in the Anavac source ........................ 146
Chapter 4. Dycor M200 Ion Source Modiﬁcations .......................... 161
I. The design and performance of the Dycor M200 source .......... 161
11. Modiﬁcation of the Dycor M200 ......................................... 167
III. Performance evaluation of the completed ion source .......... 176
Chapter 5. Conclusions and suggested future studies .................... 191
1. Possible modiﬁcations to the original Anavac EI source ......... 191
II. Differential pumping of the Nermag source ........................ 192
III. Water cooling of the Nermag source ................................. 193
IV. Improving the output of the Anavac NSI source ................. 195
V. Emission regulation for the Anavac system ......................... 196
VI. Differential pumping of the Dycor source ........................... 197
VII. Alternate ﬁlament materials .......................................... 198
VIII. Low resolution operation of the source quadrupoles ......... 198
VIX. Post-quadrupole optics optimization ................................ 199

Appendix I. A schematic diagram of the Anavac-2 remote
electronics head .................................................. 200

Appendix II. A schematic diagram of the Anavac-2 control unit .. 201

Appendix III. Experimental conditions for the optimization
studies shown in Appendices IV - XIV .................. 202

Appendix IV. Optimization of the N ermag source attached to the
Anavac quadrupole under EI conditions ................ 204

1X

Appendix V. Optimization 01‘ the N ermag source attached to the
Anavac quadrupole under EC-CI conditions ........... 205

Appendix VI. Optimization of the Nermag source attached to the
Anavac quadrupole under EC-CI conditions ........... 206

Appendix VII. Optimization of the Anavac source and quadrupole
conﬁgured for surface ionization ........................... 207

Appendix VIII. Optimization of the Anavac source and quadrupole
conﬁgured for surface ionization ......................... 208

Appendix DI. Optimization of the Anavac source and quadrupole
conﬁgured for surface ionization ........................... 209

Appendix X. Optimization of the Anavac source and quadrupole
conﬁgured for negative surface ionization .............. 210

Appendix XI. Optimization of the Anavac source and quadrupole
conﬁgured for negative surface ionization .............. 211

Appendix XII. Optimization of the Anavac source and quadrupole
conﬁgured for negative surface ionization .............. 212

Appendix XIII. Optimization of the Anavac source and quadrupole
conﬁgured for negative surface ionization ............ 213

Appendix XIV. NSI Surface studies with the Anavac source and
quadrupole .......................................................... 214

Appendix XV. The LIF spectrum of C82+ trapped in solid argon ..... 215

List of References ....................................................................... 216

LIST OF TABLES

Table Page
2.1 The properties of some surface materials commonly used

for NSI ........................................................................................ 32
2.2 Some high efﬁciency conditions for the formation of negative

ions through negative surface ionization ................................. 35
3.1 The performance of the Anavac-2 chopper ampliﬁer ............... 81
3.2 Initial optimum operating conditions for the Anavac-2 RGA .. 83

3.3. Maximum beam divergence values for the ion beam produced
by the Anavac-2 quadrupole ...................................................... 113

3.4 The dimensions and performance of several SIMION lens
models .......................................................................................... 115

4.1 The initial values of the Dycor M200 operating parameters
for the original thoriated iridium ﬁlament ................................. 165

xi

LIST OF FIGURES

9

Figure Page
1.1 The matrix isolation instrument designed for the isolation of mass

selected ions followed by spectroscopic analysis (LIF and FTIR)
2.1 An example of a clastogram for ammonia, adapted from

Mark et al.[38] ................................................................................... 15
2.2. A drawing of the conventional Nier style EI source ....................... 17
2.3. The high efficiency EI source commonly used in RGAs .................. 20
2.4. An example of a source designed for surface emission/ionization .. 30
2.5 The basic components of a quadrupole mass spectrometer ............ 37
2.6 MacSHVIION models of the electric ﬁeld produced within

hyperbolic and round quadrupole electrodes .................................. 40
2.7 The 3,51 Space stability diagram for a quadrupole mass ﬁlter ........ 45
2.8 The electrical components of the radio ﬁ'equency generation

circuit for a quadrupole mass spectrometer .................................... 53
2.9 Ion refraction during transition between regions of differing

potential ............................................................................................. 62
2.10 A block diagram illustrating a unipotential lens ............................ 63
2.11 The optimal geometry for a three element Einzel lens ................... 65
2.12 The microchannel plate detector (MCP) .......................................... 73
3.1 A functional schematic diagram of the Anavac-2 residual gas

analyzer ............................................................................................. 76
3.2 A block diagram of the Anavac-2 analyzer head ............................. 78
3.3 A block diagram of the quadrupole power supply circuit used in the

Anavac-2 RGA ................................................................................... 80
3.4 The resolution performance of the original Anavac-2 with an

air leak ............................................................................................... 85

xii

Figure Page
3.5 SIMION model of the Anavac source showing electron trajectories 87

3.6 A cylindrically symmetric model of the Anavac source showing ion

trajectories ........................................................................................ 90
3.7 A SIMION model of the ion optical changes caused by the source

housing .............................................................................................. 93
3.8 A block diagram of the source housing, source, and quadrupole 95

3.9 The mass spectrum of residual chamber gases taken by the
modiﬁed Anavac source .................................................................... 97

3.10 A mass spectrum from the Anavac source showing rf-feedback
induced shutdown .............................................................................. 101

3.11 A top view drawing of the complete Anavac-2 ion source ﬂange 103
3.12 A SIMION model of the shielded CEMA detector mount .............. 108
3.13 The rail mount system used for beam visualization ...................... 109

3.14 The beam proﬁle produced at the CEMA detector by the Anavac
source in mass scan mode ................................................................ 1 11

3.15 The ion beam proﬁle produced at the CEMA detector by the
Anavac source in operating in total, or rf-only mode ..................... 112

3.16 SIMION modeling results for the focusing of a 3 eV, O-40°
divergence ion beam ......................................................................... 1 17

3.17 SINIION models comparing the degree of aberration produced in
focusing a 0-40° ion beam using model 1 and model 2 ................... 119

3.18 SIMION models comparing the degree of aberration produced in
focusing a O-40° ion beam to 3 m using model 2 and model 4 ..... 120

3.19 SIMION models comparing the degree of aberration produced in
focusing a O-40° divergent beam using an einzel lens in accelerate-
decelerate and decelerate-accellerate modes ................................... 122

3.20 The power supply and metering scheme used to evaluate the
performance of the Anavac ion source .............................................. 125

3.21 Mass-selected ion current at the cryostat window plotted against
nitrogen partial pressure for the Anavac ion source ....................... 126

xiii

Figure Page

3.22

3.23

3.24

3.25

3.26

3.27

3.28

3.29

3.30

3.31

3.32

3.33

3.34

3.35
3.36

3.37

The Anavac source mass spectrum of an argon leak detected at
a remote Faraday plate ..................................................................... 128

The mass spectrum of propylamine taken with the Dycor M100
RGA and the Anavac source with a remote Faraday plate ............. 129

The mass-selected beam proﬁle for m/z=4 produced by the Anavac
source on the CEMA mounted at the cryostat window ................... 132

The power supply and metering system used with the N ermag
source .................................................................................................. 135

The mass spectrum of a mixture of argon and CCl4 produced using
the Anavac quadrupole ﬁtted with the modiﬁed N ermag source 137

The mass spectrum of methane at source pressures of 26 and 400

mtorr produced by the Anavac quadrupole ﬁtted with the Nermag
EI/CI source ........................................................................................ 138

An expanded view of the mass spectrum of methane produced at a
source pressure of 240 mtorr ............................................................. 140

An electron capture negative chemical ionization spectrum obtained
with CCl4 using the Anavac quadrupole and an EI/CI source ........ 142

The mass spectrum of nitrogen obtained using the Nermag source
and Anavac quadrupole with current output maximized ................ 144

The output of the Nermag source versus total ﬁlament emission 145
The experimental setup used to evaluate the emission of the
cesium glass surface ........................................................................... 147

Collector plate current versus heater current for both a bare
rhenium surface and a cesium glass bead ......................................... 148

The mass spectrum produced from NSI of CCl4 on a cesium surface,
this spectrum was taken on an HP5985 mass spectrometer ..... : ..... 150

The original Anavac source adapted for surface ionization ............. 150

A SIMION model of the Anavac source modiﬁed for surface
ionization ............................................................................................ 152

The power supply and metering setup used to operate the Anavac
surface ionization source ................................................................... 153

xiv

Figure Page
3.38 Surface ionization spectra produced by the modiﬁed Anavac ion

source .................................................................................................. 155
3.39 A comparison of the N SI rf-only output for several surfaces ........... 157
3.40 Time stability data for the production of negative ions from a

cesium glass surface with the modiﬁed Anavac source .................... 159
4.1 A block diagram of the Dycor M200 analyzer head .......................... 160
4.2 The mass spectrum produced by the original Dycor M200 RGA ..... 167
4.3 SIMION models of the Dycor M200 source before modiﬁcation and

following enclosure in a source housing ............................................ 168
4.4 The modiﬁed Dycor M200 RGA mass selected ion source ................ 170

4.5 A schematic diagram of the Dycor M200 power supply board
including the addition the addition of a resistor to allow the use of
rhenium ﬁlament wire ........................................................................ 174

4.6 Graphs of both the RF TUNE parameter versus FREQUENCY and
Faraday plate current versus LO MASS setting .............................. 176

4.7 The mass spectrum of PFTBA produced by the modiﬁed Dycor M200
source .................................................................................................. 179

4.8 The mass spectrum of PFK produced by the Dycor M200 source 181

4.9 The mass spectrum of argon produced by the Dycor M200 source
and measured at a Faraday plate at the cryostat window ............... 183

4.10 The mass spectrum of C82 produced by the Dycor M200 source and
measured at a Faraday plate at the exit of the quadrupole ............. 185

4.11 Two repeat scans of the mass specctrum of 082 produced by the

Dycor M200 source and measure at a Faraday plate at the
cryostat window .................................................................................. 187

5.1 One possible design for a source chamber to allow differential
pumping of the source region .............................................................. 192

5.2 The mass spectra of Co(CO)3N O produced by the N ermag source
and the Dycor M200 EI source ........................................................... 193

XV

Chapter 1. Introduction

One of the most powerful modern analytical techniques available is
mass spectrometry. Mass spectrometry utilizes the mass-to-charge ratio of
ions, produced through gas phase ionization of an analyte molecule, to
obtain both qualitative and quantitative information. The combination of
information about the molecular ion, and the fragment ions which are
produced by unimolecular decomposition of the molecular ion has proven
especially useful in identifying the structures of unknown compounds. The
formation of these fragment ions is generally explained using accepted
unimolecular decomposition mechanisms. These mechanisms are based
upon thermodynamically and structurally reasonable pathways through a
series of parent ion, intermediate ion, and product ion structures.

Mass spectrometrists often utilize indirect methods, such as isotopic
labeling and collision-induced dissociation, to verify the structures of these
ions.. For organometallic systems, ligand displacement reactions can be
used to identify the ligands attached to a central metal ion. By combining
these techniques, the proposed structures of many ions have been veriﬁed.
One example was the identiﬁcation of the NiC4H3+ ion produced through
the interaction of Ni“ ion with n-butane. Beauchamp and co-workers
performed a series of ion beam and ion cyclotron reactions of Ni+ with n-

butane [1]; some of the observed product ions for this reaction were:

48%» N1C2H4+ + C2H6

Nl+ + ll-C4H10 7% + NlC3H6+ 4" CH4

_3—579'_'> NlC4H8+ + H2

 

 

Two mechanisms were proposed to explain the formation of the NiC4H3+

ion, a 1,2-elimination process and a 1,4-elimination process. The proposed

ﬁrst step in the 1,2-elimination mechanism is the oxidative addition of Ni+
across a secondary C-H bond of a butane molecule, shown below [2]. This

insertion is followed by the shift of a B—hydrogen onto the metal ion to form a
butene-ion complex. Finally, the reductive elimination of H2 results in the

formation of the Ni-butene ionic complex, NiC4H8+, of unknown geometry.

1,2-elimination mechanism
1'1 PK /H
Ni+ I\lii*
I
Nl+ + C4HlO—l> CH3CHCH2CH3 _> H/C=C\H
H3C 1 CH3
Ni+

H; ; QH "1' H2
H3C CH3

The proposed ﬁrst step in the 1,4-elimination mechanism is the
oxidative addition of Ni" across the central C-C bond of a butane molecule.

This is followed by two successive B-hydrogen shifts onto the metal ion.
Finally, the reductive elimination of H2 results in the production of the bis-

oleﬁn ionic complex shown below.

1,4-elimination mechanism

H
l CH
Ni+ + c.1110 —> H3C-H2c-Ni+-CH2-CH3—>H3C-H2C-Ni+-|| 2
‘ CH2
H
HZC CH2 H2C| CH2
II-Ni"-II +H2 <— ||-Ni*-l|
Hzc CH2 HZCIli (3H2

Both of these proposed mechanisms would result in the production of ions

with identical masses. To determine which mechanism was responsible
for the formation of the NiC4H3+ ion, a series of isotopic labeling, ligand
exchange, and collision-induced dissociation experiments were performed.

The reaction was repeated using butane with deuterium atoms substituted
for the hydrogens on the terminal butane carbons. If the NiC4H3” ion was

formed through the 1,2-elimination mechanism, a neutral loss of HD or H2
would be observed. If, however, the NiC4H8+ is formed through the 1,4-
elimination mechanism the result should be a loss of D2. When Ni+ was
allowed to react with the labeled butane neutral a loss of D2 was observed;
this suggested that the NiC4H3+ ion had a bis-oleﬁn ionic structure and was

formed through the 1,4-elimination mechanism.

In previous studies the two ligand displacement reaction of
Ni(C4H4,)2+ with HCN to form Ni(HCN)2+ was determined to be exothermic
by 10 kcal [3]. However, reactions of Ni+ with HCN did not produce
Ni(HCN)n+ ionic complexes. ICR ligand exchange experiments were

performed using HCN to further investigate the structure of N iC4H8+. Nit

was again reacted with n-butane, this time in the presence of HCN, and the
rapid formation of Ni(C2H4)(HCN)+ and Ni(HCN)2+ ions from NiC4H8+ was

observed. Through double resonance experiments, these ions were linked
to the initial formation of NiC4H8+ which again indicated that the NiC4H8+

ion was formed through the 1,4-elimination mechanism. The sequential
displacement of C2H4 from NiC4H8+ also conﬁrmed that the NiC4H8+ ion
had a bis-oleﬁn structure.

In a separate study, collision induced dissociation experiments were

performed on the NiC4H8+ ion using a Fourier transform ion cyclotron

mass spectrometer [4]. Ni“ was reacted with neutral n-butane to form
NiC4H8+. All other ions were ejected from the FTMS cell, and the N iC4H8+

ion was collided with neutral argon molecules at varying collision energies.
At low collision energies the neutral loss of C2H4 was observed. As the
collision energy was increased a loss of C4H3 was observed along with the
loss of C2H4. Throughout all the collision energies investigated, the

dominant neutral loss was 02H4. These results conﬁrmed that the
NiC4H3+ ion had a bis-oleﬁn structure, and could sequentially lose one or
two of the C2H4 ligands depending upon the energy of collision.

Through the combination of these three indirect methods the
NiC4H3+ ion was conﬁrmed to have a bis-oleﬁn structure, indicating its
formation through a 1,4-elimination mechanism. These indirect methods
for structural elucidation have been used to conﬁrm proposed structures for
fragment ions formed through unimolecular decomposition as well as ions
produced through ion/molecule reactions. More conclusive results could be
obtained if the structures of ions could be determined directly through
spectroscopic methods, but this type of study is limited by difﬁculties
associated with concentrating ionic species in the gas phase due to
Coulombic repulsion. Instead, a combination of indirect methods for
structural elucidation is generally applied.

Several other indirect experimental methods exist for the
determination of ionic structures. These have recently been reviewed [5],
and are often used in combination to ensure that accurate conclusions are

drawn [6]. Some of these methods include:

Isotopic labeling studies

Ligand exchange studies/ ion-molecule reactions
Collisionally induced dissociation
Photodissociation

Kinetic energy release

Metastable ion characteristics

Isotopic labeling methods, described above, can yield useful
information, but require the use of labeled compounds and can be
complicated by internal scrambling if the ion is formed with high levels of
internal energy. Collision induced dissociation (CID) can be used to

determine both the mass and appearance potential of the collisionally

formed fragments, but the internal energy of the fragment ions is not well
deﬁned. Rearrangements can also occur in CID if large amounts of energy
are present in the fragment ions, and the CID technique requires
specialized instrumentation. Photodissociation is a powerful technique for
the determination of appearance potentials of molecular and fragment
ions. These can be used in combination with neutral heats of formation
data to infer ionic structures [7]. Unfortunately the excess energy of the
ions formed through photodissociation or other appearance potential
methods is often unknown which can complicate these structural
inferences. The observation of metastable ion peak widths and onset
energies can be used to compare the kinetic energy release and rate
constants for the degradation of ions [8]. This information can be used to
conﬁrm the presence of different structures for isomass ions, but yields
little direct structural information. Through combination, these indirect
techniques can be used as powerful techniques for investigating ion
structures.

Direct methods also exist for the investigation of ion structures;
several of these have been recently reviewed [9,10]. Microwave spectroscopy
of ions formed in plasma discharges has proven a powerful technique for
the study of the rotational absorptions bands of molecular ions [11].
Coulombic explosion experiments can be used to determine the nuclear
positions of ions through high energy collisions (MeV) of ion beams with
solid targets [12]. Photoelectron spectroscopy has been used to perform low
resolution analysis of ion structures [13]. Infrared laser spectroscopy has
recently emerged as an extremely powerful, high resolution technique for
studying ion structures [14]. This technique generally obtains infrared
absorption bands for ions using tunable infrared lasers. The laser beam is
sent coaxially through a high energy ion beam, and the infrared absorption
of ions in the beam is monitored using mode-locked detection. Mode-locked
detection is performed by applying an ac signal to the accelerating voltage of

the ion source, which allows discrimination against absorptions from

neutral molecules in the ion beam. This technique is extremely powerful,
but the wavelength region that can be scanned is limited to the tunability of
the IR laser, or the ion velocity range that can be tuned for velocity
modulation tuning. In general only a range of 0.25 cm‘1 can be scanned in
a single experiment.

Reactive species, such as ions, can also be studied using the
technique of matrix isolation. In a matrix isolation experiment a reactive
species is trapped within an inert solid matrix maintained at extremely low
temperatures. This technique, ﬁrst introduced by George Pimentel in 1954
[15], has been used to study a variety of highly reactive transient species by
electron spin resonance, Fourier transform infrared and laser induced
ﬂuorescence spectroscopies [16,17]. For ionic species, it has been stated that
“modern laser experiments have shown that some of the more detailed
infrared studies made in crystals do, in fact, produce reasonable
approximations to gas phase structures and vibrational frequencies...” [18];
thus matrix isolation spectroscopy of ions has proved to be a useful
complement to high resolution laser spectroscopy experiments. Matrix
isolation experiments are generally performed at temperatures of ~10 K,
and therefore must be conducted within a high vacuum chamber using
cryogenic techniques. The analyte sample or ion is generally co-deposited
with an unreactive matrix gas, commonly argon or neon, onto an
appropriate substrate maintained at $10 K. Ifthe matrix gas is deposited at
the correct matrix to host ratio, often 10,000:1, ions can be trapped within an
unreactive noble gas ‘cage’ and stored for extended periods. This allows the
accumulation of higher concentrations of ions than can be obtained using
gas-phase methods.

Various methods have been used to generate ionic species within low
temperature noble gas matrices and have been described in several reviews
[19,20,21,22,23]. Ions formed through these methods are generally
classiﬁed as either bound ion pairs or isolated ions (ions trapped in
separate matrix cavities). Some of the methods used to produce isolated

ions include;

Vacuum ultraviolet (VUV) photoionization [24]
y and x-ray irradiation [25]

Windowless discharge [26]

Proton radiolysis [27]

Electron bombardment [28]

Fast atom bombardment [29]

Laser vaporization [30]

Chemical ionization sources [31]

Deposition of mass-selected positive ions [32]

VUV photoionization experiments are performed by co-depositing the
sample precursor molecule and the matrix gas upon the cold substrate.
Following deposition, the matrix is bombarded with high energy VUV
photons from a mercury, hydrogen, or argon lamp. These high energy
photons can photoionize the sample molecules, forming ions in the matrix.
Gamma irradiation, x-ray irradiation, proton radiolysis, and electron
bombardment techniques are similar techniques that utilize other energy
sources to form ions during or after deposition of the matrix. Fast atom
bombardment (FAB) and laser sputtering have both been used to produce
ions that are co-deposited with the matrix gas. FAB has been used to
directly produce ions from metal targets, and laser sputtering has been
used to produce clusters from metal surfaces that are deposited and ionized
using VUV photoionization. Another method recently used to deposit both
positive and negative ions within a noble gas matrix is chemical ionization.
A mass spectrometer high pressure chemical ionization source was
adapted for production of ions in a matrix isolation experiment, with argon
utilized as the source buffer gas. From this high pressure source both
positive and negative ions were effectively trapped in an argon matrix. All
of these techniques have been successfully used to produce ions trapped

within noble gas matrices, but with all of these methods a variety of ions
can be formed from the matrix gas, precursor molecule, and contaminants
in the matrix, making the exact determination of ion structure difﬁcult.
Matrix isolation experiments have recently been performed by depositing
high energy (50-350 eV) mass-selected positive ions produced by a
quadrupole mass spectrometer [32]. This method of ion production allows
the unambiguous deposition of speciﬁc ions within the matrix for
subsequent spectroscopic analysis.

The goal of the matrix isolation project currently in progress within
the Allison and Leroi groups is to isolate mass-selected ions produced by
quadrupole mass spectrometers for spectroscopic investigation. Most
methods currently in use for the production of matrix isolated ions can
produce a variety of ionic species, at times making the identiﬁcation of
trapped species ambiguous. By trapping the mass-selected ions emanating
from quadrupole mass spectrometers we have greatly increased the
selectivity of the ion trapping process. It is also generally true that most ion
formation conditions are accompanied by the production of oppositely-
chargedicounter-ions. In processes such as VUV photoionization these
ions are assumed to result from the capture of electrons released through
ionization by electrophilic contaminant molecules, such as H20, trapped in
the noble gas matrix. The formation and/or need for these counter-ions in
the effective trapping of ions in noble gas matrices is not well understood.
We have designed an instrument that enables both positive and negative
ions to be co-deposited within the matrix gas to allow the study of the role of
counter-ions in effective ion trapping.

The instrument that has been constructed for these studies is shown
in Figure 1.1. A more detailed discussion of the design and operation of the
entire instrument is forthcoming [33], and the overall instrument will be
outlined only brieﬂy in this manuscript. The instrument is based upon an
8” diameter ultrahigh vacuum chamber constructed of stainless steel and

conﬂat ﬂanges which is pumped by a Leybold-Heraeus 1500 l/sec cryopump

 

 

' (— Matrix
A/f f 1 Gas

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cryostat Liquid
Nitrogen
Dewar Mass-Selected
x) Ion Sources
\ sprayer
11.th
FTIR / I +++ +l C10 «(EgDetecor
\ CsI @(C-o
TIOKe Window
arge t
cryopump Foe‘eidlfehroughs

 

Figure 1.1 The matrix isolation instrument designed for the isolation of
mass selected ions followed by spectroscopic analysis (LIF and FTIR).

10

to maintain clean deposition conditions. This chamber houses an Air
Products Displex 202 cryostat head which can achieve temperatures down
to 10 Kelvins. This cold head is in thermal contact with an optical substrate
upon which both ions and matrix gas are co-deposited. The matrix gas, in
this case argon, is introduced through a high precision leak valve to a
sprayer that directs the gas onto the cryostat window. Two 6” conﬂat
ﬂanges, used to mount the ion sources, are located opposite the cryostat
window as shown in Figure 1.1. Ion sources have been fabricated that can
produce low energy positive and negative mass-selected ion beams which
are directed onto the cryostat window using a series of electrostatic Optics.
Using these ion sources, mass-selected positive and negative ions are co-
deposited with matrix gas to allow the accumulation of ions for extended
periods, in some cases over a period of days. A linear motion feedthough
(LMF), not shown in the ﬁgure, can be inserted between the cryostat
window and the ion sources. An electrically isolated Faraday plate is
mounted onto the LMF and can be positioned directly in front of the optical
substrate. This Faraday plate can be inserted to measure ion current at the
cryostat window, thus allowing the operator to optimize the ion sources and
deﬂection optics for maximum ion deposition current. Following
optimization, the Faraday plate is withdrawn to allow the deposition of ions
within the noble gas matrix.

The chamber also houses window ﬂanges that allow the introduction
of infrared or visible laser light into the high-vacuum chamber, through
the optical substrate/matrix, and out of the chamber. Infrared analysis can
be performed using a BOMEM DA 3.02 Fourier transform infrared
spectrophotometer. Infrared light from one of the operator-selectable beam
paths of the BOMEM instrument passes through the matrix and is collected
at a detector installed outside the exit window mounted in-line with the
beam path. For mid-IR analysis a mercury-cadmium telluride detector is
used. A liquid helium-cooled germanium bolometer detector is also

available for far-IR studies. High sensitivity, laser induced ﬂuorescence

11

(LIF) studies were also performed using this instrument. These studies
utilized a Nd:YAG pumped dye laser located within the LASER laboratory
in the Michigan State University Chemistry Department as a wavelength-
variable excitation source. For LIF studies the infrared assembly was
removed, and the laser beam was introduced through apertures which
serve to reduce stray light, and into the matrix chamber along the same
axis as the infrared beam. The ﬂuorescence signal produced by ions
trapped in the matrix was collected through an off-axis window (not
shown). Cutoff ﬁlters were employed to discriminate against scattered
laser radiation, and the ﬂuorescence signal was detected with a PMT.

The focus of this dissertation is the modiﬁcation of residual gas
analyzers (RGAs) to produce the beams of mass-selected ions necessary for
these matrix isolation studies. RGAs are specialized high vacuum
Pressure gauges that are often employed as an alternative to ion gauges.
The are designed to measure both total and partial pressures of residual
gases within high vacuum chambers from 1x10'4 torr to 1x10'12 torr.
Partial pressure measurements are possible through the use of miniature
quadrupole mass spectrometers. RGAs generally consist of a high
efﬁciency ionization source, extraction optics, quadrupole mass ﬁlter, and a
Faraday plate detector. RGAs are much less expensive than conventional
mass spectrometers, generally costing between $5,000 and $10,000. They
are small in size, often 4 to 6 inches in length, generally operate at unit
reSolution, and can vary in mass range from 1-60 up to 1-200 daltons. To
allow their use in ultra-high vacuum environments, RGAs utilize
SDecialized high efficiency electron impact ionization (E1) sources which
can produce mass-selective sensitivities of ~2.0x10'4 A/torr. The high
efﬁciency, small size, and low cost of RGAs made them an ideal choice for
modiﬁcation to produce the mass-selected ion beams required for the
matrix isolation instrument.

Following this introduction, Chapter 2 provides a background
discussion of the ionization methods and ion sources that were used to

12

produce the ions prior to mass-selection, the quadrupole mass ﬁlter used to
mass-select these ions, and the electrostatic optical tools used to optimize
and design changes to the ion optics of these systems. Chapter 3 is a
discussion of the modiﬁcation of an Anavac-2 RGA to produce positive and
negative ions using several ionization techniques. Chapter 4 presents the
modiﬁcation of a Dycor M200 RGA to produce a beam of mass-selected
positive ions. Finally, Chapter 5 includes a general discussion of some
conclusions drawn through the work presented in chapters 3 and 4, and
some suggestions for future modiﬁcations to improve the performance of

the ion sources.

13

Chapter 2 . The theory and operation of RGA components

Jim
AW

The most common type of source used for the production of ions in
mass spectrometry is the electron impact ionization (EI) source. An EI
source is based on the ionization of neutral molecules through interaction
with a high energy (~70 eV) electron beam. The EI source was ﬁrst
introduced by Lenard in 1902 [34], and continues to ﬁnd extensive use in
many physical and analytical applications of mass spectrometry [35].

Electron impact ionization is based upon the interaction between a
high energy electron and a neutral molecule to form both intact molecular
ions and fragment ions. While the term impact implies a collision during
ionization, the respective size differences between the electron and a
molecule make an actual collision highly improbable. The actual
ionization occurs due to an interaction (on a quantum-mechanical level)
between the wave of the high energy electron and the electric ﬁeld of the
neutral molecule. When an electron is accelerated to an energy of several
electron volts, it can achieve a de Broglie wavelength of ~0.1 nm [36],
corresponding to short wave radiation. This wavelength is on the order of
the dimensions of the electron, as a result an electron of this energy will
possess a wavefunction consisting of a wide range of frequencies. As this
electron comes into close proximity with a neutral molecule, its complex
wave function can interact with the electrons of the neutral molecule.
Some portion of the frequency components present in the electron wave
will be absorbed by the neutral molecule, resulting in the transfer of
energy to the molecule. If sufﬁcient energy is transferred to the neutral
molecule an electron can be ejected, resulting in the formation of an ion.

14

Some of the different ionization mechanisms that can occur as a
result of the interaction of a high energy electron with a neutral molecule
(AB) include:

Ionization AB + es —-> AB+ + es + ee (1)
Dissociative Ionization AB + es -—> A+ + B + eS + ee (2)
Multiple Ionization AB + es —9 AB“ + es + zee (3)

Where es is a scattered electron and ee is an electron ejected by the neutral

molecule. An ion formed by EI following the ejection of an electron often
contains excess internal energy. If this energy can be effectively dissipated
within the vibrational modes of the molecule, a stable molecular ion will be
formed as shown in process (1). If, however, sufficient excess energy is
present in the molecular ion, the ion will break apart into fragments as
shown in process (2). The vibrational energy levels within a molecule are

generally much lower than the energy required for ionization; for example
the largest vibrational level is for H2“ = 0.27 eV, and the ionization

potential of H2 is 15.426 eV [37]. As a result , ions formed through EI will

often consist of a mixture of vibrationally excited states, resulting in the
formation of several different fragment ions. One of the principal
advantages of E1 is that it produces a characteristic distribution of
fragment and molecular ions (cracking pattern) that can be used to
identify and determine the structure of an analyte molecule.

The production of a molecular ion in El depends strongly upon the
amount of energy the neutral molecule absorbs through interaction with
an electron. At electron energies close to the ionization potential of the
neutral molecule EI generally produces only the molecular ion. As the
electron energy is increased more energy is transferred to the ionized
molecule, and the formation of fragment ions increases. A plot of ion
intensity versus electron energy for a given molecule is called a

Glastogram, an example of which is shown in Figure 2.1.

15

 

 

 

 

 

 

 

 

'3
6‘
E
0
$9
9.
c *2
.9
25
Q)
(I)
U)
in
O
k
U
.5 H”
I; N”
.2: .
.5 " we
7-5 W
t NH?!
(O
O.
H;
1 1% L 4 1 L r l 1 3

 

50 100 150
Electron energy (eV)

Figure 2.1 An example of a clastogram for ammonia adapted from
Mark et al. [38].

16

This clastogram shows the production of the molecular ion and fragment
ions from ammonia with increasing electron energy [38]. The onset
electron energy for the production of the molecular ion, or the ionization
energy for the intact molecule, occurs below 10 eV for NH3. At this
electron energy very little fragmentation occurs. As the electron energy
increases different fragment ions begin to appear. The electron energy
corresponding to the initial appearance of a fragment ion is called its
appearance energy. If the energy of the electron beam is increased
ﬁrrther, the relative intensities of the fragment ions become less sensitive
to electron energy. For most molecules the fragmentation pattern
stabilizes at electron energies above 50 eV. For this reason reference mass
spectra, used to identify unknown compounds from their fragmentation
patterns, are generally taken at an electron energy of 70 eV [39]. The
disadvantage to operating at these high electron energies is that it may be
difﬁcult to observe the intact molecular ion. There is often an inverse
relationship between the molecular weight of a molecule and the intensity
of the molecular ion produced using EI [40]. For this reason E1 is
considered a ‘nonsoft’ ionization technique, and often proves to be
unsuitable for the analysis of large molecules.

The interaction of a high energy electron with a neutral molecule
can produce highly excited molecular ions. If the molecular ion contains
sufﬁcient energy, and no effective pathway is present to dissipate this
excess energy, multiple ionization (3) can occur. This process generally
occurs only at high electron energies, with the partial ionization cross
section for the multiply charged ions (most often +2) rarely exceeding 1-5%
of the primary singly charged peak intensity [41].

Most EI sources are based upon the design originally devised by
Dempster [42], and further optimized to its modern conﬁguration by Nier
[43]. A block diagram of a Nier conﬁguration EI source is shown in Figure
2.2. The Nier type EI source generally consists of a source housing of

square or cylindrical geometry, depending on the desired symmetry of the

17

ion beam. A ﬁlament is mounted on one side of the source and heated
until it emits free electrons. These electrons are attracted toward the
source housing by a potential applied between the source and the ﬁlament,
commonly 70 V. Although most of these electrons are neutralized on the
source housing, a small percentage enter the source through the electron
entrance aperture. These electrons continue on their initial trajectories
through the source, and can impinge upon the collector located opposite

the electron entrance aperture as shown in Figure 2.2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Filament
. Heater Extractor/Ion
Fllament H I 1 ClCCtron focus opu'cs
entrance
Source Housing perture
‘ e c "
Repe er ° ° <29
é'ec
e C
8.9%. ............ I,
e e. '1.
escrons ,:.
e e i'
Cue“ ............ 'I
see
eee
e ee
\_e 8‘
Source
Collector I Exit
Aperture

Figure 2.2. A drawing of the conventional Nier style EI source

 

18

If these 70 eV electrons encounter neutral molecules as they
traverse the source volume, ions can be formed as previously described. A
repeller is commonly placed opposite the ion exit aperture. An electric
ﬁeld can be created within the ion source by applying a potential to the
repeller, this electric ﬁeld causes ions that are formed through interaction
with the electron beam to drift toward the source exit aperture. If the ions
pass through this aperture they can be accelerated, focused using a lens
system, and injected into the mass spectrometer.

The operating conditions of the source ﬁlament can be monitored
using several methods. If the electron current at the collector is
measured, an accurate assessment of the electron current within the
source volume can be obtained. This collector current is generally on the
order of 10-100 11A [44]. It is also possible to monitor the heating current
applied across the ﬁlament, and the total emission of electrons from the
ﬁlament to evaluate its operation. To increase the ionization efﬁciency of
the ion source a magnetic ﬁeld is often applied parallel to the electron
path. This causes the electrons traveling through the source to follow a
helical path, increasing their pathlength through the source, thereby
increasing the overall ionization probability [45].

The use of a repeller voltage to create an electric ﬁeld within the E1
source has been found to cause some mass discrimination [46]. To obtain a
more uniform sampling of the ions formed within the source volume the
source can be operated without this repeller potential [47]. Instead of
using a repeller potential to create an electric ﬁeld within the source, in
this mode of operation the entire source region is held at a common
potential and an extraction lens after the ion aperture is used to draw ions
out of the source.

The overall degree of ionization in a Nier type EI source can be
increased through: 1) Increasing the ﬂux of electrons through the source
volume, 2) Increasing the cross section of the electron beam present in the

source, and 3) Increasing the pressure of the sample within the source

19

volume. Increasing the electron ﬂux in the source will increase the
probability of an electron/neutral “collision”, but is limited by the output of
the ﬁlament, the lifetime of a ﬁlament is inversely related to its output [45].
Use of a large electron aperture will increase the cross section of the
electron beam, but ions formed in the source will then have a broader
range of initial energies because they are formed from a wider range of
positions within a region of varying electric potential. Increasing the
pressure in the source will also increase electron/neutral collision
probability, but can induce ion/molecule collisions and interfere with
operation of the mass spectrometer. The ﬁnal design and operation of this
type of source is generally some compromise that depends largely upon the
requirements of the overall instrument.

The Nier design EI source has proven effective for the production of
molecular and fragment ions from a wide range of neutral compounds,
but there are several disadvantages to its design. It has a relatively low
overall ionization efﬁciency due to small electron beam cross section inside
the source. It has been estimated that less than 1% of the neutrals that
enter the ion source are effectively converted into positive ions [36]. It may
also be difﬁcult to observe the molecular ion for certain compounds due to
the high energy of the ionization process. Certain samples can also
decompose at the temperatures required for volatilization, precluding their
use in a source of this type that requires the use of a gas-phase sample.
Finally many samples of interest cannot be volatilized at all, making their
analysis through EI impossible.

RGAs are designed to function as ultra-high vacuum pressure-
measuring devices. As such, an RGA must utilize a source that creates a
maximum number of ions at any given sample pressure. This is
especially important since it determines the ultimate vacuum pressure
that an RGA can measure (at present, RGAs are designed to measure
pressures down to 1x10'12 torr). To obtain the necessary sensitivity for

these measurements, most RGAs utilize an El source with a

20

conﬁguration that differs slightly from the original Nier design. This
design was originally utilized in the early quadrupole mass spectrometer
constructed by Paul et a1. [48], and a drawing of a typical RGA EI source is
shown in Figure 2.3.

Extractor/Focus

 

..........................

Filament Housing Source Grid
Figure 2.3. The high efﬁciency EI source commonly used in RGAs

RGA EI sources are commonly based upon a radially symmetric
geometry, which matches the acceptance of a quadrupole mass analyzer.
Similar to the Nier design, electrons are produced by heating a ﬁlament,
commonly made of tungsten, rhenium, or thoriated iridium. In the RGA
source these electrons are attracted to a highly transparent cylindrical
source grid by the electric ﬁeld between the grid and the ﬁlament. The
potential between these two elements determines the electron energy. As
the electrons travel through the grid they can ionize neutral molecules to
form molecular and fragment ions. The grid is generally biased at +3 to
+10 V, and the ﬁlament housing surrounding both the ﬁlament and grid is
generally biased at -70 to -100 V. Because of the transparent nature of the

grid, most electrons will travel through the entire source grid region.

21

After passing through this region, they slow down due to the electric ﬁeld
between the grid and the ﬁlament housing. This electric ﬁeld eventually
causes the electrons to ‘tum-around’ and pass through the source grid
again. In this fashion, an electron can make several passes through the
source grid. This greatly increases the overall efficiency for the ionization
process by increasing the pathlength of the electron through the ionization
region. By locating the ﬁlament close to the grid, most of the electrons
emitted from the ﬁlament pass through the grid region, again increasing
the efﬁciency of the source. Ions that are formed within the grid region
feel an attractive force induced by the penetrating ﬁeld of the extractor lens
(commonly held at ~40 V), and exit the grid region through the source
aperture. This extraction is similar to that observed using an external
extraction ﬁeld in the Nier source (not using the repeller), and therefore
also has the advantage of minimal mass discrimination gained through
operation in that mode.

Another advantage gained through the use of this cylindrical grid
design is the decrease in observation of ions formed by electron-induced
desorption (EID). In a Nier source a large percentage of the electrons that
travel through the source volume ﬁnally impact upon surfaces within the
source (such as the collector or the walls of the source itself). These
electrons can cause molecules adsorbed on these surfaces to desorb and
ionize. As a result, peaks can appear that are not due to species present in
the residual gases of the vacuum. The cylindrical grid design commonly
used in RGAs minimizes these electron-surface collisions. Since electrons
spend most of their time orbiting the highly transparent grid, very few
come close to the source housing. This decreases the probability of electron
collisions with the ﬁlament housing. As a result these high-efﬁciency EI
sources have a much lower background signal from EID [49].

Both the cylindrical grid and the Nier style style EI sources have
advantages depending upon the application. The cylindrical grid source
has the advantage of increased sensitivity due to the long electron path

22

through the ionization region. It also suffers less from background
signals due to formation of ions through EID. These criteria are especially
important in the analysis of residual gases to maximize the overall
sensitivity of the device and minimize background interferences at these
low operating pressures. The Nier-type source can be advantageous when
working with a system where the sample is introduced directly into the
source volume. Because the ﬁlament lies outside the source volume, fewer
electrons are generally available for ionization within the source region,
but the advantage to this design is that the ﬁlament is not directly exposed
to the sample gas. This is particularly important when working with
thermally labile samples that would decompose if exposed to hot surfaces,
or when using oxidizing samples that could attack the ﬁlament and

reduce its lifetime.

ED] .11..

The technique of chemical ionization (CI) is commonly used as a
complement to electron impact ionization in mass spectrometry. While EI
normally forms molecular ions accompanied by a large degree of
fragmentation, CI is considered a ‘soft’ ionization technique producing
little fragmentation, and is therefore particularly useful for molecular
weight determination. In CI the ionization is performed through
attachment of some charged species, commonly a proton, to the intact
analyte molecule. This interaction generally imparts a much smaller
amount of internal energy into the product ion than EI, thereby reducing
the degree of fragmentation. This effect was ﬁrst described as an
analytical technique by Munson and Field in 1965 [50].

The formation of the protonated analyte molecular ion is
accomplished through the interaction of the neutral with reagent ions
formed in a high pressure reagent gas. The two gases are introduced into

the source simultaneously, with the ratio of analyte to reagent gas of

23

approximately 1:105. Electrons are injected into the source in a fashion
similar to El. Because of the large excess of reagent gas in the source,
most of the electrons will ionize reagent gas molecules. Methane was ﬁrst
used as a reagent gas for CI, the principal ions in the mass spectrum of

methane include:

CH4 + e‘ —9 CH4+ + 2e' (4)
L—> CH3+ + H + 2e' (5)

If the pressure of the reagent gas is maintained at an extremely
high level (~1 torr), ion/molecule reactions can occur to form new ions.
Since the reagent gas predominates within the source, almost all of the
ion/molecule reactions will be between the primary ions of the reagent gas
and the neutral reagent gas. For methane the principal ion/molecule

reactions that occur are [51]:

CH4+ + CH4 -) CH5+ + CH3 (6)
CH3+ + CH4 —) CzH5+ + H2 (.7)

If the molecular ion reacts with a neutral methane molecule the neutral
molecule can accept a proton to form the ion CH5+, shown as process (6).

If the CH3+ ion reacts with neutral methane, C2H5+ can be formed through

process (7). In the case of methane these ions will not further react with
the reagent gas. This allows a large concentration of these “reagent ions”
to accumulate within the high pressure source, eventually achieving a
steady state. These reagent ion peaks, CH5+ and C2H5+, are the
predominant peaks in the mass spectrum for methane CI.

Most organic molecules have a stronger proton afﬁnity than either
CH4 or the C2H4. If a high concentration of CH5+ and C2H5+ reagent ions
is present in the source the probability of an ion/molecule reaction between

a reagent ion and an analyte molecule increases. Under these conditions

24

there is a strong probability that a analyte molecule (M) can react with a

reagent ion and become protonated, forming MH".

M + CH5+ -—> MH+ + CH4 (8)

The primary peak in the high pressure methane CI spectrum for analyte
M is generally the protonated intact molecule (M+1)+. The principal
advantage of CI is that fragmentation occurs to a much lesser degree than
with EI, and therefore much better molecular weight information can be
obtained.

Other reagent gases can be used for CI; these can be especially
useful for generating selective ionization conditions for a particular
analyte compound. The probability that an analyte molecule will accept a
proton from a reagent ion is related to the difference between the proton
affinity of the analyte and the proton afﬁnity of the molecular precursor to
the reagent ion [52,53]. This difference in proton afﬁnity can be used in the
analysis of a mixture to selectively protonate one analyte over another. It
can also be used to impart different amounts of internal energy into the
protonated analyte ion, and thereby control the degree of fragmentation
observed. If a reagent gas is selected that forms reagent ions having a
much smaller proton afﬁnity than the analyte molecule, a higher degree of
fragmentation will be observed in the analyte mass spectrum [54]. This
allows the ionization conditions to be chemically tuned to control the type of
information obtained. By properly selecting the reagent gas used in CI, it
is possible to not only selectively ionize different analyte molecules, but also
to control the degree of fragmentation observed in their mass spectrum.

Under conventional EI conditions both positive and negative ions

can be generated through the process of ion pair formation.

Ion Pair Formation AB + e' —-> A+ + B' + e‘ (9)

25

Although this ionization mechanism could potentially prove useful
for the analysis of compounds, the intensity of the negative ion signal
produced through ion pair formation in BI is generally several orders of
magnitude lower than the intensity for positive ion formation through
single ionization and fragmentation [55]. As a result, the sensitivity of an
analytical technique based on this ionization process would be quite low.

The high pressure conditions necessary for CI can also be used to
produce negative ions through electron capture negative ionization (EC-
N1) and negative chemical ionization (N01). The use of high pressure EC-
NI was ﬁrst demonstrated by Hunt, et al. in 1976 [56], and has since proven
to be an extremely sensitive technique for the analysis of certain classes of
compounds. It has become especially important in the analysis of
halogenated environmental pollutants because of its extreme sensitivity for
this class of compounds [57].

Negative ions can be formed under high pressure conditions
through several mechanisms, the three most important of which are
resonant electron capture (10), dissociative electron capture (11), and

negative ion-molecule reactions (12).

Resonant electron capture AB + e' (~O.1 eV) -9 AB‘ (10)
Dissociative electron capture AB + e' (~0.1 eV) -> A + B‘ (11)

Negative ion-molecule reactions AB + C‘ —> ABC“ (12)

Resonant electron capture involves the formation of a mi:
negative ion through the capture of a low energy electron by the analyte
molecule. This mode of ionization requires very low energy electrons
(~0.1eV) to prevent the negative molecular ion from ejecting
(autodetaching) the captured electron. Under conventional EI conditions,
where ~70 eV electrons are employed, very few electrons with this low
energy are available. If the pressure in the ion source is increased the

likelihood that an electron will collide with a molecule can be increased.

26

When such a collision takes place some of the electron’s energy is lost,
generally the energy of a “primary” electron is reduced by ~30 eV as it
ionizes a neutral molecule [58]. If several such collisions occur the
electron can lose almost all of its initial energy and form a low energy
(thermalized) electron. The secondary electrons ejected ﬁom the ionized
molecules generally have low energies, which can also be further reduced
through collisions with gas molecules. By introducing a high pressure of
“buffer gas” into the source along with the analyte, it is possible to create
conditions that will favor the production of a large number of these low
energy electrons. These thermal electrons are then available for the
production of negative molecular ions through resonant electron capture.
Excess energy gained by the molecular ions as result of the electron
capture process can also be dissipated through collisions with neutral
buffer gas molecules. If the molecular anion contains sufﬁcient excess
energy following electron capture it can fragment to form negative
fragment ions; this is called dissociative electron capture.

This mode of operation for a CI source is often called electron
capture - negative ionization (EC-NI). The formation of negative ions
through EC-N I has been shown to be strongly dependent upon the electron
affinity of the analyte, the energy spectrum of the ionizing electrons, and
the frequency of electron/molecule and ion/molecule collisions within the
source region [59]. EC-NI has proven to be an extremely selective and
sensitive technique (up to 1000 times greater sensitivity than positive CI)
for the analysis of certain compounds, especially those compounds
containing a highly electrophilic site [60].

Negative ion-molecule reactions, process (12), can also be used to
produce negative ions under CI conditions. If an electrophilic buffer gas is
selected and the source is operated at high pressure conditions, the buffer
gas molecules can capture the thermalized electrons to form reagent ions.
An example of this is the formation of Cl' using CCl4 or CH2C12 as a buffer

gas [61]. At these elevated pressures the reagent ions can then further

27

react with the analyte molecules to form molecular adduct ions. This
mode of operation is generally called negative chemical ionization or N CI.
The advantage of N CI is the simplicity of the mass spectra produced; often
only the molecular adduct ion is observed since the formation of an adduct
ion is generally a low energy process. The NCI ionization technique also
has excellent selectivity, the selectivity of the ionization depending upon
the electron aﬂinities of the reagent gas and the analyte molecule [62].

The source required to produce the high pressure conditions
necessary for CI, EC-NI, and NCI conditions is similar in design to the
Nier EI source shown in Figure 2.2. The principal modiﬁcation made to
the Nier design is the reduction of apertures in the source to reduce
conductance, allowing the creation high pressure conditions within the
source volume. The reagent or buffer gas must be present at pressures
approaching 1 torr, so the entire source housing is generally sealed except
for the electron entrance aperture and the ion exit aperture (shown in
Figure 2.2), and the size of these apertures is generally reduced. Similar
to the Nier design, the ﬁlament is located outside the source region. This
is particularly important for CI and N CI work to prevent the reagent gas,
which is often oxidative in nature, from attacking the heated ﬁlament.
The electrons produced by heating the ﬁlament are accelerated toward the
source housing by a strong electric ﬁeld induced by potential applied
between the ﬁlament and the source housing. For high pressure studies
an electron energy greater than 100 eV is commonly used to allow the
electrons to effectively penetrate the high pressure source region [63]. The
output of the ﬁlament operated under these conditions is generally only
monitored by observing the total emission from ﬁlament, not by monitoring
the collector current. This is necessary since under these high pressure
conditions only a very small percentage of the electrons that enter the
source can effectively traverse the source volume and hit the collector.

Because of the high source pressures required to form ions through

CI, EC-NI and NCI, special pumping is generally required. In most mass

28

spectrometers the analyzer region must be maintained at ~1x10'6 torr to
avoid ion/molecule collisions as the ions travel through the mass
spectrometer. The extremely high pressures present in a source operated
under CI conditions produces a high background pressure in the source
vacuum chamber. To maintain the pressure required for effective ion
formation and mass analysis, differential pumping is generally used
between the source and analyzer regions. If an aperture is placed between
the CI source and the mass analyzer and separate pumps are connected to
each chamber, a pressure differential can be maintained between the two
regions. If each region is separately pumped, the high pressure
conditions necessary for CI can be maintained in the source volume
without substantially elevating the pressure in the analyzer region. Most
mass spectrometers that are designed to perform CI utilize this type of
differential pumping.

DSEI'I'

Surface ionization is the process by which a material that has been
coated or directed onto a heated surface evaporates in an ionic state.
Surface ionization sources can be used for the production of both positive
and negative ions. These ions are generated either directly from a hot
surface, or as a result of the interaction of the heated surface with some
neutral species. Surface emission, or thermionic emission, involves the
ejection of a particle present in the matrix or substrate when the surface is
heated. Surface ionization (SI) involves the emission of charged species
from a molecule that is adsorbed on the surface. These molecules can be
either applied directly to the surface, or adsorbed onto the surface from the
gas phase. Upon heating, these adsorbed species can desorb and ionize to
form either positive or negative ions.

According to Duckworth [64], the ﬁrst observation of positive ion
emission from a heated material was made by Elster and Geitel in 1889

29

[65]. This was followed closely by the work of Beattie in 1899 in which
several salts were applied to a ﬁlament and heated, producing positive ions
[66]. The ﬁrst observation of mass-selected positive ions generated through
surface ionization was made by Thomson in 1908 [67]. He observed the
presence of a peak at m/z~27, corresponding to the production of N 3*, upon
the heating of platinum wire. In 1936 Blewett and Jones published an
extensive study in which strong currents of a variety of positive alkali ions
were produced through heating alkali aluminosilicates [68]. Surface
ionization has been used for the determination of ionization energies,
electron affinities, and isotopic ratios, as well as for generating ion beams
[69,70]. An excellent discussion the surface properties affecting the
thermionic emission and surface ionization of species has recently been
compiled in this laboratory [71]. The development of techniques which
utilize positive surface emission (PSI) has also been discussed in the
literature [72,73,74].

The overall design of a surface ionization (SI) source is shown in
Figure 2.4. The source is similar in design to the Nier conﬁguration
described earlier, with the exceptions of the placement of the ﬁlament and
the overall geometry of the slits. In a typical SI source a ﬁlament is placed
inside the source housing and coated with a substrate to produce the
desired surface (if necessary). A slit conﬁguration is often used for SI,
since this geometry is well matched to the acceptance of magnetic
instruments. The ﬁlament is oriented within the source in-line with the
source exit slit. The ﬁlament ends are connected to leads that run to a
high current power supply which heats the ﬁlament. This assembly is
also connected to a ﬁlament bias power supply that allows the ﬁlament
surface to be ﬂoated at some potential. A V-shaped repeller is commonly
mounted opposite the exit slit. As the ﬁlament is heated, ions are
produced through surface emission or surface ionization. These ions
experience a force induced by the electric ﬁeld created between the repeller
and the source housing, and migrate toward the exit slit. These ions are

 

30

further extracted and focused by the lenses following the source,
producing an ion beam. The sample in SI can be either applied directly to

the ﬁlament or introduced as a gaseous neutral through a gas inlet line.

 

 

 

 

 

Filament
Bias
Filament I I ll Extractor/ion
Heater \\ focus Optics
Repeller
G— ions 9 -. . ions

   

Gas
Inlet C

 

O

/ V Source

Coated . .
Filament Exrt Slrt

 

 

 

 

 

 

Figure 2.4. An example of a source designed for surface

emission/ionization.

Negative surface ionization (NSI) has been less frequently used than
PSI. It was ﬁrst investigated by Owen in 1904 [75], who observed an
extremely small emission of negative ions upon heating a negatively-
biased ﬁlament. In 1912 Richardson detected more intense, mass-selected

signals of negative ions from alkali iodides and bromides coated on

31

platinum ﬁlaments [76]. In general NSI has received less attention than
PSI because of the weaker emission of NSI sources, interferences due to
co-emission of electrons, and the limited number of ions that can be
produced by NSI. Recently interest had been revived in NSI for both the
generation of negative ions and the detection of neutral molecular beams
[77]. Several exhaustive reviews have been written that cover the progress
of NSI research; these are the source of much of the discussion that will be
presented here [78,79,80,81,82].

Although NSI has some limitations , there are several advantages to
using NSI as a means to generate negative ions. The NSI source, the
same conﬁguration described in Figure 2.4, is commonly quite small and
simple to operate. For atoms/molecules with high electron afﬁnities the
overall ionization efﬁciency of NSI can be relatively high. Because NSI
requires species with a strong electron afﬁnity, the technique is highly
selective and an extremely low background signal is generally obtained
ﬁom residual background gases. The ions formed through N SI generally
have a small initial transverse energy spread (~0.5 - 0.6 eV) and a low
internal energy spread, which can be important in certain applications.
Finally, relatively small amounts of sample are required for the
production of ions through N SI, especially if the sample is applied directly
to the surface.

Unfortunately there are several disadvantages to using N SI. Only a
relatively small number of elements can be effectively converted into
negative ions through NSI. NSI is only useful for producing ions from
atoms or molecules with an electron afﬁnity of > 1 eV [80]. In practice only
28 elemental negative ions have been produced using NSI [83]. In
addition, the production of negative ions through N SI is often accompanied
by the thermionic emission of electrons. This can be a problem unless
some method is used to separate the electrons from the negative ions.

Finally, the surface used for NSI can degrade and become contaminated

32

when the source is operated at high pressures or low surface
temperatures.

The production of negative ions through NSI requires the use of a
surface possessing several characteristics. It must have a low work
function to allow the species that are desorbed from the surface to pick up
electrons and form a negative ions. For effective negative surface
ionization, the surface must be maintained at very high temperatures for
extended periods, and must therefore have a high melting point. Finally
the surface must be chemically stable when exposed to the molecule that is
ionized. Several different materials have been tested for use as surfaces in

NSI [84 ,85], a few examples of such surfaces are given in Table 2.1.

Table 2.1 The properties of some surface materials commonly used for
NSI

Mace Work Function ((1)) Melting Point (K) Qper rial; ng Temp, (K)

Re 4.4 - 5.1 eV 1500 2200
LaB6 2.3 - 3.5 eV 2480 1400 1700
Tho2 1.7 - 3.1 eV 3320 1500 - 1900

Rhenium is sometimes chosen because it doesn’t form carbides, a
problem with tungsten ﬁlaments, and has a high stability toward oxygen.
Unfortunately the relatively high work function of rhenium limits its
applicability for NSI systems. LaB6 has been used for NSI work,

principally because it has a relatively low work function, and high thermal
and chemical stability [86]. A procedure for the preparation of La36
surfaces [87], and the use of these surfaces for the detection of atmospheric

iodine have appeared in the literature [88]. The principal difﬁculty with
LaB6 surfaces is that they are relatively brittle and sensitive to stress, and

that the LaB6 can sometimes react with the support ﬁlament. To avoid

these difﬁculties LaB6 surfaces have been prepared using heated graphite

disks as a support [89]. Thorium oxide-coated ﬁlaments are often used as

electron emitters, for example in ion gauges, due to their extremely low

33

work function. These surfaces have also found application in the area of
NSI. A thorium oxide coating is relatively simple to prepare, and is
commonly coated on a support ﬁlament such as rhenium, tungsten, or
iridium [90]. Thorium oxide surfaces have high melting points, can

operate at lower temperatures than other surfaces, and are relatively inert
toward oxidative compounds. Cs glass surfaces, formed from a SiOz,

A1203, and CsNOz mixture, have also been used for NSI [91], These

surfaces can be prepared relatively easily, exhibit low work functions, and
are stable at high temperatures. Geiger investigated the emission of a
series of alkali metal containing surfaces (Li, Na, Rb, Cs) that were
exposed to a constant stream of halogen molecules (C12, Br2, or 12) [92]. Of
the systems investigated, the highest efﬁciency was obtained for the Cs
amalgam with 012.

The operating conditions of the NSI source can strongly inﬂuence
the overall efficiency of the ionization process. If a gaseous sample is
used, increasing the overall sample pressure increases the collisions of
molecules with the hot surface, thereby increasing the output of the
source. Increasing the operating temperature of the surface increases
both the degree of ionization on the surface and the degree of dissociation of
molecules from the surface. However, operating at reduced temperatures
has several advantages: 1) separation of negative ions from electrons
becomes unnecessary (electron emission increases exponentially above the
threshold temperature), 2) the production of negative ions through gas-
phase electron molecule interactions is reduced, 3) surface reactions
between residual gases and the ﬁlament surface are decreased. Some
compromise is often made in determining the surface temperature for
NSI, generally toward higher temperature operation. N SI surfaces are
most often maintained at temperatures above 1500 K [80].

The selection of a precursor molecule for NSI is also important.
Generally a sample should be selected that has a relatively small M-X

dissociation energy to promote dissociation on the hot surface. For

34

example Perskey, et. al., investigated the performance of a series of
chlorine compounds under NSI conditions [84]. The negative ion output
for the series was KCl > CCl4 > CGH5C1 > 02016 which follows the M-Cl
dissociation energies for the parent molecules. Increasing the area of the
heated NSI surface can also increase the output of the source, but the
increase in output current is not proportional to the surface area increase
because of difﬁculties in focusing ions from a large area. Most often a wire
diameter of 0.1 - 0.2 mm is used as a compromise between mechanical
strength, power requirements, and durability. Ribbon ﬁlaments are
occasionally used, but tend to twist at high temperatures, reducing the
surface area exposed to the source exit slit. Finally, a large extraction ﬁeld
at the surface will increase the overall ionization of an N SI source, this is
especially important at high output currents where the surface ionization
can become space-charge limited.

NSI sources can be used with quadrupole instruments, but certain
problems may arise. Under normal operating conditions (high surface
temperatures) both negative ions and electrons are emitted from the hot
surface. The electric ﬁelds used by the quadrupole for mass selection of
ions are not completely effective at discriminating against electrons. As a
result, an NSI source often produces a background (negative) signal at the
detector for all masses. This effect is especially pronounced when the
quadrupole is operating at lower masses. To reduce this background
signal, a weak magnetic ﬁeld can be applied to the ion/electron beam
between the source and the quadrupole. Another method used to
discriminate against electrons is the use of a strong extraction ﬁeld at the
heated surface, followed by a retarding lens between the source and the
quadrupole. One drawback to this method is that if the electrons exiting
source react to form positive ions in the region between the source and the
quadrupole, these ions can be detected as a mass-dependent background

signal at the detector. A better approach is to place the retarding lens

35

between the quadrupole and the detector, thus eliminating the mass-
dependent background signal.

A comprehensive review of surfaces, surface conditions, and
precursor molecules used for NSI has been made by Kawano [82]. A few
high emciency NSI conditions for the production of C1“ are given in Table
2.2.

Table 2.2 Some high efﬁciency conditions for the formation of negative ions

through negative surface ionization

Sample Surface W 9.121111111111109 Ian Reference
0014 ThOz-W 1650 - 2100 1740 01' Persky 1968 [84]
CCl4 La36 -Re 1550 - 1700 1600 C1“ Persky 1968 [84]
012 Cs 293 293 C12' Geiger 1955 [92]
CsCl Nb 1450 1450 C1‘ Kawano 1971 [93]
KC] Th02-W 1200 - 1500 ----- Cl‘ Trischka 1952 [94]

HCl Th02-Ir ----- 1300 01- Cheung 1979 [95]

36

We:

AJntmduetien

In order to obtain partial pressure measurements for a mixture of
residual gases in a vacuum chamber, the residual gas analyzer must
separate the collection of ions produced in the source according to some
property related to their m/z value. In a magnetic mass spectrometer the
ions exiting the source enter a strong magnetic ﬁeld and are dispersed
along different trajectories according to their momenta. Most modern
RGA’s utilize a different mass selection device called a radio frequency (rf)
quadrupole to perform this mass selection. An rf quadrupole has been
deﬁned as “a dynamic mass spectrometer whose ability to transmit desired
ions and reject unwanted ions is based upon the principle of path stability”
[96]. An rf-quadrupole RGA is simply a miniaturized quadrupole mass
spectrometer. A diagram showing the principal components of a rf-
quadrupole mass spectrometer is shown in Figure 2.5.

An rf-quadrupole consists of four symmetrically arranged
electrodes as shown in Figure 2.5. Opposing quadrupole rods are
connected, and a combination of dc and rf potentials are applied to the
rods. The ions produced in the source are accelerated to an energy of 3 to
50 eV and enter the quadrupole along the Z axis of the device. The
changing potentials applied to the quadrupole electrodes create a dynamic
hyperbolic ﬁeld within the device that effects ion stability in the X and Y
planes, but not in the Z direction. As a result, ions of a limited mass range
will remain stable as they traverse the quadrupole while the unstable ions
will impinge upon the electrodes, and be ﬁltered out. The stable ions are
detected by a detector positioned at the exit of the quadrupole. The
quadrupole functioning in this capacity acts like a variable bandpass mass
ﬁlter. By varying the levels of dc and rf applied to the opposing rods, the

quadrupole can electronically sweep the bandpass region of the mass

37

(+U +V cos (0t)

 

 

 

‘\

 

(-U + V sintot)
2

 

 

 

Figure 2.5 The basic components of a quadrupole mass spectrometer, both
an end view of the electrodes and a drawing of the overall device is shown.

38

ﬁlter, allowing ions of different m/z to hit the detector. Because the
quadrupole performs its mass selection through ﬁltering rather than
dispersion the quadrupole is considered a “mass ﬁlter” rather than a true
mass spectrometer.

At present, quadrupole-based mass spectrometers represent over
80% of the mass spectrometer market [97]. Quadrupoles have become
dominant in four main areas of research: residual gas analysis, elemental
analysis, aerospace research, and organic analysis. Quadrupole residual
gas analyzers have become the standard instrument for analysis of gas
mixtures and contaminants in vacuum and surface science, thin ﬁlms
research, semiconductor manufacture, and several areas of high vacuum
spectroscopy. The unique ﬁltering properties of quadrupoles are ideally
matched to important ionization techniques in elemental analysis such as
inductively coupled plasma mass spectrometry (ICP-MS). Quadrupoles
are also utilized in the aerospace ﬁeld for upper atmosphere analysis,
astronaut respiration monitoring, and space vehicle gas monitoring
where light weight and ruggedness of the device are important. The
majority of quadrupole mass spectrometers in operation today are used in
the area of organic analysis. Quadrupoles have become important in
many areas of organic analysis through coupling with separation
techniques such as gas chromatography, liquid chromatography,
capillary zone electrophoresis, supercritical ﬂuid chromatography, and
multi-sector mass spectrometers. The combination of the separative
properties of these techniques in tandem with mass spectrometry has
produced instruments capable of solving many difﬁcult analytical
problems. '

The proliferation of quadrupole mass spectrometers can be
attributed to several properties of the quadrupole mass ﬁlter. Because a
quadrupole uses only electric ﬁelds for mass selection, there is no need for
large, expensive magnets and their associated power supplies, which

greatly reduces the mechanical complexity and cost of a mass

39

spectrometer. The ﬁltering properties of a quadrupole are based on
applied voltages which can be varied more quickly than can magnetic
ﬁelds. This allows for the fast scanning of a quadrupole, which is
important in techniques such as GC-MS and selected ion monitoring. It
also increases the ﬂexibility and ease of controlling the mass spectrometer,
making them ideal for remote operation in residual gas analysis.
Quadrupoles ﬁlter ions according to their m/z (rather than momentum)
and therefore can accept a wider range of initial trajectories and energies
than magnetic instruments. Quadrupoles have moderately high
resolution (resolution of 3000 is fairly routine) and a relatively high mass
range (up to 106 has been obtained with current technology) [96]. Finally
the mass of a stable ion exiting the quadrupole is linearly proportional to
the applied dc and rf voltages making the interpretation of detector signal
fairly straightforward.

W

The ability of a quadrupole to ﬁlter out ions of different m/z value is
based upon the motion of a charged particle in a dynamic hyperbolic ﬁeld.
A hyperbolic electric ﬁeld is the electric ﬁeld created within an array of
four hyperbolic electrodes that are arranged as shown in Figure 2.6. If the
opposing pairs of electrodes are connected and each connected pair is
biased at opposite potentials a hyperbolic electric ﬁeld will be created
between the electrodes. Figure 2.6 shows a SIMION model of the electric
ﬁeld produced when the upper and lower electrodes are biased at -100 V
and the left and right electrodes are biased at +100 V. The lines between
these electrodes map out the points of equal potential for the free space
region within the electrode array.

The force that a charged particle experiences is directed
perpendicular to these equipotential lines. A positive ion that enters this

ﬁeld will experience an attractive force toward the upper and lower

4O

325.5020 2095.635 cams vase.“ can
as: £3593 can? couscous. Emu 2.5035 23 mo 25on ZOHEHman wd 0.5me

 

41

electrodes and a repulsive force from the left and right electrodes. The
result is that the ion is strongly defocused (exponential defocusing) along
the attractive axis and experiences harmonic motion along the repulsive
axis (focusing). If the potentials on the rods are suddenly switched the
eﬁ'ect reverses. If these potentials are alternated at the correct frequency,
an overall focusing effect (toward the center of the array) will result from
the alternating squeezing and relaxing forces on the ion. This focusing
effect is called “strong focusing” [98]. The net restoring force on an ion is
proportional to its displacement from center axis and can be achieved
through the application of a sine wave ac signal to one rod pair and a
cosine wave ac signal to the other rod pair. It is this effect that makes rf-
only quadrupoles (and high order multipole devices) useful as beam
guides. As an ion travels along the Z axis of the quadrupole (see Figure
2.5) it will experience an alternating hyperbolic ﬁeld in the X and Y axes
which produce an overall focusing effect. A device based on this effect was
originally conceived by W. Paul in 1958 [48].

The presence of alternating potentials on the quadrupole rods
results in the focusing of a broad range of different m/z ions. If dc
potentials of opposite sign are applied to the opposing rod pairs at a ﬁxed
ratio with respect to the applied ac potentials, the range of stable m/z ions
will be reduced (narrowing the ‘bandpass’ of the mass ﬁlter). By
convention the y axis rod pair is deﬁned as the pair to which the negative
dc potential is applied and the x axis rod pair receives the positive
potential. This mode of operation is the basis for the use of a quadrupole as
a mass spectrometer.

The use of a quadrupole in this mode is more easily understood in
terms of the equations of motion for a particle in a quadrupole. The
potential ((1)) at any point within the free space region inside a quadrupole

can be described as:
x2y2
(I) = [U + Vcos(0)t)l

2
2r0

 

42

where: x = distance from center along x axis
y = distance from center along y axis
r0 = distance from quad centerline to rod as shown in Figure 2.5

(r) = angular frequency (27tf) of the applied ac wave
V = magnitude of applied ac wave
U = magnitude of applied dc potential

The change in potential with distance is deﬁned as the electric ﬁeld
(E) present within a region of space. By taking the partial derivative of the
potential equation above with respect to movement along each of the axes of
a quadrupole, one can determine the electric ﬁeld that the quadrupole
produces at that position. This relationship for displacement in the x(Ex),
y(Ey) and z(Ez) axes is given as:

ch ‘1’ y
EXT-"85;: -[U+Vcos((1)t)]rlg Ey=-%)7= ~[U+ VCOS((Dt)]:g

(I)
Ez='§6;= O

The sum of the electric ﬁeld components for displacement along
each axis yields the net electric ﬁeld at that position. These equations
show that the quadrupole produces a position-dependent electric ﬁeld
along both the x and y axes but not along the z axis. Therefore the ion
motion along the z axis (and consequently the ion residence time in the
quad) is independent of the potentials applied to the quadrupole rods. The
force that an ion experiences at any position within an electric ﬁeld is
simply the product of the electric ﬁeld and the charge on the ion (e). As a
result the force on an ion in the x(Fx) and y(Fy) directions can be written

as:

sz - [U + Vcos(cot)]e—§ Fy= - [U + Vcos(wt)]9§’
r0 r0

43

The acceleration that a particle experiences in any direction due to
an applied force is simply the force divided by the mass of the particle.
Therefore, the acceleration of an ion as a result of the force induced by an
applied electric ﬁeld in the x(ax) or y(ay) direction is given as:

’12]

a _ 93 _ 5c. a _ 9.21 _ _r
" 3% m , 52t m
Applying this relationship to the equations for Fx and Fy in a
hyperbolic ﬁeld yields the following equations describing the acceleration
an ion undergoes in terms of the applied potentials, the mass, charge, and

position of the ion, and dimensions of a quadrupole.

2 2
d_2x_ = -[U + Vcos(wt)]ﬂ-§ _d_2y_ = '[U + Vcos(03t)l-El—2
d t mro d t mro

These equations are forms of the well known Mathieu equation
which was originally derived as a description of vibrational motion of a
stretched membrane within an elliptical boundary [99]. The trajectory
solutions to these equations for an ion traveling along the Z axis of a
quadrupole can be deﬁned as bound or unbound. A bound solution
corresponds to an ion that adopts a stable periodic trajectory while
traveling through the quadrupole. An unbound solution corresponds to
ions that cannot adopt a stable periodic motion while traversing the
quadrupole. Ions that adopt unbound trajectories will follow a path that
exceeds the dimensions of the device (re) and impinge upon one of the
quadrupole rods.

The ability of an ion to adopt a bound trajectory and successfully
traverse the quadrupole depends on the values of the six terms in the

equations above (including mass/charge ratio). Through extended

44

analysis it has been determined that the solutions to these equations can be

described in terms of the two variables a and 51 deﬁned below [100].

_ 4eU 2eV

—(02r§m (1:03—ng

a

The actual trajectory taken by an ion as it travels through a
quadrupole is extremely complex and depends upon the initial position,
trajectory, and the ac phase of entry. To simplify the description of mass
scanning with a quadrupole a plot of a verses g is often made. Such a plot
can be used to describe the regions of x and y stability for a quadrupole with
respect to the applied dc potential (U), the applied ac potential (V), the
angular frequency of the ac waveform ((0), the charge on the ion (e), and
the mass of the ion (m). There are several regions of an a,g diagram in
which bound solutions exist for both x and y stability, the region most
commonly used for operating a quadrupole is shown in Figure 2.7.

The stability region indicated in Figure 2.7 describes a series of a
and g values for which ions can attain stable trajectories (bound solutions)
in both the x and y directions. This stability region is enclosed on the left by
the x stability boundary and on the right by the y stability boundary. If the
quadrupole is operated with dc and rf potentials in ratio corresponding to
“scan line” shown in Figure 2.7, only ions of a limited mass range will
remain stable within the quadrupole. Ions outside the stability region will
follow unstable trajectories, impinge upon the quadrupole rods, and be
ﬁltered out. The slope of the mass scan line, given as 2U/V, is adjusted to
create a small overlap with the a,q stability region. In this manner the
sharp tip of the a,g region can be used as a narrow bandpass ﬁlter. By
increasing the rf and dc levels at a constant ratio, ions of higher masses
will be shifted upward along the mass scan line and enter the narrow

stability region thus producing a mass scan.

45

   
  
 

X Stability
Boundary

/

 

Stability
Region

     
 

Y Stability
Boundary

    

  

 

 

0'2 0.4 0.6 0.3
_ 2eV

(ozrgm

 

Figure 2.7 The an space stability diagram for a quadrupole mass ﬁlter.

The slope of the mass scan line determines the range of ion masses,
or Ag, that are simultaneously stable within the quadrupole. By adjusting
the slope of the mass scan line (the dc/rf level) it is possible to change this
range, thereby altering the resolution of the mass spectrometer. Thus,
within limits, increasing the resolution of a quadrupole is simply a matter
of increasing the dc/rf ratio present on the quadrupole rods.

A quadrupole can be operated in several different modes that are
distinguished by the ratio of applied dc to rf (dc/rf ratio). The two most
common mass-selective modes of operation are constant sensitivity and
constant peak width(AM) modes. If a quad is operated with a constant
ratio of dc to rf then the m/Am for the stable ions will remain constant.
This mode of operation is commonly called constant sensitivity mode. In

46

constant sensitivity mode the quadrupole will have approximately the
same sensitivity for ions of all masses. The trade off is that as the mass of
the stable ion (m) increases, Am also increases (since UN is constant).
This causes the peak width of the quadrupole operated in this mode to
increase with increasing ion mass.

If a small ﬁxed dc potential is applied to each rod pair, it is possible
to adjust the intercept of the mass scan line such that Am will remain
approximately constant. Operation of a quadrupole in this mode has the
advantage that resolution remains fairly constant throughout the range of
the mass spectrometer. The disadvantage to this mode is that as the ion
m/z increases the sensitivity of the quadrupole decreases. The selection of
an operating mode for a quadrupole depends upon the requirements of the
application.

Another important mode of operation for RGA quadrupoles is rf-
only mode. If the dc levels applied to the quadrupole are decreased, the
slope of the mass scan line decreases and the bandpass of the mass ﬁlter
increases. If no do is present, the slope of the mass scan line will be zero
which corresponds to operation along the q axis of the an stability
diagram. Operation in this mode allows a broad range of ions to obtain
stable trajectories as they pass through the quadrupole. This mode has
proven useful for total pressure measurements in residual gas analysis,
total ion counting in GC-MS, the application of a quadrupole as a beam
guide in multi-sector instruments, and for the evaluation of source

efﬁciency.

CE'JQI'EQII'

The ultimate performance of a radio frequency quadrupole is
commonly described in terms of the highest mass it can select and the
maximum resolution that it can obtain. The maximum m/z ion that can

47

be selected by a quadrupole (Mm) can be described in terms of the ﬁeld
radius (r0), rf frequency (f), and maximum rf voltage between rod pairs
(Vm) as:

71:106vm
m:—
1?er

A similar relationship exists between the minimum peak width
(AMmin) and the ion injection energy ( Vz), the rf frequency (f), and the

length of the quadrupole rods (L) in meters.

4xro9vz
AMmin=_2_2_
f L

If these two equations are combined and the operating conditions
are optimized for both Mm and AMmin the following relationship is

obtained.

AMmin- rEV,570
Mm szm

 

Several important observations can be drawn from this relationship:

1. The ultimate performance of the quadrupole depends upon the ratio
of ro/L.

2. In theory ‘any’ mass range can be obtained with a quadrupole by
simply lowering the rf frequency, but only by sacriﬁcing resolution.

3. Reducing r0 will increase mass range, but will also decrease the ion
acceptance of the quadrupole and make source design and
contamination more critical.

4. Increasing L improves the performance of the quadrupole, but is
ultimately limited by practical construction limitations.

5. Finally, Vm is limited by the levels of rf voltage that can be produced.

48

In the design of most modern quadrupoles some compromise has
been made to allow; practical manufacture of the quadrupole electrodes,
reasonable mass range and resolution, reproducible source alignment
matching, and production of rf levels within the scope of modern
electronics. An example of typical dimensions and operating conditions

for a radio frequency quadrupole are:

Vz=5eV L=0.2m f=2MHz ro=0.4cm

These conditions can produce a quadrupole with AMmin = 0.125 daltons
and Mm of approximately Vm/9 (i.e. for rf Vp-p = 1000 Mm = 222).

The spectra produced by a quadrupole are strongly dependent not
only on the value of these parameters selected in the design of instrument,
but also on the precision of the rod manufacture and mounting, alignment
of the source with the quadrupole, the size of source and exit apertures, the
energy of the ions, and the scan rate of the quadrupole. Each of these
parameters can affect not only the shape of the peaks in the mass
spectrum produced by a quadrupole, but also their relative intensities and
the ultimate resolution of the system.

The mechanical construction of a quadrupole is extremely
important in its performance as a mass spectrometer. Most quadrupoles
are constructed using round stainless steel or molybdenum electrodes.
Although a hyperbolic rod shape is ideal, round rods are commonly used
because they are much simpler to manufacture. If round rods are
mounted so that their radius (r) is related to their spacing from the central
quad axis (r0) by the relationship r/r0= 1.148 then the ﬁeld between the rods
will approximate a hyperbolic ﬁeld to a very high degree. This effect can be
seen in Figure 2.6, in which the electric ﬁelds produced within hyperbolic
and round rods are compared using the modeling program MacSIMION.

It is extremely important that the rods of a quadrupole be in precise
alignment. If the rods lie at unequal distances from the center axis of the

49

quadrupole imperfect ﬁelds are produced within the device that can
produce higher order restoring forces on an ion. These forces can cause
peak splitting as the restoring force an ion experiences becomes non-
linear, these ﬁeld imperfections are called resonance instabilities. Field
imperfections become increasingly important as an ion is displaced far
from the z axis. This effect is often observed as mass discrimination
against higher m/z ions (note: mass discrimination in a quadrupole is
also due to the increase in number of rf cycles that an ion experiences with
increasing m/z). For example, a 1% error in the positioning of the
quadrupole rods would ultimately limit the resolution of the quad to 200
[97]. To insure excellent rod alignment the quadrupole rods are commonly
mounted in precision ground ceramic supports. Using this mounting
system the rods can be kept in alignment while the entire unit is removed
for cleaning. Quadrupoles have also been constructed with each rod
mounted individually using alumina or sapphire insulators. The
advantage to this system is that individual rods can be removed and
thoroughly cleaned. The disadvantage is that alignment jigs must be used
to ensure that the rods remain in the correct position.

Field imperfections that result in peak splitting can also originate
from sources other than rod alignment. If the quadrupole rods become
contaminated then surface charging can occur on the rods which results
in ﬁeld imperfections. Thus an increase in peak splitting with time is a
good indication that the quadrupole rods are contaminated and need to be
cleaned. Another source of ﬁeld imperfections is improper balance of the
potentials applied to the rods. This imbalance can result from improper rf
balancing or do level control which is discussed in the next section.

As the resolution of a quadrupole is increased the trajectories of
stable ions come closer to the rods of the quadrupole. Therefore at higher
resolution settings ions must originate closer to the z axis for effective
transmission. The maximum entrance aperture (D) for injected ions can

be described as:

50

D = ro(M/AM )”2

Thus by decreasing the entrance aperture it is possible to increase
the ultimate resolution of the system. The disadvantage in doing this is
that sensitivity also decreases with aperture size due to the decrease in ion
ﬂux entering the quadrupole. It is common to manufacture a quadrupole
with large apertures and compensate for this mass discrimination effect
by increasing the sensitivity of the detector, increasing the ampliﬁcation of
the detector signal, as both resolution and mass range are increased.

The alignment of the ion source with the quadrupole rods can also
inﬂuence the peak shape and sensitivity of the mass spectrometer. Ions
that originate far from the z axis of the quadrupole can follow trajectories
that will cause them to hit the electrodes even if they should theoretically
have stable (bound) trajectories. The range of initial positions and
trajectories that ions can have and still remain stable within the
quadrupole is called the acceptance of the quadrupole. If the ion source is
not in proper alignment with the quadrupole, a large number of ions will
enter the quadrupole outside of the acceptance of the quadrupole and follow
unstable trajectories. This results in peak splitting or multiple peak shape
and signiﬁcant loss of sensitivity [101].

There also exists a range of initial ion positions that can result in
poor peak shape in the mass spectrum. If an ion originates exactly along
the axes between the quadrupole rods the net force that it will experience
as it traverses the quadrupole is approximately zero. This can be seen in
Figure 2.6., the straight equipotential lines that lie along the axes between
the electrodes correspond to an electric potential of zero because the
electric ﬁeld produced by each rod pair exactly cancels. Ions that originate
at these initial positions are not properly selected by the quadrupole and
can cause peak tailing on the low and high side of peaks in the mass

spectrum [102]. To eliminate this problem cross wires are often attached

51

to the lens element directly before the quadrupole along the axis between
rods Ions that would normally enter the quadrupole from these positions
impinge upon the wires and are neutralized.

The energy of an ion along the z axis, or injection energy, inﬂuences
the sensitivity and peak shape produced by the quadrupole. For an ion to
be properly mass selected by the quadrupole it must experience a
minimum number of rf cycles. The number of rf cycles (n) necessary to

achieve some desired resolution (m/Am) is given as:

1/2
11 2 3.6 (3%)

The ion current through the quadrupole will increase as the ion
injection energy is increased, but the overall resolution of the system will
decrease. If the ions experience too few rf cycles, the mass peaks will split
and tend to tail on the low mass side of the peak.

Another consideration in the operation of a quadrupole is the rate at
which the do and rf magnitudes are varied, or the scan rate. If a
quadrupole is scanned at too high a rate, an ion that enters under stable
conditions will not remain stable during the time it spends in the
quadrupole. Thus if too high a scan rate is used the effect will be that no
ion can remain stable as it passes through the quadrupole. The practical
limit for scan rate is approximately 1000 amu/second for normal ion
injection energies [97].

The quadrupoles used in residual gas analyzers are designed to
achieve some compromise between the performance factors discussed
above and the size and portability required for remote operation and
installation. A typical RGA quadrupole consists of four round stainless
steel or molybdenum rods approximately 50 to 80 mm in length with a
diameter of 6 to 7 mm. These rods are generally ﬁxed within a precision

ground ceramic housing to allow their removal for cleaning and to ensure

52

precise alignment. The rf voltage used to operate the quadrupole ranges
from 250 to 1000 volts peak to peak at a frequency of approximately 2 MHz.
Injection energy can range from 5 to 10 eV and is often increased with
increasing mass to overcome any mass discrimination affects in the
quadrupole. For an RGA quadrupole that uses 80 mm rods with a 7 mm
radius, operates with a 5 eV ion injection energy, and has a power supply
that produces 250 Vpp at 2 MHz the ultimate performance of the system
will be Mmax = 94 and AMmin = 0.781. This produces an overall

performance of M/Am = 120.

DB'C I'Bl'E Q'i

The proper functioning of a quadrupole is dependent not only on the
physical dimensions and alignment of the quadrupole rods, but also on the
appearance of the proper electrical signals at the quadrupole electrodes.
As discussed previously, an rf quadrupole requires a mixture of do and
large amplitude rf potentials for operation. This section will discuss
generation of these signals and the circuitry required to perform scanning
functions on the quadrupole.

A quadrupole functioning in mass-selective mode requires both do
and rf signals to be applied at some controlled ratio. The rf component has
to be produced at high peak to peak levels to for the quadrupole to operate
through a high mass range, as discussed in the previous section. This
signal should have a pure sinusoidal waveform, excellent stability, and be
rapidly scannable. The dc component must be applied at a precise ratio to
the applied rf Vpp. The principal components necessary for the production
of these signals including capacitors, inductors, and transformers are
shown in Figure 2.8. A brief description of these components and their

application to quadrupole electronics follows.

53

 

 

  

 

 

h- r ~i _'
Capacitor Inductor

 

 

”AA“

 

 

 

LC Tank

 

Transformer

 

Basic quadrupole rf circuit

Figure 2.8 The electrical components of the radio frequency generation
circuit for a quadrupole mass spectrometer.

 

54

If a potential is applied across two conductors separated by an
insulator a capacitor: is formed. If the potential is removed and the two
conductors are isolated from each other then some charge (Q) will be
stored across the two conductors. The capacitance (C) of such a system is
deﬁned as the charge (Q) stored across the capacitor divided by the
potential (V) applied across the capacitor or:

C=Q/V

The unit for capacitance is the farad (F) which is deﬁned as one
coulomb/volt, a 1 farad capacitor will store 1 coulomb of charge when a 1
volt signal is applied across the two plates.

A symbolic diagram of a capacitor is shown in Figure 2.8. If a dc
potential is applied to the capacitor, current will ﬂow into the capacitor
until it is charged, after which the current ﬂow stops. If the two sides of a
capacitor are connected current will ﬂow between the two conductors until
the capacitor is discharged. Therefore when a dc signal is applied to a
capacitor the only time that current ﬂows is during instances of charging
and discharging.

When an alternating signal is applied across the plates of a
capacitor, the changing potential results in a constant ﬂow of current
through the circuit. The capacitor requires some time to charge to a given
potential. The reactance of a capacitor opposing a change in potential as
result of this charging time constant is called its capacitance reactance.
Since current is simply the rate of charge ﬂow, the current ﬂow across a
capacitor is proportional to the rate of change, the frequency (f ), of the ac
signal. Thus capacitance reactance (Xe) is given as:

Xc = l/(ZIIfC)

55

As the frequency of the applied ac signal increases, the capacitance
reactance of the capacitor decreases. The overall result of the application
of an ac sine wave signal across a capacitor is an induced cosine current
waveform, thus in an ac capacitor circuit, current and voltage are 90° out
of phase.

A capacitor is simply a device designed for its ability to hold charge.
It consists of two conductors separated by an insulator. The capacitance of
the device depends upon the capacitor area, spacing, and the dielectric
constant of the insulating material. Increasing the surface area and
decreasing the separation of the conducting plates increases the amount of
charge that can be stored across the device thus increasing its
capacitance. The dielectric constant of the insulating material also effects
the capacitance of a capacitor. Increasing the dielectric constant of the
insulator allows it to absorb more energy from the electric ﬁeld between the
two plates, thereby increasing it’s capacitance.

A device speciﬁcally designed for its ability to resist change in
current through a circuit is called an Inductor. Inductors are commonly
closely wound coils as illustrated in Figure 2.8. If an ac signal is applied
across a wire coil the changing current produces a varying magnetic ﬁeld
also shown in Figure 2.8. A change in current through the coil causes the
magnetic ﬁeld lines to expand or contract. This in turn induces a voltage
within the coil to counter the ﬂow of current. Through this action the
inductor resists change in current through the coil.

The inductance of a coil (L) can be described in terms of the coil
diameter (r) in inches, number of turns (n), and coil length (l) in inches by

the following expression:

dzn2
18d + 401

 

L(uH) =

56

The unit for inductance is the Henry, one Henry counters the emf of
1 volt when the current through the circuit is changing at a rate of 1
amp/second. The reactance of an inductor (resistance of an inductor to

change in voltage) is given as:
XL = 21th

If two coils are placed close to each other, a changing magnetic ﬁeld
produced by one coil can induce a current ﬂow in the second coil. Under
these conditions these coils are said to have mutual inductance, or mutual
magnet coupling. This effect is the basis for a device called a transformer
which is shown in Figure 2.8. The coil connected to the power supply of a
transformer is called the primary coil. An ac signal applied to the
primary coil produces an ac signal in the secondary coil. If power is
removed from the secondary coil a magnetic ﬁeld change that opposes the
primary coil magnetic ﬁeld is created. The primary coil must then draw
additional current to counteract this opposing magnetic ﬁeld. Through
this interaction ac power is transferred between the two systems without
physical connection.

If the number of turns on the primary coil differs from that on the
secondary coil then the ac voltage levels produced at the two coils will be
different. The voltage that appears across the secondary coil is
proportional to the number of turns in the coil. By selecting the
appropriate turn ratio between the primary and secondary coils a primary
ac signal can be increased or decreased to a selected level. The change in
ac level through a transformer can be determined from the following

expression:

E3 = Ep(ns/np)
where:

E3 = secondary coil EMF ns = number of turns on the secondary coil
Ep = primary coil EMF up = number of turns on the primary coil

57

Transformers are commonly constructed by winding wire around
an iron core (commonly torroidal) to increase the inductance between the
two coils. Most audio frequency and lower frequency transformers use
this design because of its superior magnetic coupling. Hysteresis and eddy
currents within the iron core reduced the performance of this design at
higher frequencies. Transformers for radio frequency signals generally
use open coils with no core to eliminate problems with hysteresis and eddy
currents.

If an inductor and capacitor are connected in series a circuit called
an LCJank is created. An LC tank is deﬁned as “a circuit consisting of
inductance and capacitance, capable of storing electrical energy over a
band of frequencies continuously distributed about a single frequency at
which the circuit is said to be resonant, or tuned” [103]. A schematic
drawing of an LC tank is shown in Figure 2.8. Resonance exists within an
LC tank through the combination of the reactances of the inductor and the
capacitor. The frequency of resonance (f) for such a circuit can be
described as:

f: 106
2mm

where: f = frequency in kilohertz (kHz)
L = inductance in microhenrys (11H)
C = capacitance in picofarads (pF)

If an ac signal of frequency (f) is applied to the system, power can be
absorbed by LC tank and will be stored through resonance between the
inductor and capacitor. The amount of current that ﬂows through the LC
tank depends upon the matching of the reactances of the inductor and
capacitor. If these reactances match exactly the current through the
inductor will be completely cancelled by an out-of-phase current through
the capacitor. As a result a very high ac level will resonate through the LC
tank. The ratio of the reactive voltage to the applied voltage is deﬁned as

58

the Q of the circuit. Therefore a high_Q circuit produces a high ac
resonant signal within the LC tank for a relatively low applied ac power.

Radio frequency circuits often consist of a radio frequency power
supply, rf transformer, and LC tank with the secondaries of the
transformer forming the inductor of the LC tank [98]. A block diagram of
a circuit used for the generation of rf levels on a quadrupole is shown in
Figure 2.8. An rf signal from the rf power supply causes the production of
a varying magnetic ﬁeld within the transformer. This changing magnetic
ﬁeld induces an rf signal within the LC tank. If the resonant frequency of
the LC tank W the frequency of the applied rf, the LC tank
will absorb the rf power. This produces a resonant rf waveform across the
capacitor formed by the quadrupole rod pairs. By selecting an rf
transformer with a large nS/np ratio and balanced inductive and capacitive
reactance, a high Q LC tank can be produced which will generate high rf
levels with low rf power consumption.

If the frequency of the rf power supply does not closely match the
resonant frequency of the LC tank, the LC tank will be unable to absorb
power from the primary rf signal. As a result the rf signal will be
reﬂected back from the transformer into the rf power supply, overloading
the electronics and possibly damaging the unit. Proper rf power
absorption can be accomplished by varying the frequency of the rf power
supply to match the resonance of the LC tank. Resonance can also be
matched by tuning a variable capacitor, installed between the quadrupole
rod pairs, that can vary the LC tank resonant frequency to match the rf
power supply frequency.

Won:
For proper mass-selective operation the rf quadrupole requires a
combination of rf and dc potentials across the two rod pairs. The electronic
circuits used to produce these signals are generally a combination of the rf

circuit described above and a dc production circuit coupled to the rf output.

59

For a quadrupole to closely produce a hyperbolic ﬁeld, the signal that

appears across the quadrupole rod pairs must posses several qualities:

1. A pure sinusoidal waveform must appear on one rod pair and a
pure cosine waveform (900 out of phase) on the other rod pair.
This rf signal must have excellent frequency stability and large
peak to peak amplitude.

A dc signal must appear on each rod pair that is accurately
radioed to the applied rf level.

The combination of applied rf and dc signals must rapidly and
reproducibly variable between differing levels.

PS9.”

The principal components of the rf generation portion of a
quadrupole circuit are shown at the bottom of Figure 2.8. The rf signal is
generated in the rf power supply shown at the far left. This signal travels
through the primaries of an rf transformer and produces a varying
magnetic ﬁeld. Rf power from this signal is received by the LC tank
consisting of the secondaries of the rf transformer and the capacitor
formed between the two rod pairs. An rf transformer is utilized for two
reasons: 1. It allows the amplitude of the rf signal produced by the rf power
supply to be increased by selecting a transformer with the required turns
ratio, and 2. By separating the LC tank from the rf power supply the dc
levels can be directly applied without specialized ﬁltering. The center of
the secondary coil of the transformer is center tapped, or connected to a
reference point. This results in the production of rf signals across the two
rod pairs that are 90° out of phase. Proper rf power absorption between the
transmitter circuit and the LC tank is maintained by varying the
transmitter frequency or changing the LC tank resonant frequency as
described above.

The dc levels are produced through rectiﬁcation of the rf signals on
each of the rod pairs. The rf signal that appears on each rod pair is fed to a
rectiﬁer (commonly a series of diodes and capacitors) that produces a dc

level proportional to the rf input. The two do signals are then fed into op-

60

amps that amplify them to the required levels. These do signals are then
fed back to the quadrupole.

To maintain constant rf levels a feedback 100p is used between the
rectiﬁcation circuit and the rf power supply. The dc outputs from the
rectiﬁer are fed through another set of op-amps back to the rf power
supply. The outputs are used to regulate both the rf amplitude on each rod
pair and the balance of rf levels between them. The rf/dc ratio and rf
balance can then be varied by changing the ampliﬁcation of the op-amps
between the rectiﬁcation circuit and the rf power supply.

The quadrupole power supply circuit is scanned by introducing an
offset signal to the reference side of the op-amps fed by the rectiﬁer. This
offset causes the op-amp output to change, which in turn induces a change
in the output of the rf power supply. The result is the production of new
rf/dc levels on the rod pairs at a controlled ratio. The offset signal can be
varied manually by varying a sensitive potentiometer, or electronically
through the production of an offset ramp (commonly 0-10 Volts) which is
often produced by the controller or by some external source. As this dc
ramp increases from 0 volts the rf and dc levels on the rod pairs will
increase, bringing different mass ions into stability in the quadrupole.
Thus a mass spectrum is produced by plotting detector signal verses offset
voltage. For quadrupole operation in constant Am mode (discussed in the
quadrupole theory section) a small opposite dc bias is required on each rod
pair. This is produced by applying a separate, constant bias signal to op-

amp reference line.

61

IIII QI'lSI D'

EB'C I'IDI'

One important aspect of the changes made to the RGAs was the
modiﬁcation of ion optical elements. Ion optical systems, or electrostatic
optical components, are a series of elements that are designed to
manipulate charged particles through the application of electric or
magnetic ﬁelds. Electrostatic lens systems employ a series of electrodes
that are arranged and biased to produce controlled electric ﬁelds. Charged
particles that are formed in, or travel through these ﬁelds will follow
trajectories in response to the forces induced by these electric ﬁelds.
Properly designed electrostatic lens systems have been used to induce
beam focusing, trapping, acceleration, deceleration, and are the primary
devices for manipulating charged particles within a mass spectrometer.
This section will discuss some basic concepts used in the design of
electrostatic optical systems, some ion optical modeling systems that can
be used to optimize these designs, and a specialized method of detection
that can be used to evaluate the ion beams produced by the systems.

If an ion is formed in a region that contains no electric ﬁelds ( a ﬁeld
free region) the path that the ion follows will be dependent only upon the
kinematics of its formation. If, however this ion passes through a change
in electric potential it will experience an acceleration due to the force
induced by that change in electric ﬁeld. One important aspect of this
process is that the trajectory that an ion follows as a result of applied
electric ﬁelds is not mass dependent. That is, different m/z ions that are
formed with the same initial energy will follow the same trajectories to
some ﬁnal point, but will arrive at different times. This is an important
property of ion optical systems for mass spectrometry, it allows the design
of ion optical systems systems that can be used regardless of the m/z of the
ion being manipulated.

62

Electrostatic lenses are a series of elements used to focus or defocus
a beam of ions within a mass spectrometer in a fashion similar to light
optics. To focus a beam of ions the lens must apply some force to the ion
beam in a direction radial to the propagation direction of the ion beam.
Strong radial focusing can be induced through the use of static multipole
lenses, but this type of lens only induces focusing in one plane. Multipole
lenses can be used in series to focus ions in all planes, as is often done in
high energy accelerators, the advantage of this type of systems is that it
doesn’t require high potentials on the elements. This type of system is too
complex for use in most mass spectrometers, instead mass spectrometer
lens systems are generally based upon the use of cylinder or disk lenses.
These lenses can induce the symmetric focusing of an ion beam, but
require the used of high potentials relative to the ion beam energy. This is
often not a limitation in mass spectrometry, particularly in quadrupole
instruments where the beam energy is generally on the order of 10 eV.

The simplest type of cylindrical or disk lens is the two element lens.
If an ion beam travels between two regions of differing electric potential, it

will experience a force not only in the transverse direction, but also in the

radial direction. An example of such a system is shown in Figure 2.9.

A F'l’] B

151' Vi

   

Figure 2.9 Ion refraction during transition between regions of differing
potential

If an ion, A+, is traveling from a ﬁeld free region in electrode A to

another another region of differing potential , electrode B, its transverse

63

velocity will change from Vi to Vf. This change in transverse velocity will

be accompanied by a change in radial velocity causing a change in the

trajectory of the ion. The initial angle of the ion relative to the transverse
axis (9i) is related to the ﬁnal angle of the ion (6f) by the equation:

Vi Sinei = Vf Sinef

The overall effect of using a two element electrostatic lens can be strong
focusing for an ion beam, but in order to effectively focus a large beam into
a small region the two element lens must accelerate the ion beam, and
utilize large apertures. Ion beam focusing with a two element lens is
generally accompanied by strong angular divergence, limiting its
applicability.

One of the most common lens designs in mass spectrometry is the
Einzel lens. The Einzel lens is a three element lens, in which the ﬁrst and
third lenses are held at the same potential and a separate bias is applied to
the central element. This lens type, often called a unipotential lens, can be
used to focus or defocus an ion beam without changing its energy. A block

diagram of a unipotential lens is shown in Figure 2.10

fl

Ions >. Ions
l 1

Figure 2.10 Block diagram illustrating a unipotential lens

 

 

 

 

64

In general any electrostatic lens that doesn’t change the energy of
an ion beam, such as an Einzel lens, will inducing focusing in the beam.
Einzel lenses can be operated in two modes, decelerate-accelerate and
accelerate-decelerate. If the central element is held at a potential that will
cause the ion to decelerate as it enters this element, the lens is said to be
operating in decelerate-accelerate mode. Ions entering the lens will slow
down and feel a defocusing effect in the region between between the ﬁrst
and second elements. As a result the ion beam will spread out in the lens.
As the ion beam enters the region between the second and third elements
is will be accelerated and experience a focusing radial force. The overall
effect on the ion beam is focusing due to the increased residence time of the
ions beam in the focusing ﬁeld. The advantage of the decelerate-accelerate
mode of operation is that the central element need only have a potential
approximately the same as the beam energy to induce focusing. When the
central element of an Einzel lens is held at a potential that causes the ion
beam to accelerate as it enters the center element, the lens is said to be
operating in accelerate-decelerate mode. In this mode the ion beam is ﬁrst
focused by the transition between elements 1 and 2, and then defocused
between elements 2 and 3. The advantage to using an einzel lens in
accelerate-decelerate mode is that the ion beam doesn’t spread out as
much within the lens, and therefore the focusing properties of the lens are
much less sensitive to ﬂaws in the mechanical structure/alignment of the
lens elements. The disadvantage to this mode of operation is that the
central element must be maintained at a potential an order of magnitude
greater than the beam energy. This can be important when working with
the high energy ion (~2 kV) beams produced by magnetic instruments.
Because of the low beam energies used with quadrupole instruments,
quadrupole instrument einzel lenses are generally operated in accelerate-
decelerate mode [98].

The design of einzel lenses for use in mass spectrometers has been
investigated by several research groups [104,105]. The focusing action of

65

an einzel lens operated in accelerate-decelerate mode has been found to be
approximately linear with increasing beam diameter. The spacing
between the three electrodes was determined to be the most critical factor
in determining the focal properties of an einzel lens. Also important are
the diameter and thickness of the central element. The focal properties of
an einzel lens have been found to be relatively insensitive to the ratio of the
diameters of the central and end electrodes. The optimum relative
dimensions determined for the design of an einzel lens are shown in
Figure 2.11. The lens is designed with the ﬁrst and third elements having
a common aperture size (b), the spacing between elements and the
aperture of the central element should be approximately 3b and 2d
respectively [106].

 

 

 

 

Figure 2.11 The optimal geometry for a three element Einzel lens.

The majority of the lenses that were used to manipulate the ion
beams produced by the modiﬁed RGAs are based upon this Einzel lens
design. The optimized geometry displayed in Figure 2.11 above was

66

reproduced as closely as possible with available ceramic insulators and
lens materials.

The performance of an ion optical system can be evaluated in a
fashion similar to light optics, in terms of its focal length, magniﬁcation,
etc. One important parameter of an optical systems is the degree of
aberration, or distortion, that a lens induces in an ion beam. Electrostatic
optic systems are notorious for the high degree of aberration they can
cause. These aberrations are generally separated into four categories,
spherical aberration, chromatic aberration, ellipticity aberration, and
deﬂection distortion.

Spherical aberrations in ion optical systems are caused by the fact
that the focusing electric ﬁelds produced by a series of electrodes are
stronger near the electrodes and become weaker toward the center of the
lens. This causes ions near the electrodes to be more strongly focused than
ions near the center of the lens, resulting in a lens that doesn’t have a
single deﬁned focal point. Instead, ion lenses generally produce a region
of strongest average focal point (minimal beam diameter) called the disk
of least confusion. Chromatic aberration, or momentum dispersion, is
caused by the sensitivity of the focal properties of an ion optical lens to the
energy of the ions being focused. If a series of ions with some energy
spread travel through an electrostatic lens the higher energy ions do not
spend as much time within the lens. As a result they experience less of
the focusing ﬁeld in the lens and are more weakly focused than ions with
lower velocities. This causes a spread of focal positions in the ion beam
that is dependent upon the spread of initial energies. This aberration can
be reduced by using an ion beam with minimal energy spread, and
carefully referencing the potentials on the lens to the ion beam energy
(source potentials). Elliptical astigmatism is an aberration that results
ﬁ'om the use of asymmetric (non-circular) lens elements, or elements that
are tilted with respect to the beam optical axis. These conditions result in

the creation of asymmetric focusing ﬁelds within the lens, and can

67

generally be identiﬁed through the observation an asymmetric ion beam
image.

The ﬁnal aberration of interest is deﬂection distortion. If an ion
beam must be deﬂected to some off-axis remote location, it is common to
bias a set of X and Y deﬂection plates to control the beam position. These
plates are positioned in series with one set controlling horizontal
deﬂection (X) and another controlling vertical deﬂection (Y). When an ion
beam is deﬂected in this fashion the linear deﬂection is accompanied by
some distortion of the symmetric beam shape due to several optical
aberrations. As a result a beam that presents a round proﬁle before
deﬂection can be distorted into an elliptical shape following deﬂection.
Some distortion due to this phenomena is unavoidable, but this effect can
be minimized by using a pair of plates for each deﬂection direction. These
plates should be operated in push-pull mode by placing an attractive
potential upon the plate the beam is deﬂected toward, and a repulsive
potential upon the plate opposite the deﬂection direction.

One ﬁnal phenomena encountered in the design of ion optical
systems is space charging. If a high enough concentration of ions is
present within an ion beam, an electric ﬁeld can be produced between the
ions that will cause the beam to expand in diameter. This becomes
increasingly important when the ion beam is strongly focused because of
the increase in the number of ions per unit volume, and can produce
spherical aberrations in the ion beam. In general, for most ion beams this
spreading due to space-charge effects is not an important factor. This is
true for the manipulation of the ion beams performed in this research.
Assuming a worst case where a 10 eV, 3 mm radius, 5_O_nA beam of
m/z=30 ions must travel the complete distance from the ion source to the
cryostat (~200 mm) the spreading of this ion beam due to space charge
effects is negligible relative to the beam diameter [106]. This is primarily

due to the low ion density of the ion beams produced by these sources.

68

BE I'lI Ql'lD' 321'

Most of the early electrostatic lens systems used in mass
spectrometry were based on the lens designs of Nier [107]. As mass
spectrometers became more sophisticated the need for better ion-optical
lens systems arose. Most early ion optical lenses were designed
empirically or using analog ray-tracing techniques [108]. The analog ray
tracing methods involved the painting of possible electrode arrangements,
in conductive paint, upon resistive paper. Potentials could then be applied
to the electrodes and the electric potential of the free space region between
the electrodes was measured directly. Using these electric potentials the
equipotential lines within this region could be mapped out and the
trajectories of ions traveling through the lens calculated. These methods
were useful for ion optical lens design, but were slow, somewhat
inaccurate, and couldn't be used to model combined ﬁelds and circular
lens geometries. With the introduction of computers, computational
methods were applied to the study of electrostatic optics.

Computation methods, especially modeling programs, have been
applied to several areas of ion-optical design. Modeling programs have
been used to calculate the optical properties of electrostatic lenses
including, object and image positions, magniﬁcations, focal points,
chromatic aberrations, and transfer efﬁciencies. These programs
generally calculate the electric potential and/or magnetic ﬁeld for free
space points within an electrostatic lens using the LaPlace equation. Ion
trajectories are calculated from these potentials using simple Newtonian
relationships, and focal points and spherical aberrations are evaluated by
integration of the paraxial ray equation. The ﬁnal result of these
calculations can include focal lengths, magniﬁcation, ﬁgure of merit, and
aberration coefﬁcients for the lens system of interest [109].

Much of the early electrostatic lens design was performed using the
results of a series of lens models calculated and tabulated by Harding and

69

Read [110]. They modeled a series of 24 possible electrostatic lenses to
compare the performance of each of these systems. The electric potential
for points within the lens systems were calculated by solving for the
LaPlace equation, and the paraxial ray equation was integrated to
determine focal points and spherical aberration for the lenses of interest.
From these results a series of equations were obtained that allow the
general determination of focal lengths, magniﬁcation, and aberration
coefﬁcients for each lens design. The overall behavior of each system could
then be evaluated for a particular application using an arbitrary
comparison term called the ﬁgure of merit.

As the availability of computer systems increased, several different
ion-optical simulation programs were developed. Complex programs such
as TRIO [111] and GIOS [112] were used in the design of several systems
including optimized einzel lens systems for ion kinetic energy
spectrometry (IKES) [113]. Other programs were developed to allow the
design of ion optical systems that could be automatically optimized using
computer algorithms. One example of this type of application was the use
of the modeling program CYLENS [114] to design a series of high
transmission cylindrical lenses that could be self-optimized through the
use of simplex optimization [115]. This type of self-optimization has also
been used to design a high efﬁciency thermal ion source that was also
optimized through simplex optimization [116].

PC based, interactive modeling programs have recently gained wide
acceptance as a tool the design of ion optics. These programs allow the
user to interactively model and evaluate ion-optical systems through the
graphics interface available on a personal computer. The most popular of
these is the program SIMION. SIMION is an electrostatic lens analysis
and design program originally developed by DC. McGilvery [117.] It
allows the user to create a 2-dimensional array of points, and to deﬁne
points within the array as either electrodes of non-electrodes. Voltages are
then assigned to the electrodes and SIMION can calculate the electric

70

potential of each of the non-electrode points in the array using a technique
called self-adjusting over-relaxation. Once these potentials have been
calculated, SIMION can then calculate the trajectories for charged
particles within the array. This trajectory information can then be used to
customize the optical system to the needs of a speciﬁc application.
McGilvery’s original SIMION program has been modiﬁed to allow its use
through an interactive graphic interface on an IBM-AT/PSZ type machine
[118,119] as well as on a MacIntosh [120]. This user-ﬁiendly interface has
made SIMION a practical design tool for anyone interested in the study of
electrostatic optics. Since its inception SIMION has been used in variety of
ion-Optical design projects including: the optimization and design of wide
acceptance-angle ion lenses [121], the design of a surface emission ion
source for a quadrupole instrument [122], time-of—ﬂight modeling [123],
and the design of a short residence time ion lens [124].

The self-adjusting over-relaxation technique used by the SIMION
program to calculate the potential of non-electrode points within its 16,000

point array has been described in detail in the literature [111]. This
technique calculates the electric potential value for a point , P0, from the

average of its four nearest neighboring points, Pl-P4 , as shown below.

P4
P1 P0 P2 Ponew = (P1 + P2 + P3 + P4)/4
P3

This calculation is an approximate solution to the LaPlace equation.
SIMION performs this calculation for all the points within the array, this
is considered one iteration. Several iterations are repeated, and as a result
the changes made by each preceding iteration slowly propagate through
the array. As the number of iterations increases, the change in the
electric potential calculated for each point decreases. The user can

determine what the maximum change should be for the ﬁnal array, when

71

this point is reached the array is considered “reﬁned”. Reﬁning using this
technique requires a large number of iterations, SIMION decreases the
number of iterations required by using an over-relaxation factor that
increases the amount each point changes in each calculation. The
optimum over-relaxation factor for each array is dependent upon the
geometry of the particular array, SIMION uses a self-testing technique to
allow the program to self-adjust the over-relaxation factor to minimize
reﬁning times. The self-adjusting over-relaxation reﬁning method
employed by SIMION greatly reduces the amount of time necessary to
reﬁne the potential array, making the evaluation of ion optical systems far
less tedious.

After the electrode array had been reﬁned, SIMION allows the user
to display equipotential contour lines within the array and model the
trajectories of ions/electrons originating outside or within the array. The
user can deﬁne voltage intervals (or use automatic intervals) for contour
display. SIMION will search through the array, locate points
corresponding to those potentials, and display lines connecting those
equipotential points. These contour lines can be used to evaluate the
magnitude and direction of the electric ﬁelds present within the array.
Ion trajectories can be deﬁned by the user including initial energy, mass,
angle, and fragmentation characteristics. SIMION determines the force
an ion experiences at any point within the array due to applied electric
and/or magnetic ﬁelds and calculates the acceleration an ion will
experience due to that force. During the initial layout of the array the user
indicates the type of symmetry present in the system, circular, planar, or
planar non-symmetric. SIMION uses this information to correctly
evaluate the force due to the potential on the electrodes in three
dimensions. The user can also determine how often the trajectory
calculations are made, the time step, to control the accuracy of the
calculation. These calculations are self-adjusting to ensure that a

trajectory will not pass over a ‘turn-around’ or zero energy point. A series,

72

or family, of these calculations can be performed, displayed on screen, and
observed at high magniﬁcation levels to allow the user to thoroughly
evaluate the optical properties of the electrostatic lens system under
investigation.

Computer modeling programs have proven extremely useful for the
optimization of ion-optical designs, but in order to fully evaluate these
systems it is also important to be able to directly observe the ion/electron
beams. Microchannel Plate Detectors (MCPs) can be used to make these
observations. An MCP is a thin glass wafer that contains an array of 104-
107 miniature electron multipliers. Originally developed as ampliﬁcation
devices for image intensiﬁers [125], MCPs can be used to detect and
visualize beams of charged particles and energetic photons. MCPs are
manufactured from extremely small, lead-doped solid glass ﬁbers that
contain an etchable core. These ﬁbers are packed together and fused into a
hexagonal array. The bound ﬁbers are then sliced to the desired plate
thickness and angle, and the etchable core is removed chemically. The
wafer is then treated to enhance the secondary emission characteristics
within the hollow channels, and the front and back faces of the wafer are
vacuum deposited with an electrode material (Inconel) [126]. A cutaway
view of a ﬁnished MCP is shown in Figure 2.12a.

The resulting MCP contains thousands of miniature ‘channeltrons’
with a typical channel diameter of 10-100 um, length to diameter ratio of
40-100, and separation of ~25 um. If a high voltage is placed across the two
face electrodes of the plate, the collision of a high energy particle with a
channel wall will produce the secondary emission of electrons within the
channel. These secondary electrons can undergo further collisions with
the channel wall producing more electrons, the resulting ampliﬁcation is
generally on the order of 1:10,000 for a single plate. MCPs can also be used
in series to increase the magniﬁcation of the detector. A series of two
MCPs can produce an ampliﬁcation factor of 104 to 107 output electrons per
incident particle. The secondary electrons produced in the MCP impinge

73

upon a detector, generally a metal anode/anode array or a high speed
phosphor screen mounted on a ﬁberoptic. The phosphor screen detection
is generally used to spatially resolve the ion beam proﬁle, while the anode
detector is generally used for high speed detection.

b) Ion

a) Plated Electrode

       
 

 

 

 

 

 

 

 

 

 

 

 

 

Glass Channels Metal Anode
Support (Cutaway View)

Figure 2.12 The microchannel plate detector (MCP).

MCPs can be manufactured with several channel geometries
including straight channel, curved channel, and chevron conﬁguration
for a pair of MCP plates (shown in Figure 2.12b). The original straight
channel design can achieve a gain of 103 to 105 for each MCP plate. The
upper range of gain for the straight channel MCP is limited by ion
feedback. The high ﬂux of electrons produced within a channel can
interact with gas molecules to form positive ions. These positive ions will
experience the strong electric ﬁeld across the channel, drift back through
the channel, and produce an afterpulse at the detector. This phenomena
is called ion feedback. Curved channel MCPs are often used to reduce the
effects of ion feedback. Positive ions drifting through a curved channel will
impinge upon the channel walls and be neutralized. This substantially
reduces detector noise, and allows curved MCPs to achieve gains in the
low 106 range [127]. The disadvantage to curved channel MCPs is that the
manufacture process required to produced properly shaped channels is

74

much more difﬁcult and costly. The Chevron conﬁguration detector,
shown in Figure 2.12b, most often consists of two straight channel MCPs.
The channels within each plate are oriented at an offset angle, typically
8°/8°, from the axis of the detector and separated by 50 - 150 um. By using
an offset angle the background noise due to ion feedback can be greatly
reduced. Under normal operating conditions a chevron conﬁguration
MCP will achieve a 104 gain across each plate for an overall gain of ~4x107.

The detection efﬁciency of MCP detectors is dependent upon both the
potential across the plate and the energy of the incident particles.
Optimum gain is generally achieved at an operating potential between 1000
and 1100 Volts per MCP plate. The detection efﬁciency of an MCP ranges
from 5-85% for ion energies of ZOO-2,000 Volts, and 60-85% for energies of
2,000-50,000 Volts. The properties and manufacture of MCP detectors have
been recently reviewed by Wiza [126], and the use of MCPs as focal plane
detectors in magnetic sector mass spectrometers has been presented [128].

Microchannel plate detectors were used extensively in the
modiﬁcation RGAs to produced mass-selected beams. Two chevron
conﬁguration MCPs with a phosphor screen/ﬁberoptic were used to
visualize the ion beam proﬁles produced by the RGAs and to determine the
angular divergence of those beams. This information was then used to
evaluate the performance of the ion optics in the RGAs, and to make
informed modeling decisions for further SIMION studies.

75

Chapter 3. Anavac-2 Ion Source Modiﬁcations
I h ' f f An -

The ﬁrst residual gas analyzer selected for modiﬁcation was an
Anavac-2 RGA produced by the Micromass division of VG-Gas Analysis
Limited. The Anavac-2 is a small RGA, designed to measure both total
pressure within a vacuum system and partial pressures of residual gases.
The minimum detectable total and partial pressures for the unit are
quoted as 5x10’12 torr and 5x10'11 torr respectively. The RGA uses the open
EI source design described in Chapter 2.I.A, a miniature quadrupole
mass ﬁlter, and a Faraday plate with a chopper ampliﬁer for detection.
The quadrupole can operate in mass-selective mode for partial pressure
analysis with a mass range of 260 daltons and resolution quoted as better
than 50% valley. The maximum operating pressure for the Anavac-2 RGA
is 1x10'4 torr. Some of the features of this RGA include: external
programmability of the mass ﬁlter, overpressure protection circuitry that
can be used to control other equipment, complete short circuit protection of
all circuitry, and program outputs voltages that allow the unit to drive an
XY recorder for recording residual gas spectra. The unit was initially
bought in 1979, and at that time the complete RGA cost only $4,500.

The Anavac-2 consists of three primary components, the analyzer
head, the remote electronics head, and the control unit. A functional
diagram of these three components is shown in Figure 3.1. The analyzer
head contains the mechanical components necessary to generate, mass
select, and detect the ions. The remote electronics head contains the
circuitry that generates the rf and dc levels necessary to operate the
analyzer head, and also the sensitive ampliﬁer required for the detection of
ions. The control unit produces the dc potentials required for the operation

of the analyzer head as well as the high current levels that heat the

76

dosh—ace new 72.32.: N-u=>a:< 2.: he 53min 93:59:92 3.552.... < ~.n 3:3,.—

    

      

...= ._.r.I<
:< t.
:. =2...

 

_Z._::__ .2. ..

     

            

     

$2: .. 2.25
.222 ._z<
.22. _._2<

.222

P—(r.
><c<z<u

 

 

zoned—FE.“

 

> 5.: E
.zH.=z< .2

r :32. .L
z» $2.2m

3...; _<Z<

               

 

_... >3. .

:. ._<.Iz._..
:.

.C:<u
231—:

. .. =...._:<
1.5.3.:—

 

£53.

   

>2v

         

.>. . >HEKZA 2A 3 ._Z<
.2 $.12. v 2. xii—1.. <.:~

 

we... z J.

  

   

.. : 5...... >.._ .

 
   

_2£!( :. .3 23:... =:I< =. ..Z<

3:22. 2. =3... _. >2

=wm>..<2< i :<:.._Zt. 1.: :2: r :22. r.

 

77

ﬁlament. The program voltages necessary to drive the remote electronics
head are also produced and controlled by the control unit.

The analyzer head contains all the components that are inserted
into the vacuum chamber. These include the ion source, quadrupole, and
detector. Figure 3.2 shows a block diagram of the overall design of the
analyzer head. The Anavac-2 source is similar in design to the source
described earlier and shown in Figure 2.3. The symmetry of the unit is
circular and Figure 3.2 shows a cross-sectional view of the components of
the Anavac-2 system. The original RGA ﬁlament, normally tungsten, is
approximately 10 mm long and oriented parallel to the Z-axis of the
quadrupole. It is spot welded to two support leads that run through high
current feedthroughs located in the endcap of the ﬁlament housing. One
of the support leads is in electrical contact with the ﬁlament housing,
allowing both the ﬁlament housing and the ﬁlament to be maintained at
the same bias potential. The outside diameter of the ﬁlament housing is
29.0 mm the inside diameter is 16.8 mm. The source grid (6.6 mm
diameter, 9.0 mm height) is located within the ﬁlament housing and is
spot welded to a 0.5 mm stainless steel (SS) circular plate with a 3 mm
center aperture. A single lens, the focus element, is positioned after the
source grid. This electrode is also a 0.5 mm SS circular plate with a 3 mm
central aperture. The ﬁnal component of the ion source optics is the
quadrupole entrance aperture plate containing a third 3 mm aperture.
The source components are held in place with threaded SS rods that are
isolated from the individual components of the source with ceramic
sheathing tubes. 1.3 mm thick ceramic spacers are used to electrically
isolate the components including the ﬁlament housing, source grid, focus
lens, and quadrupole entrance plate.

The quadrupole mass ﬁlter is housed in a SS tube housing that is 53
mm long and has a diameter of 30 mm. The entrance aperture to the
quadrupole, set by the quadrupole entrance aperture plate, is 3.0 mm and

the exit aperture is 3.0 mm. Cutaway sections in the quadrupole housing

78

allow the connection of electrical leads to the quadrupole rods. The
quadrupole rods are permanently mounted in a machined ceramic block
designed to ﬁt tightly inside the quadrupole housing. This ceramic block is
held in place by a setscrew located in the side of the quadrupole housing.
The quadrupole rods are 50.0 mm long, have a rod diameter of 6.4 mm,

and a rod spacing of 5.5 mm.

  
  
     

2 3/4" Mounting

Source FOCUS Flange

'Grid ‘E/lement Quadrupole Electron
ﬂ

Suppressor

- — — — Z‘I-Z-l-i -L-£‘:“'2‘;;l-Z'I-k2'2'2‘24';rL ~1'2' L-l-Lx’
\ .
u .5
x . 5!:
. .
. . . . .3.‘:.:._ xx . . 52'. 3.5% . . . . 34.7.

Filament

 

 

 

 

     

 

 

 

- ' Faraday
\ T / Plate
t Quadrupole uadru ole
Housing Entrance Quadrupole Q Exitp
Aperture Housmg Aperture

Figure 3.2 A block diagram of the Anavac-2 analyzer head

These dimensions yield a r/ro ratio of 1.16 which is close to the 1.148

optimum dimensions for production of a hyperbolic ﬁeld within the
electrode array. Using relationships described in Chapter 2.II.C, it is

possible to calculate the ultimate performance that could be achieved with
this size of quadrupole. For a 2 Mhz, 250 Vp-p rf signal the maximum

mass, Mm, and minimum peak width, AMmin: that could be achieved with

this system can be calculated as:

 

6 6
7x10 V
= m = 71:10 (500 V) 2 = 116 daltons

m 2 2
ero (2x106Hz) (2.75x10'3m)

79

4x109Vz _ 4x109(3 eV)

AM min: = 1.2 daltons

 

 

2 2 ‘ 2
f L (2x106) (0.050 m)2

Thus the Anavac-2 quadrupole is functioning close to the ultimate
resolution, or minimum peak width, that could be achieved with a
quadrupole of this geometry. However the mass range that is covered by
the stock Anavac-2 is approximately half of the ultimate range the
quadrupole could achieve. The probable reasons the Anavac-2 is only
designed to function from 2-60 daltons include; higher sensitivity obtained
through operating under these conditions, decreased sensitivity to changes
in mechanical alignment, and the fact that for most residual gas analysis
a 2-60 mass range is sufﬁcient to accurately monitor residual gases.

The exit aperture plate for the quadrupole also functions as a
mount, allowing the coupling of the quadrupole housing to the 2 3/4”
support ﬂange. The Faraday plate used to detect ions that pass through
the mass ﬁlter is located within the support flange along with an electron
suppressor ring electrode used to reduce background electron noise. All
the potentials supplied to the components of the analyzer head are fed
through high vacuum electrical feedthroughs located in the mounting
ﬂange.

The remote electronics head contains the circuitry used to produced
the dc and rf levels necessary for the operation of the quadrupole, and also
the chopper ampliﬁer used to detect the ion current at the Faraday plate.
The rf generation circuit consists of three primary components: the crystal
controlled oscillator that controls the rf frequency, the rf power stage that
produces the rf signals, and the control stage that controls the rf output
and produces the correct dc levels for the quadrupole. A block diagram of
the principal quadrupole power supply components for the Anavac-2 RGA
is shown in Figure 3.3

80

The crystal oscillator generates a square wave ac signal at a
frequency of 2 MHz. This signal is fed through the control stage to the rf
power stage which produces a high voltage (maximum 250 Vp,p) 2 MHz
sine wave across the primary coil of the rf transformer. The secondary
coil of the rf transformer is center tapped, producing 90° out of phase rf
signals on each of the two output leads. The output leads connect to the
high voltage feedthroughs on the analyzer mounting ﬂange through
which the rf signals are fed to the quadrupole.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

’ rf
dc output rectiﬁcation
J
. Crystal
5 E rfpower Control Oscillator
v02 g a Stage Stage (2MHz)
k
Quadrupole cou t ‘ ff

 

 

 

 

 

r rectiﬁcation

 

Figure 3.3 A block diagram of the quadrupole power supply circuit used in
the Anavac-2 RGA.

The rf outputs from each of the two output leads are also fed back to
the control stage where they are rectiﬁed to produce +dc and -dc levels.
The control stage sets the dc balance between each quadrupole rod pair,
and the rdec ratio that is produced in the rectiﬁcation step. The dc levels
produced by the control stage are fed back to the output leads connected to
each of the quadrupole rod pairs producing the dc component on the
quadrupole. These do levels can be measured at testpoints within the
remote head to verify the dc balance in the system and to indirectly monitor
the rf output. A variable capacitor (V02) is connected across the

81

secondaries of the rf transformer which can be varied to adjust the
resonant frequency of the quadrupole/secondaries circuit to match the
ﬁxed oscillator frequency. When the RGA is operated in total pressure (rf-
only) mode the dc output levels from rectiﬁcation stage of the control stage
are taken to ground, allowing the mass ﬁlter to pass all ions of m/z >9. A
complete circuit diagram of the remote head rf generation and chopper
ampliﬁer circuits is shown in Appendix I.

The remote head also houses a sensitive chopper ampliﬁer that acts
as a current-to-voltage converter for measuring ion current at the Faraday
plate. The ampliﬁer is housed in a shielded case to reduce noise from the
rf generation circuitry, and the output of the chopper ampliﬁer is fed to the
control unit for display. The control unit allows the chopper ampliﬁer to
function at several levels of ampliﬁcation that can be selected using the
RANGE switch on the control unit. More sensitive range settings result in
higher overall ampliﬁcation of the ion current, but require longer
sampling times to achieve reasonable signal/noise levels. The response
time, noise level, and overall gain of the chopper ampliﬁer at each range

setting is given in Table 3.1

Table 3.1 The performance of the Anavac-2 chopper ampliﬁer

Mr rr W W n i %f m
10~5 40 ms 0.1 109
10’5 30 ms 0.1 1010
10-7 30 ms 0.1 1011
10“3 40 ms 0.5 1012
10-9 500 ms 0.5 1013
10‘10 500 ms 1 1014

The third component of the Anavac-2 RGA, the control unit,
supplies all the remaining do and ac levels necessary to operate and
control the quadrupole. These functions include the ﬁlament circuitry,
mass program voltage circuits, and the dc levels required at the analyzer

head. The emission of the ﬁlament in the analyzer head is regulated at 2

w- nah a
:“F, "1“)

82

mA by the control unit. This emission current is measured at the source
grid which is maintained at ~+3 volts. For the stock tungsten or rhenium
ﬁlaments a current of approximately 4.35 amps is required to produce a 2
mA current at the source grid. The feedback circuit that regulates the
ﬁlament emission is also connected to protection circuit. When the RGA is
operated in ﬁlament protect mode the ﬁlament will automatically shut off
should the current from the chopper ampliﬁer exceed full scale on the
control unit meter. This protection circuitry also controls the state of two
relays that can be used to protect other equipment within the vacuum
system. This protection circuit can be disabled by operating the RGA
ﬁlament in ﬁlament override mode.

The operating mode of the quadrupole is controlled by the operating
mode switch on the control unit. The TOTAL PRESSURE position causes
the dc rectiﬁcation circuit in the remote head to be taken to ground,
allowing the detection of ions_with m/z > 9. When this switch is in the
PARTIAL PRESSURE mode the quadrupole is operating in mass-selective
mode and the detector output is displayed on the panel meter. When the
operating mode switch is is set to MASS MODE the quadrupole is
operating in mass selective mode, the panel meter displays the ion mass
selected by the quadrupole, and the detector output is available at the
program output line on the back of the control unit. The mass dial on the
unit controls the initial mass that is selected by the mass ﬁlter. This
control can be used to ‘sit’ on a mass peak, to perform a form of selected
ion monitoring, or to set the starting mass for a mass scan. The scan
switch causes the control unit to produce a program voltage that will scan
the ﬁlter upward from the initial setting on the mass dial. Two scan
speeds are available, slow (5 sec/amu) and fast ( 0.5 sec/amu), that are
selected depending upon the sensitivity required. The range of mass
values covered by the program voltage is controlled by the span control
which allows a range of 1-60 amu to be repetitively scanned. When placed

in scan mode the control unit will scan the mass ﬁlter until the span limit

83

or the upper limit of the device (m/z = 60) is reached. At that point the scan
will reset and the process will repeat. Scanning of the mass ﬁlter can also
be accomplished through the use of an external input voltage. An external
input port is available that allows the operator to scan from m/z =2-60
using a 0-10 volt dc input signal. The control unit also contains program
outputs voltage lines for both the mass scale (X) and the detector output (Y)
that can be used to drive an XY recorder. A complete diagram of the
circuitry inside the control unit is shown in Appendix II.

The initial operating parameters and performance of the Anavac-2
unit prior to modiﬁcation are given in Table 3.2. These values were used
for comparison with the modiﬁed system to evaluate the effect that design
changes had upon the performance of the RGA.

Table 3.2 Initial optimum operating conditions for the Anavac-2 RGA

Electron Energy = 70 eV Focus Potential = -40.3 V
Ion Energy = 3.0 eV Suppressor potential = -110 V
fast scan time = 61 sec slow scan time = 337 sec

ﬁlament current limit = 4.1 A
resolution = 20% valley

sensitivity in partial pressure mode: 1.2x10'4 A/mb

sensitivity in total pressure mode = 1.5x10‘3 A/mb
dc balanced at mass 60 to within 0.1 V
Mass Stability (peak top stability) drift < 0.1 amu/ 8 hours

The Anavac-2 RGA that we obtained was several years old, and the
ﬁrst study performed was an analysis of its mass resolution and
performance range. The entire analyzer head was ﬁrst disassembled and
cleaned with a mild abrasive followed by ultrasound baths of water and
acetone. Following cleaning, the unit was assembled, tested for electrical
continuity between the ﬂange feedthroughs and the analyzer elements,
and installed on a vacuum chamber. The remote head and control unit
were then installed, and the program outputs on the control unit were
connected to the X and Y inputs of an XY recorder. A Granville-Phillips
ion gauge and Ametek M100 RGA were also installed on the vacuum

84

chamber for comparison with the total pressure readings from the
Anavac-2. Following evacuation to ~8x10'7 torr, an air leak was
introduced into the chamber. Air was introduced to an ion gauge reading
of 5x10'6 torr which yielded an Ametek M100 total pressure reading of
1.6x10‘6 torr and a partial pressure reading for nitrogen of 3.5x10'7 torr.
The Anavac-2 RGA was turned on and the ﬁlament was heated to an
emission of 2 mA. The total pressure reading displayed when the unit was
set to TOTAL PRESSURE mode was 2.1x10'6 torr. The unit was then set to
MASS MODE at a sensitivity setting of x10"5 torr and scanned at a slow
setting from m/z = 2-60. The spectrum produced from the XY recorder set
at Y = 100 mV/in scale is shown in Figure 3.4. The bottom trace shows the
complete mass spectrum from m/z 2-60 and the upper trace shows an

expanded view scanning from m/z 12-20. A peak intensity of 5.2x10'12 A
for N21” at m/z = 28 was calculated from the sensitivity speciﬁcations given

in Table 3.1. Using the N2 partial pressure reading from the Ametek
M100 RGA and the quoted sensitivity of the Anavac-2 RGA in partial

pressure mode, the optimum current that the Anavac-2 should produce
with this leak is 5.6x10'11 A. This value is within an order of magnitude of
the experimental value obtained, which is good agreement considering the
experimental errors involved. Some experimental errors could include;
different pumping speeds at each measuring device, use of an older
ﬁlament on the Anavac-2, and loss of sensitivity from contaminants on the
Anavac-2 source.

The resolution performance of the Anavac-2 can be best observed in the
upper trace of Figure 3.4. This expanded view clearly shows the peaks at
m/z = 14, 16 and, 18 resulting from the air leak and residual water on the
walls of the aluminum vacuum chamber. A small peak at m/z = 17 can be
seen that is formed through the fragmentation of H20+ to form OH+. This
peak is poorly resolved, and although it is not an ideal comparison peak for
resolution evaluation, the Anavac-2 didn’t appear to be functioning at the
full 20% valley resolution given in the original speciﬁcation sheet. For the

85

.om 3 NH u SE 89¢ 5&8 05 me Be? wmwcquo ca
@265 9.82: “can: can. 13:3 :a 5m? max «-3323. 35.3.8 9: .3 agencies.“ comes—omen 35. «an 82mg

N\E
cm on ov om cm 3 o
a H ..

. D F P D ? b P b P D r P
_ ‘ 1 l4 1‘ ‘ I

’ Ll
d 1

l
a
J

 

 

 

3
mm 3

 

mg

mu

 

 

 

 

mm

86

matrix isolation studies utilizing this source, this slight loss of resolution
should not be a difﬁculty.

I ' in hn -

The ﬁrst modiﬁcation made to the Anavac-2 residual gas analyzer
was the construction of a source housing to allow the direct introduction of
sample gas into the source region. To minimize the amount of neutral
sample gas that is trapped on the cryostat surface, the source housing
must have a low conductance to the experiment chamber. A source
housing was designed that could reduce the conductance from the El
source by 97%. To accomplish this reduction in conductance, a source
housing was designed that enclosed the entire source region and reduced
the ion exit aperture. Before undertaking the lengthy manufacture of this
housing several SIMION modeling studies were performed to evaluate the
effect it would have on the performance of the El source.

The Anavac-2 source was ﬁrst disassembled and all of the stock
components were accurately measured. Using these dimensions a
SIMION model of the E1 source was constructed, this model is shown in
Figure 3.5. To evaluate the performance of the original EI source it was
necessary to ﬁrst determine the trajectories that electrons emitted from the
ﬁlament would follow in the source. From this information it was possible
to determine the region within the source grid volume that had the highest
probability of an electron/molecule collision that could lead to ion
formation. For this reason it was necessary to include in the this model
the ﬁlament and support rods shown in Figure 3.5. To include these
components the source was modeled in planar-nonsymmetric symmetry,

which required a larger number of array points, in this case 3,000 points,

87

Source Plate Focus Lena

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
   

cm
E
E Quadrupole Housing
H
In
Filament :
~
Support :
Rod E
H
E
Source ’l " :1
Grid -
Electron . m E
Trajectories ,-
d
Countour :
Lines :
Filament .2
Housing )-
.7
pg
:1
f Emu
- O V ts
67 Volts +3 Volts 40 Volts 01

Figure 3.5 SIMION model of the Anavac source showing electron trajectories.

88

than do symmetric geometries. As a result, both the accuracy of the
dimensions of the model and the symmetry of the model were somewhat
degraded. Even with this difﬁculty, this model proved useful in studying
the behavior of electrons in the Anavac-2 source.

The SIMION model shown in Figure 3.5 consists of four separate
electrodes. The ﬁlament/support rod and ﬁlament housing are held at the
same potential, in this case the normal operating potential of -67 volts. The
source electrode consists of two components, a square grid region and an
plate with an aperture, that are held at the normal source potential of +3
volts. With SIMION the user can create transparent electrode grids by
setting the electrode thickness to 1 grid unit, the source grid is modeled in
this fashion. The source plate was modeled to produce the 3 mm source
exit aperture size as closely as possible within the constraints of the array
size. The focus electrode,the third electrode element, it is a plate of
identical dimensions to the source plate. In the model shown, the focus
electrode is biased at the ~40 volt potential normally used by the Anavac-2.
The ﬁnal electrode in Figure 3.5 is the quadrupole housing. This electrode
models the quadrupole entrance aperture plate and the tube housing that
contains the quadrupole. The quadrupole entrance aperture is identical to
the source aperture, and the entire quadrupole housing electrode was
maintained at the normal operating bias of 0.0 volts.

Following construction of the electrode array, the free space region
within the array was reﬁned to determine the electric potential at each
grid point. Equipotential contour lines were then calculated at 10 volt
intervals and are shown in Figure 3.5. These contour lines help the user
to evaluate the force that a charged particle experiences due to the electric
ﬁeld created by the electrodes. Several observations can be made from
these contour lines. The curved contour line penetrating the source plate
aperture indicates magnitude and direction of the drawout potential
created by the bias on the focus electrode. The rounded shape of the

contour lines, similar to an concave optical lens, between the source plate

89

aperture and the quadrupole entrance aperture illustrate the focusing
properties induced by this three element unipotential (einzel) lens. In the
region where electrons can be produced by the ﬁlament the contour lines
are fairly parallel to the source grid. Because of this, electrons emitted
from the ﬁlament will gain a high transverse energy relative to the
quadrupole Z-axis, but little energy in the axial direction. As a result
electrons emitted from the ﬁlament will pass through the source grid on a
fairly straight trajectory, with only slight motion toward the source plate.
This allows the electrons to make several passes through the source grid
before they impact upon the source plate.

Following the modeling of the contour lines in the source, a series of
electron trajectories were calculated. These electron trajectories
originated along the length of the ﬁlament and followed the path that a
negatively charge particle would travel starting from those positions.
These electrons generally made several passes through the source grid, as
predicted from the contour lines in that region. Electrons that originated
toward the left end of the ﬁlament initially experienced less axial force
and made more passes through the source grid. Electrons formed closer
to the source plate felt a much stronger attractive potential toward the
source plate, and often only made a single pass through the source grid.
Using the information from this model it was possible to determine the
region within the source grid that an electron/molecule collision, and
subsequent ion formation, was likely to occur.

The Anavac-2 source was then remodeled, this time using a 16,000
point circular-symmetric array geometry that better modeled the circular
symmetry present within the Anavac-2 source. Symmetric geometries
also require half the number of points of asymmetric arrays, therefore
much more accurate dimensions can be modeled. Figure 3.6 shows the
16,000 point circular-symmetry SIMION model of the original source. The
same electrodes were constructed and biased, but the dimensions of this

model reproduce the real source geometry much more closely. The array

90

.motoaoonmh. :8 953% 8.5% um>m:< 2: mo Eves. 2.52:5? bﬁutvczho < wd Sang

2580 3.52.- mﬁS 9. 33> S-

    

  

 

 

\xncogma
ll , w]! .. ...-...u...u.....4u . . . . .il\\\\\\ :2
i ll“ II. ...|.,w..uu.o.nl.t..,y.l0 mwml ~65-=
, /.///%../// .

grouper“:
:2

 

  

91

was then reﬁned, and contour lines were calculated at 5 volt intervals.
Several observations can be made from the contour lines in this model.
There is less penetration of the extraction ﬁeld from the focus element into
the source region than in Figure 3.5. This is primarily due to the change
in size of source plate aperture. The model in Figure 3.5 contains an
oversized source plate aperture that was used as a compromise because of
the limited accuracy of the model array. The model in Figure 3.6 much
more accurately reproduces the real source plate aperture, and the
extraction ﬁeld produced by the focus element. The smaller contour
interval combined with the larger array size also more accurately models
the focal properties created within the source ‘einzel’ lens. The contour
lines between the source and the ﬁlament housing also show much
stronger curvature than in Figure 3.5. The model in Figure 3.6 is more
realistic because the the actual Anavac-2 ﬁlament and support rods are
relatively small and would have a much smaller effect on the electron
trajectories than is indicated in Figure 3.5. Using the model in Figure 3.6,
the trajectories of electrons originating at the position of the ﬁlament were
modeled, and the results compared to those from the model in Figure 3.5.
The potential and electric ﬁeld at each of the electron starting positions
was determined, and the trajectory information was tailored to the
individual formation conditions of each electron trajectory modeled. The
region of highest electron/molecule collision probability determined in this
fashion differed little from the results obtained in Figure 3.5.

The trajectory family function in SIMION allows the user to
combine a series of separate trajectory ﬁles into a single series of
calculations. This function was used to model a series of ions formed,
with no initial energy, within the source region of highest ionization
probability. These trajectories are shown in Figure 3.6. The trajectories of
40 individual ions, reﬂected into 82 displayed trajectories, are shown. All
of the ions that were modeled experienced sufﬁcient drawout potential to

leave the source region and pass through the quadrupole entrance

92

aperture. Ions formed toward the center of the source grid experienced
less focusing in the lens and made up the central component of the ion
beam entering the quadrupole housing. Ions that were formed in the
outer region of the source grid gained more transverse energy and crossed
over the center-line several times. These ions formed the outer component
of the ion beam entering the quadrupole housing.

Following the evaluation of these models, the design changes
involved in enclosing the ion source within a source housing were
evaluated using a third SIMION model. The model of the modiﬁed
Anavac-2 source is shown in Figure 3.7. The use of a source enclosure
required the increase in thickness of the source plate and the reduction of
the source plate aperture. The source plate aperture was reduced from the
original 3 mm to 1.5 mm, and an additional 0.5 mm thickness of electrode
was added to the source plate. The model used in Figure 3.6 was modiﬁed
to reproduce these design changes at the same accuracy level (16,000 point
cylindrically-symmetric array) used in that model. Following the
construction of the array, it was reﬁned and 5 volt contours lines were
calculated, these are shown in Figure 3.7. The penetration of the
extraction ﬁeld into the source region of this model is slightly reduced
from the original conﬁguration. The shape and placement of the contour
lines in the source lens differs little from Figure 3.6, and the contour lines
between the ﬁlament housing and the source grid remain the same, as
expected. The same family of ion trajectories was modeled for the modiﬁed
source conﬁguration. As in the previous model, all the ions that were
formed in the source region were effectively extracted and injected into the
quadrupole. Ions formed in the central region of the source grid followed
approximately the same trajectories as in the previous model. Ions that
were formed toward the outer region of the source grid experienced a
smaller drawout potential than in the Figure 3.6, and gained less
transverse energy. As a result these ions crossed the centerline less often,

and remained closer to the central axis of the ion source.

93

mummies 8.58 2: he vomsmo woman—t .8.er so.— of .‘c .355 20:45 < ham mama

33> n+
6CDOMU m23> CV: 3—0> hm-

33.8%
v8.68

 

 

 

mmocxumﬁ
395030
@033

 

94

The observed effects of introducing the source housing elements into
the SIMION model of the Anavac-2 source indicated that the modiﬁed
source should produce an ion beam that is slightly more focused than the
beam produced by the original source. The SIMION modeling indicates
that the modiﬁed source could produce approximately the same ion beam
proﬁle as the original source by reducing the potential on the focus
element to -20 volts. These results indicated that enclosed source would
still be fully functional following modiﬁcation, but that the total output of
the modiﬁed source could be slightly reduced because of the weaker

drawout potential.

i rf vl 'nfmiﬁ re

A source enclosure, or source housing, was constructed using the
information gathered from the SIMION modeling studies. This housing
was designed to contain the source grid, ﬁlament housing, and ﬁlament
assembly within a region of limited conductance to the outside chamber.
To minimize the changes to the original design, the end of the housing
containing the ion exit aperture had to be in electrical contact with the
source grid. To avoid the mounting difﬁculties associated with ﬂoating the
entire source housing, a Ceramaseal 807B9999-3 envelope was used to
serve as the body of the source housing as shown in Figure 3.8.

This envelope is a metal/ceramic/metal tube with an outside
diameter of 1.63”, inside diameter of 1.25” and length of 1.4”. Using this
envelope as the body of the source housing allowed one end of the housing
to ﬂoat at the source plate potential, +3 volts, while the other end could be
used to mount the source to a support ﬂange. The source housing also had
to allow the introduction of electric potentials and sample gas into the
source region. To facilitate this, a gold wire seal ‘ﬂange’, 1.9” in diameter,
was welded to the mounting end of the source housing. A Ceramaseal
807B8177-1 quad mini-feedthrough, rated at 5 amps dc, was welded into

95

the removable side of the ﬂange to allow the introduction of ﬁlament and
ﬁlament housing potentials. A 1/4” OD inlet line was also welded into the

removable end of the ﬂange to allow the introduction of sample gas into the

source housing.

Flange Threaded
Rod

I

Feedthrough

 

 

 

   
  

 

 

 

 

 

 

 

push-on A

 

 

rEEEEEEF‘

 

 

 

 

 

 

 

 

 

T

 

 

 

 

 

 

 

metal/ceramic/metal V
junction Tab

Mounts

Figure 3.8 A block diagram of the source housing, source, and
quadrupole.

A 1.25” plate was welded to the source plate end of the source
housing. To allow the source plate to lie ﬂat against the inside of the
housing it was necessary to do some specialized welding, which was
performed by a master welder at the cyclotron facility. After the endplate
was attached to the source housing, a 1.5 mm hole was drilled in the
center of the plate to serve as the ion exit aperture. Four 0.5” long threaded
rods were installed in the plate so that a small portion of the rod ( ~ 1 mm)
extended outside the source and the remaining rod inserted into the source

housing. These rods were positioned in alignment with the mounting rods

96

of the source components and were used to hold the source grid and
ﬁlament housing in position inside the source housing. Tab-mounts were
welded to the outside of the source plate end of the housing and to the sides
of the quadrupole housing. These tab-mounts contained ceramic
insulators and were used to physically connect the source housing to the
quadrupole while maintaining electrical isolation.

Removing the source housing ﬂange allows the user to access the
ﬁlament housing and source grid for ﬁlament replacement and source
cleaning. The source plate and ﬁlament housing are held in place by the
threaded rod which is isolated from these components by a ceramic
sheath. Spacing between these elements is set by the ceramic washers
used in the original source. After the source components are cleaned and
the ﬁlament replaced, the feedthrough connections to the ﬁlament can be
made using push-on Ceramaseal 841B2044-2 connectors, designed to carry
up to 15 amps of current. The ﬂange can then be tightened down with 8
hex screws, and the conductance of the source is then limited by the 1.5
mm ion exit aperture. By comparing all of the aperture areas in the
original source to the aperture of the source housing, it was calculated
that the source housing effectively reduced conductance from the source
region by 97%. This was accomplished without sacriﬁcing the
convenience of easy access to the source components for cleaning and
ﬁlament replacement.

Following the manufacture of the source housing, the performance
of the modiﬁed source was evaluated. The source housing and all source
components were cleaned and the ﬁlament was replaced. The source
plate, and ﬁlament housing were installed in the source housing and the
housing ﬂange was sealed. The source housing was coupled to the
quadrupole housing and the entire unit was remounted on the original
RGA ﬂange so that ion currents could be measured with the original
detection system. All the electrical connections were made using the stock

wiring, and the entire assembly was installed on a vacuum system and

97

.ooanom om>mc< verve: 2: t3 nee—3 momma 32:85 12668 .3 83.50on 3.2: 95. Wm BPJE

88
cm Om ov on cw on

b b
q .4 d u

«too

‘-

 

Va

mm

 

 

 

w—

98

pumped down. Following evacuation, the remote electronics head and the
control unit were installed and an XY recorder was connected to the
program voltage leads from the control unit. Figure 3.9 shows the mass
spectrum obtained from the background residual gases in the chamber.
The spectrum was taken at a chamber pressure of 1.5x10'6 torr measured
by an ion gauge, and the total pressure reading given by the modiﬁed
Anavac-2 was 2.0x10‘7 torr. This represents a decrease in total pressure
output of 79% due primarily to the reduction of the source aperture. The
principle components of the residual gases in the chamber are H20, N 2, Oz,
and various hydrocarbon peaks produced from the rough and diffusion
pump oil. The spectrum was taken at a sensitivity setting of x10'8 torr,
slow scan speed, and with a Y scale of 50 mV (variable). The peak series of
m/z = 12, 14, 15, 16 , 17, and 18 clearly shows that the resolution of the
quadrupole has degraded only slightly as a result of the source
modiﬁcations. The quadrupole was scanned from m/z = 2-60, and peaks
were observed through the entire range which indicated that the original

mass range of the quadrupole had also been maintained.
I ' i h l wirin

The ﬁnal installation of the modiﬁed Anavac-2 RGA required the
modiﬁcation of the analyzer head wiring. The design goal for the ﬁnished
mass-selected ion source required that the entire assembly be mounted on
a 6” conﬂat ﬂange to allow the source to be on the matrix isolation chamber
and easily removed for servicing. Another design goal was that the ion
source should be mounted in a fashion that allows it to be repositioned
under vacuum. To achieve this, the wiring that connects the source and
quadrupole components had to be ﬂexible, and long enough to allow a wide
range of movement.

Modifying most of the analyzer head unit wiring was not critical.
The dc potentials required to bias the electrodes, such as the source grid,

99

focus lens, and quadrupole housing, required only low current/low voltage
dc signals. These components were wired using ﬂexible, Teﬂon-coated
cable. The wiring to the ﬁlament needed to carry high currents at low
voltages. For these components 18 gauge copper wire was connected to the
original ﬁlament wiring, covered with ceramic thermocouple insulators,
and connected to the leads on the source housing ﬂange with Ceramaseal
880A7332-2 connectors. The most critical stage of rewiring the analyzer
head was the modiﬁcation of the quadrupole wiring.

As discussed in Chapter 2.II.D, the operation of the quadrupole
circuit depends upon proper matching of the rf power supply frequency
with the resonant frequency of the LC tank created between the
secondaries of the rf transformer and the two quadrupole rod pairs.
Changing the length of the wires connecting the quadrupole rod pairs to
the feedthrough ﬂange can change the capacitance of the circuit, causing
the resonant frequency of the circuit to change. If this frequency changes
too much, the power transmitted from the rf power supply cannot be
properly absorbed, and will be reﬂected back into the power supply. In the
case of the Anavac-2, with its automatic overload protection, this reﬂected
power causes the rf generation circuitry in the remote head to shut down.
The resonant frequency of the quadrupole circuit can be adjusted within a
limited range by adjusting a variable capacitor (VR2 in Appendix I). A
study was performed to determine the practical range of wire length that
could be used in wiring the quadrupole

This study was done using the original Anavac-2 mount and
detector to allow direct comparison with previous studies. The source was
cleaned and mounted in the source housing. This was installed on the
quadrupole housing and the entire unit was attached to the mounting
ﬂange. The original wiring was connected to all components except the
ﬁlament and quadrupole. An additional length of wire, described above,
was used to connect the ﬁlament. To study the effect of wire length upon

the behavior of the quadrupole, several lengths of wire were added to the

100

original quadrupole wiring. The additional wire lengths consisted of 22
gauge chromel A wire, twisted into the stock wire and covered by ceramic
thermocouple insulators. The opposite end of each added wire was looped
along the side of the source housing and connected to a quadrupole rod.
The unit was then installed in a vacuum chamber, with the wire loops > 6
inches from any surrounding metal to reduce the effect of the chamber on
the performance of the system. The behavior of the quadrupole with the
wire lengths; original, 14.5 cm, 25.0 cm, and 30.0 cm was observed as the
quadrupole was scanned from m/z = 2-60. With the original wire the
spectrum obtained was similar to Figure 3.9. When a 14.5 cm length of
wire was added, the quadrupole was still able to scan through its full
range without any difﬁculty. The addition of 25.0 cm of wire did change
the behavior of the quadrupole, the output spectrum appeared normal
until m/z = 32 was reached. At this point the output of the ampliﬁer went
to baseline and repeat attempts at scanning any portion of the mass
spectrum failed. When the control unit was turned off brieﬂy and then
turned back on, the normal mass scanning of values less than m/z = 32
returned. This apparent ‘shutdown’ behavior was also observed by
monitoring the dc output testpoints in the remote head. During a normal
scan the dc levels at these points increased linearly with increasing m/z.
When the ‘shutdown’ behavior occurred the dc testpoint voltages suddenly
went to ~5 volts and remained constant. This proved generally useful as
an indicator of shutdown behavior in the rf power supply. Finally, a 30.0
cm length of wire was added to the stock wire. With this wire length the
‘shutdown’ behavior was observed at m/z = 27. An example of a spectrum
displaying rf power supply shutdown is shown in Figure 3.10. This
spectrum shows a mass scan from m/z = 2-60 of an air leak. A 30.0 cm
length of wire has been added to the stock quadrupole wiring. The total
pressure reading of 2x10'6 torr was obtained from an ion gauge installed
on the system. The sensitivity scale of the control unit was set to x10‘8 torr
and the Y sensitivity of the XY recorder was 50 mV full scale. Background

101

:Bcgsnm @8335 xomnvomﬁﬁ MEBonm 8.58 2355. 2: 89a 85.50on $58 4 oﬁm anE

'0

q

.\ s2: 2

3% cm ov on em ﬁn
I

4.

 

 

 

 

102

peaks are observed in the mass spectrum for water, air, and pump oil
components from m/z = 12 to 18. Shutdown occurs at ~m/z = 27 and can be
seen as a sharp decrease in output on the leading edge of the m/z = 28 peak
(N2+). After this point the control unit had to be turned off to get mass-
selected output at any m/z value.

By tuning VR2 it was possible to bring the electronics back into
resonance. In order to obtain optimum rf power absorption, VR2 was
varied while observing the output of the dc testpoints in the remote head.
By monitoring these testpoints it was possible to determine the range of
VR2 settings through which no shutdown occurred. VR2 was then left at
the central position within this range which corresponded to optimum rf
power absorption. For the two longer wire lengths, 25 and 30 cm, varying
VR2 could keep the unit from shutting down, but because VR2 was at the
edge of its operating range no true central position was found. From these
studies we concluded that use of an additional quadrupole wire length over
15 cm should be avoided. This length limitation was used as a guideline in

designing the mount and connections for the entire source assembly.

D' n n i ml r t/ﬂn

The matrix isolation instrument was designed to accommodate two
ion sources. Two 6” conﬂat ﬂanges were installed facing the cryostat
window to allow the easy removal of each source for servicing. Following
the construction of the source housing, a 6” mounting ﬂange was designed
and constructed for the modiﬁed Anavac-2 RGA. This mounting ﬂange
had to provide: mechanical support for the ion source, electrical
feedthroughs to provide operating potentials to the source, and a gas inlet
to allow the introduction of sample gas into the source. It was also
important that the operator have the ability to change the line-of-sight
orientation of the source while the chamber was under vacuum. A

specialized pivot-mount was devised to provide this function.

103

 

Figure 3.11 A top view drawing of the complete Anavac-2 ion source ﬂange.

104

The pivot mount used to physically support the source assembly
consisted of a two point bracket (2) and a linear motion feedthrough (1)
shown in Figure 3.11. A 1.9” bracket was installed toward the center of the
mounting ﬂange. This bracket had two arms, each of which contained
setscrews which inserted inward. Depressions were drilled into the
source housing ﬂange to accept these setscrews. By inserting the
setscrews the source housing could be pinned into position, but still have
the ability to move vertically. A high- vacuum linear motion feedthrough
was installed in the mounting ﬂange in a position that allowed the
feedthrough rod to push against the bottom edge of the source housing
ﬂange. By inserting the linear motion feedthrough it was possible to move
the source upward, this motion is shown in the side view insert of Figure
3.11. The source could also be directed downward by withdrawing the
linear motion feedthrough. The weight of the source was generally
enough to cause this downward motion, but a spring could also be
installed to ensure proper positioning.

The source operating potentials were provided by two electrical
feedthroughs installed on the mounting ﬂange. A Ceramaseal submini
octal feedthrough (808B9340-1) was installed on a miniconﬂat feedthrough
mounted above the pivot mount, this is shown as component 5 in Figure
3.11. This feedthrough was used to provide electric potentials for the
source, focus, quad housing, and deﬂection optics elements (3). Flexible,
Teﬂon-coated cable, designed to carry low currents, was used to connect
the feedthrough pins to each of the components. It was also necessary to
provide high voltage connections to the quadrupole and high current
connections to the ﬁlament. As discussed previously, is was important to
minimize the changes made to the original quadrupole circuit, to
accomplish this the original 2 3/4” Anavac-2 feedthrough ﬂange was used
to provide these potentials. The high current/voltage feedthroughs on this
ﬂange were also used to provide high current signals to the ﬁlament. 18

gauge copper wire was added to the original ﬁlament wire, covered with

105

ceramic thermocouple insulators, and connected to the housing ﬂange
feedthrough using Ceramaseal 880A7332-1 connectors. These connections
could readily handle the high current loads, up to 5 amps, necessary for
proper operation of the ﬁlament. To provide connections for the
quadrupole, 10 cm of 22 gauge chromel A was added to the original
quadrupole wire. The ends of the wire were twisted together so that no
connector was added to the circuit, and the length of the wire was covered
by ceramic thermocouple insulators. This wire fed to a loop of the same
wire that connected the Opposing rod pairs as shown in Figure 3.11. For
this, and all the electrical connections, sufﬁcient extra wire length was
used to allow the source to travel through its full range of movement.

A miniconﬂat feedthrough was also used to provide sample gas to
the source housing. A 3” section 1/4” OD SS tubing was installed through
the mounting ﬂange and welded in place. A miniconﬂat ﬂange was
attached to the atmosphere side of this tubing, and a swagelock SS-400-1-1
ﬁtting was welded to the vacuum side of the mounting ﬂange. The inlet
tube on the source housing was attached to a 12” x 1/4” dia. ﬂexible SS
bellows, shown as component 4 in Figure 3.11. This bellows was looped
around and connected to the swagelock ﬁtting on the mounting ﬂange.
This arrangement allowed the direct introduction of sample gas into the
source housing, and was still ﬂexible enough to allow the source assembly

to pivot.

llE-llﬁ' llﬂi l'

53.]..11. r

Following the successful construction of an enclosed housing for the
Anavac-2 ion source, a series of ion optical elements were fabricated to
allow the manipulation of the ion beam produced by the quadrupole onto

the matrix isolation window. SIMION modeling was used to investigate

106

and optimize these designs, but to properly use SIMION, the ion beam
produced by the quadrupole ﬁrst had to be characterized. SIMION PC
version 3.0 was only designed to allow the modeling of 51,311; electric and
magnetic ﬁelds. With the introduction of SIMION PC/PSZ version 4.0,
dynamically changing electric ﬁelds could be modeled through the use of
specialized ‘user program’ routines. Theoretically this version could be
used to model the dynamic electric ﬁelds produced within the Anavac-2
quadrupole, but this would require the determination of many initial
parameters such as the rf phase angle of ion injection, transverse energy
of the ion, initial angle of ion injection, etc., that are extremely difﬁcult to
obtain. Instead of using modeling, an ion beam visualization experiment
was performed to characterize the divergence and shape of the ion beam
produced by the quadrupole. Using a CEMA detector it was possible to
directly observe the size and shape of the ion beam. By placing this
detector at varying distances from the quadrupole exit aperture, the
change in beam size with distance, or beam divergence, was determined.

A Galileo 3025 chevron conﬁguration CEMA detector was purchased
from Galileo Electro-Optics Corporation. The design of this detector
allowed the output of the CEMA to excite 3 P11 phosphor deposited upon a
ﬁberoptic, thus creating an image of the ion beam on the ﬁber optic screen.
The active area of the detector was 25 mm in diameter and the spatial
resolution was 20 um. Gain for this detector was quoted at 1x107 electrons
per incident particle with an operating potential of 1 kV across each of the
microchannel plates. In order to use this detector to visualize the beam
produced by the quadrupole, a shielded mount and movable rail system
were devised.

The front of the CEMA detector is normally operated at a potential of
~-1400 volts. If the detector was placed directly in line with the quadrupole
exit aperture, the trajectories of ions exiting the quadrupole would be
strongly inﬂuenced by this potential. To accurately determine the beam

proﬁle that the quadrupole will produce in the matrix chamber, the ion

107

beam must be allowed to traverse a ﬁeld free region before detection. To
accomplish this, a shield plate was designed and positioned before the
front plate of the CEMA detector. A circular stainless steel ringplate with
a 52 mm OD and 25 mm ID was used as the shield plate. A Buckby-Mears
~80% transparent SS grid was spot welded over the center of the plate and
holes were drilled in the plate to accept the threaded mounting rods of the
CEMA detector. This design allowed ions exiting the quadrupole to pass
through a ﬁeld free region between the quadrupole housing and the grid
plate. Ions that passed through the grid are accelerated linearly toward
the detector, and a spot was produced on the phosphor screen that
corresponded to the position where the ions had penetrated the grid. The
grid plate was attached to the CEMA detector at a distance of 4 mm from
the front plate of the CEMA. The use of the CEMA threaded rods to attach
the shield plate required that a short length of these rods, biased at -1400 V,
appear in front of the shield plate. To investigate how these rods would
affect ion trajectories, a SIMION model of the visualization experiment
was performed. This is shown in Figure 3.12.

This model shows the quadrupole housing exit plate, shield plate
with grid, and the front of the CEMA detector. In this model the shield
plate was positioned at a distance of 10 mm from the quadrupole housing.
The threaded support rods are shown extending 5.2 mm in front of the
shield plate. After the electrode array was constructed, the CEMA
electrode was biased at -1400 V and the array was reﬁned. Contour lines
were calculated at 100 V intervals, these are shown in Figure 3.12. The
contour lines between the shield plate and the front of the CEMA were
extremely linear, which indicated that an ion that penetrated the grid
would follow a linear trajectory onto the detector. The model also showed
the penetration of the electric ﬁeld produced by the support rods into the
“ﬁeld free” region. To determine whether this ﬁeld penetration would
affect the trajectories of ions exiting the quadrupole, a parallel beam of 6

eV ions was modeled exiting the quadrupole. These ions, shown in Figure

108

support
rod

shield
plate

 

 

 

 

quadrupole
exit §§§
aperture \

ion
beam

 

 

 

 

 

 

 

 

.....
.....
.....

ccccc
.....
00000
.....

ooooo
.....
.....

.....
.....
.....
.....
-----
.....
ooooo
.....
.....
.....
00000
.....
00000
.....
.....
.....
00000

 

 

 

 

 

 

 

 

 

 

 

 

 

.....
.....
.....
.....
.....
.....
ooooo
......
OOOOOOOOOOOOOOOOOOOOOOOOOOO
.....

 

ccccc
ccccc

0 volts 0 volts -1400 volts

Figure 3.12 A SIMION model of the shielded CEMA detector mount.

109

3.12, continued on their original parallel trajectories through the region
between the quadrupole housing and the shield plate. This indicated that
the penetration of the electric ﬁeld from the support rods would not
substantially affect the shape of the ion beam exiting the quadrupole.
Several similar models were evaluated, varying the distance between the
quadrupole housing and the shield plate. At distances greater than 15
mm the ion beam began to diverge due to the penetration of the electric
ﬁeld created by the support rods. To reduce this effect, the beam
divergence measurements made in this study were done at quadrupole
housing to shield plate distances 3 10 mm and a grounded copper cage was
added surrounding the support rods to provide additional shielding.

Isolation Grid CEMA

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

» -.1.<. -' «Wt‘ﬁ‘ﬁ‘ﬁ‘c‘t‘ﬁﬁm‘mﬁﬁ ........

Rail Mount

Figure 3.13 The rail mount system used for beam visualization

The shielded CEMA detector was attached to a bakeable ring and
placed on a moveable rail mount shown in Figure 3.13. The source
housing and quadrupole were attached to the rail mount by a bracket that
could slide along the length of the rail. The endcap of the source housing
ﬂange was removed to allow the sampling of the residual chamber gases
by the ion source. The entire unit was then placed in a vacuum chamber
ﬁtted with a plexiglas ﬂange facing the ﬁberoptic of the detector. The

electrical feedthroughs were connected from side ports on the chamber to

110

the source and quadrupole. External power supplies were used to provide
the operating potentials to all components except for the quadrupole. A 10
cm length of 22 gauge Chromel A was used to connect the original
Anavac-2 feedthrough to the quadrupole rods. The chamber was then
sealed, pumped down, and the variable capacitor in the remote electronics
head was adjusted to allow the quadrupole to scan through its entire mass
range.

The standard operating potentials were supplied to the source and
quadrupole, and the ﬁlament was heated with a 2.0 A current. Two
Hewlett-Packard 6516-A high voltage power supplies were used to bias both
the front and center plates. Fluke DMMs were placed between these power
supplies and the plates to measure the bias current across each plate. The
back of the CEMA was connected to ground, and the ﬁberoptic screen was
connected to a Power Designs Inc. 1543A high voltage power supply.
Normal operating potentials for the detector were: front plate = -1400 V,
center tab = -800 V, back plate = 0.0 V., phosphor screen 2 +1600 V. The
detector was slowly brought to these potentials, and a bias current < 0.3
mA was maintained, indicating proper operation of the detector. A
combined air and argon leak was added to the chamber to an ion gauge
reading of ~3x10'6 torr. The quadrupole was operated in both rf-only and
mass selected modes. The beam proﬁles for the peaks at m/z = 18, 28, and
40 were observed at quadrupole housing to shield grid distances ranging
from 1.5 to 10 mm. The images produced at the phosphor screen were
photographed on slide ﬁlm for later evaluation. Figure 3.14 shows the ion
beam proﬁle produced by the quadrupole operating in mass-selected mode
at m/z = 28. The distance between the shield plate and the quadrupole
housing in this visualization was 5.1 mm. The quadrupole was oriented
slightly off-axis from vertical, with the y rod pair (-dc) toward the vertical
direction and the x rod pair (+dc) toward the horizontal direction. The
cross shaped pattern observed was aligned with the quadrupole rod-pairs,

with a shorter arm toward the +dc poles and an longer arm toward the -dc

.355 :53 San— 5 3.53 2395. 05 .3 uSQBov <EBO Q5 3 30:30am oEPE Ewes as? Ed 394$

111

0503
Swan

 

w.” 4%,
520m /
nosmmozm ,

mat $253
@38me

112

.388 a??? .8 438 E mew—33o
8.38 2:55 05 .3 333% <EHU 23 .3 33.5.5 35th 8mm: :3 BE. ma.” 959m

 

113

poles as discussed in Chapter 2.II.A. Figure 3.15 shows the beam proﬁle
obtained with the same experimental setup, but with the quadrupole
operating in rf-only mode. The ion beam proﬁle produced in rf-only mode
should be fairly symmetric because an approximately equal force is
exerted upon the ions by each rod pair. The proﬁle observed in Figure 3.15
showed the expected uniform, semi-circular beam shape.

The ion beam produced by the quadrupole was characterized by
comparing the dimensions of the ion beam proﬁles at various distances.
The slides taken of the phosphor screen were projected, scaled to equal
magniﬁcation using the support ring as a reference, and carefully
measured. This procedure was repeated for each of the distances, and
quadrupole modes, and m/z values selected. The change in these proﬁle
dimensions with increasing distance between the quadrupole housing and
the shield plate was used to determine the size of the beam at a given
distance, and the maximum beam angular divergence. When the
quadrupole was operated in mass-selected mode two divergence values
were obtained, one along the +dc rod pair axis, and another along the -dc
rod pair axis. The results of these visualization studies are shown in Table
3.3.

Table 3.3. Maximum beam divergence values for the ion beam produced by
the Anavac-2 quadrupole.

rf-only angular divergence = 53.8°

m/z = 18 +dc rod pair axis = 19.2° -dc rod pair axis = 19.6°
m/z = 28 +dc rod pair axis = 28.9° -dc rod pair axis = 56.5°
m/z = 40 +dc rod pair axis = 33.8° -dc rod pair axis = 68.9°

114

u “31].; ggoq'7.° .90 o g_ at; o u. -o .91-..) to .

systems

Following the characterization of the ion beam produced by the
quadrupole, a series of ion optical elements were designed to allow for the
focusing and deﬂection of the ion beam onto the cryostat window. The
source housing and quadrupole were installed on the 6” mounting ﬂange,
and the unit was attached to the conﬂat port of the matrix chamber. The
distance that the ion beam must travel from the quadrupole exit aperture
to the the cryostat window was measured to be ~80 mm. The ion optical
system needed to provide the operator with ability to focus the ion beam to
an appropriate size at the cryostat window, and to allow the deﬂection of
the ion beam onto the window if the source was oriented off axis. Two
possible focal sizes were desired depending upon the matrix isolation
experiment. For LIF experiments the laser spot size is fairly small,
approximately a 4 mm diameter region is excited, and the majority of the
ion beam should be focused within this region. The infrared beam
produced by the BOMEM FTIR can be adjusted to ﬁll the entire area of the
1” diameter cryostat window, in addition the central ~3 mm of the IR beam
is not active. For these experiments the beam should be focused within the
full area of the window for optimum overlap with the infrared beam.

Several SIMION models were investigated to determine the best ion
optical system for the manipulation of the ion beams produced by the
Anavac-2 ion source. Two principle components were evaluated, an einzel
lens conﬁguration to allow the focusing of the ion beam, and a series of
deﬂection plates to control the vertical and horizontal positioning of the ion
beam. Two types of einzel lenses were modeled, a two-element system and
a three-element system. The two-element system used the quadrupole
housing as the ﬁrst element of the einzel lens with a varying potential
applied to the ﬁrst added element. In this system both the quadrupole

115

housing and second added elements were held at ground. The three-
element systems added a complete einzel lens after the quadrupole, with
this lens design the ﬁrst and third elements were held at ground and the
central element was varied. Several different versions of these two types of
einzel lenses were evaluated, ﬁve of these models will be discussed here.
The einzel lens SIMION models were constructed using a 5,400
point circular symmetry array that corresponded to a modeling scale of 0.5
mm/grid unit. Different electrode conﬁgurations were constructed and
reﬁned, and the trajectories of 3, 6, 9, and 12 eV ion beams exiting the
quadrupole were calculated. Parallel 2.5 mm diameter ion beams were
modeled to compare the ideal focal properties of the systems, and a beam of
ions with 0 to 40° angular divergence was modeled to evaluate the ability of
the system to control the beam produced by the quadrupole. Three center
einzel element voltages were determined for each model to allow
evaluation of the different systems: 1. the minimum voltage the required to
focus a parallel 2.5 mm ion beam to a point at the cryostat window (see
Table 3.4), 2. the minimum voltage required to focus all the ions in a O-40°
ion beam onto the full area of the cryostat window, and 3. the minimum
voltage required to focus all the ions in a O-40° ion beam to ~3 mm on the
cryostat window. The performance of each system was also evaluated in
terms of focal strength and ion-optical aberrations. These aberrations are
generally indicated by a poor focal spot, or ions that experience unusual
trajectories and x-axis crossovers. A model with low aberration is
generally able to focus a wider range of angular divergence
simultaneously, this is important in the manipulation of a complex ion
beam such as that produced by a quadrupole. These aberrations generally
result from an ion following a trajectory that travels close to an electrode,
resulting in a spherical aberration effect as discussed in Chapter 2.III.A

116

Table 3.4 The dimensions and performance of several SIMION lens
models.

6mm 2mm

1 2 -2.9 V
2 3 6 mm 2 mm -3.3 V
3 2 8 mm 2 mm -6.9 V
4 3 8 mm 2 mm -6.5 V
5 3 8 mm 3 mm -5.9 V

The dimensions and performance of the ﬁve models presented in
this discussion are shown in Table 3.4. Each model was described in
terms of the number of elements that were added after the quadrupole
housing, the size of the aperture in each element, the spacing between the
individual elements, and the minimum voltage that must be applied to the
central element to focus a 2.5 mm parallel 3 eV ion beam to a point on the
window. The ability of each of these models to focus a 0-40° angular
divergence ion beam is shown in Figure 3.16. The minimum voltage
required to focus all ions in a 0-40° ion beam onto the window is plotted
against beam energy in Figure 3.16a and the minimum voltage required to
focus the same beam to 3 mm is shown in Figure 3.16b. A brief discussion
of the characteristics of each of the models follows.

The ﬁrst model, model 1, was a two-element system which used a 6
mm aperture size that matched the quadrupole housing aperture. The
spacing between each element was 2 mm. With a 0-40° ion beam this
model introduced some aberration to the outermost (40°) ion as it traveled
near to the second element. The change in minimum focal voltage with
beam energy for this model was linear, and this effect was found to be
common to all lens models. This relationship proved useful in the
prediction of new focus voltages when the energy of the ion beam was
varied. This model required relatively low potentials to achieve focusing,
had a reasonably low degree of aberration, and because the lens system
was relatively short it could accept a wide range of angular divergence.

117

-50 4

1

~70 -«

Voltage .90 j

 

 

(volts) .
-llO-i
-l30-

'150 ' I ' I ' I ' I I I r I ' I v I I ﬂ

3 4 5 6 7 8 9 10 ll 12

Beam Energy (eV)

 

—9— Model 1
—0— Model 2
-—I— Model 3
—0— Modle 4
—I— Model 5

 

 

 

b) O

00

 

0

 

 

 

‘100 o
. -1
Voltage 200
(volts)

300 ~

400 -i

-500 I ' ﬂ ' I ' I . I ' I ' I ' I ' I
3 4 5 6 7 8 9 10 11 12

Beam Energy (eV)

Figure 3.16 SIMION modeling results for the focusing of a 3 eV, 0-40°
divergence ion beam; a) onto the cryostat window, and b) to a 3 mm region
on the cryostat.

118

Model 1 was found to be the best system for focusing ion beams with a wide
range of energies and relatively high angular divergence to within the full
diameter of the matrix window, making it well suited for FTIR studies.

Model2 was a three-element system that used a 6 mm aperture size
with 2 mm inter-electrode spacing. This system produced the highest
overall focusing of any of the lens systems evaluated because of the
relatively small aperture size and lens length. Model 2 also introduced the
highest degree of spherical aberration in the ions beam because the ions
traveled very near the electrode elements. Because this model required
relatively low potentials and accepted a limited range of angular
divergence, it could prove especially useful for the manipulation of high
energy beams with low angular divergence. Figure 3.17 shows a
comparison of the degree of aberration between model 1 (a) and model 2 (b).
The ﬁgure shows the trajectories of a 3 eV beam of ions with angular
divergence O-40° for each model. The minimum potential required to focus
the entire ion beam onto the window was -10.5 V for model 1 and -10.9 V for
model 2. The ion beam produced by model 1 had a lower degree of ion
crossover which indicated a lower degree of aberration. Model 2 required a
slightly higher voltage in order for the 40° ion to survive, but the overall ion
beam was more strongly focused with model 2 than with model 1.

Model 3 was a two-element system with 8 mm apertures and 2 mm
inter-electrode spacing. Because of the wider apertures, this model had
the weakest focusing of all the systems modeled and therefore required the
highest potentials to achieve focusing. This can be seen in Figure 3.16,
this lens required the highest focusing potentials for each ion beam
modeled. Model 3 introduced a relatively low degree of aberration, but
because of its weak overall focusing was found to be the least useful of the
systems investigated. It could, however, prove useful for low energy
beams with an extremely high degree of angular divergence.

Model 4 was a three-element system with 8 mm apertures and 2 mm

inter-electrode spacing. This system had the lowest degree of aberration of

119

 

 

 

 

 

 

 

':: \
quadnapole center einzel mostat
housmg lens element Window

./

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b)
E I K m
:2 w
a
V-10.9V

0V

Figure 3.17 SIMION models comparing the degree of aberration produced
in focusing a 040° divergent beam using a) model 1 and b) model 2.

120

all the lenses modeled because of the wide aperture size, but also required
high potentials to achieve focusing. Model 4 was found to be the best lens
system for the focusing of low energy beams with relatively high angular
divergence, especially when a smaller spot size is required. Its use with
higher energy beams was limited by the relatively high potentials required
to achieve focusing. Figure 3.18 shows a comparison between model 2 (a)
and model 4 (b) modeling a 3 eV ion beam with 0-40° angular divergence
that was focused to ~3 mm in each model. Because of its stronger
focusing, model 2 required only -23 V to achieve a ~3 mm beam size, while
model 4 required -85 V to focus the beam to 3 mm. Beam focusing with
model 2 was accompanied by a high degree of spherical aberration, shown
in Figure 3.18a as ion crossover points. Model 4 had a much smaller
degree of spherical aberration due to its larger aperture size, and the
focused ion beam was much more uniform, as shown in Figure 3.18b.

As mentioned in Chapter 2.III.A, einzel lenses can be operated in
decelerate-accelerate as well as accelerate—decelerate mode. In general
the operation of an einzel lens in decelerate-accelerate mode requires
much lower potentials than operation in accelerate-decelerate mode. To
determine if this mode of operation could be used to extend the ion beam
energy range of an einzel lens, the performance of model 4 operating in
both these modes was investigated. The ion beam proﬁles produced by
model 4 functioning in accelerate-decelerate and decelerate-accelerate
modes are shown in Figure 3.19a and Figure 3.19b respectively. A 12 eV
beam with 0-40° angular divergence was focused onto the full width of the
cryostat window in both models. A potential of -105 V was required to
achieve this focusing when the lens was operated in accelerate-decelerate
mode, but this mode produced highly uniform focusing. The optimum
focusing for the same lens operated in decelerate-accelerate mode was
achieved at only +18.5 V, but the ion beam experienced strong aberrations,
and loss of ion ﬂux. In general operation in accelerate-decelerate mode

was used whenever possible.

121

 

 

crossover
points

quadrupole center einzel cryostat
housmg lens element Window

 

Figure 3.18 SIMION models comparing the degree of aberration produced
in focusing a 0-40° divergent beam to 3 mm using a) model 2 and b) model 4.

122

a) -105 v
v

:. “4V

 

 

 

 

 

 

 

quadrupole center einzel cryostat
110115138 lens element Window
ion \
crossover 5;;
point , //—4
./ "'"\//

 

 

\ a

. \
ion loss--—"'—""'>

 

0 V +18.5 V

Figure 3.19 SIMION models comparing the degree of aberration produced
in focusing a 040° divergent beam using an einzel lens in a) accelerate-decelerate
mode and b) decelerate-accelerate mode.

123

The ﬁnal model, model 5, was a three-elements system with 8 mm
apertures and 3 mm inter-electrode spacing. This system had the same
overall performance as model 4, with a slightly higher degree of spherical
aberration, but stronger focusing. This lens system could prove to be a
good general lens, with some limitations when working with highly
divergent beams because of the length of the lens system. It should be
especially useful for focusing ion beams with slightly higher energy than
can be practically controlled with a system based upon model 4.

Several general observations were made from the results of these
and other modeling studies: Smaller aperture sizes generally produce
stronger focusing, but also introduced a higher degree of spherical
aberration. Use of a two-element lens design, with the quadrupole
housing as the ﬁrst lens element, required slightly higher potentials, but
allows ions to remain further away from the electrodes. This could prove
especially useful when focusing highly divergent ion beams. Increasing
the aperture size decreased the degree of spherical aberration that was
observed, but also increased the voltage required to achieve focusing.
Increasing the spacing between elements increased the focusing power of
the lens somewhat, but produced aberrations when manipulating highly
divergent beams. Finally, operation in decelerate-accelerate mode
required lower potentials than operation in accelerate-decelerate mode, but
was accompanied by very high spherical aberration and should, for our
purposes, generally be avoided.

A series of deﬂection plate modeling studies were also performed.
Some general conclusions obtained from the results of these studies
include: 1. For each deﬂection direction the use of two plates produced the
strongest and most uniform deﬂection, 2. Operating the deﬂection plates in
push-pull mode produced only slightly more uniform deﬂection than
biasing a single plate and holding the other plate at ground, and 3. The
deﬂection properties of the deﬂection plates modeled were not especially
sensitive to plate length.

124

In general, a pair of opposing 12 mm long deﬂection plates located
10 m away from the beam axis proved satisfactory for beam deﬂection.
In order to simplify the optimization of the system, only one of the two
plates used for each axis was biased and the other was connected to
ground. For best results, each deﬂection axis should have an independent
pair of opposing deﬂection plates for allow full control of the beam position.

Rather than constructing a single lens system for focusing and
deﬂecting the ion beam from the quadrupole, a series of elements were
machined to allow the user to vary the design of the system to the ion beam
of interest. All elements were cut from 0.5 mm stainless steel. Apertures
were drilled along with four holes matching the pattern on the mounting
bracket of the quadrupole housing. Threaded rod was bolted to the
mounting bracket and covered with a ceramic rod. Ceramic insulators
were then placed on the rod and used to separate the lens elements.
Electrical connections to these elements were made using bolts attached to
tabs on the side of each element. Lenses with apertures varying from 5 to
10 mm were manufactured. In general a system using two elements with
6 mm apertures and 2 mm spacing, similar to model 1, was used because
of its ability to focus a wide range of both ion energies and angular
divergence. When stronger focusing was desired a third 6 mm element
was added to reproduce model 2, and if a smaller spot size was desired

with low energy ion beams a lens system similar to model 4 was installed.

EElll 'l'l] I'llr

Following the design and construction of the deﬂection optics, the
overall performance of the Anavac mass-selected ion source was

evaluated. The original EI source was installed inside the source housing

125

and attached to the quadrupole. A focusing and deﬂection system, based
upon the design of model 1 detailed in the previous section, was attached to
the quadrupole housing mount. This optical system consisted of two
additional elements with 6 mm apertures and a spacing of 2 mm. Four
12.5 mm deﬂection plates were mounted after the focus lens, opposing
horizontal elements were mounted 2 mm from the second focus lens
element. The opposing vertical deﬂection pair was positioned 2 mm after
the horizontal deﬂection plates. The electrical and gas connections were
made between the analyzer elements and the feedthroughs as described in
Chapter 3.IV. The entire unit was installed on the right 6” port of the
matrix isolation chamber, and a Keithley 6100 electrometer was connected
to the movable Faraday plate in the chamber. External power supplies
were used to operated all the components of the analyzer except for the
quadrupole. A diagram of the power supply and metering scheme used in
this study is shown in Figure 3.20.

HP 6821
power supply

HP 6281
power supply

Tektronix 501
power supply

 

Figure 3.20 The power supply and metering system used to evaluate the
performance of the Anavac ion source.

126

A Hewlett Packard 6821 high current power supply was used to heat
the ﬁlament with a Fluke DMM connected in line to monitor heating
current. The ﬁlament bias was supplied by a Hewlett Packard 6281
regulated high voltage power supply capable of producing voltages up to
120 V. A Dynascan 2831 multimeter was connected between the bias and
heater power supplies to monitor the total emission from the ﬁlament. A
Tektronix PS501-1 regulated power supply was used to bias the source plate
and a Keithley 169 digital multimeter was connected between the grid bias
power supply and the grid element to measure source current. The 1”
Faraday plate mounted on the linear motion feedthrough was positioned
directly in front of the cryostat window. The ﬁlament was slowly heated to
2 mA total emission at a bias of -67 V, and deﬂection optics were optimized
for maximum current with the quadrupole in rf-only mode.

A study was performed to determine the best operating pressure for
the ion source. Since no direct measure of source pressure was available,
the partial pressure of an introduced gas was evaluated by observing the
primary peak in its mass spectrum produced by a Dycor M100 RGA
installed on the chamber. A plot of mass-selected ion output current
versus chamber partial pressure for the introduction of nitrogen into the
chamber is shown in Figure 3.21.

The nitrogen partial pressure was monitored using the intensity of
the m/z=28 peak in the RGA spectrum. The ﬁlament of the Anavac source
was operated at an emission of 0.5 mA. The electron energy for the study
was optimized to 70 eV, the ion energy was 12 eV, and the source lens
voltage was -40 V. The output of the Anavac source was fairly constant in
the 1-10x10'7 torr range. Within this partial pressure range a mass-
selected current greater than 1 nA was obtained for nitrogen with a
relatively low ﬁlament emission. In general the source was operated at
the lower end (~1x10'7 torr) of this range to minimize the trapping of

neutrals in the matrix.

127

 

2.0

1.5-4

     

m/z = 28
Current 10:
(nA) .

.4

  

 

 

 

0.0 I r t I Iivvv' 1 I VTTVVVI T I v Irfuv' v 1' 1 Vim

.1 1 10 100 1000
Nitrogen partial pressure (x10-8torr)

Figure 3.21 Mass-selected ion current at the cryostat window plotted
against nitrogen partial pressure in the matrix chamber for the Anavac
ion source.

The performance of the quadrupole was evaluated by taking the
mass spectrum of several samples. The samples were introduced through
a Granville-Phillips model 60 precision leak valve and an evacuated inlet
system. Figure 3.22 shows the mass spectrum obtained with an argon
leak. The chamber was evacuated to a base pressure of 2.9x10'9 torr.
Argon was introduced to a partial pressure of 5x10'8 torr and a total
pressure of 3x10“7 torr. The source was operated at an ion energy of 3 eV,
electron energy of 70 eV, a focus voltage of -40 V, and an emission of 2 mA.
The spectrum was taken at a Y sensitivity of 1” = 1x10”10 A, and the
principle peak of of the mass spectrum, m/z = 40, produced a maximum

current of 1.4 nA which exceeded full scale of the XY recorder. The small
peak at m/z = 28 was probably produced by an air leak (N2) in the inlet

system. The peaks at m/z = 18 and 20 corresponding to 1120* (from
residual water in the source) and Ar2+ were fairly well resolved. The

modiﬁed quadrupole appeared to be operating at slightly reduced

128

 

 

 

 

 

 

 

40 (off-scale)
H
20
18
28
e 4 i J. i
10 20 30 4O 50
m/z

Figure 3.22 The Anavac source mass spectrum of an argon leak. detected
at the remote faraday plate positioned in front of the cryostat WlndOW.

129

resolution, most likely due to the additional 80 mm ﬂight path of the ion
beam. Increasing the dc/rf ratio increased the resolution of the unit, but at
the cost of a large loss in ion current.

Another example of the output of the Anavac ion source is shown in
Figure 3.23. This spectrum shows a) the spectrum obtained by the M100
RGA and b) the spectrum produced by the Anavac source at the cryostat
window with the introduction of propylamine. The chamber base pressure
was 1x10'8 torr. A propylamine bulb was prepared, cleaned with 3
freeze/thaw cycles, and connected to the inlet system. Propylamine was
introduced to a total chamber pressure of 1.3x10‘7 torr. The Anavac source
was operated with an ion energy of 3 eV, electron energy of 50 eV, focus
voltage of -40 V, and ﬁlament emission of 2 mA. The maximum ion
current with this leak at an ion energy of 3 eV was ~O.1 nA. This was
increased to > 1 nA by increasing the ion energy to 12 eV. The peaks at m/z
= 30 and 44 were indicative of propylamine. The peak at m/z = 40 was due
to argon in the chamber that was present because this spectrum was taken
during a matrix-isolation experiment. The peaks at m/z = 28, 27, 26 and 25
are also due to propylamine, with some of the m/z = 28 intensity possibly
the result of an air leak. The M100 RGA spectrum showed well-resolved
m/z = 26, 27, 28, and 30 peaks while the Anavac spectrum was only able to
resolve the m/z = 28 and 30 peaks from this series. This slight loss of
resolution in the modiﬁed Anavac source is not likely to prove a limitation
in the deposition of ions unless the selection of ions only one mass unit
from another fragment in the spectrum is required. For most studies the
Anavac source resolution could be reduced further by decreasing the dc/rf
ratio to increase the output current and decrease the required deposition

time.

130

 

 

 

 

 

 

 

 

 

 

 

 

a) :r l
J. 30
l
l
'l 28
1 I
l l
: l
l
.4
i
j 27
l
4' .
. 26 ~ jﬂL_¥\’/\//
‘r 1
.L 1 ﬂ T T I I T I r I I a T I T I 5
22 24 26 28 30 32 34 36 38
m/z
b)
28
30 40
t— i i t‘- 4‘4
10 20 30 40 50
m/z

Figure 3.23 The mass spectrum of propylamine taken with a) the Dycor M100
RGA and b) the Anavac source using the remote faraday plate for detection.

131

BQEME'II'I ll'l

The CEMA detector was also used to observe the beam proﬁle
produced by the Anavac source at the cryostat window. A shielded
mounting grid/plate was constructed that allowed the CEMA detector to be
positioned in place of the cryostat window. This shield was designed to
attach to the second stage of the cryostat, and create a ﬁeld free region
between the Anavac source and the CEMA detector. The mount was
positioned so that the image on the phosphor screen could be observed
through the 2 3/4” window/ﬂange generally used to introduce the BOMEM
IR beam. The shield plate was designed to clear the inside wall of the
chamber by only ~2 cm which reduced ﬁeld penetration of the potentials
applied to the detector. A 40 mm hole was cut in the shield plate at the
normal position of the cryostat window and covered with a Buckby-Mears
80% transparent stainless steel grid. The CEMA detector was mounted
behind the hole and attached to the shield plate with electrically-isolated
threaded rods shielded by copper cages. The CEMA was wired with a 10
M9 resistor series between the two plates and the back of the second plate
was connected to ground. This allowed the CEMA to operate with only two
power supplies, one for the front plate and another for the phosphor
screen. Potentials were supplied to the detector through a high voltage
feedthrough and shielded BNC cables. The chamber was evacuated to a
base pressure of 1x10‘8 torr, and a mixed inlet leak of helium (1x10‘9 torr)
and nitrogen (8x10‘9 torr) was introduced through the source leak valve.
The source was operated at an electron energy of 70 eV, ion energy of 12
eV, and focus voltage of -40 V. The channelplate was operated with -1400 V
on the front plate and +1600 V on the phosphor screen. Figure 3.24 shows
the focused cross pattern produced when the quadrupole was selecting m/z
= 4 (helium). This ~ 3 mm focal point was achieved by applying -15.5 V

potential to the center einzel lens. In this visualization the Anavac source

132

 

Figure 3.24 The mass-selected beam proﬁle for m/z = 4 (He) produced
by the Anavac source on the CEMA mounted at the matrix window.

133

was pointed slightly below the detector, and the application of +5.1 V to the
bottom deﬂection plate was sufﬁcient to successfully deﬂect the beam onto
the detector. The unfocused rf-only beam completely ﬁlled the detector
window, but could be focused to ~ 6 mm through the application of -39.5 V
to the center einzel lens element. The mass-selected m/z = 28 beam from
nitrogen had a much higher angular divergence, and attempts to bring
the entire cross pattern onto the detector resulted in the strong defocusing
of the higher intensity, central portion of the beam. Because of this, a
relatively low central element voltage of - 52V to produced the brightest
image at the detector for the m/z = 28 beam. These studies indicated that
the lens system designed using SIMION was successful in focusing and
deﬂection the ion beam onto the cryostat window. The exact
characterization of the mass-selected ion beam at the cryostat window was
difﬁcult due to the change in beam proﬁle with increasing m/z. In
general, optimization of the ion current through iterative tuning of the
optical elements is probably the best technique to get the strongest ion
current onto the window. Because the ion beam is fairly diffuse,
optimization for FTIR studies should be done using the full size 1” Faraday
plate. In LIF studies a smaller Faraday plate could be installed to provide
a better overlap between the Faraday plate and the area of the matrix that

will be excited by the laser.

[I UNI 15'! won ‘91-. 'OH .1'- 'UIeX..-

A second source was modiﬁed for use with the Anavac-2
quadrupole. A Nermag model 1.68 EI/CI ion source was selected for
modiﬁcation to allow the production of ions using high pressure ionization

techniques. The Nermag source uses the standard Nier conﬁguration

134

with the ﬁlament residing outside of the source region. This also makes
the Nermag source useful for the production of ions from compounds
would decompose when directly exposed to the hot ﬁlament, and for the
use of highly oxidative samples. The Nermag source was designed for
high pressure use, and therefore few changes were made to its design.
The source is small and cylindrically symmetric making it ideal for
coupling with the Anavac quadrupole. A circular repeller located at the
rear of the source also serves at the sample introduction port. To allow the
use of the source with direct insertion probes, a ~1/4” aperture was located
on the side of the source housing. To decrease the conductance from the
source, this port was sealed with an aluminium plug. The quadrupole
end of the Nermag source is sealed by an endplate that can be removed for
source cleaning. To further decrease the source conductance, a series of
matching endplates were manufactured with reduced ion exit apertures.
The repeller inlet tube was also modiﬁed to allow a Teﬂon line to connect
the source inlet to the swagelock gas ﬁtting on the mounting ﬂange.
Machined ceramic plugs were installed on the side of the source housing
and countersunk to receive the setscrew pins of the pivot mount. These
ceramic insulators were used to allow the entire source housing to be
biased at the ion energy. Finally a series of additional lens elements of
varying aperture sizes were fabricated to allow for the adaptation of the
source optics to the system of interest.

A 0.4 mm aperture source endplate was installed on the Nermag
source, and the Anavac quadruple was attached to the unit along with the
original 3 mm focus and quadrupole entrance apertures. A Galileo model
4700 channeltron detector was adapted to allow it to connect to the exit
mount of the quadrupole housing and installed. The entire unit was then
installed on the 6” mounting ﬂange. The ﬁlament and quadrupole were
wired as described in Chapter 3.IV, and the source housing, repeller,
collector, and focus potentials were wired to the 8-pin mini-conﬂat

feedthrough using Teﬂon-coated ﬂexible wire. A Baratron type 170M

135

capacitance manometer head was attached to the inlet line of the
mounting ﬂange to provide a relative measurement of the pressure in the
source. The mounting ﬂange was attached to a special 11” adaptor ﬂange
that allowed it to be installed on the 11” ﬂange of an o-ring sealed chamber.
All the Nermag source studies were performed on an aluminum, o-ring
sealed test chamber that was modiﬁed from the early Enke triple
quadrupole instrument. This test chamber was pumped by a Varian M4
diffusion pump ﬁtted with a liquid nitrogen trap. Base pressure for this
chamber was generally on the order of 5x10'7 torr.

Following the installation of the source mounting ﬂange on the test
chamber, high voltage BNC feedthroughs mounted on the side ports of the
chamber were connected to the bias and output lines of the channeltron
detector. The ion source was operated using external power supplies for
all the signals except the quadrupole circuit. The power supply and

metering scheme used in these experiments is shown in Figure 3.25.

._ +1 Keithley 480
l Tektronix 501 I W

power supply
- +1»
I—l Tektronix 501
power supply gas inlet-D

repeller
I ﬁlament
+ _ Keithley
HP 6281 169 MM
power supply a

Figure 3.25 The power supply and metering system used with the Nermag
source.

 

   
   
  

 

.1,
collector

 

 

  

 

 
 
    

 

 

 

source
HP6821 aperture
power supply plate

 

 

 

136

The original ﬁlament supplied with the Nermag source was heated
by a Hewlett Packard 6821 high current power supply. Filament heater
current was monitored with a Fluke 23 DMM connected in-line with the
heater power supply. A Hewlett Packard 6281 regulated high voltage
power supply was referenced to ground and used to supply the ﬁlament
bias potential. Total ﬁlament emission was monitored by a Keithley 169
multimeter connected between the ﬁlament bias and heater power
supplies. A regulated Tektronix 501 variable power supply was used to
supply the collector and source bias potentials. Collector current was
monitored with a Keithley 480 picoammeter placed between the collector
and the Tektronix power supply. The channeltron detector bias voltage
was supplied by a Hewlett Packard 6516A high voltage power supply, and
the detector output was monitored with a Keithley 610C electrometer. The
other dc potentials required for the operation of the system were supplied
by an unregulated variable power supply constructed in—house.

The performance of the Nermag source/Anavac quadrupole
combination was ﬁrst evaluated by operating the source in El mode. The
test chamber was evacuated to a base pressure of 7.8x10'7 torr, and a
mixture of argon and carbon tetrachloride was introduced through the
inlet system. The sample pressure was increases until a capacitance
manometer reading of 90 mtorr was obtained which corresponded to
partial pressures of 9.2x10'6 and 3.7x10'6 torr of argon and CCl4. The
operating parameters of the source were optimized, and the best resolution
was obtained at ; source = +3.0 V, repeller = +16 V, lens = - 35 V, collector =
+5.02 V, x-ray plate = +54.8 V, and detector = - 2.7 kV. The ﬁlament was
biased at -67 V, setting the electron energy at 70 eV. A heater current was
applied across the ﬁlament until an total emission of 50 uA was obtained.
This corresponded to a collector current of 0.197 uA. With the Anavac
control unit in rf—only mode a current of 1.75x10'6 A was measured at the
detector. The mass spectrum obtained with a mass scan from m/z = 1-60 is

shown in Figure 3.26. The dc/rf ratio of the quadrupole was optimized for

137

.0958 aha wag—52 @3558 a 5:5 33¢ 233.628 053.8 2:
9:2. 63:69:“ 03.82353 55.80 van coma mo 9:5me a .3 85.50on mme 9E. mud 95mg

up:
cm on ov om cm 2

1%“? 3 3 2

mm om

J-o

)

1P

 

 

mm.

 

 

ov

138

highest resolution before the scan. The resulting mass spectrum showed
very good resolution, with components from Ar, CCl4, and air resolved to
approximately unit mass resolution. The peak intensity for this study was
found to be strongly dependent upon dc balance between the quadrupole
rod pairs, with the highest output occurring at a slight negative offset.
This sensitivity probably corresponded to an optimum alignment of the ion
beam in the quadrupole with the quadrupole exit aperture.

A methane leak was introduced into the source region to evaluate
the high pressure output of the source. Figure 3.27 shows the mass
spectra produced at: a) manometer pressure = 26 mtorr and b) manometer
pressure = 400 mtorr. The negative deﬂection of this mass spectrum was
due to the output signal from the detector and could not be reversed using
the XY recorder available. The source was operated under the same
conditions as the previous study, and the total pressure in the chamber in
study a) was 1.7x10'6 torr. The total ﬁlament emission was regulated at 50
uA at an electron energy of 70 eV. The output from the detector in study a)
with the control unit in rf-only mode was 1.3x10‘6 A. Spectrum a) showed
the well resolved methane pattern of m/z = 12, 13, 14, 1516 and a small peak
at m/z = 17 indicating the onset of CI conditions. In spectrum b) the total
pressure was increased to 6.2x10’6 torr which produced an rf-only current
of 3x10'6 A. This spectrum also showed peaks at m/z = 12, 13,14 ,15, and 16
along with the peaks at m/z = 17 and 29 indicative of high pressure
methane CI. All the peaks in this spectrum showed strong splitting. This
splitting was ﬁrst assigned to the imbalance of the quadrupole dc levels.
After investigation, this peak splitting was linked to the source pressure,
indicating that it is the result of ion/molecule collisions in the quadrupole.
The source pressure was reduced until this peak splitting disappeared,
and an expanded scale spectrum was taken at a manometer reading of 240
mtorr. This spectrum is shown in Figure 3.28, and clearly showed the

presence of peaks at m/z = 17 and 29 that were indicative of the formation of
CH5+ and C2H5+ in methane CI. To increase the intensity of these peaks

139

6958 “SH weasel poem—58 a 53, 635 333va 22:28 05 9:3 32.33
ESE 8v 3 93 ESE mm 3 .8 3:325 5038 a 3 28558 .8 85.58% 358 2C. bad «Suﬁ

 

:—

mm

 

 

 

 

 

 

 

 

 

 

 

 

 

3
r E
S 2
mm
1 1 . Leif L; C
mm
m
an
E
.. :5 a
L 111
I r l . + x u.
8 2. an cm 2 c

140

.5958 mozm manta Z FEES: a 5:5 v3.5 225.653. 8225. 9:. maﬁa :an ova mo
c.3303 08:8 a up Bounces 9:932: no 85.50on 358 9: do 33> cavemaxo :< mud 5.59““

m: n—

in

Eases

 

 

z

 

 

:

 

 

aw . mm

N~

t. In" lull...l.‘

141

the pressure at the manometer was slowly increased to > 1 torr and the
performance of the quadrupole was observed. Increasing the pressure did
increase the intensity of the m/z = 17 and 29 peaks, but was accompanied
by a severe loss in resolution. At a manometer pressure of ~1 torr these
ions began to dominate the spectrum, but the chamber pressure at this
leak rate exceeded 1x10‘4 torr. It was concluded that even with the
reduction of the source apertures, the conductance from the source was too
high to create true CI conditions within the source region. Increasing the
pumping speed in the chamber could improved the performance of the
system, but for this source to operate under true CI conditions the source
should be adapted for some form differential pumping that is generally
used in CI.

The performance of the Nermag source under electron capture
conditions was also investigated. The channeltron was modiﬁed to allow
for the detection of negative ions. The x-ray plate and the detector entrance
aperture were isolated, and separate leads were connected to each
element. A potential of +3.5 kV was applied to the x-ray plate, -3 V was
applied to the detector entrance aperture, and the channeltron was
operated at -2.7 kV. Negative ions exiting the quadrupole were strongly
attracted to the high positive potential on the x-ray plate and impacted
upon the plate at high energies. These high energy collisions resulted in
the production of secondary positive ions from the surface which were
detected at the channeltron. The conversion process is not well
understood, but has been shown to occur at relatively high efficiency for x-
ray electrode potentials of > +3 kV [129]. To allow the production of
negative ions the source, repeller, and focus potentials were reversed, and
the collector was held at the source potential. Figure 3.29 shows an
example of the EC-NI mass spectra produced with the Nermag source
operating under these conditions. Carbon tetrachloride was introduced
into the source to a manometer reading of 400 mtorr, and the source

parameters were optimized for ion current in rf-only mode. This

.0953 moan.“ 3:52: a 515 333357 2328 one main auto—£0833 95.80
515 353% 85.50on $9: ”83332 Eomﬁonu 03.33: 2393 28.583 2 mad 0.59m

 

 

mm.
2
M
5
ha a .F v .«r «I m
cm cm ov om cm 3 o

143

spectrum was obtained with the operating potentials: source/collector = -20
V, lens = +70 V, repeller = -45 V, electron energy = 100 eV, and total
emission = 60 uA.

The ions produced under these conditions had a transverse energy
in excess of 20 eV. This resulted in the poor resolution observed in the
mass spectrum. The maximum detector signal obtained with the control
unit operating mass-selected mode was ~ 1 nA. At source potentials below
20 eV the ion signal was lost in the baseline noise. Several mixtures of
C014 with methane and argon buffer gases were investigated in attempts
to improve the output of the unit with little success. Ammonia and
nitromethane were also used as electron capture species, but no negative
ions were observed with either of these gases. To determine the ion
current produced under these conditions prior to multiplication, the
channeltron detector was replaced with a shielded Faraday plate. The
mass-selected ion currents measured at this Faraday plate under similar
conditions were less than 1x10'12 A. These currents were too low to be
used in the matrix isolation studies of interest. The primary reason for the
poor efﬁciency of electron capture ionization using this source was
probably insufﬁcient buffer gas pressure in the source. This was
compounded by the decreased ion extraction through the small ion exit
aperture that had been installed to reduce the source conductance. To
make the Nermag source practical for use in EC-NI mode some
differential pumping system must again be devised.

Following the high pressure studies, the maximum ion output of the
Nermag source was investigated to determine its usefulness for the
production of ions from thermally labile and oxidative compounds. The
entire source unit was disassembled, and the stock ~3 mm aperture
source, focus, and quadrupole entrance plates were replaced. To optimize
the total output of the quadrupole, the 3 mm limiting aperture plate at the
exit of the quadrupole was removed and the quadrupole dc/rf ratio was

adjusted to produced minimum resolution and maximum output. A

144

shielded Faraday plate was attached to the end of the quadrupole to allow
the direct measure of ion output. The unit was wired as described
previously, and installed on the test chamber. Figure 3.30 shows a mass
spectrum produced from a nitrogen leak. The operating potentials on the
systems were; source .= +3.0 V, repeller = +16 V, lens = -35 V, collector =
+5.0 V, and electron energy = 70 eV. The base pressure of the chamber
was 2.4x10'6 torr, and nitrogen was introduced through the source inlet to
a chamber pressure of 7.1x10'6 torr. The ﬁlament was heated to 1.7 A
which produced an emission current of 14.8 uA and a collector current of
13.95 nA. The ion current at Faraday plate with the control unit operating
in rf-only mode was 1.8x10'8 A. The maximum m/z = 28 current for the
spectrum shown was 8.25x10'10 A. The resolution of the system in this
study had been substantially reduced, but the source is still capable of
resolving peaks at m/z = 28 and 32. The ion output was increased by
increasing the ﬁlament heater current. A maximum m/z = 28 signal of
5.8 nA was detected on the Faraday plate with a heater current of 2.10 A.
Figure 3.31 is a plot of ion output current versus total ﬁlament emission
under these conditions. Both the rf-only current and m/z = 28 maximum
currents are plotted. The output of the source sharply increased with
increasing ﬁlament emission from 0-50 uA. Above 50 uA emission the
source output leveled off. In general, operation of the N ermag ﬁlament in
the total emission range from 100-200 uA yielded a high, reproducible ion
output current at the Faraday plate.

In conclusion, although the attempts to produce high pressure
spectra using the modiﬁed Nermag source were fairly unsuccessful, the
operation of the N ermag source in El mode produced relatively high ion
output currents. The external ﬁlament design of this source should make
this source especially useful for the production of ions from compounds
that would decomposed when directly exposed to the hot ﬁlament.
Oxidative compounds, such as NO, could also be used with this source for

145

 

 

 

 

28
14
M L
I J t 1 #1
T I T 1 7
10 20 30 40 50 60

Figure 3.30 The mass spectrum of nitrogen obtained using the Nermag source and
Anavac quadrupole with current ouput maximized, maximum mlz 28 current = 5.8 nA.

146

much longer periods than would be possible using the original Anavac

 

  
 
 
 
 
 
 
 

 

source.
1.06—8 '-
4
8.09-9 -*
Faraday
Plate 6.06-9 -'
Current
(A)
4.09-9 4
209-94 a rf-only current
9 m/z = 28 current

 

 

 

 

 

I I I I I I l T I U U

on so 100 15'0""200
Total Filament Emission (uA)

' V I 1 I

Figure 3.31 The output of the Nermag source versus total ﬁlament
emission.

One area of current research in the Allison group is the use of
thermionic emitters for the ionization of thermally labile compounds.
Glasses containing potassium and sodium have been shown to produce
strong signals of K+ and Na+ respectively upon heating [68]. The use of Cs
glass surfaces as a means of producing negative ions through negative
surface ionization has also been investigated [71]. A series of experiments
were performed to determine if the Anavac source could be modiﬁed to
produce negative ions through NSI using these low work function Cs

surfaces.

147

To study the output from a Cs glass surface, total surface emission
experiments were performed on a vacuum rack test chamber. A specially
designed test chamber was used that contained high current feedthroughs
and could be attached to a standard vacuum rack. To compare the
emission from a Cs glass surface against the emission of a rhenium
surface two emitters were prepared. The emission of 0.007” diameter bare
rhenium wire was compared to that of a Cs glass surface deposited upon a
0.007” rhenium wire. These emitters were prepared by threading the
rhenium wire through a ~1/16” diameter, ~1/4 long 2-hole ceramic rod that
was used to separated the lengths of wire. A small (~1/16”) loop of wire
was left at one end of the ceramic tube. The Cs glass surface was prepared
from a 2 Si02 : 1 A1203 : 1.58 CsNOz mixture (by weight) that was
suspended in a slurry of acetone. This glass slurry was deposited on the
rhenium wire loop and the acetone was removed by heating with a bunsen
burner. The heat from the burning acetone also caused the glass mixture
to harden sufﬁciently to remain attached to the wire loop. Following
preparation of the two emitters, the free ends of the rhenium wire were
attached to alligator clips on the chamber feedthrough wires. To allow the
measurement to emission current from the emitters, a ~1” square collector
plate was attached to another feedthrough wire and positioned opposite the
emitters. The chamber was then assembled and evacuated to a pressure of
~1x10'4 torr.

The power supply and detection scheme used in this experiment is
shown in Figure 3.32. A Hewlett Packard 6821 high current power supply
was used to heat the emitter, and a Hewlett Packard 6281 regulated power
supply was used to supply the emitter bias. A Keithley 417 picoammeter
was attached to the collector plate lead to measure the total emission from
each emitter. A bias of -3.0 volts was applied to each emitter, and the
rhenium ﬁlament was slowly heated while monitoring emission current

at the collector plate. Approximate temperature measurements were also

148

made using an optical pyrometer. A plot of the collector plate current
versus heater current for each of the systems is shown in Figure 3.33.

 

   

 

 
  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' Keithley 417
HP 6821 Picoammeter
power supply
+
Collector
- Plate
HP 6281
power supply Rhenium
+ ‘ \ Filament
Vacuum Rack
.1 Loop

Figure 3.32 The experimental setup used to evaluate the emission of the
cesium glass surface.

6000 "
. —9— Bare ﬁlament emission (nA)

5000 . —'°_ Cs Bead Emission (nA)

 

 

 

Plate

Current
(nA)

 

 

 

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
Heater Current (A)

Figure 3.33 Collector plate current versus heater current for both a bare
rhenium surface and a cesium glass bead formed on a rhenium wire.

149

Emission from the bare rhenium wire began at a heater current of
~1.8 A and increased strongly at ~2.2 A which corresponded to a surface
temperature ~1255 °C. The output of the bare wire steadily increased from
2.2 to 3.0 A with a collector current of 2 uA produced at 3.0 A (temperature
~ 1750 °C). Emission from the cesium glass surface began at 1.6 A and
increased strongly at 2.0 A, Which corresponded to a wire temperature of
~1100°C. The emission from the cesium surface increased more rapidly,
with a collector plate current of ~6 uA produced with a heater current of
only 2.5 A (wire temperature ~1420 °C). The earlier onset and increased
overall emission from the cesium bead indicated that this surface had a
lower overall work function than the bare rhenium surface. A leak of
acetonitrile was introduced into the system and the experiment repeated.
With acetonitrile in the system the onset of emission from the cesium
surface occurred at a lower heater currents and produced a stronger
signal, indicating the possible production of negative ions from NSI.

To determine if the emission signal from the cesium surface was
caused by negative ions, or simply electrons, a similar experiment was
performed using a mass spectrometer. Previous work in the Allison lab
had resulted in the modiﬁcation of several mass spectrometer direct
insertion (DI) probes to accept this type of bead for K+IDS studies. The
modiﬁed DI probes contained electrical feedthoughs and leads to allow the
biasing and heating of the bead support ﬁlament. Another cesium bead
was prepared and spot welded to the leads on the end of the DI probe tip.
The probe was inserted into the EI/CI source of a Hewlett Packard 5985
quadrupole mass spectrometer and aligned with the exit aperture of the
source. The HP5985 used a Galileo 4700 series channeltron detector that
had been previously modiﬁed to allow detection of negative ions. This was
operated at an x-ray plate bias of +3.5 kV and a channeltron bias of -2.2 kV.
The base pressure in the source region was 3x10'6 torr. Carbon

tetrachloride was introduced through the sample inlet to a source

150

chamber pressure of 6x10'5 torr. A bead bias of -3.0 V was applied, and the
total ion current at the detector was optimized at a bead heater current of
2.3 A. The position of bead was also adjusted to create a maximum total
ion signal at the detector. The quadrupole was scanned from m/z = 0-200,
and the mass spectrum stored. Figure 3.34 shows the average of 290 scans
with a total ion current of 6010. The only ions observed from m/z = 0-200
were at 35 and 37 resulting from the formation of 3501' and 37C1'
respectively. This NSI production of strong signals of Cl' from the cesium
surface indicated that these cesium surfaces could prove useful for
negative ion beam formation using the Anavac source. Other gases were
investigated for negative ion production including acetonitrile and
nitromethane, but the strongest negative ion output was obtained with
CCl4.

Following these studies, the original Anavac EI source was
modiﬁed for use as a surface ionization source. This was accomplished by
removing the source grid plate and installing a mount for the cesium
bead. A high current feedthrough from a gas chromatograph thermal
conductivity detector (TCD) was spot welded in the center of the ﬁlament
housing endplate as shown in Figure 3.35. The free ends of the cesium
bead were spot welded to the inside leads of the TCD feedthrough, allowing
the bead to be positioned directly over the source exit aperture. The outside
leads of the TCD feedthrough were connected to two of the pins on the
source housing ﬂange feedthrough. The ﬁlament housing could be biased
separately from the bead using another lead on the source housing
feedthrough ﬂange.

A SIMION modeling study was performed to determine the best
operating conditions for the Anavac NSI source. The dimensions of the
source assembly were measured and a 9,200 point cylindrically symmetric
array was used to construct a model of the source electrodes. This array
was reﬁned, and the potentials on each element were optimized to produce
high ion transmission for 3 eV negative ions emitted from the bead

151

 

 

 

 

 

35
Relative _
Intensity 37
llllIll—TTITIIIITll—IITWIIII”[Trl‘l llll lllllllllllllllllllllllll‘l‘l—I’l
O 10 20 30 m/z 40 50 60 70

Figure 3.34 The mass spectrum produced from NSI of CC14 on a cesium
surface, this spectrum was taken on an HP5985 mass spectrometer.

 

    

Thermal Conductivity
Detector Feedthrough

 

 

 

 

Figure 3.35 The original Anavac source adapted for surface ionization.

152

surface. The resulting SIMION model is shown in Figure 3.36. The best
compromise between ion transmission and ion energy was obtained with
the potentials: ﬁlament housing = -9.0 V, bead = -3.0 V, source plate = 0 V,
and lens = -40 V. The source plate was maintained at 0 V to simulate the
use of a picoammeter to monitor the source plate current. The model
Figure 3.36 shows 5 V contours from 0 to -40V and 2 V contours from 0 to
10 V. Strong focusing ﬁelds were created by the source einzel lens similar
to that observed in previous source models. The ﬁeld lines behind the bead
illustrated the ﬁeld created by applying a -9 V potential to the source
housing. This ﬁeld helps push ions toward the eidt aperture of the source.
The ﬁeld line in front of the bead showed the surface extraction ﬁeld
present in the system which aids in the detachment of an ion from the
bead surface. The trajectories of ions emitted along the bead surface are
also shown in Figure 3.36. Ions formed along the sides of the bead failed to
exit the source aperture, but ions formed along the front of the bead
successfully exited the source aperture and were injected into the
quadrupole.

Using these modeling results, a study was performed to optimize the
operating conditions of the Anavac NSI source. The output of K” from
potassium glass surface is very strong, and this system was selected to
allow the easy optimization of source conditions. Once these operating
potentials were determined they were reversed for use with NSI. A
potassium glass bead was prepared using the methods described earlier.
A 0.007” rhenium ﬁlament was looped through a 1/4” length of 2-hole
ceramic rod and spot welded to the TCD feedthrough. A mixture of
potassium glass, 1 K20 : 1 A1203 : 2 Si02 by weight, was suspended in an
acetone slurry, deposited on the wire loop, and heated with a bunsen
burner. The grid plate was removed from the source and the ﬁlament
housing and endplate containing the bead were installed, allowing the
ﬁlament housing to be electrically isolated. In this conﬁguration the bead

surface was located directly over the source exit aperture at a distance of

153

ﬁlament source lens quadrupole
hmxsmg pla - 1 housing

CODIOUI‘S

 

ion
trajectories

 

3:, : ...................................
.9,o Volts 40 Volts 0.0 Volts
0.0 Volts

Figure 3.36 A SIMION model of the Anavac source modifed for surface ionization.

154

~ 2 mm. The entire source was assembled, wired, and installed on the test
chamber. External power supplies were used to bias the components, the
diagram of the power supply and metering setup used in this study is
shown in Figure 3.37.

HP 6281 power supply

HP 6821
power supply

Tektronix PS501

”we" supply Keithley 417
picoammeter

 

Figure 3.37 The power supply and metering setup used to operated the
Anavac surface ionization source.

A Hewlett Packard 6821 high current power supply was again used
to head the bead ﬁlament and a Fluke 23 DMM was used to monitor heater
current. A Tektronix 501 regulated power supply provided the bead bias
and a Dynascan 2831 multimeter was connected between the bead bias
supply and the emitter to measure total bead emission. A Hewlett Packard
6281 regulated power supply was used to supply the ﬁlament housing
potential. A Keithley 417 picoammeter was connected to the source plate to
monitor source current, and a Keithely 610C electrometer was connected to
the shielded Faraday plate at the end of the quadrupole to measure ion
current produced by the quadrupole. Other operating potentials were
supplied by an unregulated variable voltage power supply not shown.

155

The bead support ﬁlament was heated and the ﬁlament housing and
lens potentials were optimized for maximum rf-only output with a bead
bias of -3.0 V. The control unit was then switched to mass scan mode and
a mass spectrum was taken from m/z = 1-60. The spectrum recorded on
an XY recorded is shown in Figure 3.38a. The operating conditions for
this study were: source = +3.0 V, ﬁlament housing = +9.0 V, and lens = -40
V. The ﬁlament heater current was 0.95 A which produced a total
emission of 0.972 uA, and source current of 140.7 nA. The rf-only signal
obtained prior to this scan was 1.31x10‘8 A and the maximum m/z = 40
current of 1.46x10'9 A was measured. This scan was obtained with the 3
mm quadrupole exit aperture in place, the dc balance on the quadrupole
offset with +dc 10% above -dc, and the dc/rf ratio set to a central position
within the range of VR2. A series of source output measurements were
performed at different heater currents, and a close relationship between
source current and mass-selected output current was observed. This
relationship proved useful for monitoring the output of the source during
matrix isolation experiments. The strong output from this potassium
surface allowed the easy optimization of source operating conditions, and
could prove useful if matrix isolation studies are performed that require a
strong current of K” ions.

The optimum source operating conditions determined using the
potassium emitter were reversed for use in negative surface emission. A
cesium bead was prepared by applying 2 8102 : 1 A1203 : 1.58 CsN02
mixture suspended on a acetone slurry to another 0.007” Re wire loop.
This was spot welded to the TCD feedthrough and installed in the source.
The unit was attached to the test chamber and the same power supply and
metering setup was connected, with the polarity of the outputs reversed.
The chamber was pumped down to a base pressure of 1.6x10'6 torr and a
leak of 0014 was introduced through the source inlet to 5x10‘6 torr. The
ﬁlament was heated and the source operating potentials were optimized
for maximum rf-only output. The optimized operating conditions were:

 

 

 

 

 

 

 

 

 

 

391‘»
a)
41K+
6 I 1?) 20 L 3b 4b ' ~o so
m/z
b ) 35Cr
37C['
0 3 12) 2'0 ' 3o : «to F so ' R
m/z

Figure 3.38 Surface ionization spectra produced by the modiﬁed Anavac ion source
showing a) the output from potassium glass, and b) the output from Cs glass with CCl4

157

bead = -2.0 V, ﬁlament housing = -13.7 V, and lens = +43 V. The control
unit was switched to mass scan mode and the mass spectrum was
obtained, which is shown in Figure 3.38b. This scan was taken at a scale
of 1” = 0.1 nA, with a heater current of 1.9 A which produced a source
current of 10 nA, and total emission of 0.1 uA. Pressure and heater
current optimization studies were performed, and a mass-selected output
in excess of 1 nA was produced. Once again the current at the source plate
proved to be a good indicator of both the rf-only and mass-selected outputs
from the source. At higher ﬁlament emission levels the total ﬁlament
emission also proved to be a good indicator of source output, and was

somewhat more stable indicator than source plate current.

 

—'9— Cs bead initial

-8 —I— Cs bead exposed

 

 

 

 

 

10
—9— Cs bead 02W

10 -9 —9— Bare Re ﬁlament

10 -10
rf-only 10 .11
Current . . ..

(A 0 °
) 10 '12 . °
10 -13
10 -14 I if I ' I ' I ' I f I ' I ' I ' r ' I ' l
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3
Bead Heater Current (A)

Figure 3.39 A comparison of the NSI rf-only output for several surfaces.

After several hours of exposure to CCl4 the maximum output from

the source decreased to ~02 nA. This decrease in performance was

accompanied by a decrease in source current. Increasing the ﬁlament

158

accompanied by a decrease in source current. Increasing the ﬁlament
heater current only increased the mass-selected ion output nominally. To
attempt to regenerate the surface, the aged bead was heated and exposed to

oxygen at a pressure of 3x10’7 torr for 10 minutes. The output of the
surface under a CC14 leak was then tested. Finally the output of the fresh

Cs surface, aged Cs surface, and 02 treated surface were compared to that

of a bare rhenium wire installed in the source and operated under the
same conditions. The results of these studies are shown in Figure 3.39
which shows the rf-only output for each surface plotted against ﬁlament
heater current. The highest output was observed with the fresh cesium
bead. After several hours of exposure to CC14 the output from the same
surface decreased to the level shown as ‘Cs bead exposed’. This same
surface was exposed to oxygen which improved its output at higher heater
currents and is shown as ‘Cs bead 02 treated’ in Figure 3.39. Finally, a
bare 0.007” rhenium loop was studied and its output observed to be much
lower than any of the cesium surfaces.

A study was also performed to determine the long term stability of
the N SI source using a cesium surface, the results of this study are shown
in Figure 3.40. The source was operated at: bead bias = -2.5 V, ﬁlament

housing = -10.0 V, and lens = +35 V. Prior to this study a new cesium bead
was installed, heated, and exposed to CCl4 for ~ 1 hour. The base pressure

of the system was 1.6x10'6 torr and CC14 was introduced to 5x10’6 torr. The

ﬁlament heater current was maintained at 2.6 A to simulate unmonitored
operation of the source. The output of the source was monitored for over
one hour and the source current, rf-only current, and m/z = 35 current
were measured at 10 minute intervals. Figure 3.40 shows plots of a) rf-
only and source current with time, and b) m/z = 35 and source current
with time. The source current was somewhat unstable with time, and the
rf-only and m/z = 35 currents reﬂect this change in total surface emission.
Although the source output was not very stable with time, mass-selected

ion currents in excess of 0.2 nA were produced over the length of the study.

159

This study was repeated, and by manually regulating the source current
both the rf-only and m/z = 35 currents could be maintained at a relatively

constant level.

These preliminary studies indicate that negative surface ionization

will prove useful for producing mass-selected negative ion beams for

matrix isolation experiments. Further studies to improve the output and

stability of these systems need to be made.

Some areas that could be

investigated to improve the performance of these sources include:

1.

Investigating the inﬂuence of bead alignment with the source
aperture.

Studying the use of other types of surfaces such a LaB6 and Th-Ir
with other gases and/or solids.

. Investigating the effect of treating the surface with oxygen as a

means to regenerate the surface while under vacuum.

Changing the ion exit aperture size to increase the total ion
output from the source

Investigating the use of a ﬂoatable picoammeter to allow the
application of a bias to the source plate which could be used to
create stronger electric ﬁelds at the surface of the bead.

100

 

 

 

 

   

 

 

 

 

 

 

 

 

 

1 013-8 3000
a 3 rt' -only Current (A)
9 Source Current (nA)
8.0c-9 '
25C!)
rf-0n]y 5-08'9 Source
current 2000 Current
(A) 4.0e-9 (DA)
0 a 1500
2.0.9 I
0.0e+o . . . 1 , . , . r - r . , . , 1000
O 10 20 30 4O 50 60 70
Time (minutes)
b) 8.0e-10 - F 3000
, mlz=35 max current (A) .
o .
70:40 4 Source Current (nA)

_ source
[NZ-33 current
current (HA)

(A)
2 “'10 I Y I v I v I v I r I v T v I”)
0 10 20 30 40 50 60 70

Time (minutes)

Figure 3.40 Time stability data for the production of negative ions
from a Cs glass surface with the modiﬁed Anavac source monitoring
a) rf-only and b) m/z = 35 currents with time and source current.

161

Chapter 4. Dycor M200 Ion Source Modiﬁcations

I. 1‘ 0‘ ° 0.... O‘IOIH-J ‘ o ‘ U 0, U NH

The second residual gas analyzer selected for modiﬁcation was a
Dycor Ametek model M200 RGA. This RGA was designed to analyze
background residual gases from pressures of 1x10'3 to 1x10'11 torr. The
Dycor M200 has a larger quadrupole than the Anavac-2 and produces
higher amplitude rf signals which allow it to mass-select ions from m/z 1-
200. The source sensitivity of the unit is quoted at 2.0x10'4 A/torr for
nitrogen, taken at a resolution of 10% valley. The minimum peak width
the Dycor M200 can achieve is 0.5 amu at 10% valley with a drift of <01
amu per hour following warmup. This RGA was originally purchased for

~$5,000, and was selected primarily due to its low cost and extended mass

  
   
    

range.
maChined electron 2 3/4"
source ceramic suppressor mounting
repeller plate ﬂange

   

...............
............

..................................................
........................................
................................
.............................

...............
..................
...................
...............
...........
.....

 
  
  
 

quadrupole

focus support Faraday

wires plate

ﬁlament
electrical

feedthroughs

Figure 4.1 A block diagram of the Dycor M200 analyzer head.

162

The RGA consists of four units; the analyzer head, the power
supply, the electrometer/ampliﬁer remote head, and the control unit. The
analyzer head includes all components that are inserted into the vacuum
chamber. It was designed to mount on a 2 3/4” conﬂat ﬂange, and the

source, quadrupole and detector components were originally housed in a
~160 mm long tube ﬁtted with 2 3/4” conﬂat ﬂanges. The total length of the
unit from the feedthrough ﬂange to the end of the source was 171 mm, and
the diameter of the system was 34 mm. The components of the source
include the: electron repeller, source grid/plate, focus lens, quadrupole,
electron suppressor, and Faraday plate. These are shown in Figure 4.1.
All the components except for the Faraday plate are connected to the
feedthroughs in the mounting ﬂange by 1/16” stainless steel rods. These
rods provide both physical support and electrical contact for each
component. The unit is disassembled by loosening the connectors that
couple these support rods to the electrical feedthroughs in the mounting
ﬂange.

The Dycor M200 uses an El source similar to the Anavac-2 source,
but of a more open design. The source components include the repeller,
ﬁlament, source plate, and focus lens shown in Figure 4.1. Instead of a
solid ﬁlament housing surrounding the ﬁlament, the Dycor M200 uses a
repeller grid ( 32 mm diameter, 41 mm length) that is biased negative with
respect to the ﬁlament to create an electric ﬁeld that causes the electrons to
travel toward the source grid. This open grid design is used to increase
conductance into the source and thereby increase the source sensitivity for
residual chamber gases. The ﬁlament is mounted inside the repeller grid
on two ﬁlament support rods connected to feedthroughs on the mounting
ﬂange. The ﬁlament is a loop ~ 20 mm in diameter that is spot welded to
the two support rods. This design effectively gives the source two
ﬁlaments. If one arm of the ﬁlament loop burns out, the other side can
still operate, extending the period between ﬁlament replacement. This

ﬁlament loop design also increases the total emission surface of the

163

ﬁlament and better matches the circular symmetry present in the source.
The source plate/grid is similar in design, but larger than, the source
plate/grid in the Anavac-2 source and normally is positioned with the grid
inserted inside the repeller cage and the source plate 2.4 mm from the end
of the repeller. The wire grid is cylindrical with a 14.5 mm diameter and
length of 26 mm. It is spot welded to the source plate which is a 0.5 mm
thick cylindrical plate with an outside diameter of 32 mm and has a
central 6.3 mm diameter ion exit aperture. The focus lens is positioned 2.4
mm from the source plate and is also a 32 mm diameter cylindrical plate
with a central 6.3 mm diameter aperture. The focus lens also has a set of
ﬁne cross wires spot welded across the central aperture. These are
installed to eliminate ions that would enter the quadrupole from those
initial positions. The cross wires are aligned so that they lie along the axes
between the quadrupole rods, by eliminating these initial ion entry
positions the amount of peak tailing in the mass spectrum is reduced as
discussed in Chapter 2.II.C.

The Dycor M200 quadrupole is much larger that the Anavac-2
quadrupole with a rod length of 114 mm, rod diameter of 10.0 mm, and rod
spacing of 6.0 mm. The four rods are mounted in machined ceramic
mounts that also contain holes through which the component support rods
are inserted, thereby further ensuring the proper positioning of the source

components. These quadrupole dimensions yield a r/ro of 1.67, which is

close to the ideal value of 1.148 discussed in Chapter 2.II.C. The
quadrupole is normally operated with a maximum of 1000 Vp_p, 2.27 MHz
rf signal and an average ion energy of 7 eV. The ultimate performance

that can be achieved with system can be calculated using the theoretical
equations described in Chapter 2.II.C for maximum mass (Mm) and

minimum peak width (AMmin) as:

__ 7X106Vm _ 7x10602000 V)
m _

 

 

.. = 332 daltons
2 2 2
fzro (2.27x106Hz) (3.0x10'3m)

164

9 9
4x10 V
AMmin= Z = 4m (7 eV) =o.42 daltons

2 2 2
f L (2.2711106) (0.114 m)2

 

 

These calculations indicate that the Dycor M200 quadrupole is
operating well below its ultimate maximum mass, but at approximately
the best resolution that could be achieved. This is most likely some
compromise between sensitivity, construction costs, mechanical tolerance
in quadrupole manufacture, and user needs.

An electron suppressor plate is mounted after the quadrupole and
serves as the quadrupole exit aperture. This component is a 32 mm
diameter cylindrical plate with a central 6.3 mm aperture, and is
normally biased at -120 Volts to suppress electron noise from the source.
The ions are detected by a shielded Faraday plate that is mounted within
the 2 3/4” mounting ﬂange.

The power supply unit provides dc voltages to the analyzer head,
generates the rf and dc signals used to operate the quadrupole, and
contains the power supplies and emission control necessary for the
operation of the ﬁlament. It also houses the analog-to-digital converters
that convert the analog signals from the electrometer head and analyzer
into the digital format for manipulation and display by the control unit.
The power supply can produce a 1 to 5 V potential, across the ﬁlament
leads, which is regulated to produce an emission current (monitored at the
source plate) determined by the control unit. The rf power supply is
frequency-variable from 2 to 3 MHz, controlled by the control unit, and fed
to the electrometer head through a BNC cable. All other potentials are fed
to the electrometer head and analyzer through a ribbon cable.

The electrometer/ampliﬁer remote head attaches to the 2 3/4”
mounting ﬂange on the analyzer head. Most of the operating potentials for
the analyzer simply pass through this unit with some added dc capacitive

ﬁltering. This remote head also houses the sensitive, optically coupled

165

ampliﬁer for detecting ion currents at the Faraday plate, and the rf
transformer used to increase the amplitude of the rf signal produced by the
power supply unit.

The control unit is used to display and control the operation of the
entire system. It is microprocessor controlled, and contains a CRT that
allows the user to display and change the value of a series of operational
parameters that are used to control the power supply unit. It also allows
the operator to store two 600 point intensity ﬁles, or spectra, which can be
displayed, printed, or uploaded through an R8232 port. The R8232 port can
also be used to modify the value of the operational parameters of the RGA,
making remote control and data acquisition possible with a personal
computer. The control unit operation parameters control all the functions
of the RGA, a description of their function is given below and a list of their
initial values is given in Table 4.1.

TOTAL PRESSURE - turns on the total pressure mode for the RGA, in

this mode the RGA will switch to rf-only mode between scans and also
display the ﬁlament resistance, voltage, current, and RF TUNE values

LO MASS - sets the lower limit of the scan range for the quadrupole
HI MASS - sets the upper limit of the scan range for the quadrupole

SAMP/AMU - sets the number of stored data points per amu for the
mass scan

DWELL - sets the sampling time for each data point in the mass
spectrum

EMISSION - sets the ﬁlament emission regulation level, monitored at
the source grid, and displayed as ELECT CUR on the control unit

EL ENERGY - sets the potential difference between the source grid and
the ﬁlament rods, which varies if the source potential changes with m/z

FOCUS - sets the potential energy applied to the focus lens

REPELLER - sets the potential applied to the repeller surrounding the
source

166

FREQUENCY - sets the rf Frequency of the power supply, in TOTAL
mode the user can observe the RF TUNE parameter to monitor reﬂected
rf power

FIL PROT - sets upper pressure limit for overpressure protection of
ﬁlament

ANALOG MODE - scans the quadrupole between LO MASS and HI
MASS values, records and displays analog intensity vs m/z spectrum

BAR MODE - scans quadrupole between LO MASS and HI MASS values,
records analog intensity vs m/z spectrum, integrates points within 1/4
amu of peak mass for bar intensity display.

TAB MODE - selected ion monitoring for up to 12 masses

 

LO RES and HI RES - sets rf/dc ratio on quadrupole at the low and high
ends of mass spectrum

LO POS and HI POS - allows user to align mass spectrum with the
display mass axis

LO SENS and HI SENS - sets ion energy (1 to 10 eV) at the low and high
ends of the mass range, ion energy is varied linearly between these
values during a mass scan allowing the user to operate the unit in
constant sensitivity or constant peak width modes

Table 4.1 The initial values of the Dycor M200 operating parameters
for the original thoriated iridium ﬁlament.

EMISSION = 1.0E-03 EL ENERGY = -70.0 V FOCUS = -31.0 V
REPELLER = -72.0 V FREQUENCY = 2.27E+06 RF TUNE = 6.552

LO RES = 4048 HI RES = 2427 L0 SENS = 6.00
HI SENS = 8.30 FIL PROT = 1.0E-03 torr FIL = ON
FIL VOLTS = 2.0 V FIL CUR = 2.9 A FIL RES = 0.68

A study was performed to evaluate the resolution and mass range
performance of the Dycor M200 RGA prior to modiﬁcation. Both the Dycor
M200 and another Dycor M100 RGA (1-100 m/z range) were mounted on
the matrix isolation chamber. The chamber was evacuated to a base
pressure of 4x10'9 torr, and a leak of perﬂuorotributylamine was

introduced into the system to a pressure of 5.8x10'6 torr measured by the

167

M100. The M200 was set to DWELL = 1.0 seconds, and 6 SAMP/AMU and
scanned from m/z = 1-200 in ANALOG mode. A plot of the mass spectrum
produced by the M200 is shown in Figure 4.2. The highest mass that was
observed in the spectrum was m/z = 181, but the unit appeared to operate
properly to m/z = 200, at which point the scan was automatically reset.
The unit produced ~20% valley resolution throughout the entire mass
range, with a slight increase in peak width at higher masses. The peak
width increase at higher masses resulted from the operation of the
quadrupole with the ion energy increasing from 6 eV at m/z = 1 to 8.3 eV at
m/z = 200.

IIM' inthrM2

Following evaluation of the original performance of the Dycor M200
RGA, the analyzer head was modiﬁed to produce a beam of mass-selected
ions. To allow the introduction of sample gas directly into and minimize
conductance from the source region, a source housing was designed
similar to the source housing used with the Anavac-2. The design was
based upon results from SIMION modeling studies of the Dycor M200 EI
source. The source elements were carefully measured, and a 16,000 point
circular symmetry SIMION electrode array was constructed that closely
reproduced these dimensions. The SIMION array was reﬁned, and the
standard operating potentials for the source were applied to the electrodes.
Figure 4.3a shows the reﬁned SIMION model of the original source with
the repeller = -72 V, source grid = +6.0 V, focus lens = -31 V, and
quadrupole at 0.0 V. Equipotential contour lines were displayed to
illustrate the penetration of the extraction ﬁeld from the focus element into
the source grid and the focal properties of the lens system. The trajectories
of electrons formed at the ﬁlament position were modeled to determine the
region within the source grid of highest electron density. The trajectories
of 75 ions formed within this region are shown in Figure 4.3a. All the ions

 

168

.ccmﬁﬂbsnmbeoscSQ .5.“ <Um com—2 .59:— 353.3 2: .3 @8355 52.5...on 9.2:: 25. N6. char.—

cow of cm:

#3

 

o:

co:

Hm:

cm —

ce—

 

 

cw

O:

we

_..

 

38:35
2,523

169

0.0 Vols

 

     
    
   

Ion Trajectories

l

 

 

 

 

 

 

 

 

 

 

Electrodes

 

 

   
      

 

 

 

 

 

 

 

 

 

 

Equipotential
Contour Lines
13160110“ Source Focus

Repeller Grid Lens
\ .x s .\\\

Grid \
\tiS:

ﬁzz/g- .5. ll
., .
..
“'— J \\
‘~

I, — t I ‘ ‘_. _
’{rr//
// '/ ‘///

 

 

b)

 

 

 

 

 

 

 

 

 

 

 

    
  
 
 

F

-72 Volts I j
x |

-31 Volts

    

 

+6 Volts
\

 

 

 

 

 

 

 

 

 

 

 

 

0.0 Volts

Figure 4.3 SIMION models of the Dycor M200 source a) before modiﬁcation
and b) following enclosure in a source housing.

170

that were modeled experienced sufﬁcient drawout potential from the focus
element to be extracted from the source and injected into the quadrupole,
producing the ion beam proﬁle shown. A second SIMION model was
constructed based on this model, but with a reduced ion exit aperture and
additional lens thickness that represented the placement of the source
within a source housing. The model, shown in Figure 4.3b, was reﬁned,
the elements were biased at the same operating potentials, and the contour
lines and trajectories calculated for the previous model were recalculated
and displayed. The contour lines in Figure 4.3b clearly shows a reduction
in the penetration of the extraction ﬁeld due to the smaller ion exit
aperture in this model. The ion trajectories calculated for this ”model were
almost identical to those calculated in Figure 4.3a, indicating that
enclosing the source should effect the ion beam injected into the
quadrupole only slightly. These results were similar to those observed
earlier in the SIMION modeling studies of the Anavac-2 source housing.
Following the SIMION modeling of the Dycor M200 EI source, a
source mount was designed and constructed to enclose the E1 source, turn
around the unit so that the ion beam was directed inward toward the
matrix experiment, and feed the electrical and gas lines to the various
analyzer components. Figure 4.4 shows a mechanical drawing of the
completed analyzer chamber with the analyzer components installed. The
EI source housing was constructed using a Ceramaseal 807A0296-1
metal/ceramic/envelope for the housing body. The upper end of this
envelope was welded to a 4 1/2”-to-6” conﬂat conversion ﬂange, labeled as
the support ﬂange in Figure 4.4. A 0.5 mm thick stainless steel plate was
welded to the lower end of the envelope and a 6.3 mm aperture was drilled
in the center of this plate to serve as the ion exit aperture. Threaded rods
extending 10 mm inside the source housing were silver soldered into this
plate. The support rod attached to the Dycor source plate was sheared off,
and the source plate was bolted onto these threaded rods. This
conﬁguration allowed the source grid/plate to be mounted within the

171

 

 

Fig. 4.4 The modiﬁed Dycor M200 RGA mass selected ion source. Components
include; (1) enclosed source housing, (2) source grid, (3) support ﬂange,

(4) feedthrough ﬂange, (5) gas inlet, (6) source electrical feedthroughs,

(7) quadrupole rods, (8) clamp rings, (9) support rods, (10) ﬁlament,

(11) electron repeller, (12) original RGA mounting ﬂange, (13) focusing/deﬂection
optics, (14) multi-pin feedthrough, (15) ceramic mount for quadrupole rods.

172

enclosed source housing while adding only the additional thickness of the
housing endplate to the ion optical system.

A 4 1/2” conﬂat ﬂange was used as a mount for the repeller,
ﬁlament, and sample inlet lines. A miniconﬂat ﬂange connected to a 1/4”
stainless steel tube was welded into the center of this ﬂange to serve as the
sample inlet line. Four Ceramaseal 809B8030-1 high current BNC vacuum
feedthroughs were mounted around the inlet line, and arranged to
reproduce the original positioning of the ﬁlament and repeller components
in the Dycor RGA. The support rods for the ﬁlament and repeller were
shortened by ~50 mm, and connected to these feedthroughs using the
original RGA connectors. The repeller was positioned to extend 84 mm
from the face of the 4 1/2” ﬂange. When the unit was inserted, this
distance closely reproduced the original 2.4 mm spacing between the
repeller and the source plate. Ceramic sheaths were placed over the upper
half of the ﬁlament support rods to insulate them from the repeller, a
0.005” rhenium wire loop was spot welded to the rods, and they were fed
through the repeller and connected to the feedthroughs on the 4 1/2” ﬂange.
The fourth feedthrough on this ﬂange was connected to the source plate
using ﬂexible, Teﬂon-coated wire. With these connections completed, the
feedthrough ﬂange was carefully installed on the 6” mounting ﬂange and
tightened down. This mounting system closely reproduced the 2.4 mm
spacing between the repeller and source plate elements, while enclosing
the source in a low conductance housing. The potentials to the source
plate, repeller, and ﬁlament could be fed via BNC feedthroughs in the 4 1/2”
ﬂange, and sample gas could be directly introduced into the source
housing.

Two 1/4” stainless steel support rods were installed in the mounting
ﬂange as shown in Figure 4.4. These were used to support two sets of
clamp rings that ﬁt around the ceramic quadrupole mount. The focus
lens was installed through the quadrupole ceramic, and both components

were inserted inside the clamp rings. Tightening two setscrews, mounted

 

 

173

in the clamp rings, allowed the quadrupole ceramic to be locked into a
position that placed the focus element at a distance of 2.4 mm from the
bottom plate of the source housing. A 193 mm long chamber was
constructed to house the source and quadrupole. Each end of this chamber
was welded to rotatable 6” conﬂat ﬂanges, with the support ﬂange
attaching to the upper ﬂange, as shown in Figure 4.4, and the lower ﬂange

connecting the unit to the matrix isolation chamber. Two ports, a 2 3/4”

and a miniconﬂat port, were installed on opposite sides of this chamber.
The 2 3/4” port was positioned opposite the quadrupole connection rods at a
distance that closely reproduced the original, ~ 75 mm, distance from
these rods to the Dycor mounting ﬂange. The support rods on the
quadrupole were shortened, and the original Dycor connectors were
modiﬁed so that these lines could be sent out the 2 3/4” port where they were
connected to the feedthroughs on the original Dycor mounting ﬂange. The
focus element was also wired through this feedthrough ﬂange using
ﬂexible Teﬂon-coated wire. This connection scheme closely reproduced the
original wiring used to supply the rf and dc potentials to the quadrupole
rods, minimizing the change in the resonant frequency of this quadrupole
circuit.

A Ceramaseal 808B9340-1 8-pin miniconﬂat electrical feedthrough
was mounted on the opposite port and the feedthrough pins were
connected to ﬂexible Teﬂon-coated wires to supply potentials to the lens
system. A series of deﬂection optics were fabricated that were designed to
mount at the exit end of of the ceramic quadrupole mount as shown in
Figure 4.4. Circular einzel lens elements were machined from 0.5 mm
stainless steel stock to an outside diameter of 32 mm. Elements with
several center aperture sizes were machined including; 1/4”(6.3 mm),
7/32” (5.5 mm), 3/16”(4.8 mm), and 9/32” (7.1 mm). Two sets of four
deﬂection plates were manufactured that were 1/2” wide and 1’2” and 3/4”
long. All of these elements were designed to attach to threaded rods

installed through the quadrupole ceramic mount. These threaded rods

174

were isolated from the lens components by glass sheath rods, and the
distance between components was determined by the ceramic spacers used
to separate the individual elements.

The source ﬁlament used following modiﬁcation was 0.005”
rhenium wire, this was chosen because of its low cost and availability. A
minor modiﬁcation was made to the control board of the power supply unit
to allow it to create the higher heater current necessary to operate with
this higher work function material. A 3 kn resistor was added parallel to
R22 on the power supply control board which is shown circled in Figure
4.5. This modiﬁcation allowed the ﬁlament to operate at higher current
levels than before modiﬁcation, which were necessary to heat the rhenium
wire to the 1-3 mA emission levels for normal operation of the RGA.

It was also necessary to patch into the ribbon cable connecting the
power supply to the electrometer head to allow the source, ﬁlament, and
repeller lines to connect to the BNC feedthroughs on the 4 1/2” feedthrough
ﬂange. A bud box was installed which connected to the ribbon cable
connector on the back of the power supply unit. All connections except for
the repeller, source, and ﬁlament lines were fed through this box to the
ribbon cable. The repeller, source and ﬁlament lines were connected to
high voltage BN C feedthroughs on the box. High voltage BNC cables were
made that connected to another bud box attached to the source chamber.
This box contained 0.01 uF capacitors used as noise ﬁlters on the source
components. One end of the BNC cables fed through this box was
connected to the power supply patch box, and the other to the feedthroughs
on the source feedthrough ﬂange. These modiﬁcations allowed the remote
electrometer head to attach to the original Dycor mounting flange,
supplying the rf/dc potentials to the quadrupole and the dc potential to the
focus element, while the the patch boxes allowed the repeller, ﬁlament,
and source potentials to be fed through BNC cables to the removable
feedthrough ﬂange.

 

6:3 Soﬁa—c 835;» hood mo em: 9: 32? 8 829:8 .8363 a mo
55:.me 05 $5332: v.83 >395 .833 oomz Sean 2: mo Encmmmc 23:853. < was 6.59%

mutumoeuomm m_ - wit! :51!!! WE. “HM. Haiti...»
m

3.8-:03!!!
IF!
_

 

 

 

 

 

 

 

 

 

175

 

325?! 3c IE3. “
- iii; - -1. -. :-E I--. .r--l._l

g
a: nail - . . - mmﬁuﬁuﬁﬁ.”*ai

.: 2 - i. . at -.
ms We . “is

l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I
ll
1,'
1"
ll
.1
51523..
,Wy—‘ﬁ—t—
a
ﬁ

 

A i - - . ' I

43....
o .5

- a . III... I. ~ .6!
an: as: ”i 3 t. it u... p u
on. . ii- 8 t . . L». .9. a u
w d. c n 1.1. u A .7.” “3 73 H... .3? 2m...
. o m a n ll! 0. no .5433 P. _ amm‘.
»
o
.
.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- ,0. l 0. .\° .— ‘o. 11». .so~o I“ v
i l O. . .u ...s . so. lib O
«a a“. a. 2“} Hi- 0" - NH.._am:— . ImMP— Tn .
n H .. ...H..H a 3...... a... s m. 25.. . 1
a. 9. A; owes. . :.HH. — 2:— M...- ‘ gown-0 , — . o
in ’3‘.
- if- .. .. .. : ...---:.. 2
-II : .3 a ($.25. .0..." ..ﬂ4~.fﬂd 5......
has. .0. at? ...t. J .P? .x 4%. a 1W...
1 . c .
....:.- . .3... .......... . .. a a ”19+...
8 as. o. 1 i 3.0 S v. a . . .0: It . . .
oh.- .no 1 a. ‘ v :‘o CVVSW...‘ m T 1V? 1. ~ .7
:u .3 “6 .3 : 0.1 u . ..- . . «2 .1
4m Ms... 2 . «um 2“ .. am .3. .. -. .. u . 3
~ I 6 t O 0‘: o 1 :1 Qt.“— . an o w
.mmd, . . m “l r. :7... 3 a." » .n.... 9‘. t.“
a. .53.! Hi. 39 . t. -tﬁe .. _
.3 3... . .1 1: . 4 ._ .7 ad
a Q o . s
L .— .mmll can" u «“3: mm sn.‘ - 0 —

 

 

 

 

 

.2.

 

176

 

Following fabrication of the components of the source mount, a
0.005” rhenium ﬁlament was spot welded to the ﬁlament support rods and
the source feedthrough ﬂange was installed. The quadrupole and focus
lines were connected to the original Dycor ﬂange, and a shielded Faraday
plate was installed after the quadrupole. The Faraday plate was connected
to a miniconﬂat electrical BNC feedthrough ﬂange mounted on the
miniconﬂat port of the source housing. A Granville Phillips model 60 leak
valve was installed on the source inlet line, and the entire assembly was
connected to the left 6” port on the matrix isolation chamber. The chamber
was evacuated, and the electrometer head and BNC lines from the patch
boxes were connected to the source mount. A Keithley 610C electrometer
was connected to the miniconﬂat BNC port to monitor Faraday plate
current.

Both the power supply and control unit were switched on, and the rf
frequency of the system was Optimized. With the ﬁlament off, the unit was
switched to total pressure (rf-only) mode which caused the control unit to
display the value of the RF TUNE parameter. This parameter is a
measure of the rf power that is reﬂected back from the rf transformer into
the power supply. The RF TUNE parameter can vary from 0 to 10, and in
the original conﬁguration the unit had an RF TUNE of 6.552. With an rf
frequency of 2.27 MHz the value of RF TUNE displayed on the control unit
following modiﬁcation was 10.0, indicating a high level of reﬂected rf
power and poor rf power transfer to the quadrupole. The frequency of the
rf power was varied at the control unit using the FREQUENCY parameter
while observing the value of the RF TUNE parameter. A plot of RF TUNE
verses FREQUENCY is shown in Figure 4.6a. The FREQUENCY
parameter was varied from 2.27 MHz to 2.35 MHz, and a minimum RF
TUNE value of 6.493 was obtained at a FREQUENCY of 2.308 MHz. This

 

 

177

a)

 

 

 

6 I I I I ' I ' I I I j I ' I ' ﬁ
2.27 2.28 2.29 2.30 2.31 2.32 2.33 2.34 2.35
FREQUENCY
b) 2e—11 -

  
    
    
 
  
 

Oe+0 ..

faraday -2e-11-

plate I
-4e-11 4

current .
(A) -6e11-‘

~8e—11‘
-1.0e-10-+

-1.2e-10-4

 

 

' I ' I 1 I I I ' I ' I ' I ' I ' I ‘ I I I ‘ I

135791113151719212325
LOMASS

Figure 4.6 Graph of both the RF TUNE parameter versus FREQUENCY (a), and
the faraday plate curent vs LO MASS setting (b) for the modiﬁed DYCOR M200.

 

178

RF TUNE value was lower than the original setting, indicating that the rf
reﬂected power had been decreased through the quadrupole wiring
modiﬁcations and that the unit should operate properly.

After the rf frequency was optimized, the total mode was switched
off, and the unit was switched to mass scan mode. The ﬁlament was
turned on, and set to an EMISSION level of 1 mA, the source was operated
with EL ENERGY = 70.0, REPELLER = -72 V, and FOCUS = -31 V. The L0
MASS parameter was set to 1 and the HI MASS parameter set to 100. As
soon as the ﬁlament was switched on, a signal of -1.4x10'10 A was detected
at the Faraday plate, indicating that electrons from the ﬁlament were
passing through the quadrupole and hitting the Faraday plate. In the
original Dycor these electrons were eliminated by placing a suppressor
plate before the Faraday plate and biasing it at -120 V. The modiﬁed M200
source no longer utilized any electron suppression, and a study was
performed to determine if the quadrupole could discriminate against
electrons at high mass settings. The L0 MASS parameter was varied
while observing the sign and magnitude of the signal at the Faraday plate.
A plot of the Faraday plate current verses LO MASS setting obtained from
this study is shown in Figure 4.6b. The L0 MASS parameter was varied
from 1 to 25 which determined the initial mass that the quadrupole was
selecting. A negative current, indicating electrons hitting the Faraday
plate, was measured for L0 MASS settings less than 10. Above LO MASS
setting of 10 the Faraday plate signal was relatively constant, with a slight
positive deﬂection at m/z = 14. This indicated that the quadrupole was
effective at discriminating against electrons at settings above m/z = 10, and
the positive deﬂection at m/z = 14 was determined to be the result of a small
air leak (N+ and N22+) in the source inlet that was later sealed. As a
general rule this source should not be used to mass-select ions lower than
m/z = 10 unless a suppressor plate is installed after the quadrupole, or the
presence of a background electron current can be tolerated.

 

 

179

The ability of the system to produce mass-selected ions was
evaluated by introducing samples into the source, scanning the
quadrupole, and observing the current at the Faraday plate. The base
pressure of the chamber was 31:108 torr. Argon was introduced through
the leak valve to a total chamber pressure of 4.51:10'7 torr. The Dycor M200
was scanned at a dwell of 2.0 sec with 10 samp/amu and the ion current
was detected by a Keithley 610C electrometer which was used to drive an
XY recorder that used a time sweep for the x axis.. The source was
operated with 1 mA emission, 70 eV electrons, repeller = -7 2 V, and focus =
-31 V. This produced a maximum signal of 11.0 nA at the Faraday plate
for m/z = 40. The argon leak was removed, and nitrogen was introduced
into the source to a chamber pressure of 4.0x10'7 torr. With the source
operating under the same conditions, a maximum mass selected signal of
10.0 nA was measured at the Faraday plate for m/z = 28.

The nitrogen leak was removed, and a leak of
perﬂuorotributylamine (PFTBA) was introduced to evaluate the resolution
and mass range of the modiﬁed Dycor M200 source. A PFTBA bulb was
cleaned and prepared with three freeze/thaw cycles. This was connected
to the source inlet which was evacuated to 50 mtorr. The bulb valve and
source leak valve were opened to introduce PFTBA into the source to a
chamber pressure of 4x10‘7 torr. The source was operated under the same
conditions, and the quadrupole was scanned from m/z 15-190 at a dwell of
2.0 seconds and 10 samples/amu. The spectrum recorded at the XY
recorder is shown in Figure 4.7. The bottom trace shows the scan from
m/z = 15-105 at a sensitivity of full scale = 10 nA, the central trace show a
scan from m/z = 90-160 at full scale = 1 nA, and the upper trace shows a
scan from m/z = 150-190 at a sensitivity of full scale = 0.1 nA. The
maximum current at the Faraday plate in this spectrum was 2.23 nA at
m/z = 69. The spectrum shows well-resolved peaks through the range
from m/z = 15-140, above this range the peak width increased dramatically,
but ion peaks were observed above m/z = 170. The intensity of the ion peaks

 

180

 

 

.233va6 2: ac ﬁxo 05 as @3555 32m >333“ a an @9532: new
8.58 com—z SHED REES: 05 .3 @8353 oEEa—aasntaohosﬁa .3 82.50on mng 25. Ev ousmmh

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OOH om om ch ow on 3v an ON
“ w w i. m w u .1 w
aw
om~ Own amu can a: COM ca
'P n p p b p n
l a l n q q -
m m m m H m
2: 2: c: 8" 2: cm”

P P p p
I q d u 1
.Il l’lll’l llll‘ll l<l|l'l/\\Ill l4

181

in the upper mass range was reduced slightly relative to the PFTBA
spectrum taken with the original Dycor M200 RGA. To further evaluate
the mass range performance of the unit, the PFTBA leak was replaced
with a leak of perﬂuorokerosene introduced through the source inlet to a
pressure of 4.7x10'7 torr. The resolution of the unit was reduced from the
original HI RES value of 2427 to a value of 2200 to allow better observation of
high mass peaks in the mass spectrum. The source was operated under
the same conditions, and the quadrupole was scanned from m/z = 20-200 at
a dwell of 1.0 second and 15 samp/amu. The spectrum recorded on the XY
chart recorded is shown in Figure 4.8. The maximum current obtained in
this scan was 2.22 nA at m/z = 69. This spectrum clearly shows peaks up
to m/z = 181 indicating that the mass range of the Dycor M200 had not been
degraded by the design modiﬁcations. The resolution at higher mass
values was slightly reduced due to the change in H1 RES setting, but
approximate unit mass resolution was still observed. This resolution was
higher than required for most matrix isolation studies, and in most cases
the HI RES and LO RES settings could be decreased to increase ion output
from the source.

The ability of the source to produce a mass-selected beam at a remote
Faraday plate was evaluated by removing the Faraday plate from the end
of the quadrupole, installing a focusing/deﬂection optics systems after the
quadrupole, and monitoring the ion current at the Faraday plate mounted
on the linear motion feedthrough 203 mm from the end of the quadrupole.
The lens system installed after the quadrupole was based on the design of
model 1 from the SIMION deﬂection optics studies in Chapter 3.V.B.
Three lens elements were used to function as an einzel lens. The ﬁrst two
elements had central aperture of 6.3 mm and the third element had an
aperture of 7.1 mm. The spacing between these elements was set with 2
mm ceramic insulators, and the ﬁrst and third elements were connected

to ground. Two pairs of 1/2” long deﬂection plates were placed after the

einzel lens. These were offset with the horizontal pair preceding the

 

 

 

182

 

 

.838 comz .894— 9: 3 noozcoa mammoumxonoscumq .3 82.53% meE 9E. wé 959nm

 

 

 

 

 

 

 

 

 

d8
can am.— cm— a».
_*~_
NE.
=0. :2 a: 2.”— $2 a".
- - _ n P
J l<lll : «ll/\-ll,-: l ill/A lI/\.- la-|<\lll</\\Illlltl.
«wgm
Hmu— u~\p~_
s. c... e I... a a a. 3.. a
.~\\ .nmﬁl

 

on:
a: ‘

 

 

 

am

9

_ 3335
n 3.3.3.

n

a

$5.2:—

. r “ES—oz

 

3:.

325:.—

.3. 3:53.

 

is:

183

vertical pair, and the spacing between each plate pair was also set at 2
mm. This system was installed on the quadrupole ceramic, and ﬂexible
Teﬂon-coated wire was used to connect each element to a feedthrough pin
on the 8-pin miniconﬂat feedthrough. The unit was then installed on the
left 6” ﬂange of the matrix isolation chamber and the chamber was
evacuated. The linear motion feedthrough was inserted to 0.95” which
positioned the 1” diameter Faraday plate mounted at the end of the
feedthrough directly in front of the cryostat window, at a distance to 203
mm from the lens system. The lead from the Faraday plate was fed
through a BNC miniconﬂat electrical feedthrough and connected to a
Keithley 610C electrometer used to measure ion current and drive an XY
recorder.

The base pressure in the chamber was 3.0x10'9 torr, and a leak of
argon was introduced through the source inlet to a chamber pressure of
6.8x10‘7 torr. The electrometer head and BNC cables were connected to the
source and the unit was turned on. The source was operated at: repeller =
-72 V, electron energy = 70 eV, focus lens = -31 V, and ﬁlament emission =
1 mA. The unit was set to scan over m/z = 28 and the potentials on the
deﬂection optics were optimized for maximum ion current at the Faraday
plate. The optimum values were; center einzel lens = -25 V, lower
deﬂection plate = +1.4 V, right deﬂection plate = 1.6 V, all other elements
were kept at 0.0 V. The quadrupole was scanned from m/z = 1-60 at a dwell
of 250 msec with 15 samples/amu. The spectrum recorded on the XY chart
recorder is shown in Figure 4.9. This spectrum was recorded at a
sensitivity of full scale = 1 nA and the maximum signal for this scan was
0.64 nA at m/z = 40. The ion optics were re-optimized, and a m/z = 40
signal maximum greater than 1 nA was obtained. This spectrum showed
some decrease in resolution compared to earlier studies where the
Faraday plate was placed at the end of the quadrupole.

A comparison study was performed to evaluate the loss in resolution

that occurred when the Faraday plate was located 203 mm from the exit of

 

 

184

 

 

6395335 05 .3 .68 93 :3: 88 now .3335? 343%.”. 9: .3 39$ 5 30:2:ch 933 .3328

.5 H a .3 “38338 can 8.50m com—2 3:0 95 .3 38:on cow; .3 85.58% 32: 25. m.v 83mg

$8
8 on 3. cm an S

p P p u
d ‘ . it

4 <1

‘-

 

on

 

 

 

ov

185

the quadrupole. A shielded Faraday plate was installed on the end of the
quadrupole and connected to one of the pins on the 8-pin miniconﬂat
feedthrough. The unit was installed on the diffusion pump test chamber
and evacuated to a base pressure of 1x10'6 torr. A bulb of C82 was

prepared, using three freeze/thaw cycles to remove volatile contaminants.
This was connected to the source inlet, evacuated, and C82 was introduced
into the source chamber to a pressure of 3.0x10'6 torr. Ion current was
measured by a Keithley 610C electrometer, connected to the Faraday plate
feedthrough, which was used to drive an XY chart recorder. The source
was operated at 1 mA emission, repeller = -72 V, focus -31 V, and with 70
eV electron energy. The quadrupole was scanned from m/z = 10-100 at a
dwell of 250 msec with 15 samples/amu. Figure 4.10 shows the mass
spectrum produced at the XY recorder. The lower trace was made at full
scale = 10 nA and the upper trace was taken at full scale = 1 nA. The
spectrum shows that the quadrupole was able to effectively resolve ion
separated by two mass units, but true unit mass resolution was not
observed. This was probably due to the fact that the HI RES parameter was

set to 2100 for this scan. The quadrupole was able to resolve the important
peaks in the spectrum of CSg, of particular signiﬁcance was the ability to

mass-select the peaks at m/z = 76 and 78, the molecular ion of CSZ. This

was important for the CSZ+ ion deposition studies that were conducted

using this source.

The unit was then removed from the test chamber, and the Faraday
plate was taken off the quadrupole ceramic mount. The three element
(apertures = 6.3, 6.3, and 7.1 mm) lens system with four (1/2” long)
deﬂection plates described earlier was installed after the quadrupole and
the elements were connected to the 8-pin feedthrough miniconﬂat ﬂange
leads. The unit was installed on the left 6” port of the matrix chamber, and
the chamber was evacuated to a base pressure of 8.1x10'8 torr. The linear
motion feedthrough was inserted to 0.95”, placing the Faraday plate
directly in front of the cryostat window. The Faraday plate feedthrough

186

.3335va5 2: .3 :3 of .3 33:2: 32a tan—5.5.. a .3 3.533: can
ooh—om CONE 39am «5 a: @8333 ova—2&3 39:3 .3 83.56on 39: 9E. oﬁv 85mg

 

 

 

 

 

 

 

 

 

 

N
8. 8 8 2 8 \E 8 E. on 8 E
u u a x x a u x x 1.
m3 -
em
3.
2 N
Q: 8 cm on 8 \E 8 S S cm 2
w w i. w u m A " Ti.
3 5 3 2 S
C c _ x \
2. ”a

 

wm

 

 

 

 

 

 

Ao_3m-b3v on WI 3.37:3 3. i.

187

was again connected to a Keithley 6100 electrometer which was used to
drive an XY chart recorder. The 082 bulb from the previous study was
connected to the source inlet, the inlet was evacuated, and the source leak
valve was opened until a total pressure of 3x10'6 torr was measure in the
chamber. This corresponded to a C82 partial pressure peak at m/z = 76 of
4x10‘7 torr measured by the Dycor M100 RGA mounted on the chamber.
The source was operated under the same conditions, and the quadrupole
was scanned from m/z = 10-100 at a dwell of 250 msec and 15
samples/amu. The potentials applied to the deﬂection optics were
optimized for ion current at m/z = 76 current, and the optimum settings
were: center einzel lens = -25 V, right deﬂection plate = -1.8 V, bottom
deﬂection plate = -1.58 V, and all other elements = 0.0 V. Figure 4.11
shows two consecutive mass spectra recorded on the XY recorder at a
sensitivity of full scale = 10 nA. the maximum signal measured at m/z =
76 was ~5.6 nA. These spectra showed that the resolution of the system
had been degraded by placing the Faraday plate 203 mm away from the
quadrupole. This loss in resolution indicated that the system in its current
conﬁguration was no longer capable of resolving ions that were separated
by only a single mass unit, and this must be taken into account when
planning matrix isolation studies. Fortunately the source was able to
resolve ions that differ by 2 mass units, as indicated by the effective
separation of ions at m/z = 76 and 78, the molecular ions of C82. This
resolution proved sufﬁcient for most matrix isolation studies, and was
occasionally degraded further to increase ion current.

To evaluate the shape of the ion beam produced by the modiﬁed
Dycor M200 source, the shielded CEMA detector mount described in
Chapter 3.VI.B was installed in place of the cryostat head. This allowed
the visualization of the ion beam proﬁle produced by the source at the same
position as the cryostat window. The output from the phosphor screen was

observed through the 2 3/4” window ﬂange normally used to introduced the
FTIR beam. A 10 M9 resistor network was installed across the CEMA

188

 

335.695 05 no .58 95 89a 8.: now .3353 “Swab“. 05 .8 8:8,. 5 @3555 33m .333“ a 3 @83on EB
8.38 88% .8th one .3 3035.5 35:56 cognac me 85.58% mde .2: we 368 among 259 3.... 0.59m

 

 

 

 

 

 

 

 

 

 

 

N}:
2: am an E. cc cn av an :N o.
— b n p n n n n p h
. u a - u q u . . a
5 3 J
an an
NM
3.
E N}:
2: ca 8 S. ow em cw on cm 3
m w 4.. x x x w x . x
we
xm
3.
mm

 

 

 

 

2.

 

189

plates and the rear plate was connected through a Fluke 23 DMM to
ground. The high voltage operating potentials were supplied to the CEMA
detector from a high voltage feedthrough ﬂange using shielded BNC
cables. The CEMA was normally operated with -1400 V on the front plate
of the detector and +1800 V on the phosphor screen. These conditions
produced a current at the DMM of less than 0.1 mA indicating proper
operation of the channelplate detector.

Several lens systems were investigated for use with the Dycor M200
source. Einzel lenses using three 6.3 mm apertures, a 6.3 and two 7.1 mm
apertures, and two 6.3 mm apertures followed by a 7.1 mm aperture
systems were installed and visualized with the CEMA. Two versions of
each of these systems were installed, one with 2 mm and another with 4
mm inter-element spacing. Each of these systems were followed by a set of
four offset, 1/2” long deﬂection plates. The beam proﬁles of these six
systems were evaluated for m/z = 4 (He+), m/z = 28 (N2+), and m/z = 40
(Ar+). The potential applied to the central einzel lens element was varied
from -120 to + 120 V, and none of the lens systems were able to focus the
entire mass-selected ion beams within the 1” diameter of the detector. This
most likely due to the fact that the source ion beam travels a much larger
distance, 203 mm, from the end of the quadrupole to the window than with
the Anavac-2 source. As a result a potential that correctly focuses some
portion of the complex ion beam exiting the quadrupole, will defocus
another portion. By operating the einzel lenses in accellerate-decelerate
mode, a clear cross pattern was produced with each of the lens systems,
but some portion of the beam was lost. The most intense and clearly
focused cross patterns was produced using the 6.3, 6.3, 6.3 mm aperture
system with 2 mm inter-element spacing. The ﬁrst and third elements of
this system were held at ground and the center element was biased at -60.5
V. To determine whether these values actually produced the highest ion
current at the matrix window, the CEMA detector was removed and the

linear motion feedthrough was inserted to 0.95” to position the 1” Faraday

190

plate in front of the cryostat window. With the 6.3, 6.3, 6.3 mm aperture, 2
mm spacing einzel lens, and two sets of offset 1/2” long deﬂection plates
installed, the potentials applied to the lens elements were optimized for
mass selected ion current (m/z = 40 with an argon leak) at the Faraday
plate. Several such studies were repeated during the process of
conducting matrix isolation studies with this lens system. The best
average optimal settings were: center einzel lens element = -25 V, right
deﬂection plate = -1.8 V, bottom deﬂection plate = -1.58 V, and all other lens
elements maintained at 0.0 V. At these settings mass-selected ion
currents of more that 1 nA were maintained with only minor adjustment

for several hours during the course of the matrix isolation studies.

 

 

 

191

Chapter 5. Conclusions and suggested future studies

The ion sources constructed through the modiﬁcation of the Anavac-
2 and Dycor M200 residual gas analyzers have proved reliable for the
production of mass-selected ion beams. These devices have been used in
several matrix isolation studies with good success. We have successfully
trapped and spectroscopically observed via laser induced ﬂuorescence
spectroscopy mass-selected ions within a low temperature argon matrix.
An example of the LIF spectrum of C82+ isolated in an argon matrix is
shown in Appendix XV. The LIF technique is much more sensitive than
is infrared analysis; IR spectroscopy requires the accumulation of higher
numbers of absorbers for effective detection. To minimize the time
required for the IR experiments it would be desirable to increase the mass-

selected ion current produced by these sources.

IP'm°°n

Several possible studies and could be performed to determine which
modiﬁcations would increase the ion output from the modiﬁed RGAs. The
total ion ﬂux from the original Anavac-2 EI source was reduced by
approximately 79% by enclosing the unit in the sealed source housing.
Most of this current loss was probably due to the reduction of the ion exit
aperture. If a higher ﬂux of neutrals from the source could be tolerated,
increasing the source housing aperture size should result in an increase
in ion current at the cryostat window. It might also be useful to install a
repeller electrode within the cylindrical grid. This could be used to
increase the electric ﬁeld within the ionization region, thereby increasing
the total output of the source. Disadvantages to using this type of repeller
include: difficulty in installing a repeller in the small cylindrical grid

region, increased ion energy due the higher electric potentials within the

 

192

ionization region, and difficulties in electrically isolating and biasing the
repeller electrode. The Anavac source was also used for the deposition of
NO+ from NO. The 0.005” rhenium ﬁlament installed in the source was
able to survive in the highly oxidative NO environment for only a few
hours. If the Anavac EI source is used under chemically reactive

conditions, a more robust ﬁlament material, such as thoriated-iridium
should be used.

II ' i i

The modiﬁed Nermag EI/CI source proved unacceptable for use as a
CI source with the present vacuum system. However, it was able to
produce very strong mass-selected ion beams when coupled to the Anavac
quadrupole and operating in E1 mode with the original 4 mm apertures in
place. Differential pumping could be employed to allow this source to
produce ion beams under high pressure CI conditions . A source chamber
could be constructed to enclose the source assembly. The new chamber
could be placed between the 6” source ﬂange and the matrix chamber 6”
ﬂange as shown in Figure 5.1.

By installing a separate pump, possibly a turbomolecular pump,
and an aperture plate after the quadrupole, this type of chamber could
greatly reduce the ﬂow of sample gas from the source into the matrix
chamber. The aperture plate would have to be sealed to the chamber
walls; it would allow ions to pass through a small conductance-limiting
aperture located close to the quadruple entrance aperture, while
discriminating against the ﬂow of neutrals from the source region into the
matrix isolation chamber. Because this aperture plate would have to be
permanently attached to the walls of the chamber, the pivoting feature of
the source could no longer be utilized. This should not prove to be a
difﬁculty; the source could be permanently positioned off-axis and the ions

deﬂected onto the cryostat window using deﬂection and focusing optics.

193

Also, the differential pumping provided by this type of chamber would
greatly reduce the total ﬂux of neutrals into the matrix chamber, possibly
making off-axis orientation unnecessary. One disadvantage to using this
type of differential pumping would be that the ion beam would need to
travel a greater distance between the quadrupole exit aperture and the

cryostat window, which could result in some loss of ion current.

 

   
 
 
  

source aperture
chamber / plate
‘22“;°\L deﬂection
ur o
t / optics Matrix
1:" Isolation
source I I _ Chamber
ﬂange /)h conductance
mount limiting
aperture

 

 

 

Figure 5.1 One possible design for a source chamber to allow differential
pumping of the source region.

I W i

Matrix isolation studies were performed using the Nermag source
with the Anavac quadrupole in the matrix isolation of Co“ from the
Co(CO)3NO precursor. Early attempts to form Co+ using the Dycor M200
EI source were unsuccessful because the compound decomposed when
directly exposed to the hot ﬁlament. The maximum current of Co‘“ (m/z =
59) produced by the Dycor M200 source was only ~5x10'11 A, this mass

spectrum is shown in Figure 5.2. The modiﬁed N ermag source with 4 mm

 

 

 

194

 

.3353 329?. 2.3 as accessed 3% $3.8m 05 .E @3633 one? ace. on... .AEoSoé Quezon com—z you»:
one 23 33315233. 33.3.3858 35.82 05 53, 68:33 OZHAOOVoO «a 85.58% mme 2E. Nd magma

om" cm—

. d

p:
a:

N
a? 8 \E o.» om. ow om cw

3 ﬂ \IJ

 

vv

«5 —.o u 3.6: :3

 

 

 

 

a...
S
N E
An we we \ mm Jm w_ .l@

 

<= 3 u 28m :3
on ma

 

 

an

195

apertures produced much higher currents of Co+ because the sample was
not directly exposed to the heated ﬁlament. An example of the output from
the Nermag source is also shown in Figure 5.2; the m/z = 59 current in
this spectrum was ~ 0.5 nA. When this source was initially turned on a
strong Co+ signal was observed, which decreased with time. This falloﬁ‘
was attributed to the slow radiative heating of the Nermag source by the
ﬁlament which increased the degree of sample pyrolysis. To increase
output from this source with such labile compounds, the system could be
adapted to allow the source to be water-cooled. Inlet and outlet water
feedthrough lines could be installed in the 6” mounting ﬂange, and
coupled to swagelock ﬁttings on both the vacuum and atmosphere ends of
the lines. A short length of ~1/8” stainless steel tubing could be brazed to
the body of the Nermag source and terminated with swagelock ﬁttings.
Teﬂon lines could be used to connect the source housing to the swagelock
ﬁttings on the ﬂange, while maintaining the necessary electrical isolation
between the two components. Using this system, water could ﬂow through
the inlet line, remove heat from the source, and ﬂow out through the outlet
line, cooling the body of the Nermag source. If this type of cooling system
is used, care must be taken to ensure that water leaks into the vacuum

system are prevented.

IVImrvi N r

Several methods could be investigated to improve the negative ion
output current from the Anavac source operating in NSI mode. In the Cs
glass bead NSI studies performed, the Cs bead wire was spot-welded to the
TCD feedthrough and inserted into the Anavac source housing. This
method of mounting made it difﬁcult to precisely align the Cs glass
surface with the source exit aperture. Poor alignment of these elements
could cause most of the negative ions emitted from the surface of the bead

to be neutralized on the source aperture plate, resulting in low negative ion

196

output from the source. Alternate bead mounting methods should be
investigated to ensure proper alignment of these elements. Several
alternate NSI surfaces and neutral molecules have been successfully used
by other researchers to produce high currents of negative ions as
mentioned in Chapter 2.I.C. Low work function surfaces such as
thoriated iridium or LaB6 could be adapted for use in the NSI source as

alternatives to the Cs glass system used in these studies. Other neutral
precursor molecules, such as KC1(S) coated on the NSI surface or C12(g),
could also be studied to determine if they could be utilized for the
production of high currents of negative ions through NSI. The output of
the Anavac NSI source may also be limited by the size of the ion exit
aperture. Increasing the size of this aperture could result in a stronger
output of negative ions, but would also increase conductance from the
source. Finally, the use of a stronger extraction ﬁeld at the NSI surface
could be investigated. By placing a bias on the source aperture plate, the
electric ﬁeld at the surface of the N SI bead could be varied. Increasing this
surface extraction ﬁeld may increase the current of negative ions exiting

the ion source.

V, Emissisg rsgulatign f9]: th Anavac systsm

One of the advantages to using the Dycor M200 mass-selected ion
source is that the modiﬁcations made to this unit allowed the original
ﬁlament emission controls to be retained. The ion sources adapted for use
with the Anavac quadrupole were connected to external power supplies
which had to be regulated manually. This was not a difﬁculty for the
source optimization studies, but became a limiting factor when the sources
were operated for the extended matrix isolation experiments. To make
these sources practical for such experiments, regulated power supplies
should be constructed. To design these supplies the method of ﬁlament

heater current regulation must ﬁrst be determined, or the power supplies

 

 

 

197

must be designed to function using several modes of emission regulation.
The original Anavac EI source could be operated by regulating ﬁlament
heater current through monitoring the total emission of the power supply,
or by monitoring the current of electrons at the source grid. The original
Anavac control unit ﬁlament regulation circuit could be adapted for use if
the ﬁlament heater current was monitored at the grid element. The
ﬁlament on the Nermag source could regulated by monitoring total
emission from the power supply or current at the collector, depending
upon whether the source is operated in CI or E1 modes. The ﬁlament in
the Anavac NSI source could be regulated through monitoring total
ﬁlament emission, or by monitoring the electron/negative ion current at
the source plate. The source plate current proved to be a more sensitive
measure of total ion output, but varied with changes in sample pressure

within the source.

KI D'ﬂi .1 . El] 12

The modiﬁed Dycor mass-selected ion source would also beneﬁt
from modiﬁcations, similar to those shown in Figure 5.1, to provide
differential pumping. During a normal matrix isolation experiment the
pressure within the deposition chamber was generally maintained below
5x10'7 torr. To maintain these pressures, the leak into the Dycor source
was controlled at a level that produced a partial pressure reading of 2x10'7
torr for the sample gas. At these low pressures the maximum mass-
selected ion current that could be measured at the cryostat window was ~1
nA. Studies performed at higher source pressures indicated that stronger
ion currents could be maintained, up to ~5 nA, if the ion source pressure
was increased. By implementing differential pumping between the E1
source and the matrix chamber, higher pressure source conditions could
be utilized without substantially elevating the pressure in the matrix

isolation chamber. This could be accomplished by installing a pump on

198

the source chamber housing and sealing this chamber with a plate
positioned between the ion source and the quadrupole. Such differential
pumping of the source chamber would remove most of the sample gas
lealn'ng from the ion source, while still allowing the mass-selected ion

beam produced by the quadrupole to be directed onto the cryostat window.

WWW

To lower the operating temperature of the Dycor source a lower work
function ﬁlament material, such as thoriated iridium, could be used in
place of a rhenium ﬁlament. This type of ﬁlament is more difﬁcult to
prepare, but operates at a lower temperature and is more resistive to
oxidation. Pre-fabricated Th-Ir ﬁlaments mounted on the Dycor ﬁlament
support rods can be purchased directly from the manufacturer. To
increase the output from the Dycor source the ﬁlament could also be
operated at higher emission levels, but this could reduce the lifetime of the
ﬁlament. In general this parameter, along with the repeller voltage,
electron energy, and focus voltages were optimized at regular intervals

during a matrix isolation experiment to maintain maximum ion current.

A ___o'. 7‘ or-.. 09' . 01 o, h o_ ' o .o_1 001'

To further increase the mass-selected ion output from both of the
sources, the operation of their quadrupoles could be modiﬁed. The remote
electronics head of the Anavac system allows the user to decrease the
resolution of the quadrupole. This is accomplished by varying VR2 on the
remote head, which determines the dc/rf ratio on the quadrupole. The
setting of VR2 could be varied to decrease the resolution of the quadrupole,
resulting in a strong increase in the ion current at the matrix window.
Increasing the ion energy, and removing the 3 m exit aperture on the

Anavac quadrupole housing could also be used to increase the ion current

199

at the cryostat window. The resolution of the Dycor unit could be reduced
by decreasing the HI RES and LO RES settings on the control unit. Ion
energy could be increased using the LO SENS and HI SENS parameters,
and the operation of the unit in BAR or TAB modes could also be
investigated.

lQEE-ll"'2'

Finally, modiﬁcations to the ion optics after the quadrupole should
be further investigated to determine whether the ion current at the cryostat
window can be increased. The Dycor source, in particular, could beneﬁt
from improved transfer of the mass-selected beam from the quadrupole to
the cryostat window. The energy and m/z dependence of the mass-selected
ion beam produced by Dycor quadrupole should be investigated by
visualizing the ion beam produced after the quadrupole using the CEMA
detector. It is possible that the use of a multistage optics system could
increase efﬁciency of the transfer of the ion beam onto the cryostat window.
A lens system which employs two sets of einzel lenses, coupled by a tube
element for physical support, is one system that might be considered.
Although it would be complex, an rf-multipole beam guide could also be

used to transfer the mass-selected beam onto the cryostat window.

200

duo..— uomcchoﬂo 36:8.— méa>a=< 05 .3 898.5. 33:85“; < ._ £233..

 

O; .0; o.

 

Ya iig‘I-Moiui

 

.! I 11 o!!!‘ I {Ii
vlllllal ill. .. I. .. t .

n
q

* I

 

\

 

 

‘I.

nxU 3N hsﬂama(\ 'shzla 4“, .
'al v.5! N 01.7.3232.

1. 10.".

"30. all ‘L .0 .IIoS

g. 0.0! O ‘00I'

m

 

 

.._

 

 

~1
' I ,IHQ‘OX‘

\

 

 

 

’l
3

 

 

wo-o—q—o
._. 2

5

 

i—f

 

 

 

 

m .

 

 

 

 

 

 

. _

.n _._

 

I48 3.. U...
I,“ w \\
.I . 7. 9‘“ a.
. o .l
a s3 Pb); . ‘J ‘sl/
1.6 ‘.2 lap. ‘Vo .0. — '5. g
' .- I.
.o .. I.
.. a I“.
.. .. ‘ ‘-

 

1......3:
.qtf. .21....

l_ as“?

 

 

 

a

7.... ._ .

*4»! I
'."l l .-

_

 

.
. u..'l

. TL .._.
l .:

M4.4. ).

 

 

 

 

 

 

 

lav—(3.2.) .7

IWW . -. .0? p

 

z-'o° iv;

no. - ‘.‘§l..! 09>

 

 

I.:.
é...

 

 

 

 

 

201

:35. 35:8 55: «-2323 05 no Esau—*5 2.3—=28 < .= £2.92“:

/

%

J
1.. if! ‘ ‘;.rﬂ& ,}\H‘ .nw

ﬂ .1 _‘

.1 . Iii e)

a:

 

202

Appendix III - Experimental conditions for the optimization studies
shown in Appendices IV - XIV

Appendix

IV

VI

VII

VIII

IX

Nermag source with Anavac quadrupole and channeltron detector
0.5 mm aperture exit plate on Nermag source, argon leak = 5.0

mtorr, chamber base pressure = 5.7x10'7 torr, total chamber

pressure = 6.3x10'7 torr, initial voltages: source = +3 V, repeller =
+15 V, collector = +5.02 V, Lens = -40 V, ﬁlament = -67 V, x-ray plate
= +50 V, detector = -2.7 kV, 50 uA total ﬁlament emission

Nermag source with Anavac quadrupole and channeltron detector
0.5 mm aperture exit plate on Nermag source, EC-NI mode, CCl4

leak , chamber base pressure = 5.2x10'7 torr, initial voltages: source

= -20 V, repeller = -45 V, collector = -20 V, Lens = +50 V, ﬁlament = -

120 V, x-ray plate = +3.5 kV, detector = -2.7 kV, detector aperture = -3
V, 50 uA total ﬁlament emission

Nermag source with Anavac quadrupole and channeltron detector
0.5 mm aperture exit plate on Nermag source, EC-NI mode, C0,,
and 100:1 Ar:CCl4 leaks, chamber base pressure = 5.2x10'7 torr,

initial voltages: source = -20 V, repeller = -60 V, collector = -20 V,
Lens = +50 V, ﬁlament = -120 V, x-ray plate = +3.5 kV, detector = -2.7
kV, detector aperture = -3 V, 50 uA total ﬁlament emission

Anavac surface source and quadrupole with channeltron detector
surface ionization with KNO3 bead, chamber pressure = 1.6x10'6

torr, initial voltages: bead = +3.0 V, ﬁlament housing = +9.0 V, lens
= - 40 V, x-ray plate = +50 V, detector = -2.7 kV,ﬁ1ament I = 0.8 A

Anavac surface source and quadrupole with channeltron detector
surface ionization with KNO3 bead, chamber pressure = 1.6x10'6 torr,

initial voltages: bead = +3.0 V, ﬁlament housing = +9.0 V, lens = - 40
V, x-ray plate = +50 V, detector = -1.2 kV,ﬁlament I = 0.8 A

Anavac surface source and quadrupole with Faraday plate detector
surface ionization with KNO3 bead, chamber pressure =

2.5x10’6 torr, initial voltages: bead = +3.0 V, ﬁlament housing = +9.0
V, lens = - 40 V, source = 0.0 V, ﬁlament current = 0.8 A

Anavac surface source and quadrupole with Faraday plate detector
negative surface ionization with cesium bead, chamber base

pressure = 1.4x10‘6 torr, CCI4 chamber pressure = 5.0x10'6 torr,

source pressure = 18 mtorr, bead = -2.0 V, initial ﬁlament housing =
-8.5 V, lens = +43 V, source = 0.0 V, ﬁlament current = 2.35 A

203

Appendix III (continued)

Anaemiix

XI

XII

XIII

IV

Anavac surface source and quadrupole with Faraday plate detector
negative surface ionization with cesium bead, chamber base

pressure = 1.4::10'6 torr, CCl4 chamber pressure = 6.3x10‘6 torr,

source pressure = 20 mtorr, bead = -2.5 V, initial ﬁlament housing =
-11 V, lens = +43 V, source = 0.0 V, ﬁlament current = 2.2 A

Anavac surface source and quadrupole with Faraday plate detector
negative surface ionization with cesium bead, chamber base

pressure = 1.4x10'6 torr, CCl4 chamber pressure = 6.3):10‘6 torr,

source pressure = 20 mtorr, bead = -2.5 V, initial ﬁlament housing =
-11 V, lens = +43 V, source = 0.0 V, ﬁlament current = 2.2 A

Anavac surface source and quadrupole with Faraday plate detector
negative surface ionization with cesium bead, chamber base

pressure = 1.41:10'6 torr, CCl4 chamber pressure = 5.0x10'6 torr,

source pressure = 18 mtorr, bead = -3.0 V, initial ﬁlament housing =
-8.5 V, lens = +43 V, source = 0.0 V, ﬁlament current = 2.2 A

Anavac surface source and quadrupole with Faraday plate detector
negative surface ionization, compating old Cs surface with cesium

bead after exposure to oxygen at 2.0x10'6 torr, chamber base
pressure = 1.0x10'6 torr, CCl4 chamber pressure = 5.8x10’6 torr,

source pressure = 18 mtorr, bead = -2.5 V, initial ﬁlament housing =
-11 V, lens = +35 V, source = 0.0 V

204

Appendix IV - Optimization studies for the Nermag source with 0.5 mm source aperture
coupled to the Anavac quadrupole, ion current measured with a channeltron detector.

m/z = 40
current

(x10nA)

m/z=40
current
(x10-8A)

 

m/z = 40 current versus repeller volatage

 

 

 

 

 

 

 

 

3 ‘ argon leak = 5 mtorr source pressure
7-1
6.
5.
4-1
3-1
d
2.1
1-1
0°ﬁI IfIrI'I I'ITIII l'l
111213141516171819 20 2122
Repeller Volt.
Maximum m/z = 40 current versus lens
3.0
2.5
2.0
1.5
1.0
0 5 R+201ntensity (E-8A)
0 R+15 Intensity (Ii-8A)
0-0 I T I ' I ' I I I ' I I I 1
-70 -60 -50 4o -30 -20 -10 0

Lens Voltage

205

Appendix V - Optimization studies for the Nermag source (with 0.5 mm source
aperture) coupled to the Anavac quadrupole, ion current was measured with a
channeltron detector. The-spectrum was taken under EC-CI conditions using
carbon tetrachloride.

EC-NI optimization of rf-only current versus
source pressure of carbon tetrachloride

  
  

 

 

 

 

 

 

 

 

4.00:,8 -
soot-,3 -
rf-only
Current 2,006-3 .
(A)
1.00e-8 4
0.00e+0 1 I I f I ' I ‘ 1
o 200 400 600 800 1000
Source Pressure (mtorr)
Detector output vs source pressure
Detector
Current
(A)
El Rfonly (A)
0 W222 (A)
10.10 I ' I f I ' I Y I ' j
o 200 400 600 300 1000

Source Pressure (mtorr)

206

Appendix VI - Optimization studies for the Nermag source (with 0.5 mm source
aperture) coupled to the Anavac quadrupole, ion current was measured with a
channeltron detector. EC-CI conditions were maintained using carbon
tetrachloride and argonzcarbon tetrachloride mixtures.

rf-only current vs source pressure for pure CC14

12“ ‘ and 100:1 Ar:CCl4

 

    
   
   

 

 

 

 

100 ~
‘ - 100:1rfdiff
80 d 0 pure CC14 rfdiff
rf-only

Current 60 ..
(0A) <
.10 .
20 <
0 -1

 

I ' I ' I f T " I ' I

, . .
0 100 200 300 400500600700 800 9001000
Source Pressure (mtorr)

pure CC14 vs ArzCCl4 100:1, m/z=35 negative ion signal
5 1

 

' a 100:1m/z=35
0 pure CC14 m/z35

 

 

m/z = 35 3 ‘
Current ‘
(HA) 2 .

 

 

o-l 'IrI'T'IiIfI

01002003004005006007008009001000
Source Pressure (mtorr)

207

Appendix VII - Optimization studies for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a potassium bead,
and ions were detected by a channeltron detector.

Source and rf-only current vs ﬁlament housing voltage

 

 

 

 

 

 

 

40 - 10 '5
35 J —G— Source 1 (nA)
‘ —O— rf-only l (A)
30 - I
25 - .
Source 1 .‘ rf-only
Current 20 -1 current
(nA) 1 (A)
15 4
10 «
5 1
l s _ -
0 ' f ' I ' I I I . I a J u n 10
2 4 6 8 10 12 14 16 18 20
Filament Housing Voltage (V)
rf-only current versus lens voltage
rf-only
current
(A)
OC+O ' I 1 I ‘— T ﬁ— I Y I ﬁ' I

 

 

-70 -60 -50 «lo -310 -20 -10 0
Lens Voltage (V)

208

Appendix VIII - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a potassium bead,
and ions were detected by a channeltron detector.

Source current and rf-only current versus heater current

 

 

 

 

 

1000 10'7
-8
100 10
. - 10'9
10 . Detector
Source 10'10 Current
Current 1 (A)
(nA) 10.11 rf-only
.1 10.12
.01 —°— SourccHnA) 10.13
—.'— Dctectorl(A)
.001 ° ' I v I ' T v I I I10-14
04 0.6 0.8 1.0 1.2 1.4 1.6

Heater I (A)

209

Appendix IX - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a potassium bead,
and ions were detected at a faraday plate mounted after the quadrupole.

rf—only and source current vs heater current

     

 

 

 

 

 

 

 

 
 

 

 

 

250 - . 2012-8
4 P
2004 ,
1.5c-8
1
Source 150 ‘ rtlonly
Current Current
(nA) ‘ )- 1.0e-8 (A)
100 -1
.- 5.0-3-9
50 d —0— source I (nA)
—9'—' rf-only l (A) h
0 , A , . ' 0.0c+0
0.3 0.9 1.0 1.1 1.2 1.3
Heater Current (A)
rf-only and m/z = 39 current vs heater current
2.0e-8 - - 2.0
l.5e-8 -l 1.5
rf-only , m/Z=39
Current Current
(A) 1.0e-8 - -1.0 (nA)
5.0e-9 - - 0.5
—a— rf-only I (A)
0 —°—" m/s=39 I (nA)
0.0e+0 , . , . , . . ' . , . 0.0

 

 

0.3 0.9 1.0 1.1 12 1.3
Heater Current (A)

210

Appendix X - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a cesium bead biased
at -2.0 V, with a leak of carbon tetrachloride. Ions were detected at a
Faraday plate mounted after the quadrupole.

rf-only and m/z = 35 current vs time

 

 

 

 

 

 

 

 

4.00-10 1 _ 6.0e-ll
I
- 5.5e-ll
3.5e-10 -1 .
rf-only ‘ - 5.0e-ll
current . m/z=-35
(A) 3.0e-10 m 1- 4.5e-ll Current
L (A)
l » 4.0.3-11
2.5c-10 ‘
—a— rf-only I (A) .. 3.5e-11
—°— mz=-35 I (A) _
2.0c-10 . I . r r I . I . l . 3.0e-11
0 10 ' 20 30 40 50 60
time (min)

source and rf-only currents vs ﬁlament housing voltage

 

 

 

 

 

 

 

300 ! 4.06-10
—9— source I (nA) 3.5e-10
250 F -—0— rfonly I (A)
- 3 3.0c-10
Source 200 I " 2 Se 10
Current ‘ ° I lef-only
(nA) 150 - U .. 2.0510 urrent
. ., .. > (A)
100 ~ 1.5e-10
P - 1.06-10
50 d . . '
n r - 5.06-11
i c C r
0 . ~ . . . . . , ﬂ 0.0e+0
-12 -ll -10 -9 -8 -7

Filament Housing Voltage (V)

211

Appendix XI - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a cesium bead biased
at —2.5 V, with a leak of carbon tetrachloride. Ions were detected at a
Faraday plate mounted after the quadrupole.

source and rf-only currents vs ﬁlament heater current

 

 

 

 

 

 

 

 

   
  
 
 
 
  
  

 

 

 

 

 

 

500 .1 - 5.00-10
‘ —6'— source I (nA)
—0— rf-only I (A)
source rf-only
current current
(nA) (A)
1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3
ﬁlament heater current (A)
source and rf-only currents vs ﬁlament housing voltage
300 - - 4c- 10
—0— source I (nA) . 4e- 10
250 ‘ —9— rf-only I (A)
- 3e-10
200 - '
Source - 2e- 10 rf-onl
Current ‘ 1 C y
(n A) 150 . _ 26,10 urrent
. ~ (A)
- le-lO
100 ~ .
d b 16.10
50 1 ,
- 5e-ll
o . a . f . 0e+0

 

v I I I I ' I ' I
.13.0 -12.5 .12.0 -11.5 -11.0 -10.5 -10.0 -95 -9.0
Filament Housing Voltage (V)

212

Appendix XII - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a cesium bead biased
at -2.5 V, with a leak of carbon tetrachloride. Ions were detected at a
Faraday plate mounted after the quadrupole.

rf-only and m/z=35 current vs manometer pressure

    

 

 

 

 

 

 

 

 

 

 

 

 

60°10 -« 1.2e-10
soc/10 - _ 1.0010
rf-only 4.0e-10 -1
. - 8.0.5.11
current m/z=-35
(A) 3.0e-10 - current
. _ 6.03.11 (A)
2.0:,10 -
—a— -4.oe-11
1.0e-10 q rf-only I (A)
l —o— m/z=-35 l (A) »
o.o¢+o I . , . f . . . . . 2.0e-11
0 10 20 30 40 so
manometer pressure (mtorr)
source and rf-only currents vs manometer pressure
180 - r 6.0e-10
I
160- h - 5.0c-1O
4.0e-10
source 1407 — f I rf—onl
current ‘ ° —9— some 1 (M) * Yt
. .. . (A)
100 - .. - 20:40
so - . 1.1.0:-10
50 - . j . , . , . , . . 0.0e+0
0 10 20 30 4o 50

manometer pressure (mtorr)

213

Appendix XIII - Optimization study for the Anavac source and quadrupole.
The source was conﬁgured for surface ionization using a cesium bead biased
at -3.0 V, with a leak of carbon tetrachloride. Ions were detected at a
Faraday plate mounted after the quadrupole.

source and rf-only currents vs filament housing voltage

 

 

 

 

4.0¢+2 1 . 1- 5.06-10

I l

35°” "j —9— Source I (nA)
1—4.0e-10
3.0e+2-' —"— rf-onlyHA)

J ' rf
source 2.0“2 1 ° _ -only
current . 2 06 1 current

(nA) l.5e+2-4 ° F ' ' (A)
1.0e+2~ c

1 v "1.0C'10

5.0e+l -‘ .,

. .. F

0.0c+0-°°°-—.-,.,.,.,-Tﬁ,.,-,-,o.oc+0

 

 

 

-150 -145 -140 -135 -130 -12.5 -120 -115 -110 -105 -100 -9.5
ﬁlament housing voltage (V)

214

Appendix XIV - Optimization study for the Anavac source and quadrupole
the source was conﬁgured for surface ionization using a cesium bead biased
at -2.5 V, the aged Cs surface (no 02) and the same surface after exposure

to oxygen (w/02) are shown.

source vs ﬁlament heater current

 

 

 

1000
100 El no 02 Source I (nA)
0 w/02 source I (nA) J
10 *
Source
current 1 -
(nA) ~
.1
.01 9
.001 f 1 I I I f ' v I f I r I r I r I 1 r f I

 

1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3
Filament Heater Current (A)

rf-only current vs ﬁlament heater current

46-10'1

 

a no 02 rf-only I (A)
‘ 0 w/OZ rf-only I (A)

 

 

 

36-10 -1
rf-only

current
(A) 25-10 «

 

u
————_ _

06w 0‘ I I I I I I I I

. .
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3
Filament Heater Current (A)

 

 

 

215

—.nN

 

new; mom—om E comma: +mm0 mo 82.8QO «:4 25. 5% 56:33..

m.NN

0:011 . .,III.'..

_

muminzu>§>
Anaconaozhv
h.NN n.NN
_ L _ _

.IIIIII III! III 1I.II'I|II.III.II.II III. III-I1 I.

3358 oo..o£.mcoon£u_3 .mcovnxsoo

mommﬁz J<Zo_m -221

 

 

can

own

Ohm

can

own

com

O—o

ONm

Gnu

MISNBINI

PS”

10.

11.

12.

13.

14.

15.

16.
17.

216

LIST OF REFERENCES
L.F. Halle, R. Houriet, M.M. Kappes, R.H. Staley, and J .L.
Beauchamp, J. Am. Chem. Soc., 104, 6293 (1982).
DB. Jacobson and BS. Freiser, J. Am. Chem. Soc, 105, 5197 (1983).
M.M. Kappes and R.H.J. Staley, J. Am. Chem. Soc., _1_Q_4, 1813 (1982).
DB. Jacobson, and BS. Freiser, J. Am. Chem. Soc., 105, 736 (1983).
KL Busch, G.L. Glish, and SA. McLuckey, Mass Spectrometry/ Mass
Spectrometry: Techniques and Applications of Tandem Mass
Spectrometry, VCH Publishers Inc, New York, 1988, 107-151.
J .L. Holmes, Org. Mass Spectrom., 2_Q, 169 (1985).
J .D. Morrison, in M.T.P. International Review of Science, Physical
Chemistry Series One, Vol. 3: Mass Spectrometry, A. Maccoll, ed.,
Chap. 2, Butterworths, London (1972).
PC. Burgers and J .L. Holmes, Org. Mass Spectrom., 12, 123 (1982).

J. Berkowitz and KC. Groeneveld, eds., Molecular Ions; Geometric
and Electronic Structures, Plenium Press, New York, 1980.

J .P. Maier, Chem. Soc. Rev., 1'1, 45 (1988).

RC. Woods, Ian and Cluster Ion Spectroscopy and Structure,
J .P. Maier, ed., Elsevier, New York, 1989, 2758.

Z. Vager, R. Naaman, and ER Kanter, Ion and Cluster Ion
Spectroscopy and Structure, J .P. Maier, ed., Elsevier, New York,
1989, 1-26.

M.E. Jacox, J. Phys. Chem. Ref. Data, 11, 269 (1988).

T. Amano, Phil. Trans. R. Soc. Lond. A, 324, 163 (1988).

E. Whittle, D.A. Dows, and CC. Pimentel, J. Chem. Phys., 22, 1954
(1943).

M.C.R. Syrnons, Chem. Soc. Rev, 1.3, 393 (1984).
R.N. Perutz, Chem. Rev., 85, 97 ( 1985).

18.

19.

20.

21.
22.

30.
31.
32.

33.

35.

J .V. Coe and R.J. Saykally, Ion and Cluster Ion Spectroscopy and
Structure, J .P. Maier, ed., Elsevier, New York, 1989, 131-154.

L. Andrews, Advances in Infrared and Raman Spectroscopy, R.J.H.
Clark and RE. Hester, eds., Vol. 7, Heyden, London, 1980, 59-85.

M.J. Almond and A.J. Downs, Advances in Spectroscopy, R.J.H.
Clark and RE. Hester, eds., Vol. 17, John Wiley & Sons, New York,
1989, 107-185.

L.B. Knight Jr., Acc. Chem. Res., 19, 313, (1986).

L. Andrews, Absorption Spectroscopy of Molecular Ions and
Complexes in Noble-Gas Matrices, L. Andrews and M.

Moskovits, eds., Elsevier, Oxford, 1989.

L. Andrews, Ann. Rev. Phys. Chem., 30, 79 (1979).

D. E. Milligan and ME. Jacox, J. Chem. Phys, ﬂ, 1952 (1969).

J .R. Morton and K.F. Preston, Inorg. Chem., 2_4, 3317 (1985).

C.A. Wight, B.S. Ault, and L. Andrews, J. Chem. Phys., 65, 1244
(1976).

L. Andrews, J .M. Gryzbowski, and R. 0. Allen, J. Phys. Chem., 19,
904 (1975).

L.B. Knight J r., J .M. Bostick, J .R. Woodward, and J .J . Steadman, J.
Chem., Phys., 18, 6415 (1983).

L.B. Knight Jr., P.K. Miller, and J .J . Steadman, J. Chem. Phys., &1,
4587, (1984).

L.B. Knight Jr. and J .O. Herlong, J. Chem. Phys., 91, 69 (1989).
J. Hacaloglu and L. Andrews, Chem. Phys. Lett., 160, 274 (1989).

D. Forney, M. J akobi, and J .P. Maier, J. Chem. Phys., 9_Q, 600
(1989).

MS. Sabo, Ph.D. Dissertation, Michigan State University, 1990.
P. Lenard, Ann. d. Phys., 8, 149 (1902).

T. D. Mark and G. H. Dunn, eds., Electron Impact Ionization,
Springer-Verlag, Vienna (1985).

36.

37.

38.
39.

40.

41.
42.

43.

45.

46.
47.
48.
49.
50.
51.
52.
53.

218

M. E. Rose and R. E. W. J ohnstone, Mass Spectrometry for Chemists
and Biochemists, Cambridge University Press, Cambridge, England
(1982).

T. D. Mark in Gaseous Ion Chemistry and Mass Spectrometry,
Jean H. Futrell, ed., John Wiley & Sons, New York (1986) p 61-93.

T.D. Mark, F.Egger, and M. Cheret, J. Chem. Phys, 81, 2629 (1977).

K. Biemann, Mass Spectrometry: Organic Chemical Applications,
McGraw Hill, New York (1962).

F. Pichou , R. I. Hall, M. Landau, and C. Scherman, J. Phys., B16,
2445 (1983).

L. J. Kieffer and G. H. Dunn, Rev. Mod. Phys., 88, 1 (1966).
A.J. Dempster, Phys Rev., LI, 316 (1918).
A. O. Nier, Rev. Sci. Instrum., 11, 212 (1940).

H. E. Duckworth, Mass Spectrometry, Cambridge University Press,
1958, 37-41.

J. T. Watson, Introduction to Mass spectrometry, Raven Press, New
York 1985.

T. D. Mark, Int. J. Mass Spectrom. Ion Phys., 48, 125 (1982).

A. O. Nier, Rev. Sci. Instrum., 18, 398 (1947).

W. Paul, H. P. Reinhard and U. von Zahn, Z. Phys., _1_5_2, 132 (1958).
J. H. Batey, Vacuum, 31, 659 (1987).

EH. Field and M.S.B. Munson, J. Amer. Chem. Soc., 81, 3289 (1965).
M.S.B. Munson and EH. Field, J. Amer. Chem. Soc., 88, 2621 (1966).
F. Hunt , Adv. Mass. Spectrom., 8, 359 (1973). I

B.L. J elus, B. Munson, and C. Fenselau, Biomed. Mass.
Spectrometry., .1, 96 (1974).

F. H. Field, Chemical ionization mass spectrometry, Accounts Chem.
Res., 1, 42 (1968).

 

55.

56.

57.
58.

59.

61.

62.

63.

65.
66.
67.

69.

219

C. E. Melton, Principles of Mass Spectrometry and Negative Ions,
Marcel Dekker Inc., New York, 1970.

D. F. Hunt, G.C. Stafford, F.W. Crow, and J .W. Russell, Anal.
Chem., 48, 2098 (1976).

EA. Stemmer and RA. Hites, Anal. Chem., 51, 684 (1985).

GE. Klotz, Fundamental Processes in Radiation Chemistry, P.
Ausloos, Ed., John Wiley and Sons, New York, 1968, p 40.

RC. Dougherty, Anal. Chem., 83,, 625 (1981).
J .H. Bowie, Mass Spectrom. Rev., 8, 161 (1984).

HR Tannenbaum, J .D. Roberts, and RC. Dougherty, Anal. Chem.,
41,49 (1975).

M. Deinzer, D. Grifﬁn, T. Miller, J. Lamberton, P. Freeman, and V.
Jonas, Mass Spectrom., 9, 85, (1982).

AC. Harrison, Chemical Ionization Mass Spectrometry, CRC Press,
Boca Raton, Fla, (1983).

HE. Duckworth, Mass Spectroscopy, Cambridge University Press,
Cambridge, 1960, p 35.

J. Elster and H. Geitel, Ann. Phys., Chem., 81, 315 (1889).
J. C. Beattie, Philos. Mag., 48, 97 (1899).

J .J . Thomson, Proc. Cambridge Philos. Soc., 15, 64 (1908).
J .P. Blewett and E.J. Jones, Phys. Rev., 89,, 464 (1936).

J .C. Blais, A. Brunot, and M. Cottin, Int. J. Mass Spectr. Ion Phys.,
22, 71 (1976).

R.Grosser, KG. Heumann, and W. Schindlmeier, Adv. Mass Spectr.,
J .F.J . Todd, ed., B, 1057 (1985).

D.Bombick,Ph.D. Dissertation, Michigan State University, 1986.
J .B. Moreau, Rev. Phys. Appl., 3 , 286 (1968).
NI. Ionov, Prog. Surf. Sci., 1, 237 (1972).

74.

75.
76.
77.

78.
79.
80.

81.

82.

83.

85.

86.
87.
88.

89.

90.

91.
92.

220

T. Fujii and H. Jimba, Int. J. Mass Spectrom. Ion Phys., 19, 221
(1987).

G. Owen, Proc. Cambridge Philos. Soc., 12, 493 (1904).
O.W. Richardson, Phys Rev., 34, 386 (1912).

M.J. Cardillo, M.S. Chou, E.F. Green, and DB. Sheen, J. Chem.
Phys., 54, 3054 (1971).

H. Kawano, Shinku, Vacuum, 28, 1 (1980).
H. Kawano, Shinku, Vacuum, 28, 229 (1980).

H. Kawano, and FM. Page, Int. J. Mass Spectrom. Ion Phys. , 58, 1
(1983).

H. Kawano, Y. Hidaka, and EM. Page, Int. J. Mass Spectrom. Ion
Phys. ,80, 34 (1983).

H. Kawano, Y. Hidaka, M. Suga, and EM. Page,Int. J. Mass
Spectrom. Ion Phys. , 8Q, 77 (1983).

H. Hotop and WC. Lineberger, J. Phys. Chem. Ref. Data, 4, 539
(1975).

A. Persky, E.F. Green, and A. Kuppermann, J. Chem. Phys., 49, 2347
(1968).

J .C. Blais and G. Bolbach, Int. J. Mass Spectr. Ion Phys., 2_4, 412
(1977).

J .M. Lafferty, J. Appl. Phys., 22, 299 (1951).
L.J. Favreau, Rev. Sci. Instr., 88, 856 (1965).

H. Kishi and H. Kawano, Int. J. Mass Spectr. Ion Phys., &, 301
(1988).

N. Kashihira, E. Vietzke, and G. Zellerman, Rev. Sci. Instrum., 48,
171(1977).

S. Dushman, Rev. Mod. Phys., 2, 381 (1930).

D. Bombick and J. Allison, J. Chrom. Sci., 21, 612 (1989).
W. Geiger, Z. Phys.,1_40, 608 (1955).

 

93.
94.

95.

97.

98.

99.

100.

101.

102.

103.

104.
105.
106.
107.
108.
109.

110.

111.

221

H. Kawano, J. Chem. Phys., 18, 3471 (1971).

J .W. Trischka, D.T.F. Marple, and A. White, Phys. Rev., 85, 136
(1952).

J .T. Cheng and S. Datz, J. Chem. Phys, 11, 1814 (1979).

J .F.J . Todd and G. Lawson, Int. Rev. Sci. Mass Spectrometry,
Physical Chemistry Series two V5, A. Maccoll and AD.
Buckingham, eds., Butterworths, London, 1975.

PH. Dawson, Mass Spectrometry Reviews, 1986, John Wiley & Sons
Inc., 5, 1 (1986).

Extrell Corporation 1978 short course, theory and operation of
quadrupoles.

P. E. Miller and M. B. Denton, J. Chem. Ed., 88, 7 (1986).

D.W. McLachlan, “Theory and Applications of Mathieu Functions”,
Clarendon: Oxford, 1947.

FM. Mao and J .H. Leck, Vacuum, 31 , 669 (1987).

N. Ierokomos, 20th ASTM Meeting on Mass Spectrometry and Allied
Topics, 221, Dallas, Texas.

The ARRL handbook for the Radio Amateur, B.S. Hale, ed.,
American Radio Relay League, Newington, Ct, 1989.

RC. Wilson and GR. Brewer, Ion Beams, John Wiley and Sons, 1973.
J .E. Delmore, Int. J. Mass Spectrom. Ion Processes, 58, 151 (1984).
P. Grivet, Electron Optics, Pergamon Press, New York, 1965.
A.O.C. Nier, Rev. Sci. Instrum., 18, 336 (1947).

LA. Dietz, Rev. Sci. Instrum., 38, 235 (1959).

Hsi-men, and Chi-yeh, Advances in Electronics and Electron Optics,
J. Ximan, ed., Supplement 17, (1986).

E. Harding and RH. Read, Electrostatic Lenses, Sydney and
Machenster, 1974.

T. Matsuo, Jpn. J. Appl. Phys., 23, 1518 (1984)

 

112.

113.

114.

115.

116.

117.
118.

119.

120.

121.
122.

123.
124.
125.
126.
127.
128.

129.

222

H. Wollnick, J. Brezina, and M. Berz, Nucl. Instrum. Methods A,
258, 408 (1987).

R.J. Ryba and G.J. Mains, Int. J. Mass Spectrom. Ion Physics, 42, 169
(1982).

A. Vertes, P. Juhasz, and L. Matus, Int. J. Mass Spectrom. Ion
Processes, 13,, 109 (1986).

A. Vertes, P. Juhasz, L. Balazs, M. de Wolf, and R. Gijbels, Int. J.
Mass Spectrom. Ion Processes, 84, 255 (1988).

HT. Shaw, S. Niemeyer, and CF. Hunt, Int. J. Mass Spectrom. Ion
Processes, 88, 317 (1986).

DC. McGilvery, Ph.D. Dissertation, LaTrobe University, 1978.

DA Dahl and J .E. Delmore, presented at the 34th annual A.S.M.S.
conference 1986 (pg 304).

D. A. Dahl, J. E. Delmore, and A. D. Appelhans, Rev. Sci. Instrum.,
81, 607 (1990).

DC. McGilvery and R.J.S. Morrison, MacSIMION, PTY LTD,
Clayton, Victoria, Australia, 1988.

J.E. Delmore, Int. J. Mass Spectrom. Ion Processes, 58, 151 ( 1984).

HS. McKown, D.H. Smith, and BL. Sherman, Int. J. Mass
Spectrom. Ion Phys., 81, 39 (1983).

Wright, M. C., Inst. Phys. Conf. Ser., 94, 363 (1989).

J .E. Delmore, Int. J. Mass Spectrom. Ion Processes, 81, 121 (1985).
WC. Wiley and CF. Hendee, IRE Trans. Nucl. Sci., N - , 103 (1962).
J. L. Wiza, Nucl. Instrum. Methods, 182, 587 (1979).

J .G. Timothy and BL. Bybee, Rev. Sci. Instrum., 48, 202 (1977).

CE. Grifﬁn, H.G. Boettger, and DD. Norris, Int. J. Mass Spectrom.
Ion Phys., 18, 437 (1974).

Guor-Rong Her, Ph.D. Dissertation , Michigan State University,
1985.