..

.‘7

..
i

, .ﬂ..-

Y. . ‘ jv'v'v'r‘ . '. " '
. i H .‘1' V . . -'.'" ’ . ‘ ”’1’” '11 z"'T-""11‘1'|€

:"14 11121211 1:11'1111' 11,11,1- 111'};
111 “1113'

1“ .—"
1 11111

f‘.‘
F.
q- A'-

._.I '.0

-.‘-... _'

:2:
..- 0-»
v- 4‘

--.a~

 

ltd—’1.
.‘-

.1.”

J."—

‘1

1.331 . 21%
1. .
. 111111"m~"11" 1 1 1 1- 1 1:.'

.3.1'1
1111 I" ‘ f ”In, 1. 1M! :-‘1:.s ' 3311' i
.4: ‘ - $531. .~'.~1:»11‘-. ..
' 11 31.3111. ' . '1 1111.1..111 .11

33-7
sLigﬁg
. .
$35323, .
‘J .’ ‘._. :11

 

Q
n

.
ir'v "
£2

Ilrféi lg (“:1 ‘
.i- 1113111111 111.31% £111 £13“;
J J" ' '1 1“!“

:51 '1'.

. H- h'?

‘ [,1 €111,“ 1"1'1'n'ti'11‘r r

. <1 ‘ ' ' '7'. W

_ 31m. 1 , .qnlggu111'
2.! 1‘: 11.1111. "

111531;};-

1‘.
"1* 1".1 311'!”
3 111 ‘

r "1

      
  
  

 

J

 

.gw 3:
‘1

44.332 '4 :— qﬂ-

 

"r’

-d. a
v

’11

'1

.

I?

1

L.

.V . ,1 .
'F 3: :
-: ”b:-

' '7

—.- X ‘¢-n
NA-..or u
- aviw 7 rd
..1 .3 cl

.
v

"v"

I_ -}

‘3raez-EI.-.r.:‘-.~r;

‘II
V 1:535:

 

5.3.
h.

.~

296)}. ‘ V 1 _ ‘
Ix:.'!',,':-1 .‘ '. ._ j’. 11‘ "11.3.”
. '1 '1 o ' q' ' : ' . ’ '11)}1 "1
11‘11‘ ' c' ‘ ’1.;-'

jit.‘

"L

49a

...‘.¢-J-.
. ' I 4—
.

w
“3:"-

 

. - —.-¢ - ."
>1-v.-‘.°?‘ ‘
» tam"...
“w"? -
1‘ ‘w

;.__‘....-_:5
z}:

1
I I
'!

|j
11111011

    

‘ “1""
Id‘ﬂiawuhll 110111141... 1!

3 llllllllllllllllllll

(iQQX)

This is to certify that the

dissertation entitled

NON-LINEAR ION ACCELERATION FOR IMPROVED MASS
RESOLUTION IN TIME-OF-FLIGHT MASS SPECTROMETRY

presented by

Benjamin DeCamp Gardner

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Chemistgy

a ' Major professlor T
Date W;

”(Iii-n. ‘0‘ 1' A * F1 lnrr I: 1 - . 0‘1 1

 

 

 

LIBRARY

Michigan State
University

 

 

 

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

 

_ _‘uupu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

use mm.m.i4

 

NON-LINEAR ION ACCELERATION FOR IMPROVED MASS
RESOLUTION IN TlME-OF-FLIGHT MASS SPECTROMETRY

By

Benjamin DeCamp Gardner

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry
1997

ABSTRACT

NON-LINEAR ION ACCELERATION FOR IMPROVED MASS
RESOLUTION IN TlME-OF-FLIGHT MASS SPECTROMETRY

By

Benjamin DeCamp Gardner

The continued advancement of chromatographic separations techniques serves as a
primary driving force for improved detector technologies. The capacity to rapidly
separate complex analyte mixtures and achieve temporally narrow eluent peaks poses a
serious challenge to even the most sophisticated detectors. To date, the mass
spectrometer, which provides a second axis of information for multi-dimensional
analysis, has been the detector of choice for the most discriminating analyses. Given the
diminishing time ﬁ'ames, the time-of-ﬂight mass spectrometer currently stands as the
most appropriate detector technology, due to its combination of hi gh-speed data
acquisition capability, multi-channel detection speciﬁcity and superior detectability. Yet,
if the timeof-ﬂight mass spectrometer is to remain relevant to the emerging
chromatographic techniques, fundamental improvements to basic mass spectrometer

design will continue to be required.

In response to the challenges posed by the increasing chromatographic information
density, research at Michigan State University has resulted in the development of a time-

of-ﬂight mass spectrometer and an accompanying data acquisition system that provides

mass spectral acquisition rates of up to 500 Hz and fully accommodates the requirements
of high-performance gas chromatography-mass spectrometry. This dissertation
documents some of the developments of this time-of-ﬂight mass spectrometer system
along with performance characteristics. It further examines the fundamental limitations to
the achievement of mass resolving power and speciﬁcally delineates a method
compensating for one of these limitations, that posed by the initial spatial distribution of
ions along the optical axis within the time-of-ﬂight ion source at the moment of creation
in, or capture by, the applied electrical ﬁeld. For the ﬁrst time, the limitations to the
model of space focusing using two-ﬁeld linear ion acceleration developed by Wiley and
McLaren .[1] in 1955 is experimentally veriﬁed. A superior model based on a
combination of linear and non-linear acceleration is theoretically evaluated and
experimentally veriﬁed using a novel time-of-ﬂight mass spectrometer designed to allow
the probing of ion acceleration behavior in selected ion source regions with site-speciﬁc
laser-induced photo-dissociation. The data show that, in extended-length ion sources, a
greater than seven-fold improvement to resolving power may be expected by using non-
linear ion acceleration. Finally, the relative merits of non-linear ion acceleration are
discussed along with the potential application to various analytical techniques in which

analyte introduction necessitates the use of a relatively large ion optical volume.

 

l. Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150.

Copyright © by
BENJAMIN DECAMP GARDNER

1997

”2-4

ACKNOWLEDGMENTS

While only one person puts his or her name on the cover, the research that results in
endeavors such as this is invariably the product of many hands. My work is no exception
to this, and there are many individuals to whom 'I am deeply grateful for their help. First,
I am indebted to Dr. Watson for his guidance in my academic progress through the Ph.D.
program. Dr. Watson has been very supportive of the unusual arrangement that led to my
association and collaboration with Dr. Holland, and has many times ﬁlled in important
gaps in my professional development. To Jack Holland I am especially grateﬁil. Jack, you
taught me to view science in a unique way and to pursue my instincts, even when the
model I proposed was controversial. I will always be grateful for your support and
encouragement. Your unconventional guidance, both in academia and industry, has
helped lay a solid foundation for me, and your friendship has been priceless. To quote a

time-wom phrase, “I owe you dinner.”

Many others have also played a role in my success: Ron Tecklenburg, Raimund Grix,
Chris Enke, George Yefchak, and David Sparkman have each been fruitful collaborators
and good ﬁ'iends, as has Richard Berkley at the EPA, and Don Patterson, John Barr, and
James Grainger at the Centers for Disease Control in Atlanta. Still others like Mike
Davenport, Mel Micke, and Melinda Beming have each been of incredible assistance to
me in the Mass Spectrometry Facility. My co-workers, including Ken Roth, Dave

McLane, Paul Vlasek, Jennifer Johnson, and I have certainly shared many good times in

the laboratory. Jennifer, especially, has been a close friend, and continues to encourage
me, even from a distance. Jim Munz, an extraordinary machinist in the Physics
Department machine shop, is responsible for the vacuum welding and creating the
mechanical parts which I merely assembled. I also thank Slava Artaev and Kevin McNitt
at Meridian Instruments (now at LECO Corp.) for their friendship, both while and since I

worked there.

To my family and ﬁ'iends away from the halls of academia, at my church and
elsewhere, words just aren’t sufﬁcient for me to express my thanks for your support and
personal friendships. We have shared much in our lives, and I know that mine is much

richer because of it. I thank God for each of you.

vi

TABLE OF CONTENTS

LIST OF TABLES x
LIST OF FIGURES xi
CHAPTER 1
GENERAL INTRODUCTION 1
Introduction 1
General Principles of GC—MS 2
The Problem 3
A Historical Perspective 9
The MSU Connection 11
Part One of the Plan - Instrumentation For Time Array Detection 12
Part Two of the Plan — Improvements to Mass Resolving Power 13
Research Goals 14
Chapter References 15
CHAPTER 2
THEORY AND INSTRUMENTATION DEVELOPMENT IN TIME-OF-
FLIGHT MASS SPECTROMETRY 17
Introduction 1 7
Early Development of TOFMS 20
The Need for Resolving Power 21
Deﬁnitions of Resolving Power and Mass Resolution 23
Problems Limiting Resolving Power 25
Methods for Improving Resolving Power 29
The Wile y/McLaren Ion source ‘ 29
Subsequent Ion Source Methods for Improving Resolving Power 31
Sector Focusing 32
Dynamic Fields 33
The Ion Minor 33
The Effect of Ionization Method 35
Summary ' 37
Chapter References 39
CHAPTER 3
THE MICHIGAN STATE UNIVERSITY M-TOF 42
Introduction 42
Duty Cycle Requirements 43
The M-TOF 45

Ion Storage and the Gn’x Source 46

vii

Subsequent Ion Sources 50

Further Developments 57
Summary ' 58
Chapter References 59
CHAPTER 4
NON-LINEAR ION ACCELERATION 60
Introduction 60
The Wiley/McLaren 1“t Order Space Focus Model 60
Ion motion through an electric ﬁeld 61
Space Focusing 65
Ideal Space Focusing 69
Dynamic Non-linear Fields 69
Static Non-linear Fields 70
Non-parabolic Non-linear Fields 73
Experimental 82
Experimental Apparatus 82
Electric Field Calculations 91
Experimental Method ' 92
Evaluation of linear conﬁguration 93
Evaluation of Non-linear Conﬁguration 99
Expectation Using A Multiple-ﬁeld Non-linear Approximation 99
Performance Using Calculated Design 101
Performance Using Empirical Voltages 104
Effect of Deliberate Perturbation 108
Attempts at Evaluating A Grid-free Non-linear Conﬁguration 110
Summary and Signiﬁcance 116
References 119
CHAPTER 5
PROSPECTS FOR FUTURE IMPROVEMENTS TO RESOLVING
POWER 120
APPENDIX A

RAPID ANALYSIS OF VOLATILE FLAVOR COMPOUNDS IN
APPLE FRUIT USING SPME AND GC-TIME-OF-FLIGHT MASS

SPECTROMETRY 122

APPENDIX B

PROGRAM LISTINGS ‘ 129
TOFSORC3 129
NONLINEAR 132

LINEAR 146

viii

APPENDIX C
ON THE MASS-INDEPENDENCE OF NON-LINEAR ION
ACCELERATION 147

ix

Table 3-1.

Table 3-2.

Table 4-1.

Table 4-2.
Table 4-3.

LIST OF TABLES

Typical voltages used in operating the M-TOF-

Comparison of various ion source conﬁgurations. For each cross
sectional view, the gray lines represent filaments, while the dark

lines represent ﬁeld cages and electron beam focusing electrodes. .....

Analytical parameters used to calculate the voltages and electric
ﬁelds required to provide linear acceleration.

46

...57

95

 

Electric field ratios used with linear acceleration experiments. ..........

Calculated and experimentally derived voltages and electric ﬁeld
values applied to ion source electrodes conﬁgured in non-linear
acceleration mode.

...95

104

Figure 1-1.
Figure 1-2.

Figure 2-1.

Figure 2-2.

Figure 2-3.

Figure 2-4.

Figure 2-5.

Figure 2-6.

Figure 3-1.
Figure 3-2.

LIST OF FIGURES

Reconstructed total ion chromatogram (RTIC) and ﬁve spectra
illustrating the skewing of decaﬁuorotriphenylphosphine (DFI‘PP)
mass spectra when acquired on a scanning quadrupole mass
spectrometer.

The effect of an inadequate spectral acquisition rate in
reconstructing a true chromatographic proﬁle. The gray line
represents the actual chromatographic event while the black line

represents the integration of the preceding acquisition interval. .........

Conceptual diagram . of m/z analysis by time-of-ﬂight mass
spectrometry.

 

Comparison of ion temporal distribution under low and high
resolution conditions. For time-of-flight mass spectrometry, in
contrast with other mass spectrometric forms, as the temporal
distribution is narrowed, the SW increases.

Measurement of mass resolving power by a) the mass method, and
b) the time method.

....7

18

22

24

 

Models of ion behavior patterns leading to diminished resolving
power, where ions initially located beyond s, (.) or possessing
thermal energy (0,0) are compared with a reference ion (0).
Arrows indicate direction and relative magnitude of thermal
energy. Relative positions are shown at three times: that of ion
formation (1.), arrival at the space focus plane (1,), and later in the
ﬁeld-free region (r). Adapted from [10,11]

Comparison of single (A) and dual ﬁeld (B) ion accelerating
methods. The dual ﬁeld method allows the space focus plane to be
located beyond the 25, position. -

Single ﬁeld ion mirror illustrating the ﬂight path taken by two iso-
mass ions having different total energies. The ion with the greatest
energy penetrates the mirror the farthest before being reﬂected
toward the detector, and thus experiences an increased flight path
as well as a temporal delay.

10

36

 

M-TOF mass spectrometer.

 

Cross sectional view (a), and potential diagram (b), illustrating the
operation of the original Grix/Wollnik ion source---

xi

47

-49

Figure 3-3.
Figure 3-4.

Figure 3-5.
Figure 3-6.
Figure 4-1.

Figure 4-2.

Figure 4-3.

Figure 4-4.

Figure 4-5.

Figure 4-6.

The Grix M-TOF ion source.

 

where, with a 12

R=1500. -- -

 

 

M-TOF detectability using the Grix ion source. .....

 

Trend for ion population growth curves.

McLaren -

 

linear-ﬁeld acceleration. --

 

path contains the non-linear potential.

 

is calculated by successive approximation.

 

region.

 

<1
Demonstration of M—TOF resolving power for a) residual gas ions
nsec FWHM, R=920 and b)
tris(heptaﬂuoropropyl)-8-triazine where, with a 30 nsec FWHM,
-52
-53
55
Two-ﬁeld acceleration geometry implemented by Wiley and
-62
Illustration of ion behavior at the space focus plane using single
_ 68
Potential proﬁle of a non-linear parabolic ﬁeld. The entire ﬂight
72
Graphical illustration of the method for calculation of the
appropriate non-linear- ﬁeld required for ideal space focusing
where the detector is some distance from the ion source exit and a
. linear acceleration region is included. Each electric ﬁeld segment
74
Non-parabolic non-linear acceleration scheme employing a linear
acceleration region in combination with a non-linear ionization
-- - 76
Calculated electric potential and ﬁeld gradients required to
achieve complete space focusing at speciﬁc ion ﬂight times for m/z
79

Figure 4-7.

Figure 4-8.

Figure 4-9.

56, D=153 cm, d=0.50 cm, for a) t=30 msec and b) t=15 msec. ..............

Calculated electric potential and ﬁeld gradients required to
achieve complete space focusing at speciﬁc drift region distances

for ml: 56, t=8 msec, d=0.50 cm, for a) D=5 cm and b) D=50 cm ..........

Calculated electric potential and ﬁeld gradients required to
achieve complete space focusing with speciﬁc ion source
acceleration region dimensions, for m/z 56, D=153 cm, t=5 msec,

80

 

assembly.

 

 

 

 

where d= a) 0.05 cm, b) 0.5 cm, and c) 2.5 cm. 81
Time-of-ﬂight mass spectrometer; conﬁguration and vauum

83
Figure 4-10.Laser optical and ion optical assemblies. The ionization spot is
positioned between ion source electrodes by adjustment of the

moveable stage along the optical rail. 86

Figure 4-11.Ion source assembly. Measurements shown are in cm. 88
Figure 4-12. Resistive divider networks used for a) linear and b) non-linear ion

90

source conﬁgurations.

xii

Figure 4-13.Typical mass spectrum of Fe+ isotopes using the linearly
conﬁgured ion source. The resolving power is ~180.

94

 

Figure 4-14. Arrival time distributions for Fe* ions originating from various ion
source locations. The ion source was conﬁgured for linear
acceleration, with 1639 V applied to the acceleration region and
1887 V applied to the source backplate

Figure 4-15.Theoretical and experimental ion arrival time data for m/z 56,
D=153 cm, d=0.648 cm, at three acceleration ﬁeld/ionization ﬁeld
ratios.

96

98

 

Figure 4-16. Comparison of ideal and actual potentials applied to ion source

conﬁguration, and electric ﬁelds generated in both cases. ...................

Figure 4-17.Calculated ion arrival-time distribution for the multi-gridded ion

source, beginning at the ionization/acceleration region boundary ......

Figure 4-18. Theoretically predicted and experimentally determined ion arrival

times for m/z 56 ions originating in each ion source position.
D=153 cm, d=0.648 cm.

 

Figure 4-19.Theoretically predicted and experimentally determined ion arrival
times for m/z 56 ions originating in each ion source position.
D=153 cm, d=0.648 cm.

 

Figure 4-20. Arrival time distributions for ions originating at the various ion
source ionization positions, where the ion source voltages were
empirically adjusted to provide optimal performance.

 

Figure 4-21.Effect of two :l:3 V perturbations to electrode 6 on time-of-ﬂight
for ions originating from ionization positions 6-9.

 

Figure 4-22.Equipotential surface plot of electrode voltages applied to the grid-
free ion source. Each equipotential surface exhibits cylindrical
axial symmetry, and each is located farther back along the optical
axis than desired

Figure 4-23. Plot of axial potential and the theoretical electrode potentials
required to achieve the optimal axial potential proﬁle for the grid-
free non-linear ion source conﬁguration. Electrode potentials were

determined using the Simion ion optical modeling program. ..............

Figure 4-24.Equipotential surface plot of ion source potentials where electrode
voltages have been adjusted to achieve a nearly ideal axial
potential proﬁle.

 

Figure 4-25. Comparison of resolving power obtained using linear (a), and non-
linear (b), ion acceleration

xiii

.100

.102

103

106

107

109

111

113

115

118

CHAPTER 1

GENERAL INTRODUCTION

Introduction

One of the great milestones in analytical chemistry has been the development and use
of gas chromatography-mass spectrometry (GC-MS). Widely used, GC-MS is applied in
a diverse and ever-widening array of disciplines, ranging from the analysis of foodstuffs
and medicine, to petroleum, geochemistry, and environmental monitoring. It is the
foremost tool for the deﬁnitive analysis of volatile complex mixtures and, because of its
unparalleled speciﬁcity, it has become the dominant referee method in analyses where

ambiguous results have been obtained by more classical analytical procedures.

GC-MS represents the combination of two analytical techniques; gas chromatography
(GC), which is a separations technique, and mass spectrometry (MS), which is an analysis
technique. The strength of the GC-MS combination is attributable to the multi-
dimensional capability created by both the high separation power of the gas
chromatograph and the detection capacity of the mass spectrometer. The gas
chromatograph separates the components of a volatile sample mixture and delivers
relatively pure analyte to the mass spectrometer for analysis, while the mass spectrometer
in turn analyzes each component with the speciﬁcity and sensitivity needed to provide

both qualitative and quantitative analytical information.

Instrumentation for GC-MS has undergone a tremendous evolution over the past three

decades. The need to evaluate increasingly complex samples, along with increases in

sample load, continues to drive the development of improved chromatographic
technology. Modern chromatographic column technologies continue to yield higher-
resolution separations in shorter time spans. Unfortunately, the concomitant temporal
reduction in chromatographic peak width has not been accompanied by an increase in the
detection bandpass of commercially available mass spectrometers. Consequently, these
mass spectrometers are no longer the optimal chromatographic detectors for applications

requiring both high chromatographic resolution and complete analytical analysis.

Given the current trend toward higher resolution chromatographic analyses, GC-MS
will remain the predominant analytical methodology for the most modern applications
only if improved MS technology is developed. As will be shown below, time-of-ﬂight
mass spectrometry (TOFMS) stands as the only detector technology presently capable of
acquiring data at rates accommodating current and emerging chromatographic methods

while retaining both high selectivity and sensitivity.

General Principles of GC-MS
The technique of gas chromatography, as applied to the analysis of organic

compounds, is predominantly a form of partition chromatography in which the mobile
phase is a gas and the stationary phase is a liquid coating on a column wall or other
immobilized support. Once a sample mixture is injected into the carrier gas and
volatilized, it is transported through the column where the chemically distinct sample
components separate as a consequence of differences in their partitioning between the
two phases, exiting the column sequentially in a process referred to as elation. As the
chromatographically separated sample components emerge from the column end, they

enter the vacuum of the mass spectrometer.

ef
Ch
de
6X1
COI
Sig

l0n

In the MS technique, sample molecules are ionized by one of several methods, the
most common being electron bombardment. Upon ionization, many of the ions fragment
from molecular size into various smaller ions and neutral species. The abundances of
these ions are then determined as a function of their mass-to—charge ratio (m/z) and
plotted in a histogram as relative abundance vs. m/z. This histogram, referred to as a mass
spectrum, is a “ﬁngerprint” of the compound, and is often unique. The abundances of
ions originating from the same compound may also be summed to represent the quantity
of that compound in the mixture. A plot of total ion abundance vs. time yields a
chromatogram similar to that obtained by a single channel detector with each peak

representing a particular component.

The Problem

Since the chromatographic behavior is ultimately characterized from reconstruction of
MS data, the analytical performance of GC-MS depends on the capacity of the mass
spectrometer to analyze the eluting chromatographic components at a rate sufﬁcient to
provide both qualitative and quantitative information with minimum uncertainty. The
effect of an insufﬁcient analysis rate is manifested in two ways; by spectral skewing and
chromatogram distortion. Commercial mass spectrometers used as chromatographic
detectors operate in a scanning mode where information about ion abundance is obtained
exclusively and sequentially over the m/z range being examined. This information is then
compiled into a single mass spectrum. Spectral skewing occurs if this ﬁnite scan time is
signiﬁcantly long in comparison to the rate of change in analyte concentration within the

ionization chamber of the mass spectrometer. Here the relative abundances of the various

ions formed in the ionization process will be misrepresented, and a corrupt or skewed

mass spectrum will be generated.

The relationship between the qualitative appearance of the mass spectrum and the rate
of change in analyte concentration is illustrated in Figure 1-1. For unidirectional
scanning, in this case from low m/z to high 212/2, 3 mass spectrum acquired while the
analyte concentration is increasing produces a spectrum having the intensities of peaks
for higher m/z ions enhanced, whereas a spectrum acquired while the analyte
concentration is diminishing exhibits relatively enhanced peak intensities for ions of low
m/z. In both cases the qualitative appearance of the 'mass spectrum is misrepresented,
making compound identiﬁcation difﬁcult, or even impossible. The correct mass spectrum
is collected either during a time when the analyte concentration remains constant or when
the scan speed is rapid compared to the elution dynamics, eliminating or at least

minimizing any effect of changes in analyte concentration.

Fit
[he
503

Iota-alt,

b) 198

Iguana,

 

198

 

 

6)

 

442

 

442

 

 

198

442

 

6)

bung“,

 

 

 

198

 

 

 

 

RTIC of DFTPP

 

 

 

 

198

.II

Figure 1-1. Reconstructed total ion chromatogram (RTIC) and ﬁve spectra illustrating
the skewing of decaﬂuorotriphenylphosphine (DFTPP) mass spectra when acquired on a

scanning quadrupole mass spectrometer.

"a

",9.

I61
G;
Ihr
inl
Shl
re;
6X;

otl

be
Sca
ma

ultl

The second problem, that of chromatographic distortion, relates to the reconstruction
of the chromatogram. The optimal chromatogram used for quantitative purposes is
generated from a continuous measurement of the analyte signal. A chromatogram
generated from a scanning mass spectrometer, however, is constructed by summing the
total or selected ion abundances measured during a single scan and assigning this value to
a discrete elution time point. A spectral generation rate that is relatively slow compared to
the temporal width of the eluting analyte will yield a distorted chromatogram which

poorly describes the true elution proﬁle of the sample.

The effect of an inadequate scan rate is illustrated in Figure 1-2. The chromatogram
reconstructed from the scan-acquired data mistakenly identiﬁes only a single non-
Gaussian shaped peak and incorrectly deﬁnes its position on the time axis. Furthermore,
the mass spectrum assigned to the apex of the reconstructed peak contains ion abundance
information from two analytes. For isolated, Gaussian-distributed peaks, it has been
shown that 10—12 data points must be collected across the peak in order to sufﬁciently
represent the peak area for quantitation [1]. Many more data points are required when the
exact elution time must be determined, and when the analyte must be distinguished ﬁ'om

other partially co-eluting species [2].

The obvious solution to avoid spectral skewing or chromatographic distortion would
be to scan at a sufﬁcient rate. Unfortunately, the more rapidly a mass spectrometer is
scanned, the less time is available for monitoring the ion abundance of each m/z in the
mass range and the concomitant decrease in signal-to-noise ratio (S/N) diminishes the

ultimate detection limit of the instrument. For reasonable levels of sensitivity and SW, the

Total Ion Intensity

  

True Chromatogram

Reconstructed

 

/ Chromatogram

 

 

_/

 

 

 

I I r 131'—

Spectral Acquisition

Figure 1-2. The effect of an inadequate spectral acquisition rate in reconstructing a true
chromatographic proﬁle. The gray line represents the actual chromatographic event while
the black line represents the integration of the preceding acquisition interval.

maximum possible rates are: 1—2 scans/second for sector instruments, 2—5 scans/second

for quadrupoles, and 5—10 scans/second for ion traps.

Two alternate methods are available which avoid this scan rate/perfonnance tradeoff.
First, the effective scan rate of a mass spectrometer can be increased by reducing the
mass range being scanned. This has the effect of increasing the sampling rate of the ions
within the range without diminishing the S/N. The second method incorporates the use of
selected—ion monitoring (SIM), where the abundance of only a single or few select m/z’s
is monitored. This method provides a nearly continuous chromatographic proﬁle and an
enhanced detection limit compared with scanning, due to the increased S/N that arises
from spending more time on the selected m/z measurements. As the mass to charge ratios
chosen for monitoring must be pre-selected, however, the method is limited to the
analysis of known compounds. For both methods, the primary drawback is that little or
no information is available to identify unknown or unexpected compounds since most of

the mass spectral information has been sacriﬁced.

The ideal solution to both the skewing and reconstruction problems would be to
incorporate a non-scanning MS technique that simultaneously monitors the entire
selected m/z range. By this method, known as array detection, the mass spectrometer
would provide the increased S/N associated with SIM over the entire mass range being

examined and the qualitative appearance of the mass spectra would be preserved.

it

We

lim
of
MCI

SEqL

A Historical Perspective

Given that current commercial GC-MS systems utilize MS techniques relying
exclusively on scanning rather than array detection methods, it is interesting to note that
the feasibility of GC-MS was ﬁrst demonstrated in 1959 using an instrument that was
capable of array detection—the time-of-ﬂight mass spectrometer [3]. This mass
spectrometer could monitor ions at rates in excess of 2000 per second and, as an array

detector, eliminate any possibility of spectral skewing.

The major limitation to the technique at that time was the absence of a suitable data
collection method. Mass analysis in time-of-ﬂight mass spectrometry is accomplished by
measuring the arrival times of ions striking the detector at the end of their “ﬂight”
through the mass spectrometer. The brief duration of the generated signal, referred to as a
transient or transient waveform, may be as little as 100 us for the entire mass range
observed and only 10 us for the ions representing a single m/z. The only electronic device
available to monitor the output signal in these time-scales in 1959 was the oscilloscope,

and mass spectral information could only be recorded by photographing the oscilloscope

display.

By the early 19603, a method of data collection using time—gated boxcar integration ‘
was developed. This method is known as time-slice detection, whereby a small portion of
the information available in each transient is measured and recorded. By creating a ﬁxed
time delay prior to data collection from successive ion source extractions, the abundance
of a single m/z could be continuously monitored. Alternatively, by incrementally
increasing the delay prior to data collection from successive extractions, data for a

sequentially generated (i.e. scanned) mass spectrum could be obtained [4]. The

10

implementation of time-slice detection reduced the effective spectral generation rate
enough to enable the possibility of data recording by conventional methods. As a result,
however, the advantages of array detection were sacriﬁced as the instrument was
converted to a scanning device. The lack of a suitable data collection method combined
with poor overall instrument performance and small dynamic range led to the eventual
abandonment of TOF in favor of other MS techniques for GC applications, and few

publications reporting the use of GC-TOFMS have since been reported [5—13].

Other mass analyzers, namely the magnetic sector and the quadrupole, have been more
successfully coupled with GC. Magnetic-sector mass spectrometers were ﬁrst coupled
with GC in the early 1960s [14], and were capable of acquiring mass spectral information
at rates ranging 0.1—1 scans per second. These rates were sufﬁciently fast to accurately
detect packed-column eluates and slow enough to be recorded with a galvanic chart
recorder. The development of the quadrupole mass spectrometer in the late 19608 helped
generate signiﬁcant interest in GC—MS by providing an economical instrument which had
adequate sensitivity and mass resolution and was readily adapted to computers for
instrument control and data acquisition [15]. The 1—2 scans per second acquisition rates
available with these mass spectrometers made them well suited for use with open-tubular
capillary columns which yield eluting times of only a few seconds. By the mid 19703,
GC-MS systems predominantly used quadrupole technology with acquisition rates of up
to 5 scans per second and adequate sensitivity. More recently, the development of the
quadrupole ion trap has further increased the spectral acquisition rates available by

scanning mass spectrometers, with rates of 2—10 scans per second readily achievable.

d:

CC

int

118]

Chi

0pc

tech

deter

Sufﬁc

11

The MSU Connection

Researchers at Michigan State University were ﬁrst confronted with the limitations of
commercial GC-MS technology in the early 19805 while investigating urinary organic
acid metabolites in humans [16]. The experimental protocol in this “metabolic proﬁling”
research yielded highly complex samples, components of which often partially co-eluted
with other species. The initial investigations utilized packed columns in which the major
components could be effectively analyzed because their eluting times were long relative

to the mass spectrometer scan time.

A more thorough analysis of the urinary metabolites became possible as high-
resolution capillary columns became available. Many more components could be
chromatographically resolved using these columns, but the possibility of using MS
detection was reduced as a consequence of the shorter elution times. While known
components could be identiﬁed by chromatographic retention index using GC alone, the
identiﬁcation of unknown compounds, which required the obtaining of structural
information by MS, was precluded in cases where the unknown analyte was temporally

narrow or partially co-eluting with other compounds.

These researchers published a paper in 1983 entitled “Mass Spectrometry on the
Chromatographic Time Scale: Realistic Expectations” describing the difﬁculties inherent
to chromatographic detection by mass spectrometry [17]. In this publication the various
types of mass spectrometers were examined and the attributes of scanning and array
techniques for chromatographic detection compared. The authors concluded that array
detection was the only method capable of collecting complete mass spectra at rates

sufﬁcient to provide both unskewed spectra and accurate chromatographic representation

12

for high-resolution capillary column GC-MS. Time-of-ﬂight mass spectrometry, using its
inherent capacity for array detection, was identiﬁed as the best suited candidate, having a
potentially high spectral generation rate in addition to a wide dynamic range, adequate
mass range, and reasonable cost. Consequently, MSU researchers devised a plan to
accomplish GC-TOFMS with the goal of achieving a IOO-Hz spectral generation rate

over the range m/z 50-500 [18].

Part One of the Plan - Instrumentation For Time Array Detection

The ﬁrst part of this plan addressed the same limitation that TOFMS faced when it
was ﬁrst used for GC-MS in 1959; namely the lack of appropriate data acquisition
hardware. Maximal usage of the ion abundance information available in each TOFMS
transient is obtained only when the entire generated transient signal is recorded; an

acquisition method known as time-array detection (TAD).

Research at MSU has resulted in the design and construction of an integrating
transient recorder (ITR) capable of continuously acquiring up to 10,000 complete
transients per second [19,2021]. The ITR provides both high-speed data acquisition and
time-array detection across the entire selected mass range. Furthermore, several
successive transients can be summed in a time-locked registry to yield an integrated
spectrum which is then electronically stored. This signiﬁcantly reduces the overall data
storage space requirement while maintaining the desired spectral generation rate for GC-
MS and improves the S/N of each spectrum by a factor equal to the square root of the

number of transients summed.

m:

ear

Wit

Far

GC

13

Part Two of the Plan - Improvements to Mass Resolving Power

The second part of the MSU plan involved the development of a time-of-ﬂight mass
spectrometer capable of providing mass spectral information of sufﬁcient quality to
enable the use of TAD. The analysis of gas phase ions by TOFMS necessarily
incorporates spatial and energetic variations which potentially diminish the mass
resolving power. This resolving power, which is described more thoroughly in chapter
two, is the capacity to distinguish ions of differing m/z and is increasingly difﬁcult to
achieve at higher m/z, eventually limiting the effective mass range of the mass
spectrometer. Many early time-of-ﬂight instrumental conﬁgurations utilized ion focusing
methods for achieving sufﬁcient mass resolution that are, unfortunately, mass-dependent.
These methods resolve only a small portion of the analyzed mass range in each transient.
However, the use of TAD requires mass-independent ion focusing, as ion abundance

information over the entire mass range is collected with each and every transient.

Furthermore, these mass-dependent focusing techniques typically employ a low
gradient electric ﬁeld to initially extract ions from the ionization chamber, creating a
situation where the effect of collisional broadening, due to the interaction of the ions with
neutral molecules, further diminishes the achievable mass resolving power [22]. An
adequate TOF mass spectrometer for GC-MS using TAD should provide at least unit
mass resolution by a mass-independent ion focusing means over the entire mass range in
each and every transient and minimize the effects of the collisional broadening inherent
with gas chromatographic sample introduction. Research in the MSU Mass Spectrometry
Facility has led to the development of such a TOF mass spectrometer for use in

GC-MS [23].

14

Research Goals

As the development of the technology required to fulﬁll the plan initiated in 1983
nears completion, consideration must now be given to extending the use of TOFMS to
other separation techniques such as liquid chromatography (LC) and capillary zone
electrophoresis (CE) which share the same requirements for adequate mass spectrometric
detection as GC, but utilize sample introduction techniques that are distinctly different.
Speciﬁcally, these methodologies produce large sample volumes in the gas phase and

thus place stringent requirements on ion source design and performance.

This dissertation describes, along with the development of the time-of-ﬂight mass
spectrometer designed for use in GC-MS with TAD, the development of an ion source
which allows the study of the basic phenomena limiting mass resolving power; especially
those phenomena which would be deleterious when sampling large ion volumes.
Chapter 2 contains a discussion of the fundamental principles of mass analysis by time-
of-ﬂight mass spectrometry and the major instrument conﬁgurations that have been
developed to provide improved mass resolving power. A description of the novel TOFMS
system collaboratively developed at MSU is contained in Chapter 3 along with discussion
of ion source developments and limitations. The core research of this author, the
development of an ion acceleration model and an ion source that compensates for spatial

variations in initial ion position is described in Chapter 4.

15

Chapter References

l.

10.

11.

12.

13.
14.
15.
l6.

17.

18.

Enke, C. 6.; Watson, J. T.; Allison, J .; Holland, J . F. In Proceedings of the 3 7"
ASMS Conference on Mass Spectrometry and Allied Topics, Miami Beach, Florida,
May 21—26, 1989, 32.

Chesler, S. N.; Cram, S. P. Anal. Chem. 1971, 43, 1922.
Gohlke, R. S. Anal. Chem. 1959, 31, 535.

Gohlke, R. S. Anal. Chem. 1962, 34, 1332.

Ebert, Jr., A. A. Anal. Chem. 1961, 33, 1865.

McFadden, W. H.; Teranishi, R.; Black, D. R.; Day, J. C. J. Food Sci. 1963, 28,
316.

D’Oyly—Watkins, C.; Hilhnan, D. E.; Winsor, D. E.; Ardrey, R. E. In Dynamic
Mass Spectrometry, Price, D.; Williams, J. E., Eds., Heydon & Son Ltd.: London,
1970; Vol. 1, pp 163—174.

Kirchgessner, W. G.; DiPasqua, A. C.; Anderson, W. A.; Delaney, G. V.
J. Forensic Sci. 1974, I 9, 313.

Dobson, R. L. M. Energy Res. Abstr. 1986, 11, Abstr. No. 41188.

Dobson, R. L. M.; D’Silva, A. P.; Weeks, S. J .; Fassel, V. A. Anal. Chem. 1986, 58,
2129.

Erickson, E. D.; Enke, C. G.; Holland, J. F.; Watson, J. T. Anal. Chem. 1990, 62,
1079.

Watson, J. T.; Schultz, G. A.; Tecklenburg, R. E.; Allison, J. J. Chrom. 1990, 518,
283.

Wilkerson, C. W.; Colby, S. M.; Reilly, J. P. Anal. Chem. 1989, 61, 2669.
Lindeman, L. P.; Annis, J. L. Anal. Chem. 1960, 32, 1742.

Finnigan, R. E. Anal. Chem. 1994, 66, 969A.

Holland, J. F.; Leary, J. L.; Sweeley, C. C. J. Chrom., 1986, 397, 3.

Holland, J. F.; Enke, C. G.; Allison, J.; Stults, J. T.; Pinkston, J. D.; Newcome, 8.;
Watson, J. T. Anal. Chem. 1983, 55, 997A.

Allison, J .; Holland, J. F.; Enke, C. G.; Watson, J. T. Anal. Instrum. 1987, 16, 207.

19.

20.

21.

22.

23.

16

Holland, J. F .; Newcome, B. N.; Tecklenburg, R. E.; Davenport, M.; Allison, J.;
Watson, J. T.; Enke, C. G. Rev. Sci. Instrum. 1991, 62, 69.

Enke, C. G.; Holland, J. F. United States Patent No. 4,490,806, 1984.

Holland, J. F .; Enke, C. G.; Davenport, M. R.; Janow, L. W. United States Patent
No. 5,367,162, 1994.

Futrell, J. H.; Tieman, T. 0.; Abrarnson, F. P.; Miller, C. D. Rev. Sci. Instrum.
1968, 39, 340.

Grix, R.; Tecklenburg, Jr., R. E.; Watson, J. T.; Holland, J. F.; Yefchak, G. E.;
Gardner, B. D.; Allison, J .; Enke, C. G.; Wollnik, H. In Proceedings of the 39th

ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, Tennesee,
May 21—26, 1991, 154.

CHAPTER 2

THEORY AND INSTRUMENTATION
DEVELOPMENT IN TIME-OF-FLIGHT MASS
SPECTROMETRY

Introduction

In the most common form of time-of-ﬂight mass spectrometry, ions created in the
vacuum of an ion source are accelerated to constant energy and allowed to drift along an
evacuated ﬁeld-free ﬂight tube where they separate into groups according to their mass-
to-charge ratio (m/z), as shown in Figure 2-1. The time required for an ion extracted from
the source to reach the detector is measured and used to calculate mass. This relationship
between time (or velocity) and m/z can be explained in the following manner. The force

F, exerted on an ion by an electric ﬁeld is given by
F=qE=zeE (2—1)

where E is the electric ﬁeld given in volts/cm, and q, the charge on the ion (in coulombs)
is the product of e, the elementary unit of charge, and z, the number of unit charges. The
kinetic energy gained by an ion is equal to the work W performed on the ion by the

electric ﬁeld, and may be represented as
KE=W=F-d (2-2)

where d is the distance the ion travels through the electric ﬁeld. Consequently,

substituting the previous equations, it is shown that

17

18

/ Electron Beam

 

 

 

 

 

 

(ok/ Ions
I i, Q 0
8E3 0 .
030: O 00 —_
o O i O O O
—\'/_ .
Ion Source Flight Tube Detector

Figure 2-1. Conceptual diagram of m/z analysis by time-of-ﬂight mass spectrometry.

19
KE = zeEd (2-3)

The velocity of an ion may be found by incorporating the classical equation for kinetic

energy

KE = — mv2 (2-4)

where m is the ion mass, and v is the velocity. Equations 2-6 and 2-7 can be combined

and rearranged to solve for velocity

 

2 Ed
”Ii 28 (2-5)
m

which upon further rearrangement, and substitution of distance and time for velocity, it

can be 'shown that

D L".

t = JZeEd Z (2-6)

 

where t is the time of ﬂight, and D is the ﬂight distance from the ion source to the

detector surface. This can be ﬁrrther simpliﬁed to
= k — (2-7)

where k is the collection of constant values. Empirically, the measured ﬂight time t is

near

m
tmcas = tojfset + 1“]; (2'8)

20

where tome, is the offset time to account for any delays imposed by the measurement

circuitry. This ﬁrst-order relationship between time and Jm/z provides for a simple

calibration of the time-of-ﬂight mass spectrometer.

Early Development of TOFMS

The concept of time-of-ﬂight mass spectrometry was ﬁrst envisioned in 1946 by
Stephens [1] as a means for monitoring chemical systems in which the composition was
changing rapidly as a function of time. The ﬁrst development of a TOF instrument was
reported in 1948 by Cameron and Eggers [2], and was followed by reports on other
instruments, including those developed by Keller in 1949 [3], by Wolff and Stephens in
1953 [4] and by Katzenstein and Friedland in 1955 [5]. These instruments provided high
spectral generation rates of up to 5000 Hz, but exhibited poor sensitivity and were unable
to distinguish between ions greater than m/z 100. TOFMS became a practical technique
only through the innovations of Wiley and McLaren in 1955 [6] which extended the mass
resolving power, and a commercially produced instrument based on their design became

available from the Bendix Corporation in 1956 [7].

In spite of these developments however, TOFMS quickly became known as a
technique of limited capacity, ideal only for some custom applications, and generally fell
out of favor in the analytical laboratory as other types of mass spectrometers having
higher resolving power and greater detectability were developed. Expanded interest in
TOFMS did not revive until the mid 19703 with two developments; 1) plasma [8] and
laser desorption techniques in which, through pulsed ionization of an analyte on a planar

surface, some of the factors limiting mass resolving power could be eliminated, and 2) the

21

ion mirror or “reﬂectron” by Mamyrin [9] which could compensate in a mass-

independent manner for the energetic limitations to mass resolving power.

The Need for Resolving Power

Throughout the development of TOFMS, the improvement of mass resolving power
has been a primary objective. The need for this is three-fold. First, the resolving power
determines the extent to which ions having different m/z may be differentiated by the
mass spectrometer, which in turn affects the extent to which an analyte can be deﬁned by
its mass spectrum. Secondly, the m/z range is largely affected by the resolving power. For
gas chromatographic applications a mass range requirement of 1—600 u is desirable, and
GC applicable analytes can range up to 1000 n. For liquid chromatography, analytes of up
to 2500 u are observed and for other applications such as laser desorption, a mass range
of 10,000 it or more may be required. Finally, with time-of-ﬂight there is a unique
relationship between resolving power and detectability. Other forms of mass
spectrometry enhance the resolving power by removing ions having a slight difference in
m/z or in energy than those ions of interest. Consequently, ion transmission is decreased
and detectability is diminished. In time-of-ﬂight mass spectrometry there is a direct
correlation between resolving power and detectability, meaning that as the resolving
power is increased, the temporal distribution of each iso-mass ion population becomes

more narrow and there is actually an increase in S/N as shown in Figure 2-2.

22

Ion Current

High
Resolution

 

 

 

Low
Resolution

 

 

time

Figure 2-2. Comparison of ion temporal distribution under low and high resolution
conditions. For time-of-ﬂight mass spectrometry, in contrast with other mass
spectrometric forms, as the temporal distribution is narrowed, the S/N increases.

23

Deﬁnitions of Resolving Power and Mass Resolution

There are several related deﬁnitions deﬁning the term resolving power, R, but only
two methods, as illustrated in Figure 2-3 are common. The ﬁrst method, or “mass

method,” is derived from a plot of ion intensity versus m/z, and is deﬁned as

R = 5 (2-9)

where m is the m/z of the peak being examined (2 = 1) and Am is deﬁned as

 

Am: M (240)

where don,” _ m is the measured distance between adjacent m/z peaks and d, is the peak
width at an amplitude deﬁned as a percentage of the peak height. Typically, this

amplitude is deﬁned as 50%, or full width at half maximum (F WHM).
The second method, or “time method,” is derived from a plot of ion intensity versus
time, and is deﬁned as

R =— (211)

where t is the average ion ﬂight time of the m/z peak, and At is the arrival time
distribution at a deﬁned percentage of the peak height. In practice, when equivalent peak

height percentages are used, the results of equations 2-1 and 2-3 are equal.

24

 

 

 

 

a) F— drm-r) — rn —_'>i
dm ‘—> 523.9%.
\ Va’iIe
. . ’ n
m m+1

 

 

 

 

 

9) k a t >'
At
“ rh

Figure 2-3. Measurement of mass resolving power by a) the mass method, and b) the
time method.

25

Two other terms are also used in deﬁning m/z differentiation: resolution and unit mass
resolution. Resolution is inversely related to resolving power. A mass spectrometer
having a resolving power of 500 has a resolution along the mass axis of one part in 500.
The term unit mass resolution is used to deﬁne the maximum m/z for which the peak is
separated from adjacent peaks. The separation is determined by the percent valley or

overlap between the peaks; typically 10%.

Problems Limiting Resolving Power

In principle, nearly inﬁnite resolving power could be attained if all ions in each
iso-mass group would arrive at the detector simultaneously. And while this ideal
circumstance could be achieved if 1) all the ions possessed identical (or no) internal
energy or motion; 2) all started from the same point, and 3) all were accelerated to
identical kinetic energy, any one of these conditions is rarely achieved and the resolving
power is consequently diminished. Therefore, some mechanism or combination thereof
must be used to compensate for the effects of variations in ion starting conditions in order

attain sufﬁcient resolving power.

For a gaseous sample, there is both a distribution of initial ion starting positions and a
thermal Maxwell-Boltzrnan distribution of ion energies (~0.1 eV) that is manifested as a
distribution of ion detection times within each iso-mass group. The manifestations of
these distributions are often referred to as the space and energy eﬂects. The energy effect
also has a distinct detrimental component known as the tum-around eﬂect, whereby those
ions having a thermal energy vector directed against the accelerating ﬁeld travel away

from the ion source exit before being turned around. This results in ions of the same m/z,

26

same initial position, and same magnitude of thermal energy, exiting the ion source at

different times and consequently leading to a difference in arrival-times at the detector.

The three ion starting parameters are modeled in Figure 2-4. The effect of each case
can be compared with the reference ion of the same mass located at the initial position so,

with no internal energy. The energy, U,o. applied to the reference ion is
USO : 30E (2‘12)

which represents the total kinetic energy of the reference ion. It is assumed that ions
originating from the same plane, orthogonal to the ﬂight axis, receive equal kinetic

energy when the accelerating ﬁeld is applied.

In the ﬁrst case S, an ion initially located at a position so + 8s, receives a total energy
directly proportional to its initial position when the accelerating electric ﬁeld is applied.

The energy applied to this ion will differ ﬁ‘om Us, by the relationship

U, = MU... (2-13)
s0

Ions located farther from the ion source exit receive greater energy than those ions closer
to the exit and overtake the less energetic ions at a distance 2s, from the ion source exit; a
phenomenon known as space focusing. The location along the ﬂight axis where this
occurs, in this case 2:0, is traditionally known as the space focus plane and represents the
position along the ﬂight path where R,, that portion of the resolving power attributable to
space focusing, is the highest. (The fact that the space focus plane is not an inﬁnitely thin
plane, but rather a region where ions exhibit a temporal minimum is the basis for the

research described in Chapter 4.)

27

In the case of U+, an ion having some initial energy 8U directed toward the detector is
compared to the reference ion located in the same plane. While both ions receive a total
energy of U, as a consequence of their location within the source, the ion having the
initial thermal motion leaves the source with a ﬁnal energy of U, + 5U, achieves a greater
velocity and arrives at the detector before the reference ion. Methods correcting for this
variance are known as energy focusing methods, discussed below. In the special case of
U—, the initial energy EU is directed away ﬁom the detector, and the ion initially travels
against the electric ﬁeld before being turned around and accelerated out of the source. For
two ions of the cases U+ and U— where both ions have the same magnitude of internal
energy but in opposite directions, each ion will leave the source with identical energy.
However, the U— ion will arrive at the detector later due to the tum-around time in the ion

source.

The resolving power is independently degraded by each of the starting parameters,
with the overall resolving power representing the combined effect of the space and energy

resolutions. The total resolving power, R, is deﬁned as

_‘_+ 1 + 1 + 1 (244)
RS RU RU R

- electronic:

 

 

.1.
R

4.

While not a fundamental parameter of ion motion, the term R,,,c,,o,,,.c, is included here to
acknowledge that in practice, if all ion space and energy effects can be corrected or
eliminated, the ultimate limitation to R is that imposed by the speed and precision of the

detector and electronics circuitry.

28

 

 

 

 

<— IO" Field-free Drift Region 3
Source
t, t, t,
Us 0 U8 0
o o-» 0—»
Ideal U U
o o—» 0—+
Space Focus
Plane U: .. .
0 O—>
S 168483-9923. U”
o»
. -—3

 

 

U+

 

 

Combined

 

<0

 

 

 

Figure 2-4. Models of ion behavior patterns leading to diminished resolving power,
where ions initially located beyond so (.) or possessing thermal energy (0,0) are
compared with a reference ion (0). Arrows indicate direction and relative magnitude of
thermal energy. Relative positions are shown at three times: that of ion formation (to),
arrival at the space focus plane (t,), and later in the ﬁeld-ﬁ‘ee region (t2). Adapted ﬁ'om

[10,11].

29

Methods for Improving Resolving Power
The Wiley/McLaren Ion source

The ﬁrst major improvements to resolving power were accomplished by Wiley and
McLaren [6], who incorporated two design features that yielded vastly signiﬁcant
improvements to both mass resolving power and overall mass range. Their ﬁrst
innovation was the development of the two-ﬁeld ion source. Earlier ion sources used a
single electric ﬁeld for ion extraction that imposed a tradeoff between energy and space
focusing. In order to maximize energy resolution, it was desirable to use a high electric
ﬁeld strength and accelerate the ions to their ﬁnal velocity quickly. However, this
required an axially short ion source geometry, and as the detector must consequently be
placed only a short distance (2s0) ﬁ'om the ion source in order to achieve a space focus
condition, little time was available for mass separation. The total ﬂight time could only
be increased by either lowering the electric ﬁeld strength (and consequently the energy
resolution), or lengthening the ﬂight path by moving the detector well beyond the space
focus region. Since the dominant parameter limiting resolving power was the initial
energy spread, the latter option was required. Wiley and McLaren found that by using a
two-ﬁeld ion source, illustrated in Figure 2-5, the space focus region could be located
farther than 2s, from the ion source at a distance which is a function of the two
accelerating electric ﬁeld strengths. Thus, while the low amplitude ﬁrst accelerating ﬁeld
slightly reduced the energy resolution, the ability to achieve both space focusing and an

increase in the total ﬂight time for all ions yielded an overall increase in R.

30

 

 

 

 

 

 

 

 

A) Ion Source Drift Region
I
I
I +.
l Space Focus
5:“ I r Plane
E l
.9 I
o .
a” I
I
I
I .
|
B)
I I .
l l g
I I
i I Space Focus
E I l / Plane
‘0" '.
c I I g
.9 I I i
8 ' ' .
I I
l l E.
I I -
I I
I

 

 

 

 

Figure 2-5. Comparison of single (A) and dual ﬁeld (B) ion accelerating methods. The
dual ﬁeld method allows the space focus plane to be located beyond the 2s0 position.

31

Wiley and McLaren’s second innovation provided additional focusing by introducing
an adjustable time delay between the formation of ions and the application of the
accelerating ﬁeld. During this time, ions move to new locations in the ion source due to
their thermal energies and, upon extraction, receive total kinetic energies dependent on
these new locations. This energy focusing method, known as time-lag focusing,
essentially attempts to transform the energy distribution of the initial ion population into
a spatial distribution, thus reducing the temporal effect of the energy distribution at the
space focus plane. Unfortunately, the optimal time lag is mass dependent, requiring that
the lag time be varied with successive extractions in order to record a wide mass range.
However, using these methods, Wiley and McLaren increased the mass resolving power

to ~300, with the tum-around effect becoming the dominant limiting factor.

Subsequent Ion Source Methods for Improving Resolving Power

Since the work of Wiley and McLaren, a number of approaches have been pursued in
an attempt to compensate for the initial energy spread in electron impact ion sources.
Sanzone proposed a multiple-ﬁeld ion source to provide both space focusing and
compensation for the nun-around effect, but theoretical analysis indicated that no
improvement to mass resolving power could be expected beyond that provided by the
Wiley-McLaren two-ﬁeld model [12]. Marable and Sanzone later attempted to minimize
the tum-around effect with the technique of impulse-ﬁeld focusing (IFF) [13], in which a
short duration, high amplitude electric ﬁeld pulse was applied to the ionization region just
prior to the extraction pulse. This initial pulse added a bias to the thermal energy
distribution of the ion population, causing all ions to move toward the ion source exit

prior to extraction and, in essence, reduce the effect of the initial energy distribution on

32

mass resolving power. While Sanzone and coworkers obtained some initial results using
this technique and were able to achieve a mass resolving power of 450, pursuit of this
technology was limited by an inability to create voltage pulses of sufﬁciently short

duration and large amplitude [14].

Sector Focusing

Other methods for energy focusing have been proposed using magnetic sectors [15]
and electric sectors [16—18]. Several of these designs have been constructed, including
one incorporating a toroidal electric sector and two ﬁeld-free regions [19], and another
incorporating four tandem electric sectors and having a reported resolving power of

~700 [20,21].

An electric sector has also been used in energy ﬁltering [22]. In this instrument, an ion
beam generated from a continuous source was directed into an electric sector. A narrow
energy range of ions was then directed into a time-of-ﬂight mass analyzer, converted into
discrete ion packets using a beam modulation technique, and mass analyzed. The use of a
continuously generated ion beam eliminated the tum-around effect associated with pulsed
ionization and the ﬁltering minimized the total energy spread. However, while this
technique yielded resolving powers of up to 900, ion transmission was diminished. In an
effort to regain ion transmission, beam modulation was used to introduce ion packets into
a time-of-ﬂight instrument directly from a continuous ion source [23] without energy
ﬁltering. While transmission improved, this instrumental conﬁguration achieved a

resolving power of only 700.

33

Dynamic Fields

In contrast with using static electric or magnetic ﬁelds, it is also possible to use
dynamic ﬁelds, whereby the potential applied to some portion of the ﬂight path is varied
over time. Muga [24] proposed a velocity compaction technique in which a synchronized,
continuously increasing potential is applied to the drift region subsequent to ion
extraction. This increases the velocity of all ions as they travel toward the detector.
However, those ions farthest from the detector in a given iso-mass packet experience a
greater increase in energy, allowing them to catch up to those ions closer to the detector,
and thus decrease the temporal distribution of the ion packet. Yefchak [25] modeled a
similar technique called dynamic ﬁeld focusing (DFF), where the synchronized potential
was increased only on the latter portion of an initial drift region. Modeling of this
technique indicated that a resolving power of >2000 should be achievable in a linear
time-of-ﬂight mass spectrometer. Kinsel and Johnston [26] proposed the technique of
post source pulse focusing as a simple variation of time-lag focusing where a voltage
pulse could be applied to a short ﬁeld-free region located after the ionization and
acceleration regions. Properly timed, the pulse could be applied once the ions of interest
had entered the region and thus provide discrete focusing. The technique was postulated

to increase mass resolving power to >1000.

The Ion Mirror

Probably the singularly most effective innovation for improving mass resolving power
comes from the development of the reﬂectron or ion mirror mass-independent energy
focusing device ﬁrst reported by Mamyrin in 1973 [27—29]. This device, illustrated in

Figure 2-6, is simply a series of electrostatic diaphragrns that provide a retarding electric

34

ﬁeld with enough potential to “reﬂect” ions. Ions with different kinetic energies penetrate
the mirror to different depths before being turned around. While all ions leave the mirror
with exactly the same magnitude of energy they entered with, those ions possessing
greater energy travel farther into the mirror and experience a time delay as a result of the
increased ﬂight path, compensating for their higher velocity in the ﬁeld-ﬂee region.
Instruments incorporating an ion mirror in the ﬂight path produce a second space-focus
plane having a higher resolving power than the ﬁrst space-focus plane due to the

additional energy focusing.

The simplest type of ion mirror is the single-ﬁeld mirror that contains a homogeneous
electric ﬁeld. Ions penetrate the electric ﬁeld and are eventually repelled back into the
ﬁeld-free drift region. The single-ﬁeld ion mirror provides what is known as ﬁrst-order
focusing; a reference to the mathematical expressions used to describe focusing. The
more commonly used two-ﬁeld mirrors, are capable of second-order focusing, which
yields a comparatively higher resolving power. Wollnik [30] described a grid-free, and
thus non-linear-ﬁeld mirror capable of yet higher-order focusing along with increased ion

transmission.

The addition of the ion mirror in time-of-ﬂight instrumentation has greatly improved
the achievable resolving power and has been incorporated into a variety of instrument
conﬁgurations. Mamyrin was initially able to achieve a resolving power of ~3000 using a
2-stage mirror. Schlag incorporated a 2-stage mirror into a multi-photon ionization time-
of-ﬂight instrument and achieved a resolving power of 3900 [31]. Della Negra used an
ion mirror with a 252Cf ionization time-of-ﬂight instrument and achieved a resolving

power of 2500 [32]. Schlag later achieved a resolving power of 10,000 in a multi-photon

35

ionization instrument [33]. Yang and Reilly achieved a resolving power of 11,000 using
an ion mirror in a laser desorption time-of-ﬂight instrument where the laser was directed
from behind the sample [34]. Standing [35], who reported a resolving power of 3200 for a
linear secondary ion mass spectrometry (SIMS) instrument, was able to further increase
resolving power to 10,000 by incorporating a single-ﬁeld ion mirror. Using a two-ﬁeld
mirror in a laser ionization time-of-ﬂight instrument, Bergman and co-workers were able

to obtain a resolving power of 35,000 [36].

The Effect of Ionization Method

The dramatic increases in resolving power obtained when incorporating the ion mirror
are partly due to the fact that the ionization techniques employed minimize either the
initial spatial or energetic distribution of the ions being examined. Ionization techniques
such as SIMS [37], 252Cf-plasma desorption [38], and matrix-assisted laser desorption
ionization (MALDI) [39], ionize solid phase samples on a planar surface and thus
minimize the initial spatial distribution of ions. The tum-around time is also eliminated
because the surface matrix prevents ion motion against the accelerating ﬁeld. Resolving
powers of several thousand have been reported using these techniques in linear time-of-
ﬂight instruments, with resolving powers increased to 10,000 or more when combined

with an ion mirror.

In another approach, molecules entering the ionization chamber are largely collimated
into a molecular beam, reducing the range of thermal motion orthogonal to the beam axis
to well below that expected ﬁ'om the typical Maxwell-Boltzrnan distribution. Ions are
then formed in the beam using laser ionization and extracted along an axis perpendicular

to the molecular beam. Thus, the initial spatial distribution is minimal due to the highly

36

 

 

Figure 2-6. Single ﬁeld ion mirror illustrating the ﬂight path taken by two iso-mass ions
having different total energies. The ion with the greatest energy penetrates the mirror the
farthest before being reﬂected toward the detector, and thus experiences an increased
ﬂight path as well as a temporal delay.

37

focused laser beam and the signiﬁcant reduction in the energy spread yields a higher

resolving power. Using this technique, Reilly reported a resolving power of ~47OO [40].

Techniques such as those employed with gas and liquid sample introduction methods
result in gas phase ionization where there is a distribution of both initial ion locations and
kinetic energies, producing a need for both space and energy focusing if resolving powers
similar to the planar ionization methods are to be approached. While the addition of ion
mirror optics provides excellent increases in resolving power when the space or energy
effect is encountered alone, the combined effects cannot be simultaneously compensated
for using a single static ion optical device [41,42]. One possible solution would be the
incorporation of multiple ion optical devices that would independently compensate for
distinct components of the initial ion conditions. It is the intent of this dissertation to
examine the feasibility of using another solution, a non-linear electric ﬁeld within the ion
source to accelerate the ions over the region of initial ion location, in order to provide
improved space focusing and, consequently, an increase in resolving power. This
development has the potential, when combined with other focusing techniques such as the
ion mirror and impulse-ﬁeld focusing, to provide complete correction for the initial ion

behavior of gas phase samples.

Summary

This chapter outlines the basic theory of time-of-ﬂight mass spectrometry and the
major achievements that have resulted in improved mass resolving power in time-of-
ﬂight instrumentation. While these improvements have dramatically increased the
resolving power in instrumental conﬁgurations that ionize samples from planar surfaces

and other similarly conﬁned ionization volumes, only relatively low resolving powers are

38

currently achievable when analyzing gas phase samples such as those produced in gas or
liquid chromatographic analyses. An ion optical mechanism for overcoming one of the
primary limitations to resolving power, the initial spatial distribution, is the focus of this

dissertation.

39

Chapter References

10.
ll.
12.
13.

14.

15.
16.
1.7.

18.

19.

20

Stephens, W. E. Bull. Am. Phys. Soc. 1946, 21, 22.

Cameron, A. E.; Eggers Jr., D. F. Rev. Sci. Instrum. 1948, 19, 605.

Keller, R. Helv. Phys. Acta, 1949, 22, 386.

Wolff, M. M.; Stephens, W. E. Rev. Sci. Instrum. 1953, 24, 616.

Katzenstein, H. S.; Friedland, S. S. Rev. Sci. Instrum. 1955, 26, 324.

Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150.

Wiley, W. C. Science, 1956, I24, 817.

Mcfarlane, R. D.; Torgerson, D. F. Int. J. Mass Spectrom. Ion Phys. 1976, 21 , 81.

Mamyrin, B. A.; Karataev, V. I.; Shmikk, D. V.; Zagulin, V. A. Zh. Eksp. Teor. F iz.
1973, 64, 82.

Yefchak, G. E. Doctoral Dissertation, Michigan State University, 1990.
Cotter, R. J ., Anal. Chem. 1992, 64, 1027A.

Sanzone, G. Rev. Sci. Instrum. 1970, 41, 741.

Marable, N. L.; Sanzone, G. Int. J. Mass Spectrom. Ion Phys. 1974, I3, 185.

Browder, J. A.; Miller, R. L.; Thomas, W. A.; Sanzone, G. Int. J. Mass Spectrom.
Ion Phys. 1981, 37, 99.

Poschenrieder, W. P. Int. J. Mass Spectrom. Ion Phys. 1971, 6, 413.
Bakker, J. M. B. Int. J. Mass Spectrom. Ion Phys. 1971, 6, 291.
Poschenrieder, W. P. Int. J. Mass Spectrom. Ion Phys. 1972, 9, 357.

Kilius, L. R., Hallin, E. L., Chang, K. H., Litherland, A. E. Nucl. Instrum., Meth.
1981, I91, 27.

Steffens, P., Niehuis, E., Friese, T., Greifendorf, D., Benninghoven, A. J. Vac. Sci.
Technol. A 1985, 3, 1322.

Sakurai, T.; Matsuo, T.; Matsuda, H. Int. J. Mass Spectrom. Ion Phys. 1985, 63,
273.

21.

22.

23.

24.

25.

26.

. 27.

28.

29.

30.

31.

32.
33.
34.

35.

36.
37.
38.

39.

40

Sakurai, T.; Fujita, Y.; Matsuo, T.; Matsuda, H. Int. J. Mass Spectrom. Ion Phys.
1985, 66, 283.

Pinkston, J. D.; Rabb, M.; Watson, J. T.; Allison, J. Rev. Sci. Instrum. 1986, 5 7,
583.

Yefchak, G. E.; Schultz, G. A.; Allison, J .; Enke, C. G.; Holland, J. F. J. Am. Soc.
Mass Spectrom. 1990, I, 440.

Muga, M. L. Anal. Instrum. 1987, 16, 31.

Yefchak, G. E.; Enke, C. G.; Holland, J. F. Int. J. Mass Spectrom. Ion Phys. 1989,
87, 313.

Kinsel, G. R.; Johnston, M. V. Int. J. Mass Spectrom. Ion Phys. 1989, 91, 157.

Karataev, V. 1.; Mamyrin, B. A.; Shmikk, D. V. Sov. Phys. Tech. Phys. 1972, I6,
1177.

Mamyrin, B. A.; Karataev, V. 1.; Shmikk, D. V.; Zagulin, V. A. Sov. Phys. JE T P
1973, 37, 45.

Mamyrin, B. A.; Shmikk, D. V. Sov. Phys. JET P 1979, 49, 762.

Grix, R.; Kutscher, R.; Li, G.; Griiner, U.; Wollnik, H. Rapid Comm. Mass
Spectrom. 1988, 2, 83.

Boesl, U.; Neusser, H. J.; Weinkauf, R.; Schlag, E. W. J. Phys. Chem. 1982, 86,
4857.

Della-Negra, S.; LeBeyec, Y. Int. J. Mass Spectrom. Ion Phys. 1984, 61, 21.
Walter, K.; Boesl, U.; Schlag, E. W. Int. J. Mass Spectrom. Ion Phys. 1986, 71, 309.
Yang, M.; Reilly, J. P. Int. J. Mass Spectrom. Ion Phys. 1987, 75, 209.

Tang, X.; Beavis, R.; Ens, W.; Lafortune, F.; Schueler, B.; Standing, K. G. Int. J.
Mass Spectrom. Ion Phys. 1988, 85, 43.

Bergrnann, T.; Martin, T. P.; Schaber, H. Rev. Sci. Instrum. 1989, 60, 792.
Chait, B. T.; Standing, K. G. Int. .1. Mass Spectrom. Ion Phys. 1981, 40, 185.
Macfarlane, R. D.; Torgerson, D. F. Int. J. Mass Spectrom. Ion Phys. 1976, 21 , 81.

Hillencamp, F., Karas, M., Beavis, R. C., and Chait, B. T. Anal. Chem. 1991, 63,
1 193A.

41

40. Opsal, R. B.; Owens, K. G.; Reilly, J. P. Anal. Chem. 1985, 57, 1884.
41. Stein, R. Int. J. Mass Spectrom. Ion Phys. 1974, I4, 205.

42. Stein, R. Int. J. Mass Spectrom. Ion Phys. 1994, I32, 29.

CHAPTER 3

THE MICHIGAN STATE UNIVERSITY M-TOF

Introduction

While the integrating transient recorder project described in Chapter 1 was being
completed, attention was focused on developing a time-of-ﬂight mass spectrometer that
would complement the capacity for high-speed data acquisition. Previous attempts at
developing time-of-ﬂight instrumentation, while successfully illustrating novel concepts,
had failed to provide both mass independent focusing and the detectability required for
cutting-edge chromatographic analyses. An acquired linear CVC 2000 [1] time-of-ﬂight
instrument utilizing a two-ﬁeld Wiley/McLaren ion source geometry and time-lag
focusing was capable of generating only scanned mass spectral information and exhibited
relatively poor detectability [2]. A bearn-deﬂection time-of-ﬂight instrument successfully
demonstrated the performance of the ITR, but with only marginal detectability [3]. A
proposed DFF modiﬁcation to the CVC 2000 could possibly have provided sufﬁcient
mass resolving power along with a concomitant increase in detectability, but the project

was dropped due to the high costs projected for completion of the electronics circuitry.

During this period, one of the Mass Spectrometry Facility directors delivered a lecture
at an American Society for Mass Spectrometry meeting [4] in which the research progress
to date was described as well as the fundamental requirements for adequate time-of-ﬂight
instrumentation compatible with the ITR. This lecture led to a meeting with Professor
Hermann Wollnik, who described a time-of-ﬂight mass spectrometer [5] developed in his

laboratory in Giessen, Germany, in which a novel electron ionization source was

42

43

combined with an ion mirror for improved detectability and mass-independent ion
focusing. It was reported that resolving powers near 20,000 had been achieved with this
instrument at a base pressure of 1x10’'0 torr. Subsequently, a collaboration was
established whereby a time-of-ﬂight mass spectrometer would be sent to the MSU Mass
Spectrometry Facility for adaptation as a chromatographic detector and an ITR would be
sent to the laboratory in Germany. The collaboration was initiated by an exchange of
post-doctoral associates, with Dr. Ron Tecklenburg ﬁrst traveling to the Wollnik
laboratory to learn the performance and operation of the time-of-ﬂight mass spectrometer,
and Dr. Raimund Grix arriving at MSU soon after the mass spectrometer was delivered.
A second collaboration with Meridian Instruments, Inc., enabled Dr. Grix to remain

working in the MSU Mass Spectrometry Facility for several years.

Duty Cycle Requirements

The time-of-ﬂight mass spectrometer developed by the Wollnik research group
incorporated two unique features. First, the ion mirror providing mass independent
focusing was of a gridless non-linear ﬁeld design and was reported to possess high-order
focusing properties [6]. Second, and eventually more important to the success of GC-

TOFMS, was a reported capacity for ion storage between source ion extraction events.

The signiﬁcance of these two features lies in the potential effect they have on duty
cycle. There are two types of duty cycle relevant to mass spectrometry. One of these, the
scanning duty cycle, is an assessment of the efﬁciency of measurement along the mass
(m/z or u) axis. For scanning mass spectrometers, this is the ratio of the mass range being
detected at any instant to the total mass range of the analysis. For example, if the

resolution of measurement is 0.2 u and the range being scanned is 500 u, the scanning

44

duty cycle is 0.04% and at any given instant only 1/2500 of the information along the
mass axis is being measured. Previous TOF instruments, with their mass-dependent
focusing, employed time-slice detection (described in Chapter 1) and were thus
essentially scanning instruments. With the availability of the ITR, m/z information could
be acquired over the entire mass range of interest for each transient, signiﬁcantly
increasing the duty cycle. This potential could only be realized however, if ions over the
entire mass range were simultaneously in temporal focus, and the ion mirror provided this

feature.

The second duty cycle, the sampling duty cycle, is a measurement of the efﬁciency of
sample utilization. This is an important consideration for TOF instruments since they
employ a cycle of ion generation, instantaneous ion extraction and subsequent m/z
analysis. Previous commercial TOF instrﬁents used pulsed ionization, typically 1 usec,
followed by a 99 psec analysis, yielding a duty cycle of only 1%. This is adequate for a
steady state sample, but for dynamic samples, such as a chromatographic eluent, severe
undersampling can occur. Even for an instrument using continuous ionization, most of
the ions formed during the 99-usec extraction and analysis time are lost to the vacuum
system and neutralization processes. The ion source technology developed by Dr. Grix in
the Wollnik laboratory was reported to have ion storage times of up to several msec with
storage efﬁciencies of up to 20% [5]. Additionally, the process providing the ion storage
also caused “ion bunching” that lowered both the initial ion energy and spatial
distributions in the source and enhanced the mass resolving power; also contributing to

improved detectability.

45

That both of these features were simultaneously required is apparent when it is
considered that the overall duty cycle of a TOF instrument is the product of both the
scanning and sampling duty cycles. In the example above, 1/2500 x 1/ 100 = 0.0004%,
which is signiﬁcantly lower than the duty cycle available with other forms of mass

spectrometry. For a TOF instrument to be suitable in GC—MS, a signiﬁcantly higher duty

cycle was required.

The M-TOF

The time-of-ﬂight mass spectrometer initially delivered to MSU was quickly replaced
by an improved design more . amenable to chromatographic sample introduction. This
instrument, referred to as the M-TOF (where the M stands for “mirror”), illustrated in
Figure 3-1, was the result of a collaborative effort involving a number of professors, post-
doctoral associates and graduate students. This authors’ responsibilities included the
mechanical design and engineering. The M-TOF consisted of a vacuum system having
two essentially separate chambers to facilitate differential pumping between the relatively
higher pressure ion source region and the lower pressure analyzer and thus preserve the
mass resolving power. Both chambers were vacuum pumped using 180 L/sec
turbomolecular pumps, each backed by an isolated roughing pump. The gridless ion
mirror was attached to the lower ﬂange of the analyzer chamber, while the custom built
channel-plate detector assembly was attached to the upper analyzer chamber ﬂange. Half
of the ion source was mounted to the upper ﬂange of the source chamber, along with the
electrical feedthrough, while the remainder of the ion source was mounted to the inside of
the ion source chamber. An Einzel lens and deﬂection plate assembly were located in the

connecting tube between the two chambers. A viewport on the front of the ion source

46

chamber facilitated positioning of the chromatographic column end, which entered the
vacuum chamber through a separate feedthrough/ﬂange assembly. The instrument had a
1.65-m overall ion ﬂight path length. The system was operated with the analyzer drift
region at ground potential, and the ion source conﬁgured to provide approximately

1000 V total ion energy. Typical ion optical element voltages are listed in Table 3-1.

 

 

 

 

 

 

 

 

 

 

 

 

Electrode Potential (Volts)
G3 820
G3 when mlsed 950
‘ (32 820
Filament 755
' Pushers 760
G1 255
G0 0
Einzel Lens -6OO
Mirror Back 1000 ,

 

 

Table 3-1. Typical voltages used in operating the M-TOF.

Ion Storage and the Grix Source

A cross-sectional view of the cylindrically symmetric ion source developed in the
Wollnik laboratory is shown in Figure 3-2. Electrons generated by an annular ﬁlament are
directed into the center of the source between two gridded electrodes (G2 and G3) by two
ring electrodes called “pushers”. During ion storage, the potential applied to G2 and G4
are similar and slightly positive with respect to G3. This provides a potential well in the

vicinity of G3 where ions could be retained between ion extractions, and the

47

Ion Source

     
 

Interface

\

 

Gaseous Sample =>

Detector

 

T => Signal to ITR

 

 

Ion Flight Path \

Turbo
Pump

/ Ion MIITOI'

 

 

 

llIllIlIIk

 

 

Figure 3-1. M-TOF mass spectrometer.

48

potential depth of the well could be varied to optimize this trapping capability. An
extraction pulse applied to G3 repels the stored ions through G2 and G1 into the analyzer

for m/z analysis.

Early in the development of the ion source at MSU by Dr. Grix, questions arose
challenging the hypothesis that G3 was the location of ion storage. This was partially due
to the fact that, in principle, the limited transmission of the grid material should have
removed a large fraction of the generated ions as they pass through the grid multiple
times during storage. Additionally, it was observed that ion detectability remained high
even when the potential applied to G3 during ion storage was higher than that of G2 and
G4. It should thus be impossible for the ion storage to occur at G3 under these conditions.
Subsequent experiments with and without the G3 electrode in the ion source yielded
similar detectabilities and storage characteristics, verifying that G3 was not the essential
storage element and that the ion source did not function as originally described by the

inventors [7].

A more likely explanation for ion storage was that a potential well was being formed by
the continuously generated electron beam. It was observed that ion storage and
detectability were signiﬁcantly more dependent on the potentials applied to the pusher
electrodes and the alignment of the pusher/ﬁlament assembly. Early evidence for electron
beam ion storage mechanism was ﬁrst provided by Studier [8] in 1963, in a paper
describing a time-of-flight ion source wherein, using continuous ionization, a BOO-fold
increase in detectability was achieved over that attained using pulsed ionization. Further
support came from Baker and Hasted [9], and consequently, the Studier conﬁguration

was used as the model for further ion source developments in the facility.

49

 

 

  
 

8)
Filament
GCBI {/F—l G1
-.-QPtl°.a_'._
Axis
‘— K/ll‘l
Pushers
b)
Extraction
a > i
E /
q, Storage3
*5 ': 2
D.

 

 

 

G4 G3 62 G1

Figure 3—2. Cross sectional view (a), and potential diagram (b), illustrating the
operation of the original Grix/Wollnik ion source.

50

The second source developed by Dr. Grix utilized an advanced design, that eliminated
the previous G3 electrode and retained the annular ﬁlament and electron beam pushers. In
this design however, the repelling electrode (now G3), the ﬁlament and the pushers were
mounted on a removable ﬂange and the G2, G1 and G0 electrodes were mounted in the
ion source chamber. This ion source is depicted in Figure 3-3, and illustrates the heated
conduit through which GC capillary column was inserted. Optimal chromatographic

detectability was obtained when the column end was ﬂush with the G2 electrode surface.

The M-TOF performance [10] with this Grix ion source was, at times, phenomenal.
Mass resolving powers of 1500 were achievable, as illustrated in Figure 3-4, and an
ultimate detectability of 5 frnol bromobenzene (780 pg) was attained, as illustrated in
Figure 3-5. However, it eventually became clear that when operational, rather remarkable
results could be attained, and when not, as was often the case, the ion source required
extensive attention. This attention could apparently only be provided by Dr. Grix, and
even under his expert care, it often took many hours of mechanical adjustment and

electrical tuning of both the ion source and mirror to effect a functional recovery.

Subsequent Ion Sources

When the limitations to the Grix/Wollnik ion source conﬁgurations were fully
recognized, a series of experiments aimed at achieving a more reproducible and stable ion
source conﬁguration was begun. These experiments were conducted primarily by this
author and Dr. George Yefchak who, after graduation from MSU, was employed by
Meridian Instruments to guide aspects of the research that would eventually lead to a
commercial product. Dr. Yefchak had previously developed a theoretical model

describing the ion storage process [7‘], and conducted initial experiments suggesting

51

Cage GS

\

Filament —> o

 

 

K
o / Pushers

 

 

‘62

/
Capillary
Inlet

       

GO 4—Insulator

Figure 3—3. The Grix M-TOF ion source.

52

Low Mass High Mass

 

 

 

 

 

 

 

12000« 4°°°i
' 28
10000‘ “466
1 3WD:
.. 8000‘ Q .
,5 4000‘ .5
l. 32 1000 467
2000- . ”W
22.5 i'ééiﬁés'atd - fads 31.0 - - ' 79115' - - -910
Flight Time (us) Flight Time (us)
(a) (b)

Figure 3—4. Demonstration of M-TOF resolving power for a) residual gas ions where,
with a 12 nsec F WHM, R=920 and b) tris(heptaﬂuoropropyl)-8-triazine where, with a 30
nsec FWHM, R=1500.

53

 

77

Intensity

 

 

”'1 " 4' 1 Viiflf‘l'lqili‘ . I
37 41 45

5 fmol Bromobenzene

156

i
l 1 “i 1, 1“1 1'
1’ "‘ ‘

 

49 53 57 61

Flight Time (us)

Figure 3-5. M-TOF detectability using the Grix ion scurce.

 

54

that the storage process may be mass-dependent, as well as pressure
dependent [11,12]. These results led to a comparison of ion storage capabilities for the
various ion source conﬁgurations and formed the basis for benchmark performance

evaluations.

The theoretical model postulates that the number of ions, n, in the storage region

follow the rate equation

dn
Esz—kLn (3-1)

where t is time, and k. and kL are the formation and loss rate constants, respectively.

Integration of Equation 3-1 yields
_ ff. ‘kL’
n(t) _ (1 — e ) (3-2)
kL

It is important to distinguish between ion storage, as used in describing time-of-ﬂight ion
source behavior, and ion trapping. With ion trapping, a given ion may have an extended
residence time within the trap volume, such as is observed in ion trap mass spectrometers.
With ion storage, no evidence of extended ion residence times is observed, either through
evidence of ion-molecule reactions or otherwise. The concept merely describes a steady-
state, or “leaky bucket” ion population model in which enhanced storage levels are
achieved when ion loss rates are minimized. A theoretical plot describing this model is
shown in Figure 3-6. An increase in the observed ion population occurs as the storage
time between ion source extractions is increased, until a steady-state ion population is

achieved.

55

 

Ion Intensity

 

 

 

 

Extraction Period
Figure 3-6. Trend for ion population growth curves.

S6

The experimentation with non-Grix style ion sources began by evaluating a rather
small unmodiﬁed F innigan quadrupole ion source that, while exhibiting a low resolving
power of ~50, showed a capacity for ion storage. This author then constructed a series of
larger ion sources through which to explore the actual mechanism of ion storage and to
develop a more rugged and reliable mechanical design. A summary of results obtained
with these sources is given in Table 3-2. A variety of ion source physical parameters were
tested, including several having multiple linear ﬁlaments and several incorporating an
isolating “cage” that provided a ﬁeld-free region around the ion optical volume. Of the
various conﬁgurations explored, those having a more extensive cage and those having
more ﬁlament surface (i.e. more electron ﬂux) showed the best storage characteristics,
and exhibited the best resolving power and detectability. The best achieved detectability
was 1.6 pg bromobenzene, observed using a time-synchronized oscilloscope for detection

readout.

While the use of linear ﬁlaments and the ﬁeld isolating electrodes provided a more
robust ion source conﬁguration, the reliability of these ion sources remained marginal and
exhibited day to day ﬂuctuations in performance. Shortly after these experiments were
completed however, in a collaboration with Meridian Instruments, Inc., it was ascertained
that the inclusion of rare-earth magnets and additional electron beam steering electrodes
had a signiﬁcant positive impact on performance consistency. This knowledge became

instrumental in developing the ion sources presently used.

57

 

Geometry l/kL k, R

i 7
E E 0.19 11,900 1300
a) i '
xmxx‘.
E. E 0.32 1,700 1200

E 0.17 650 900
__CL_

E 0.08 2,400 900
d) 's. ~.

0.07 1 ,000 600

 

 

 

 

 

 

 

-- -- 600

 

 

 

 

 

 

Table 3-2. Comparison of various ion source conﬁgurations. For each cross sectional
view, the gray lines represent ﬁlaments, while the dark lines represent ﬁeld cages and
electron beam focusing electrodes.

Further Developments

As work progressed in developing the ion source technology required for GC-TOFMS,
funding priorities shiﬁed at the National Institutes of Health, which was the primary
federal funding agency sponsoring this research. Consequently, funds for this research
signiﬁcantly diminished. However, Meridian Instruments, Inc. had acquired the licenses
for the ITR from MSU and, being interested in developing a commercial GC-TOFMS
product, continued to ﬁnancially support much of the research. This research led to the
production of a viable GC-TOFMS system, the commercial rights to which were sold to
LECO Corporation in 1995. A research project by this author, utilizing the LECO PCB-

650 time-of-ﬂight mass spectrometer, is included as Appendix A.

58

Summary
The research conducted at the MSU Mass Spectrometry Facility has yielded a

signiﬁcant advance to the state of the art in mass spectrometric detection for cutting-edge
gas chromatographic analyses. A gas chromatographic analysis record of 55 compounds
in 14 seconds has been achieved, and a number of terms are now an accepted and

common part of the scientiﬁc nomenclature.

While there was no consensus regarding the mechanism of ion storage, and while the
equipment developed in the laboratory exhibited a low working efﬁciency, further
engineering led to the development of a powerful commercial product. As such, the entire
project represents a major success for all who were involved: MSU, the members of the
Mass Spectrometry Facility, the National Institutes of Health who sponsored the research,
and for the American model of moving good ideas from the academic setting to the
broader public arena. It was also this work in ion source development that prepared this

author for developments described in the next chapter.

59

Chapter References

10.

11.

12.

CVC Products, Inc., Rochester, New York.

Erickson, E. D.; Enke, C. G.; Holland, J. F.; Watson, J. T. Anal. Chem. 1990, 62,
1079.

Schultz, G. A.; Charnberlin, B. A.; Sweeley, C. C.; Watson, J. T.; Allison, J.
J. Chrom., 1992, 590, 329.

Enke, C. G. In Proceedings of the 3 6" ASMS Conference on Mass Spectrometry and
Allied Topics, San Francisco, CA, June 5-10, 1988, 363.

Grix, R.; Griiner, U.; Li, G.; Stroh, H.; Wollnik, H. Int. J. Mass Spectrom. Ion Proc.
1989, 93, 323.

Grix, R.; Kutscher, R.; Li, G.; Griiner, U.; Wollnik, H. Rapid Comm. Mass
Spectrom. 1988, 2, 83.

Yefchak, G. E. Doctoral Dissertation, Michigan State University, 1990.
Studier, M. H. Rev. Sci. Instrum., 1963, 34, 1367.
Baker, F. A.; Hasted, J. B. Phil. Trans. Roy. Soc. London, 1966, 261, 33.

Grix, R.; Tecklenburg, Jr., R. E.; Watson, J. T.; Holland, J. F.; Yefchak, G. E.;
Gardner, B. D.; Allison, J .; Enke, C. G.; Wollnik, H. In Proceedings of the 39’"

ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, Tennesee,
May 21—26, 1991, 154.

Yefchak, G. E.; Puzycki, M. A.; Allison, J .; Enke, C. G.; Grix, R.; Holland, J. F.;.
Li, G.; Wang, Y.; Wollnik, H. In Proceedings of the 38“ ASMS Conference on Mass
Spectrometry and Allied Topics, Tucson, AZ, June 3—8, 1990, 540.

Seeterlin, M. A. Doctoral Dissertation, Michigan State University, 1993.

CHAPTER 4

NON-LINEAR ION ACCELERATION

Introduction

As energy correcting devices such as the ion mirror continue to develop and improve
in their focusing properties, the major detriment to resolution arises from energetic
variations due to spatial displacement and temporal variations due to ion turn-around.
While ion tum-around can be virtually eliminated with the newer surface ionization
techniques, it remains a major variable in methods employing three-dimensional
ionization volumes. In these applications, the effect of ion tum-around can be reduced by
higher source extraction ﬁelds. This is a common practice which, unfortunately,
exacerbates the spatial displacement problem. Additionally, many recent applications
encourage the increase of either the ion source dimension itself or the source
dimension/total ﬂight distance ratio, a situation which also emphasizes the need for better
spatial corrections. Consequently, the major objective of this study is to evaluate prior
efforts to correct for temporal variations due to displacement along the optical axis at the
moment of ion formation, to postulate a non-linear ﬁeld model to accomplish this end,

and to conﬁrm the efﬁcacy of the model with experimental results.

The Wiley/McLaren 1“ Order Space Focus Model

Wiley and McLaren’s innovations using the two-ﬁeld ion source brought recognition
of a location along the ﬂight path where the temporal spread of an iso-mass ion packet is

minimal, which in terms of space focusing is the optimal location for ion detection. This

60

61

location since known as the space focus plane, and the strengths and limitations of the
Wiley and MeLaren innovations along with the model on which they are based can be

evaluated by examining the processes that determine when the ions arrive at the detector.

Ion motion through an electric ﬁeld

Figure 4-1 illustrates the two-linear-ﬁeld ion source model implemented by Wiley and
McLaren[1]. An ion with initial energy U0, beginning at a position 3, axially located
within the ion source, will increase its energy as it is accelerated through the electric
ﬁelds until it achieves a ﬁnal energy U, upon leaving the ion source. The time required
for an ion to reach the detector can be determined using the analytical equations for

motion and force.

The kinematic equation for motion in space along a straight line with constant

acceleration has the basic form

1
x=x0 +vot+§at2 (4-1)

where x is position, x0 is the initial position, v0 is the initial velocity, a is acceleration and
t is time. Since the distance d = x—xo, Equation 4-1 can be rearranged to

1
Eat2 +v0t—d=0 (4-2)

and solved for t using the quadratic equation solution to yield

-v iE/v2+Zad
t= ° 0° (43)

where only the + second term will be applied since the second term is always larger

62

Ionization Region
/‘/— Acceleration Region Detector \
i .

«s-llﬁ—d—D4 D

 

Space Focus Plane (SFP)\

Potential

 

 

 

Field-free Drift Region

Figure 4-1. Two-ﬁeld acceleration geometry implemented by Wiley and McLaren.

63

than the ﬁrst and time is always positive. Equation 4-3 can be applied to time-of-ﬂight
analysis by substituting for v and a using the following ftmdamental relationships. Given

that

1 2
U = va (44)

where U is the ion total kinetic energy, m is mass and v is velocity, the equation can be

solved for velocity, v, giving

-22 4,
v— m (-)

Likewise, the analytical equation for force,
F = ma = qE (4'6)

can be solved for acceleration, a, to form
a = — (4-7)

(The term q is used here rather than ze to preserve the Wiley/McLaren nomenclature,
which includes dimensional units of cm, 11sec, eV, and mass units u). Substituting the

equations for a and v into Equation 4-3, yields an expression for t that is a function of U0

and initial ion position

t(Uo,d) = 1.02[%](,/U, +qu - JU—o) (4-8)

which forms the basis of the equations Wiley and McLaren used to compute the ﬂight

time for ions accelerated using the two-ﬁeld ion source.

64

The ﬂight time through the ﬁrst accelerating ﬁeld in the two-ﬁeld ion source is thus

J—

2m
qE.

 

)(W —JU—o) (4-9)

1, =1.02[

and is proportional to the square root of the energy applied to an ion plus or minus the
square root of the energy already possessed by the ion depending on whether the initial

energy is with or against, respectively, the applied energy.

As the ion travels through the second electric ﬁeld it acquires additional energy. The
time of ﬂight through the second electric ﬁeld is determined in similar fashion to that of

the ﬁrst electric ﬁeld, and is thus

 

 

t, =1.02(‘(:?](,/U0 +qu, +qu, -,/UO +qu,) (4-10)
(1

which, as was shown with Equation 4-9, is proportional to the square root of the energy

applied to the ion minus the energy already possessed by the ion.

The time tD required to transit the ﬁeld-free region D ﬁom the ion source exit to the

detector is a function of the total ion energy, and is thus

072 0J2
m = 1.02 m (4-11)
2,]00 + qu, + qud ZJU

 

1,, =1.02

Finally, the total time required for an ion beginning at position 5 to reach the detector

is the sum of each of these times

t=t3+td+tD (4-12)

65

Space Focusing

Space focusing depends on the fact that an ion located closer to the exit in the ion
source acquires less energy, and consequently a lower velocity, than an ion located farther
back in the ion source. The less energetic ion exits the source ﬁrst but is subsequently
overtaken by the more energetic ion at the space focus plane. While the ion arrival time at
the space focus position is a function of total ion energy and m/z, ideal space focusing

requires that the arrival time be totally independent of initial ion position 3.

Mathematically Described

In evaluating an ion possessing no internal energy (U0=0) and beginning at position s
from an ion population centered about so, there should ideally be no change in arrival
time at the space focus plane with respect to s.

(:3 = 0 (4-13)

0.50

Unfortunately, when using linear, homogeneous electric ﬁelds to obtain Equation 4-12 as

derived above, the result of Equation 4-13 is not zero.

Evidence for this can be easily obtained using the equations derived above to evaluate
two ions beginning from two different axial locations, SI and s2, in a single-ﬁeld ion
source. Since the time required for both ions to anive at the space focus plane should be
equal according to the Wiley/McLaren model, the equations for the total ﬂight time of
both ions (Eq. 4-12) can be set equal to each other. The combined equation can then be

solved for D.

D = 2,/s,s, (4-14)

66

Likewise, since the two ions being evaluated must travel the same distance D, the
same equations can also be rearranged to solve for D and set equal to each other. By

subsequently solving for t, one obtains

t= 2'" S——‘——— (4 15)
ﬁfT-ﬁ'
In Equations 4-14 and 4—15 it is seen that both D and t depend on the values of SI and
s2, which would not be the case if, under the conditions of constant linear acceleration,

dt/ds=0 at the space focus position.

' Wiley and McLaren evaluated deviations from ideal space focusing in the following

manner. Assuming that U0=0 and s=s0, the following substitutions

U = qsoE, + qud (4-16)
= 50E, +dEd (4-17)
0 soE,

can be implemented into Equation 4-12, yielding

 

m 2%
tU , =1.02 — 2 k + d+D 4-18
(ow %A*f;nﬁ; E ( )

The location of the space focus position for ions beginning at s=soi‘/26s can be

determined by setting the derivative of t with respect to s equal to 0, to obtain

D: 2sMk,/_E1— ﬁrs] (4-19)

67

Deviations from ideal space focusing can then be determined using a series expansion of

t(0,s) about so.

Graphically Illustrated

Graphically, ion behavior using linear ﬁeld acceleration is illustrated in Figure 4-2.
Here, ﬁve ions having equal m/z and no internal energy are accelerated through the ion
source with the total energy acquired by each ion being a function of the initial location.
It can be seen that ions 1 and 2, originating from positions closest to the ion source exit,
arrive at an axially co-planar location earlier than the remaining ions, and at a location in
front of the deﬁned space focus plane. Likewise, those ions originating from positions
farthest ﬁ'om the ion source exit (ions 4 and 5) achieve a co—planar relationship beyond
the deﬁned space focus plane and at a later time. Only those ions originating from
positions near the center of the ionization region arrive at the prescribed space focus
plane co-temporally. Never are all of the ions either co-temporal at a given plane or co-
planar at a given time. However, as postulated by Wiley and McLaren, under conditions
where the spatial displacement is minimal, a reasonable ﬁrst order approximation of

space focusing can be achieved.

This single ﬁeld model was used by Wiley and McLaren to evaluate and describe
space focusing, and provided the theoretical basis for the two-ﬁeld source used in their
mass spectrometer design. The addition of the second, shallow-gradient electric ﬁeld in
the ionization region allowed them to place the space focus plane at the detector surface,
even with the detector located several meters from the source. The gain was a signiﬁcant
increase in resolving power, although the temporal characteristics of the space focus

plane were diminished from that of the single-ﬁeld source.

68

 

 

V

hlon Sour >14 Drift Region

 
  
     
    

 

 

 

 

 

 

I :-
1 i
E E Space Focus
E 1 ‘ Plane
.7! .3 1 f
E 1 i
9. 1 ‘
0 1
Q. I
|
l
i I
2 1
1
1+3, —>¢—2s, Distance
0 o
o
t 8
e o
e.
t o
. o
.0
t e
’t
t G) .
Q?
t 3’ @

 

Figure 4-2. Illustration of ion behavior at the space focus plane using single linear-ﬁeld
acceleration.

69

The fact that the single-ﬁeld model provides better space focusing than the two-ﬁeld
model if further validation of the conclusion that linear ﬁelds can provide only a
ﬁrst-order approximation of perfect space focusing. This is intuitively understandable
since linear physical displacements produce linear energy variations which result in
second order velocity changes that linearly affect the resulting time measurements. Thus,
with linear ﬁelds the space focus plane is in reality not a well deﬁned plane. The region
of minimum temporal spread is displaced along the ion optical axis in relation to the
position of origin in the ion source, and the overall width of the temporal minimum
dimension increases with the range of the physical displacement along that axis in the ion
source. Clearly, to achieve the higher levels of resolution desired for many modern

applications of TOFMS, an alternate acceleration scheme must be employed.

Ideal Space Focusing

Given that linear accelerating ﬁelds do not provide complete space focusing, and that
there is a non-linear relationship between energy and ion velocity, it is reasonable to
propose that some form of non-linear acceleration may provide improved space focusing.
Indeed, there have been several attempts at incorporating non-linear acceleration into
TOF mass spectrometer design, and these can be classiﬁed as either dynamic or static in

nature.

Dynamic Non-linear Fields

In the dynamic non-linear ﬁeld model, the ﬁeld imposed on the ions is static over
space at any given moment and may be homogeneous. However, as the ﬁeld is
temporally changing, the force applied to the ions is effectively non-linear in time. The

ﬁrst reported use of a dynamic ﬁeld occurred with the IFF technique, described in

70

Chapter 2. With IFF, a two-step-in-time electric ﬁeld was generated in the ion source,
with the ﬁrst relatively high amplitude pulse rapidly accelerating the ions prior to
application of the extraction ﬁeld. The velocity compaction technique proposed by
Muga [2], the dynamic ﬁeld focusing technique proposed by Yefchak [3], and post-
source pulse focusing [4] also represent the incorporation of non-linear ﬁelds in which
additional time-dependent acceleration is applied at a location well beyond the ion source
where the ions have achieved a substantial separation in space. This has the advantage of
allowing the additional acceleration to be tailored to a limited mass range, as opposed to

IFF, where all ions are accelerated using the same ﬁeld.

Static Non-linear Fields

In the static ﬁeld model, non-linear ﬁeld gradients are held constant in space
throughout the time of ion transit. A number of references exist using non-linear post-
source ﬁelds to correct for the initial energy spread [5—10], however little has been
published regarding the use of non-linear static ﬁelds for space focusing. Stein [1 1]
proposed, but did not experimentally verify, that the only ﬁeld capable of providing
perfect axial space focusing would be one in which the force was directly proportional to
the distance from the detector. Rockwood presented information at several scientiﬁc
conferences [12,13] describing a TOF mass spectrometer using an “inverted perfectron”
parabolic potential accelerating ﬁeld for space-time focusing of electrospray generated
ions. A diagram illustrating a series of gridless diaphragrns along the entire length of the
ﬂight path can be found in the author’s extended abstract, but no detailed description or
evaluation of the perfectron has been published. Hulett and co-workers [14] developed a

quadratic potential time-of-ﬂight mass spectrometer for measuring the m/z of ions

71

generated ﬁom the interaction of organic molecules and positrons in a Penning trap. The
ions thus formed were accelerated by the application of a quadratic ﬁeld that extended to
the detector and varied with the square of the distance ﬁom the detector. This single non-
linear ﬁeld conﬁguration is a special case of non-linear acceleration that takes advantage
of the second order relationship between energy and velocity. By placing the detector at
the end of the acceleration region, a simple quadratic ﬁeld was suitable in compensating
for initial position. However, a signiﬁcant disadvantage for this conﬁguration was that
the relatively short ﬂight path (61 cm) yielded a relatively short ﬂight time.
Consequently, the spectrometer had a resolving power of approximately 10, and ions
above mass ~100 were completely unresolved. A second disadvantage for this design was
that only the farthest 20% of the ionization volume provided sufﬁcient energy to the ions

for detection.

It is apparent that the single parabolic ﬁeld, illustrated in Figure 4-3, can provide
complete space focusing. However, the two shortcomings described above make such an
accelerating ﬁeld inappropriate for a practical, functioning mass spectrometer. First,
while in principle, iso-mass ions beginning from any axial point along the ﬂight path of
the parabolic ﬁeld should arrive at the detector simultaneously, those ions beginning from
positions close to the detector lack sufﬁcient energy to produce secondary electrons on
the detector surface and thus diminish detectability. Practically, only a portion of the
parabolic ﬁeld will provide the required energy. Unfortunately, this is also the region
having the greatest electric ﬁeld gradient, and thus the greatest dispersion as well.
Second, parabolic designs do not provide a sufﬁciently long ﬂight path for adequate

separation between ions of successive m/z, and do not allow for the inclusion of

Potential

 

72

Detector
Space Focus Location \

Ion Source Exit

1- - ~ - II- a a .
, . _. . ,. . . Y
r-m—_—_‘wu.‘_.w~m_____— ..._.- .-. -_-. . — ..

 

Ionization/Acceleration Region Distance

Figure 4-3. Potential proﬁle of a non-linear parabolic ﬁeld. The entire ﬂight path

contains the non-linear potential.

73

additional energy correction devices such as the ion mirror along the ﬂight path. As will
be shown below, an operationally more practical design would include a linear
acceleration ﬁeld to increase overall ion energy for better and more uniform detection, as
well as a non-linear region to compensate for initial ion position. This conﬁguration alters
the ideal quadratic ﬁeld of the ion source described by Hulett into various polynomial
forms which are a function of ion optical parameters and structural dimensions. The
following theoretical modeling and experimental results will conﬁrm that complete space
focusing can be realized with the combination of a non-linear ﬁeld in the ionization

region follower by a subsequent linear accelerating ﬁeld.

Non-parabolic Non-linear Fields

Addition of a linear acceleration region into the overall source design alters the non-
linear ﬁeld in the ionization region ﬁ'om parabolic to polynomial. As illustrated in Figure
4-4, the particular non-linear electric ﬁeld proﬁle required in a given ion source and mass
spectrometer design can be determined by evaluating a series of linear electric ﬁelds
which, if relatively short in comparison to the overall ion source geometry yield an
effectively non-linear electric ﬁeld. First, using Equations 4-8 and 4-11, the linear
accelerating ﬁeld required to provide the expected time of ﬂight for ions of a given m/z,
beginning at the ionization/acceleration region boundary, must be determined. Next, a
second, linear electric ﬁeld segment is calculated for the same ions beginning ﬁ'om a pre-
determined distance within the ionization region. Using Equations 4-9, 4-10, and 4-11,
the correct second electric ﬁeld accounts for the fact that the ions will also be accelerated
by the previously calculated ﬁrst electric ﬁeld. Subsequently the third linear electric ﬁeld

segment is calculated in similar fashion for the same ions, beginning slightly farther back

74

Acceleration Region

 

 

 

 

 

 

Ionization Region -|
7'. Detector
l
I
' 1
1 2
1 a
| SFP \i
A I E
“a, T 1 '
. :
l
E I
E l
o
‘5
o.
l
I
I
I
l
I
l
l
l
I
l
_ l
Distance

Figure 4-4. Graphical illustration of the method for calculation of the appropriate non-
linear ﬁeld required for ideal space focusing where the detector is some distance from the
ion source exit and a linear acceleration region is included. Each electric ﬁeld segment is
calculated by successive approximation.

75

in the ionization region, incorporating an equation similar to Equation 4—10 that accounts
for the third electric ﬁeld and the fact that the ion will be subsequently accelerated by the
both of the previously calculated electric ﬁelds. Determination of the nth electric ﬁeld
occurs in the same manner. As the time of ﬂight for an ion beginning anywhere in an ion

source having n ﬁelds is

 

 

- 1 I .
U +qu.E. —\[U +q 2): E
[J 0 __n j j 0 z.” k k
1: EE 2} .1- E I: 1 + ID (4_20)
q H 2 ZxE

 

 

where the indexes i, j, k and I, serve as computational counters, the ﬁnal electric ﬁeld
segment n can be easily determined. Equation 4-20 is merely a continuation of the
derivation of Equation 4-8, and the validity of this approach to non-linear ﬁeld
determination was recently veriﬁed by Yefchak [15]. As shown in Figure 4—5, the
idealized non-linear ﬁeld proﬁle is sigrnoidal. The non-linear ﬁeld in the region farthest
ﬁom the ion source exit retains a quasi-parabolic shape, with a diminishing rate of change
in the electric ﬁeld moving toward the exit. This is followed by an inﬂection in the
electric ﬁeld proﬁle and subsequently, a narrow region just before the
ionization/acceleration region boundary, where the rate of change in the electric ﬁeld

increases sharply.

76

Acceleration Region
!onization Region 4“

 

 

I4
i‘

Detector —\

   

SFP \g

  

Potential

 

 

 

Distance

Figure 4-5. Non-parabolic non-linear acceleration scheme employing a linear
acceleration region in combination with a non-linear ionization region.

77

In order to evaluate the electric ﬁeld parameters for non-linear ion acceleration with a
given ion source conﬁguration, a computational program based on Equation 4-20 was
written in the CH programming language. This program, listed in Appendix A as
TOFSOURC3.CPP, calculates the electric ﬁeld and absolute axial voltages required
within the ionization region to provide ideal space focusing. These ﬁelds are determined
using a successive approximations algorithm where the value for a given electric ﬁeld
segment is repeatedly modiﬁed until the desired ﬂight time for a given ion is achieved to
within a speciﬁed tolerance. The program is designed to allow the incorporation of
multiple ﬁelds, with the accelerating region having a distinct axial dimension and the
ionization region having mono-incremental ﬁeld segments. Operation of the program
occurs by inputing values for the total ﬂight time t, drift region distance D, accelerating
region length d, step size between ionization region increments, mass, and charge. The
program is then compiled and an output data ﬁle generated, in txt format, listing the axial

position and electric ﬁeld requirement.

Three major parameters, ﬂight time, drift distance, and acceleration region length,
effect the overall ion source theoretical voltage proﬁle. Figure 4-6 illustrates the effect of
varying the total ﬂight time, using an ion source conﬁguration having a 0.50 cm
accelerating region length and a 153 cm drift length. As the required ﬂight time is
reduced ﬁ'om 30 nsec to 15 nsec, the required potential within the ion source increases, as
expected. Irnportantly, the overall shape of the non-linear ionization region potential is
speciﬁc to the physical dimensional parameters, with only the amplitude being directly
related to the desired ﬂight time. The effect of varying the physical dimensions is

illustrated in Figure 4-7 and Figure 4-8. The effect of varying the distance between the

78

ion source exit and the space focus plane at the detector under the condition of constant
ﬂight time is illustrated in Figure 4-7. For a ﬂight time of i=8 usec and an acceleration
region distance of d=0.50 cm, an increase in drift region distance requires a greater initial
ion source potential, as expected. Additionally, the overall potential proﬁle in the
ionization region is signiﬁcantly different for each case. In the D=5 cm case, the
inﬂection point in the potential proﬁle occurs close to the ion source exit at ~s=1 cm. For
the D=50 cm case, the inﬂection point is signiﬁcantly farther back in the ion source at
~s=8 cm. The presence of an inﬂection point in each proﬁle is typical of non-linear
acceleration when a secondary linear accelerating region is included. Finally, the
dependence of the potential proﬁle on the linear accelerating region length is illustrated in
Figure 4-8 where, as the accelerating region length is increased, the potential proﬁle of

the non-linear region becomes less sigrnoidal and more parabolic in function.

As indicated above, the theoretical calculations, even under broad parameter
perturbations, conﬁrm that complete space focusing should be able to be achieved with a
non-linear ion source region followed by a linear acceleration region. Practically, this
appears to be achievable over a range of physical dimensional parameters, which implies
that the design of the potential proﬁle as well as the ion optical parameters could be
tailored to meet speciﬁc mass spectrometer design and performance requirements. This

ﬂexibility represents a signiﬁcant advantage over strictly parabolic-ﬁeld designs.

79.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8)
1200 1000
«~ 900
1000 ~~ Potential
«» 800
A 800 1. 700
E ’5
3 r- 600 3
£3; a
g 600 1 1~ 500 3
a
in C
E + 400 .3
'o 400 L a.
+ 300
1 200
200 1r
. . 1' 100
Electric Field
0 4 t i 1 r t 1 0
3.5 3 2.5 2 1.5 1 0.5 0
Ion Source Posltlon, s (cm)
b)
5000 4000
4500 4’ , 1’ 3500
Potential
4000 1
-- 3000
A 35001»
E ,_ ‘3
3 3000 1 2500 :5
Q z.
s 2500 T ._ 2000 E
3 E
5 2°00 " «~ 1500 ‘6
“O a.
1500 -~
-~ 1000
1000 ..
500 .. . . J " 50°
Eleg’tnc Flelg
O + t r E . t r 0
3.5 3 2.5 2 1.5 1 0.5 0

Ion Source Posltlon, s (cm)

Figure 4-6. Calculated electric potential and ﬁeld gradients required to achieve
complete space focusing at speciﬁc ion ﬂight times for m/z 56, D=153 cm, d=0.50 cm, for
a) t=30 msec and b) t=15 msec.

80

 

   
 

 

 

 

 

     

 

 

 

 

 

8)
50 500
45 4- 450
40 «~ » 400
35 «» E 350
E ’4?
2 30 « 300 ‘7;
3 2,
° 25 «— Potential 250 7.-
Z. :2
0 c
2 20 ~~ «~ 200 .3
g o.
15 ~» E 150
10 + "l' 100
5 l + 50
0 E E E E E E E 0
14 12 10 a 6 4 2 0
Ion Source Position, 8 (cm)
b)
2500 4000
l—
T 3500
200° " Potential
» 3000
E 3 -
2 1500 ~» 250° %
£3 a
O _
a r- 2000 g
3 1000 5
; «~ 1500 ‘6
U o.
-» 1000
500 «-
. . «- 500
Electnc Field J
0 E E E E E E E 0
14 12 10 8 6 4 2 0

Ion Source Position, 5 (cm)

Figure 4-7. Calculated electric potential and ﬁeld gradients required to achieve
complete space focusing at speciﬁc drift region distances for m/z 56, t=8 msec, d=0.50
cm, for a) D=5 cm and b) D=50 cm.

81

 

     
 
  

 

 

 

 

 

 

 

   
 

 

 

 

 

 
   
 

 

 

 

 

8)
700 1200
A 600 «- w 1000 A
g 500 l 800 g
7. - “ °
= 400 .1 Potential 2,
° 600 7.
a 300 E z
0 C
E 400 g
> 200 "" . . o
'0 100 4 EleCtnC FlGld t. 200 0'
0 E E E E E E E wl O
14 12 10 8 6 4 2 0
Ion Source Poeltlon, s (cm)
b)
120 1200
A 100 «— ~ 1000 A
E a
g 30. ~800 L;
a Potential 3,
a 60 1~ 1* 600 2
a ‘c‘
g 40 E E 400 3
> O
u 20 ~~ E 200 °‘
0 E E E E E E E 0
14 12 10 8 6 4 2 0
Ion Source Position, 5 (cm)
C)
140 1600
A 120 -4- 1" 1400
E, 100 w . . 1- 1200 E
3 80 Electnc Field l1ooo .9,
i -r 800 g
8 6° " Potential taco g
\ 3. 0
g 4° «- 400 o.
20 .1 E 200
0 E a E E E E E 0
14 12 10 8 6 4 2 0

Ion Source Position, 5 (cm)

Figure 4-8. Calculated electric potential and ﬁeld gradients required to achieve
complete space focusing with speciﬁc ion source acceleration region dimensions, for m/z
56, D=153 cm, t=5 msec, where d= a) 0.05 cm, b) 0.5 cm, and c) 2.5 cm.

82

Experimental

A mass spectrometer suitable for evaluating the theoretical models of linear and non-
linear ion acceleration was designed and fabricated in house by this author. The system
consists of a linear time-of-ﬂight mass spectrometer with an open-geometry ion source
through which laser induced photo-ionization can be achieved at speciﬁc locations within
the ionization region of the source to selectively probe and evaluate ion acceleration
behavior. The ion source consists of multiple closely spaced ring electrodes electrically
biased using voltage divider circuits to generate linear ﬁelds and, by means of a series of

short, linear, electric ﬁeld segments, an approximation of non-linear ﬁelds.

Experimental Apparatus
Vacuum System

The mass spectrometer, illustrated in Figure 4-9., was designed as a test platform for a
variety of ion source conﬁgurations. The ion source chamber consisted of a stainless steel
6-way cross (6 34” tube diameter, MDC Manufacturing Co.) with Conﬂat style ﬂanges,
and was vacuum pumped using a 800 US (N2) diffusion pump (Edwards, model 160).
The analyzer region consisted of a 2-meter length of 6 3/4” diameter stainless tube vacuum
pumped using a 322 US (NZ) diffusion pump (Edwards, model 100). Both diffusion
pumps were backed by a single roughing pump (Alcatel, model 1020). A 1 V2” diameter
aperture was located between the ion source housing and analyzer region to facilitate
some measure of differential pumping between the two vacuum regions during analysis.
Vacuum was measured using two Bayer-Alpert style vacuum gauges, one located at the

ion source housing and one located near the throat of the analyzer region diffusion pump,

using a Hewlett Packard 5945A

83

$580me :53?» 98 scream—mace £82:onon mmaE 235.3685 .mé 939....

 

 

 

 

 

 

 

_ 1 H

mean.
56.55

 

 

 

 

 

 

 

 

 

850w co. A
3st E::om> / T

 

 

 

 

 

 

_

 

 

 

 

 

 

 

 

£95. £5 e a?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\_

 

 

 

 

 

[_

 

[

 

 

 

 

ﬂ.
Wt

/ 88¢qu oEmm \

 

_J._E

Figure 4-9.

84

gauge controller. Under normal operating conditions with no analyte being introduced
into the vacuum system, the base pressures achieved were 2><10'7 torr and 9x10‘8 torr for

the ion source and analyzer regions, respectively.

Detector

The detector, located at the end of the analyzer region just beyond the diffusion pump
ﬂange, consisted of a micro-channel plate chevron assembly (Galileo Electrooptic)
mounted on a Conﬂat style ﬂange. All electrical connections were made through this
ﬂange, and the entire detector assembly was electrically shielded from the analyzer region
using a 6.50” diameter screen, gridded to facilitate vacuum pumping, and a 1.50”
diameter center aperture covered with 90% transmission electroformed grid mesh
(Buckbee Mears) to minimize ion transmission loss. The voltage applied to the detector

for the experiments described below was —1350 V.

Ion Source Platform

The ion source mounting platform, as shown in Figure 4-10, consisted of a 6.00 inch
diameter brass stage mounted 6.25 inches from the ion source ﬂange. The stage had a
1.06” diameter mounting ring for ion source attachment located on the ﬂange side, and an
Einzel lens and deﬂection plate assembly located on the analyzer side. All electrical
connections for the ion source, Einzel lens, and deﬂection plates were made through the

ion source ﬂange.

Laser and Optical System

Gas phase analyte molecules were photo-ionized using a laser and a series of optical

elements to direct and shape the beam. The system consisted of a Quanta-Ray GCR—ll

 

Nd?
prm
outp

dura;

A
optic
adjus
cylinc
betwe
target
5-mm
throng
chamt
Was m
in Figt
laser b

Optical

85

Nd:YAG laser with a model HG-Z harmonic generator and model DHS-Z splitter set to
provide a laser wavelength of 355 nm. The laser was ﬁred at a rate of 10 Hz with an
output power of 45 mJoules as measured with an Ophir laser power meter. The ﬁring

duration was approximately 5—6 nsec.

An optical rail, mounted on an ion source chamber ﬂange, housed the beam guiding
optical components, which consisted of two high-reﬂection aluminum mirrors on
adjustable mounts and a vertically oriented cylindrical lens on a ﬁxed mount. The
cylindrical lens focused the beam into a slit image which made it narrow enough to pass
between the ion source electrodes and increased the photon density in the space being
targeted for ionization. The adjustable mirrors provided the ﬂexibility for guiding the
5-mm diameter collimated beam from the laser, located on a separate portable cart,
through the cylindrical lens and ultraviolet transmitting viewport into the ion source
chamber. One mirror and the cylindrical lens were mounted on a sliding assembly which
was moveable along a single axis oriented in parallel with the ion optical axis, as shown
in Figure 4-10. The optical path could thus be adjusted so as to place the focal line of the
laser beam at the axial center of the ion source while precisely positioning it along the ion

optical axis at pro-selected points of ionization.

86

 

 

 

/ Mounting Platform

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__ Einzel Lens
Ion Source \ . / , _
E
- - To
Ionization Spot Detector
J
_ L
L Li
Laser Beam \ Deﬂection Plates
l
__-_.._-._--_--_. -._ _ _--_--_--_--_--_.._.._ _ .Yareartl._
Quartz VIeWport Atmosphere l
l Cylindrical Lens
f 1 AI Mirror
+1 —>
t 1
W X
Moveable Stage Optical Rail

Figure 4-10. Laser optical and ion optical assemblies. The ionization spot is positioned
between ion source electrodes by adjustment of the moveable stage along the optical rail.

Ion Sc

The
from 0.1
machint
diaphra‘
within 1
used or
except :
the bacl
The 501
except :

region r

Samp
The

(approx
needle

which c
i0n SOL

inlCl’lOC

POWe.
The

VOllage

87

Ion Source

The ion source electrodes consisted of 3.00” outer diameter diaphragms constructed
from 0.017” thickness stainless steel sheet, cut into rings by the MSU physics department
machine shop using a wire EDM. Two inner diameters were used; 0.50” for the
diaphragms at each end of the accelerating region and 1.75” for the diaphragms used
within the ionization and acceleration regions. Electroformed grid mesh identical to that
used on the detector assembly was spotwelded over the opening in each diaphragm
except for one in the accelerating region which was left grid free. The ﬁnal electrode at
the back of the ion source was solid. The ion source assembly is shown in Figure 4-11.
The source electrodes were mounted in a stack separated by 0.100” ceramic spacers
except for the spacer separating the accelerating region electrode and the ﬁrst ionization

region electrode, which was 0.030” thick sapphire.

Sample Inlet

The sample introduction assembly was constructed using a small glass sample vial
(approximately 2 mL volume) with a 1/4” diameter neck swaged to a vacuum capable
needle valve. The valve was attached to a 12” length of Mt” diameter stainless tubing,
which completed the assembly. Sample was introduced to the vacuum environment of the
ion source housing by inserting the stainless tubing through a direct probe vacuum

interlock and opening the needle valve.

Power Supplies

The detector power supply was a Bertam Associates, Inc. Model 603-50N 5 kV high
voltage supply. The Einzel lens was supplied by a Gamma High Voltage Research high

voltage power supply, model RCIO-SN/V M 5 kV . The deﬂection plates were powered

88

 

 

 

 

 

 

 

 

 

 

Ionization Acceleration
Region (1 Region
,[Q __ Electrode
Bil or on an oh no or on E3 Designation
5 / Grid Material
/ Ion Source Exit

 

 

”

N

\ Ionization Position

 

 

 

 

 

 

 

 

 

 

 

U—OOGZ
—0.275
—O.578

0880*

1182—)
» 1.485—>
<——1.7e7——>

N—— 2.089 —-——>
2.370 ———>‘

Figure 4-11. Ion source assembly. Measurements shown are in cm.

 

 

 

 

 

89

using a Hewlett Packard 6207B DC 0-160 V, 0—0.2 A power supply. Ion source voltages
were supplied using a Glassman model ER, 3 kV, 100 mA high voltage power supply.

The output of this power supply was connected to the voltage dividers described below.

Voltage Dividers

Two voltage dividers were constructed using a resistor network. The linear divider, as
shown in Figure 4-12a, supplied a linear electric ﬁeld to both the acceleration electrodes
and the electrodes in the ionization region. The voltage applied to the electrode dividing
the ionization and acceleration regions could be varied using a SOO-kQ potentiometer,
allowing the ratio of the two electric ﬁelds to be varied and thus enabling the space focus

location to be optimized at the detector.

The non-linear divider, shown in Figure 4-12b, also used a resistive network with
ﬁxed resistors providing the base divisions, and potentiometers included for the

ionization region electrodes to allow ﬁne tuning of the required voltages.

Transient Recorder

Data were collected using a 500-MHz, 2.5-GS/sec Tektronix TDS 620B, two-channel
digital real-time oscilloscope. The oscilloscope was used in a variable delay mode set to
observe the dominant ion peak, with a l6-shot moving average of successive laser pulses
being recorded over a 2-usec time window. Each data set was saved as an ascii ﬁle and

downloaded to diskette for subsequent analysis.

9O

 

 

 

500 m
15tum *%

 

 

 

 

 

 

 

Figure 4-12. Resistive divider networks used for a) linear and b) non-linear ion source
conﬁgurations.

Ll

f1:

f0

‘1

91

Electric Field Calculations
Linear Conﬁguration

A program calculating the ion arrival-time distribution expected when using a linear
ﬁeld ion acceleration model was written in MathCad 4.0 [16] and used to compute the
arrival-time proﬁles expected over a series of ionization/acceleration electric ﬁeld ratios
for the instrument conﬁguration described above. This program is included in
Appendix A as LINEAR.mcd. Given a set of analytical parameters (m/z, d, D,
accelerating voltage, total voltage), the program calculates the ﬁght time for a selected
ion beginning at the acceleration ﬁeld/ionization ﬁeld boundary as well as anywhere
along the ionization region axis. It then generates a graph of arrival time vs. initial ion

position as well as a tabular output ﬁle.

Non-linear Conﬁguration

A second program calculating the series of electric ﬁelds required to provide the
approximation of ideal space focusing was also written using MathCad 4.0 and is
included in Appendix A as NONLINEARmcd. This program is based on Equation 4-20,
and is similar to the program TOFSORC3 in that it calculates both the electric ﬁeld and
voltage required for each ion source segment. However, this program calculates the series
of linear electric ﬁelds required to provide co—temporal ﬂight times for ions beginning
from a pro-selected set of inter-electrode positions. Both the electrode spacing and the
targeted inter-electrode ionization position are speciﬁed prior to calculation. The program
subsequently calculates the arrival times expected for ions originating over the entire
axial space between electrodes. Each of the mass spectrometer dimensional parameters

are included in the program (analyzer path length and ion source dimensions) as are the

92

analytical parameters (mass, charge, and accelerating voltage). The program ﬁrst
calculates the ﬂight time of a given ion having no internal energy and located along the
analytical axis of the ion source at the acceleration/ionization ﬁeld boundary. The
program then successively calculates the optimal electric ﬁeld required for each
ionization region given the optimal electric ﬁelds calculated for all previous ionization

region segments.

Experimental Method

The experiments were conducted using the laser to photo-dissociate gas phase analyte
molecules located within the designated ionization region of the ion source. Fe(CO)s was
chosen over an organic molecule as the analyte due to its combination of relatively high
ionization cross section, relatively low ﬁrst ionization potential (7.9 eV) [17], and its
volatility at room temperature. The time-of-ﬂight mass spectrum of femtosecond laser
photo-induced dissociation of F e(CO)5 at 71=400 nm has recently been reported [18], and

Fe+ was described as the predominant ion species after 500 femtoseconds irradiation time.

The analyte was leaked into the ion source housing through the direct inlet assembly,
yielding vacuum levels of 3x105 torr in the ion source housing and 6x10'6 torr in the
ﬂight tube. The laser was then ﬁred at 10 Hz with the moveable mirror and cylindrical
lens set to aim the laser beam between two of the ionization region electrodes. The arrival
time of the dominant peak in the spectrum was measured using the oscilloscope. Once the
oscilloscope had collected a minimum of 16 transients and averaged them, the laser beam
was blocked, the data recorded, and the mirror and lens assembly moved a pre-measured
distance to allow the sampling of the ionization space between the next electrodes. Nine

locations within the ion source were sampled in each data set.

93

The ion peak shown in Figure 4—13 exhibits a ﬂight time corresponding closely to the
expected average arrival time of singly charged “Fe, which accounts for 91.8% of the
isotopic abundance of Fe. For an accelerating voltage of 1629 V and total potential of
1887 V the expected arrival time was 21.35 usec starting ﬁ'om ionization region
position 5. The measured arrival time under these conditions was 21.20 nsec. Also
observed are smaller peaks at 20.82 nsec and 21.39 usec which correspond to 54Fe and
57Fe and account for 5.8% and 2.1% of the Fe isotope abundances, respectively. While the
system remained uncalibrated during the analyses and thus any electronic timing offsets
were unaccounted for, the experimental goals did not require explicit knowledge of any

ion' peak identity. It was sufﬁcient to monitor the same peak throughout the experiment,

regardless of the m/z.

Evaluation of linear configuration

Three sets of data were collected with the ion source conﬁgured to provide linear
electric ﬁelds at three different ionization/acceleration electric ﬁeld ratios. The expected
arrival time distributions were calculated by the LINEAR program, using the analytical
parameters listed in Table 4-1. A minimum acceleration ﬁeld potential of 1600 volts was
required to achieve an observable detector signal, and thus served as the lower limit for
all experiments. The data correspond to the measured arrival time of the peak apex for
m/z 56 ions. Example graphs of the original recorded data are shown in Figure 4-14 and
illustrate the difference in arrival times observed among the various sampled ionization

regions.

94

0.07

 

56Fe+

0.06 ..

0.05 --

 

:0
2

 

 

0.03 -~

Intensity (Volts)

0.02 *-

 

0.01 E 54Fe+

57Fe+
0 MAM-y)

1*

20.50 20.70 20.90 21.10 21.30 21.50 21.70

 

 

 

 

Time-of-ﬂlght (nsec)

Figure 4-13. Typical mass spectrum of Fe+ isotopes using the linearly conﬁgured ion
source. The resolving power is ~180.

95

 

 

 

 

 

 

 

 

Mass, m 56 u
Charge, 2 1

Drift Distance, D 153 cm
Acceleration Region Distance, d 0.648 cm
Total Potential, V1 1887 V

 

Table 4-1. Analytical parameters used to calculate the voltages and electric ﬁelds
required to provide linear acceleration.

Using the analytical equations for ion motion (Equation 4-12), the predicted arrival
time distributions of m/z 56 ions, along with the experimentally obtained data, are plotted
in Figure 4-15 as a function of initial ionization location and ionization

region/acceleration region ratio. These ratios are listed in Table 4-2.

 

 

 

 

AcceleratingVoltage Back Plate Voltage Electric Field Ratio
1605 1887 21.96
1629 1887 24.36
1710 1887 37.27

 

 

 

 

 

Table 4-2. Electric ﬁeld ratios used with linear acceleration experiments.

The theoretical traces exhibit a curvature due to the arrival time distribution expected
when using a multi-linear-ﬁeld ion source conﬁguration. The magnitude of this curvature
is a function of the electric ﬁeld ratio which controls the axial location for space focusing.
The curve corresponding to the electric ﬁeld ratio of 37.27 exhibits an increase in ion
arrival time as the initial position is moved back in the source, which is a consequence of

the space focus region being positioned beyond the detector surface. The curve identiﬁed

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.5
0.4% .—
0.3 ‘ 7
E
O
.>, 6
Q n...“
m
c
93
c
- 0.2 5
4
01 3 MA .A u-.__
._ .2 _ _. - A 2
0 1 A
' l l l | I
20.5 20.6 20.7 20.8 20.9 21 21.1 21.2 21.3 21.4

Arrival Time (nsec)

Figure 4—14. Arrival time distributions for F e+ ions originating from various ion source
locations. The ion source was conﬁgured for linear acceleration, with 1639 V applied to
the acceleration region and 1887 V applied to the source backplate.

97

as having an electric ﬁeld ratio of 21.96 shows a generally opposite trend where, while
the ion arrival time increases initially as the position moves back in the ionization region,
a maximum arrival time occurs fairly close to the ionization region exit and the ion
arrival time subsequently decreases as the initial position is located farther back in the
source. This trend of ion arrival time decreasing with increased initial position is expected
when space focusing occurs some distance before the detector location. The optimum ion
arrival time distribution, shown in Figure 4-15 as the 24.36 ratio, exhibits the minimum
absolute arrival time range over the entire ionization region, with a total temporal
distribution of 227 nsec. This is achieved when the ion source electric ﬁeld ratio is

optimized, and corresponds closely to the optimal ratio predicted by Wiley and McLaren.

In each case, the experimentally derived data agree closely in overall curvature with
the theoretical data. This strongly supports the ﬁrst-order space focus model and is the
ﬁrst data to conﬁrm the predictions of ion behavior originally proposed by Wiley and
McLaren. Only two minor discrepancies are present, and neither of these diminish the
strength of the data obtained. First, the arrival times for the experimentally derived data
were on average approximately 160 nsec earlier than predicted, which might be
attributable to either an offset in the electronics circuitry or a discrepancy between the
measured and actual ﬂight—path distance. To aid in the comparative evaluation, this offset
was added to the arrival time observed in each data set. Second, it is also observed that
while the experimental data exhibit the same overall curvature as the theoretical data, the
experimental data appear to be shifted to the leﬁ . This could easily be attributed to any
misalignment of the laser beam, which if not oriented orthogonal to the ion source, would

result in the ionization position being axially displaced.

98

 

 

 

 

21.5
I C
I o
I
-~ 21.4
I o
I
° «- 21.3
I
o
-- 21.2
I
o
-- 21.1
‘5
0
O
3
ﬂ
2
-. 21 2
’1?
&
I 9
0
.§
l-
«- 20.9
.. 20.8
+ Ratio = 37.27 “ 2°]
+ Ratio = 24.36
- 20.6
+ Ratio = 21 .96
. 4. 1 4 20.5
2.5 2 1.5 1 0.5 0

Ionlzatlon Position, 3 (cm)

Figure 4-15. Theoretical and experimental ion arrival time data for m/z 56, D=153 cm,
d=0.648 cm, at three acceleration ﬁeld/ionization ﬁeld ratios.

99

Evaluation of Non-linear Conﬁguration
Expectation Using A Multiple-field Non-linear Approximation

The multiple-ﬁeld ion source used for these experiments represents an approximation
of a truly non-linear ﬁeld design. This design is utilized because it is simple to construct
and allows suitable ﬂexibility necessary for initial evaluation. The assumption is made
that using multiple linear ﬁeld segments will not signiﬁcantly alter the required ideal non-
linear potential proﬁle, and that the overall performance will not be signiﬁcantly
diminished as a result. This assumption is based on the premise that the linear ﬁeld
segments are relatively short in comparison with the axial ion source length and thus
impose only minimal deviation. A comparison of the ideal and actual potential proﬁles is
shown in Figure 4-16. The ideal proﬁle was generated using the TOFSORC3 program
with 0.0001-cm increments over the 2.499-cm ionization region length. The voltage
proﬁle used in the evaluation was calculated with the NONLINEAR program using the
ionization region electrode spacing dimensions of between 0.124 and 0.303 cm, as
described previously. As shown, the potentials in both cases are virtually identical. More
revealing however, is the comparison of electric ﬁeld gradients, also shown in Figure 4-
16. The electric ﬁeld proﬁle present in the experimental design is a step-function with the
step boundaries occurring at the ion source electrode positions. The difference between
the ideal and actual electric ﬁelds is most pronounced over the ﬁrst ionization region, and
it is here that the greatest experimental deviation resulting from a non-ideal electric ﬁeld

may be expected.

100

 

 

    
  
 

 

 

 

 

 

2500 2500
2000 «» J— 2000
’5‘ Potential —/
é’, 1500 -» "150° ’3
2 2:
O O
2 a:
3 a
'9 =2
2 «3
.3 1000 4» «1000 o.
2
I.“
Electric Field —\
500 "F q»- 500
o l I r I i I 0
3.5 3 2.5 2 1.5 1 0.5 0

Ion Source Poslton s (cm)

Figure 4-16. Comparison of ideal and actual potentials applied to ion source
conﬁguration, and electric ﬁelds generated in both cases.

101

The effect of ionization position offset can also be examined by evaluating the
temporal deviation expected for ions beginning at non-ideal locations along the ionization
region axis, as plotted in Figure 4-17. While the non-linear ion source conﬁguration
yields equal ﬂight times for ions beginning at nine discrete locations within the ionization
region, there is a distribution of ion ﬂight times for ions beginning from other axial
locations. The pattern of temporal ﬂight-time distributions for ions originating across
each inter-electrode space is similar to that observed using the linear two-ﬁeld ion source.
In this case however, the magnitude of the temporal distribution is reduced as a
consequence of the electrode spacing being short in comparison with the overall
ionization region length. The most signiﬁcant temporal distribution occurs at the ﬁrst

ionization space where the ideal electric ﬁeld has the greatest rate of change.

Performance Using Calculated Design

The voltage proﬁle used in the non-linear ion source conﬁguration was designed to
provide an average ion arrival time similar to that obtained using the linear electric ﬁeld
conﬁguration. The physical dimensions described above were entered into the Mathcad
program NONLINEAR which calculated the necessary voltage for each electrode, listed
in Table 4-3. These voltages were then set on the voltage divider as measured, using a

handheld digital voltage meter.

A plot of the theoretically predicted and experimentally determined arrival time
distributions using the calculated required voltages is shown in Figure 4-18. Ideally, there
should be no deviation in ion arrival time due to initial ion location if the electric ﬁelds

have been properly selected. The experimental data show an ion arrival time range of

102

 

21

20.9

20. 8

Time-of-flight (nsec)

20,7

20.6

20. 5

 

 

 

2.5 lon Source Posltlon, s (cm) 0

Figure 4-17. Calculated ion anival-time distribution for the multi-gridded ion source,
beginning at the ionization/acceleration region boundary.

103

 

 

 

 

 

 

21.05
~r 20.95
~~ 20.85
'5‘
3
H :7
' a
~~ 20.75 -
I I C.
o.-
?
0
.§
.—
1~ 20.65
-~ 20.55
I v . l 20.45
2.5 2 1.5 ‘I 0.5 0

Ionization Posltlon, s (cm)

Figure 4-18. Theoretically predicted and experimentally determined ion arrival times for
m/z 56 ions originating in each ion source position. D=153 cm, d=0.648 cm.

104

only 63 nsec, which is a 3.6x improvement in variance over the best case observed using
linear electric ﬁelds. Furthermore, if the end points are discarded, the ﬂight time variance
is reduced to 32 nsec, which is a >7 x temporal improvement. This data set alone appears

to validate the expectation of the non-linear acceleration proposal.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Calculated Values Corrected Values
Electrode Potential (V) Potential (V) Difference (V)
0 1600 1600 0
1 1641 1642 1
2 1689 1699 10
3 1 727 1 730 3
4 1762 1764 2
5 1 794 1796 2
6 1824 1824 0
7 1852 1854 2
8 1880 1879 -1
9 1903 1901 -2

 

Table 4-3. Calculated and experimentally derived voltages and electric ﬁeld values

applied to ion source electrodes conﬁgured in non-linear acceleration mode.

Performance Using Empirical Voltages

 

In order to improve the performance of the non-linear acceleration ion source, the
voltages supplying each electrode were adjusted using the potentiometers located on the
voltage divider. Beginning with ionization region position 2, the voltages of electrodes
2-9 were individually adjusted to set the ion arrival time from each ionization region to be

20.764 us. The experimentally determined optimal voltages are listed in Table 4-3.

105

The difference between these sets of voltages ranges from —2 to +10 V. The predicted
ion arrival times and the measured arrival times using the empirical voltages is illustrated
in Figure 4-19, and a plot of the ion arrival time distributions at each ionization position
is shown in Figure 4-21Figure 4-21. There are several possible causes for the minor
discrepancies between the experimental and calculated voltages. The 10 V difference
observed for electrode 2 is partially a consequence of the performance from ionization
position 1, which consistently yielded arrival times lower than those resulting from the
other ionization positions. This discrepancy could not be corrected, and limitations of the
voltage divider precluded the possibility of providing to the other electrodes voltages that
would yield identical ion arrival times to that of ionization position 1. Thus, the voltage
on electrode 2 was altered to ensure that the arrival time associated with ionization
position 2 could duplicated throughout the rest of the ionization region. A second
contributor to the voltage discrepancies is that physical distortion in the electrodes or grid
material may have altered their axial location. A displacement of only 2 mm corresponds
to an ~2.5 V deviation in expected voltage, which is within the voltage error observed.
Visual observation of the ion source after conducting the experiments revealed slight
imperfections of approximately 1 mm in the grid material on several electrodes. Overall,
the difference between the expected and required voltages necessary to achieve complete
space focusing are relatively insigniﬁcant. Complete space focusing can be readily
achieved using an ion source designed to provide non-linear ion acceleration. The fact
that minor adjustments can be made that compensate for electrical or mechanical
imperfections also has practical signiﬁcance to the possibility of constructing mass

spectrometers implementing non-linear ion acceleration.

106

 

 

 

 

21.05
-~ 20.95
«~ 20.85
’5
O
Q
35
“
WI :
or 20.75 .9
c.
h
I 9
0
.§
.—
-~ 20.65
«— 20.55
4 . . . 20.45
2.5 2 1.5 1 0.5 0

Ionlzatlon Posltlon, s (cm)

Figure 4-19. Theoretically predicted and experimentally determined ion arrival times for
m/z 56 ions originating in each ion source position. D=153 cm, d=0.648 cm.

107

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.5
_-__9__,.I~‘M\~_ L r
“m— “ L .—
0.3M _
E
T)
a?
2
Q
0.2m - ﬂ
w/N- At
0.1 3 .2. A
2 44.2- _
12.. A - I
00 l l l l

 

 

20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9
Arrival Time (nsec)

N

21.1

Figure 4—20. Arrival time distributions for ions originating at the various ion source

ionization positions, where the ion source voltages were empirically adjusted to provide
optimal performance.

108

Effect of Deliberate Perturbation

The non-linear ion source conﬁguration was ﬁrrther evaluated by adjusting the voltage
of a single electrode within the ionization region in order to alter the electric ﬁelds within
the source and consequently, the time-of-ﬂight for those ions traversing the altered
electric ﬁelds. The experiment was conducted by altering the voltage on electrode 6 by
+3 V in one case and —3 V in a second case, and measuring the change in ﬂight time for
ions beginning in ionization regions 6—9. The effect of the perturbation on ion arrival
time is plotted in Figure 4-21 as ion arrival time versus ionization position and is
compared with the theoretically predicted ideal arrival time distribution as well as the
predicted arrival time distribution using the perturbed electric ﬁelds. The close agreement
between the theoretically predicted non—ideal behavior and the experimentally obtained
data is further evidence supporting the hypothesis of non-linear acceleration.
Furthermore, the minimal difference between the ion arrival times for the non-ideal and
ideal conﬁgurations suggests that the non-linear arrangement is relatively insensitive to

voltage ﬂuctuations, minimizing the need for power supply voltage stability.

109

 

 

W - 21.05

~~ 20.95

.. 20.85

 

 

.. 20.75

1
ll

 

 

 

 

4- 20.65
«~ 20.55
Ideal
—I— +3 V
+ -3 V
. . . . 20.45
2.5 2 1.5 1 0.5 0

lonlzatlon Position, s (cm)

Figure 4-21. Effect of two i3 V perturbations to electrode 6 on time-of-ﬂight for ions
originating ﬁ'om ionization positions 6—9.

110

Attempts at Evaluating A Grid-free Non-linear Conﬁguration

While the multi-gridded ion source provides strong evidence supporting the
requirement of non-linear acceleration for complete space focusing, the series of linear
electric ﬁelds created within the ion source only approximate a truly non-linear potential
gradient. An ideal non-linear ion source conﬁguration requires that the potential proﬁle
and the electric ﬁeld within the ionization region vary smoothly as a function of axial
location, and the only way to achieve this conﬁguration is with a grid-free, open ion

source geometry.

In an attempt to evaluate such a conﬁguration, the ion source described above was
modiﬁed by removing the grid material attached to ionization region electrodes 1—8. The
electrodes spanning the acceleration region remained gridded as before. The non-linear
electric ﬁeld required for this conﬁguration was determined using the TOFSORC3
program, where the linear ﬁeld segments calculated by the program were 0.0001 cm in

length; signiﬁcantly small in relation to the overall ion source axial geometry.

The elimination of the electrode grids poses a severe constraint to achieving the
appropriate electric ﬁeld potential proﬁle due to the non-homogeneous nature of the
desired non-linear ﬁeld. The manifestation of this non-homogeneous nature is illustrated
in Figure 4-22. Here, the electric ﬁeld proﬁle, represented in a topographical format as
equipotential contour lines, is plotted over a cross-sectional outline of the ion source. This
proﬁle was generated using the Simion[19] ion optical modeling program, which
calculates electric ﬁeld gradients and models predicted ion trajectories in a given
electrode geometry. The voltages applied to the ionization region electrodes in the Simion

optical model were those predicted by the TOFSORC3 program. As shown in the ﬁgure,

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Electrode/i; / / / / R
g > > > > > g > >
N v N Ix '-
§8%§§§§§ §
i / Grid
I I I I IK
Equipotential \xRequired Equipotential
Line \ Axial Position

 

 

 

 

 

 

 

I

Figure 4-22. Equipotential surface plot of electrode voltages applied to the grid-free ion
source. Each equipotential surface exhibits cylindrical axial symmetry, and each is

located farther back along the optical axis than desired.

 

 

 

 

 

 

 

112

the equipotential electric ﬁeld lines corresponding to the electrode voltages are not
strictly orthogonal to the ion source axis, and exhibit a curvature where the axial location
of the applied potential is located farther from the ion source exit than required, and the
potential at the axial position corresponding to the electrode coordinate is lower than

expected.

In order to achieve the required non-linear potential proﬁle along the longitudinal axis
of the ion source, the voltages applied to the ionization region electrodes must be
signiﬁcantly different from the predicted voltages in order to compensate for the electric
ﬁeld inhomogeneity and redistribute the potential gradient along the optical axis. A graph
of the ionization electrode voltage proﬁle required for the instrumental conditions used in
these experiments is plotted in Figure 4-23. As shown, the difference between the
required applied electrode voltage and that required along the optical axis is the greatest
at the position corresponding to electrode 1, where the applied potential of 3940 V is over
two times greater than that of the desired axial potential. This condition places an electric
ﬁeld strength of ~20kV/cm between electrode 1 and electrode 0 (held at 1600 V), which
is far greater than the insulating capacity of the ion source can withstand. Consequently,
every attempt at implementing this grid-free non-linear acceleration conﬁguration, even
under conditions where the overall accelerating potential was greatly reduced, resulted in
arcing between the electrodes and the formation of electrically conducting traces on the
insulating ceramics. For this reason, no successful evaluation of the grid-free design was

accomplished using this ion source.

113

 

4000

0 3500

   
  
    
 

Applied electrode
potential K

e 3000

e 2500

Required axial

potential .. 2000

 

Potential (Volts)

« 1500

e 500

 

 

 

115 3 215 2 ‘L5 1 (L5 0

Ion Source Position (cm)

Figure 4-23. Plot of axial potential and the theoretical electrode potentials required to
achieve the optimal axial potential proﬁle for the grid—free non-linear ion source

conﬁguration. Electrode potentials were determined using the Simion ion optical
modeling program.

114

A second important consideration in the design of a non-linear grid-free ion source
relates to the axial proﬁle of the equipotential surfaces. Any deviation of these surfaces
from axial orthogonality results in an axially non-parallel ion trajectory, with the
consequent axial dispersion of ions within the ion source reducing ion transmission
through the ion source and ultimately, signiﬁcantly reducing the number of ions reaching

the detector. A suitable ion source design must minimize this cause of ion dispersion.

A plot of the equipotential surfaces theoretically present with the ion source
conﬁgured with the electrode potentials illustrated in Figure 4-23 is shown in Figure 4-
24. In comparison with the electric ﬁeld plot illustrated in Figure 4-22, the application of
the calculated electrode potentials signiﬁcantly reduced the difference between the
required axial potentials and those present along the various optical axis positions.
However, this increased accuracy was accompanied by an increase in curvature of the
equipotential surfaces. This condition is most pronounced near electrode 1, where the

electric ﬁeld gradient is the steepest.

In ion source design, there is a tradeoff between equipotential surface curvature and
the problem of sustainable electric ﬁeld gradients described above. If the inner diameter
of electrode 1 were reduced, the voltage required to provide the appropriate axial
potentials would also have been reduced. In principle, the electrode 1 inner diameter
could be reduced to the 0.50” diameter of the ion optical volume, which in this case is
determined by the inner diameter of the accelerating region electrodes. However, the
result would a signiﬁcant increase in the equipotential surface curvature and a decrease in

ion transmission. Ion trajectory modeling indicated that for the mass spectrometer

115

 

 

 

 

 

Figure 4-24. Equipotential surface plot of ion source potentials where electrode voltages
have been adjusted to achieve a nearly ideal axial potential proﬁle.

116

conﬁguration used in these experiments, suitable ion transmission could only be achieved

using a relatively large ionization region diameter.

Summary and Signiﬁcance

The dependence of ion time-of-ﬂight on initial ion position when using linear ion
acceleration has been documented by experimental data using a two-ﬁeld ion source
model ﬁrst described by Wiley and McLaren. The experimental results, along with the
theoretical analysis, validate their claims that linear ﬁelds provide a ﬁrst-order
approximation of space focusing. Theoretical calculations and experimental data also
indicate that for static ﬁeld conﬁgurations, the independence of time-of-ﬂight from initial
ion position requires the use of a non-linear acceleration ﬁeld. This may occur either
exclusively or in concert with a subsequent linear ﬁeld, so long as the non-linear ﬁeld is
applied over the region spanning the axial range of initial ion positions. The concomitant
use of non-linear and linear ﬁelds combines the best features of the completely parabolic
non-linear model and the Wiley/McLaren two-ﬁeld linear model: the non-linear ﬁeld
provides complete space focusing, and the two ﬁelds allow the space focus plane to be
located beyond the end of the ion source. This ﬂexibility cannot be achieved using either

strictly non-linear accelerating ﬁelds or two linear ﬁelds alone.

While the achievement of a completely non-linear potential proﬁle is problematic, it is
possible to achieve an adequate approximation of a non-linear ﬁeld using a series of short
linear ﬁelds, so long as those ﬁelds are relatively close in relation to the overall axial ion
source dimension. When compared with the two-ﬁeld linear conﬁguration, the data show
that this ion source provides a signiﬁcant improvement to space focusing across the entire

ionization region with a slight temporal distortion near the ion source exit where the

117

electric ﬁeld gradient is the steepest. In normal operation this region is not utilized for ion
generation or capture, and this distortion will be of little or no consequence in the

exploitation of non-linear acceleration schemes.

The signiﬁcance of space focusing using the ion source conﬁguration described in this
chapter is illustrated in Figure 4-25. Here, the raw ion abundance information obtained
from sampling each ionization position under conditions of linear and non-linear
acceleration, displayed previously in Figure 4—14 and Figure 4-20 respectively, is shown
summed. The integrated ion arrival time proﬁle obtained using linear acceleration, shown
in Figure 4-25a, displays a resolving power of 50, while the proﬁle obtained using non-
linear acceleration, shown in Figure 4-25b, displays a greater resolving power of 180.
Additionally, there is a concomitant ~3.5x greater signal to background ratio using non-
linear acceleration. Thus, for signiﬁcantly long axial ion extraction volumes, the use of
non-linear acceleration yields a signiﬁcant improvement to resolving power and
detectability over linear acceleration conﬁgurations. For TOFMS conﬁgurations
employing energy focusing devices such as the ion mirror in addition to non-linear

acceleration, an even greater gain in resolving power would be expected.

118

 

 

0.6

 

 

 

 

 

 

 

 

 

 

 

a) b) “Fe‘
0.5 0.5
0.4 0.4
2‘ f.‘
O
2, a
g 0.3 E 0.3
m m
C C
9 9
E .E
0.2 0.2
56Fe+
0.1 0.1
or v 0 b
20.5 20.7 20.9 21.1 2L3 21.5 20.2 20.4 20.6 20.8 21 21.2
time-of-ﬂight (11sec) time-of—ﬂight (nsec)

Figure 4-25. Comparison of resolving power obtained using linear (a), and non-linear
(b), ion acceleration.

10.

11.

12.

13.

14.

15.

16.
17.

18.

19.

119

References

Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150.
Muga, M. L. Anal. Instrum. 1987, I6, 31.

Yefchak, G. E.; Enke, C. G.; Holland, J. F. Int. J. Mass Spectrom. Ion Phys. 1989,
87, 313.

Kinsel, G. R.; Johnson, M. V. Int. J. Mass Spectrom. Ion Phys. 1989, 91 , 157.
Glashchenko, V. P.; Semkin, N. D. Sov. Phys. Tech. Phys. 1987, 32, 665.

Grix, R.; Kutscher, R.; Ki, G.; Grtiner, U.; Wollnik, H. Rapid Comm. Mass
Spectrom. 1988, 2, 83.

Bandura, D. R.; Makarov, A. A. Int. J. Mass Spectrom. Ion Phys. 1993, 127, 45.
Steadman, J .; Syage, J. A. Rev. Sci. Instrum. 1993, 64, 3094.

Vlasak, P. R.; Beussman, D. J .; Ji, Q.; Enke, C. G. J. Am. Soc. Mass Spec. 1996, 7,
1002.

Flory, C. A.; Taber, R. C.; Yefchak, C. E. Int. J. Mass Spectrom. Ion Phys. 1996,
152, 177.

Stein, R. Int. J. Mass Spectrom. Ion Phys. 1974, I4, 205.

Rockwood, A. L. In Proc. of 34“ ASMS Conf 0n Mass Spec. and Allied Topics,
Cincinati, OH, June 8—13, 1986, 173.

Rockwood, A. L.; Udsetlr, H. R.; Gao, Q.; Srrrith, R. D. In Proc. of 42"” ASMS Con]?
0n Mass Spec. and Allied Topics, Chicago, IL, May 29—June 3, 1994, 1038.

Hulett, L. D., In; Donohue, D. L.; Lewis, T. A. Rev. Sci. Instrum. 1991, 62, 2131.

Flory, C. A.; Taber, R. C.; Yefchak, C. E. Int. J. Mass Spectrom. Ion Phys. 1996,
152, 169.

Mathsoﬁ, Inc., Cambridge, Massachusetts, 02142.
Norwood, K.; Ali, A.; Flesch, G. D.; Ng, C. Y. J. Am. Chem. Soc. 1990, 112, 7502.

Banares, L.; Baumert, T.; Bergt, M.; Kiefer, B.; Gerber, G. Chem. Phys. Lett. 1997,
267, 141.

Idaho National Engineering Laboratory, Idaho Falls, Idaho, 83451.

CHAPTER 5

PROSPECTS FOR FUTURE IMPROVEMENTS TO
RESOLVING POWER

For many established and developing GC-MS and LC-MS techniques, the task of
coupling the exit end of the chromatographic column to the mass spectrometer represents
the most difﬁcult obstacle to the successful implementation of the combined technique. In
each case, the emerging analyte greatly expands as it enters the vacuum environment,
placing stringent requirements on the pumping required to maintain the vacuum, and
complicating any attempt to analyze the entire eluent volume. Time-of-ﬂight ion source
conﬁgurations requiring an axially narrow initial spatial distribution are thus ultimately
limited in their application, and the reduction or elimination of the narrow spatial
requirement should yield an increase in applicability and detection efﬁciency. The use of
non-linear ion acceleration stands as a viable solution addressing TOFMS instrumental
conﬁgurations where the initial spatial distribution represents the primary limitation to
achievable resolving power. Furthermore, in instrumental conﬁgurations employing
capture of previously ionized analytes, as is the case with electrospray and atmospheric
pressure ionization, the axial dimension at the moment of ion capture can introduce
energetic variations which are amenable to the velocity correcting realized by non-linear
ﬁeld acceleration. Also, the recent popularity of attempts to miniaturize time-of-ﬂight
analyzers places added emphasis on space focusing as the ratio of the ion source
dimension to the total ﬂight path is often increased, exacerbating the spatial distribution

problem.

120

121

While the limitations of initial energy spread and tum-around time are not addressed
in this work, correcting devices already addressed in the academic literature could be
employed to provide additional enhancements to resolving power. It is anticipated that
the combination of the dual non-linear/linear ion acceleration along with impulse-ﬁeld
focusing and non-linear ion mirror ﬁelds may provide complete correction for each of the
fundamental limitations to mass resolving power. For future applications, it is clear that
the implementation of non-linear acceleration will extend the ultimate performance of

TOFMS instrumentation.

Appendix A

122

J. Agric Food Chem 1007. 45, 1001-1007

Jun Song,1 Ben D. Gardner,t John F. Holland) and Randolph M. Beaudry”

Department of Horticulture, Department of Chemistry, and Department of Biochemistry,
Michigan State University,East Lansing, Michigan48824

1N1

Rapid Analysis of Volatile Flavor Compounds in Apple Fruit Using
SPME and GC/Time-of-Flight Mass Spectrometry

 

Solid-phase microextraction (SPME), time-compressed chromatography (TCC), and time-of-ﬂight
mass spectrometry (TOFMS) were examined for their suitability and compatibility for rapid sampling,
separation, and detection of apple ﬂavor volatiles. Flavor-contributing volatile compounds were
found to have relatively high partition coeﬁcients on a 100 um thick coating of polydimethylsiloxane
(PDMS) on a SPME ﬁber. The time required to saturate the PDMS coating was highly volatile-
dependent, varying from less than 2 min to greater than 30 min. However, the response of this
system was linear rn the ppb to ppm range when the adsorption duration was standardized. 'lhe
speed of the TOF mass ter permitted identiﬁcation and quantiﬁcation of compounds having
chromatographic peek widths of only a fraction of a second. 111s unskewed nature of fragmentation
patterns obtained allowed individual component spectral characterization of unknown compounds
even when not fully chromatographically separated. Thus, the time required for chromatography
couldbe reduced by anorderofmagnitude without loss inanalytical performanceﬁ)’ Typical analysis
ﬁmesforcomplexmixhrreswueZ—ﬁminascomparedwichO-BOminrequiredfor stendardpurge-

and-trap analyses.

 

Keywords: SPME (solid- phase microatmction), time-command

massspectmmetiy; hcodspacesompling;

INTRODUCTION

Aromaisoneofthemostimportantqualitycriteria
of fruit and vegetable products, and a considerable
amount of research has been devoted to developing
analytical techniques for characterizing aroma-produc-
ing compounds. Both qualitative and quantitative
infomationisdesiredinordertomonitorproduceﬁavor
quality and ripeness and to provide quality control for
fresh and processed products. Progress in aroma re-
search is impeded, however, by slow sampling, separa-
tion, and detection methodologies. Fruit aroma re-
searchissubjecttoanumberofconstraintsthatmake
morerapidanalysisdeeirahle. Theseasonalnatureof
beworkedoutquicklyandeﬁcientlyinordertomeet
research‘timewindows’esteblishedbythecommodity.
multiple emerimental treeunente as part of a statisti-
cally relevant design, so sample numbers for most
experiments are relatively large (tens to hundreds), and
ottenanentirestudymustbecompletedinalimited
time. In M aroma biosynthesis studies, especially
those evaluating the response to metabolic precursor
hedingaromaproducﬁrmistbnamicwithahorthinute)
and long-term (hour) temporal compute. As a result,
sampling mustberepetitive and spanawiderengeof
time intervals. Furthermore, an analytical system
should be suﬁciently ﬂexible to work with both ﬂow-
throughandstaticsystemshavingabroadrangeof
components. Volatile aroma compounds are typically
analysedusingeitherdirectheadspace (Neubeller and

 

" Author to whom correspondence should be ad-
dressed.

' Department of Horticulture.

* Department of Chemistry.

1 Department of Biochemistry.

Wi-OSBIW-O WC: $14.“)

partition coeﬁidatcpmge-and-tmp; volatiles,

_ y; trim-afﬁrm
° :13?pr

Buchloh, 1978) or dynamic headspace purgeend- -trap
methods (Streif, 1981; Werkhoﬂ' and Bretschneider,
1997).Whiledirectheadspeceana1yeisislargelycon-
ﬁned to higher concentration samples, purge-and-trap
sampling can enhance sensitivity by enriching volatile
components on a polymer bed. The dynamic headepace
purge-and-trap method has been thoroughly standard-
ised by Buttery at at. (1989), Mattheis ct cl. (1991), and
SongandBangerth(1993)forroutineanalysisoftomato
and apple fruit volatiles. However, this technique rs
mums-consuming, and prone to methodological
diﬁiculties. During analysis, the sample preparation
step for purge-and-trap analyses rs frequently the most
time-consumingendlaborintensivestep. Itisalsothe
primary point of analyte loss from the matrix.

An alternate sampling methodology, solid-phase mi-
aoextrection(SPME),hasthepotentialtoreducethe
time investment in sampling and should work well in
combination with rapid separation and detection sys-
tems. SPMEhssbeenappliedtotheanalysisofvolatile
and nonvolatile compounds (Arthur and Pawlissyn,
1990)ingeseousand liquid samples andtoanslyse
flavorin fruitjuice beverages, in vegetableoils (Yang
and Peppard, 1994), and in orange juice (Steﬁ'en and
Pawliszyn, 1996). SPMEhasbeenehowntobeasimple
and eﬁ'ective sampling method providing a linear re-
sponse to concentrations covering four orders of mag-
nitude (Arthur ct aL, 1992; Louch et aL, 1992). It is
also rapid, providing complete extraction and transfer
of volatile organic compounds in under ﬁve min (Gard-
ner st (11., 1995).

The recent development of a gas chromatographic
detectoruﬁlisingTOFMSwithtimearraydetection
(TAD)hastheanalyeisspeedneededtotakeadvantage
of the rapidity of SPME sampling technology. This
instrument system additionally enables the use of time-
compressed chromatography ('I‘CC) whereby temporally
unresolved components can be characterised by their

ormmcrmm

123

11102 J. Am'c. FoodChan, Vol.45, No. 5, 1997

unique mass spectra. TCC permits a signiﬁcant reduc-
tion in the time invested in separation. To date, SPME
in conjunction with TCC and TOFMS for analysis of
ﬂavor volatiles in horticultural produce has not been
evaluated. In this paper, we report on SPME sampling
characteristics (speed, variability, linearity, saturation
kinetics, and matrix eﬂ'ects) in combination with rapid
GC techniques (short, narrow bore columns, high linear
ﬂow rates, and high temperature ramp rate) and a TOF
mass spectrometer for detection as a system for rapid
and quantitative analysis of aroma volatiles in apples.

EXPERIMENTAL PROCEDURES

Volatile Aroma Standards Preparation Volatile aroma
compoundswcrepurchasedfromSigmaCo. andF'lukaChemi-
cal Corp. Standard mixtures were prepared with 0.3-1.0 pL
of butyl acetate, ethyl 2-methyl butanoate, hexyl acetate,
l-butsnol, l-hexanol, and 8-methyl-6-heptcn-2-cne in specially-
built gas-tight 4.4-L glass volumetric ﬂasks ﬁtted with a
tapered ground glass stopper containing a gas-ﬁght Mininert
valve (Alltedr Assoc, Inc., Deerﬁeld, IL). The mixed standards
samplewasappliedtoasmallpaperﬁlterdiskinthenettof
theﬂaskanddroppedtotheﬂaskbottom. 'l‘heﬂaakwas
sealed,andtheliquidmaterialwas allowedtovaporiseto
provideknownheadspaceconcentrationsrangingfrotho 10

uLL“.

”LEW. ASPME(Supelco Co.,Belle-
fonts, PA) ﬁber coated with polydimethylsiloxane (PDMS, I
cmlong,100;¢mthicknees)wasusedtocollectandconcentrate
volatiles by virtue of its sorption characteristics (Arthur and
Pawliszyn, I990). The SPME device consisted of a retractable
ﬁberencloeedinasheath. Duringsampling,afterentering
the sample curtainer, the ﬁber was extended, exposing the
sorpticnsurface. 'I‘heﬁbcrwasthcnreu‘actedpriorteremoval
ﬁom the sample container. The ﬁber was preconditioned at
260'Cfor1-2h. Varioussampleexposureﬁmeswereused
foradsorpﬁomandthevelatilesweredcsorbedfromtheﬁber
for90sat250°Cintotheglass-lined,splitlcssiniectorportof
aGC(HP-6890, Hewlett Packard Co.). AllSPME samplings
werecaniedoutat23°Cintrip1icateunlcssotherwisenoted

SPIESotw-otionﬁnctics. SPMEﬁberewereexposedtc
thebeadspaeeof4.4-Lﬂaskscontainingamixtrrreofbutyl
acetate and beryl acetate at a concentration of approximately
6pLL". Fiberswereheldintheatmosphcreoftheﬂaskfor
speciﬁclengthsofﬁmerangingﬁ'om6to480s. Analyseewere
carriedoutusingaGCwithaﬂameionisaﬁcndctectorWID).

Todeterminesaturationkinetiaincomplexmixturesof
volatiles, apple fruit (Moles domestics Borkh. cv. Golden
Delidous)storedbr2-3monthsat0°0wcrewarmedtoroom
temperature (23 'C). Intact apple fruit (300-460 g) were
placedinaS-Lﬂaskwithpuriﬁedairpaseingthmghata
ﬂowrateof26-30mLmin". Samplingofvolatilecompounds
wasaccompliahsdbyplaeingtheSPMEﬁberintoagla-tee,
ﬁttedwitha'l‘dcn-linedhalﬂioleseptummﬂtedrAssoclm,
Deerﬁeld,na)locatedattheoutletcfthcﬂaak. Allthe

' gash'neswcsecomposcdof‘l‘eﬂon.

Apple fruit volatiles readied steady-state concentration in
theﬂukwithht-GhaﬂcrwhichtimeSPMEﬁberswere
caposedtothevolatilesemanaﬁngfromtheﬂaskforvarying
lengthsoftimeupteﬂmin. Itesponsedeterminatiomwere
madethreeﬁmeeforeachexposureperiod. Whilenumcrous
volatileswerehackeddataarereportedforbutylacetate,
butanol, beryl acetate, hexyl 2-methyl butanoate and a-far-
nesenauthasecompormdsappearedteexhibitthefullrange

m

cfresponses
SPMEResponsaLirwat-ity. Agasmixturecontainingbutyl
acetate, ethyl 2-methyl butanoate, and hexyl acetate was
forthe standards described above. A stream of
volaﬁlee-ﬁeerritrogengas(I-2mLL")waspassedthr-ough
theﬂasktodilutethcvolaﬁlceandtogenerateawiderange
of volatile concentrations. The ﬂask was periodically sealed,
andtheconcentrationofthevolatilesintheﬂaskwas
determined by comparison to the gas standards described
previously. Aloo-pLgassamplewasremovedbyusingagas-

Songstal.

tight syringe (Hamilton No. 1810 with stainless steel needle)
and directly iniecbd to a gas chromatograph (Carla AGC
Series 400) equipped with a FID, which was used to measure
volatile vapor concentrations. The GC was equipped with a
packed column (10% DEGS-PS, 80/100 mesh Supelcoport, 3.3
m long, 3 mm id.) and was maintained at 140 °C. Helium was
usedasthecarriergasataﬂowrateonOmLmin“. Some
volatiles, such as hexyl acetate and Somethyl-6-hepten-2-cne,
were found to adsorb to the steel needle (data not shown), and
care had to be taken in optimizing the sampling procedure.
Pumping the syringe plunger 16—20 times (which apparently
saturated the adsorption sites inside the needle) gave a
mnsistent GC response, such that the standard’s response had
a coefﬁcient of variation of only 2—6%. By this means, any
eﬁ'ect of needle adsorption was removed from GC response.

Alter establishing the gas concentrations in the dilution
ﬂask, the TOFMS response was determined using the SPME
ﬁber with a 6-min sampling time (see below). TOFMS
response was correlated with volatile concentration.

SPME Partition Coefficient. The partition coeﬁcient (K) of
the SPME ﬁber coating was determined for standard mixtures
(6- 10 pL L“) of butyl acetate, ethyl 2-methvl butanoate, hexyl
acetate, butanol, hexanol, and 6-methylo6-hepten-2-onc. The
SPME ﬁber was exposed to the volatiles for sufﬁcient time
(8-8 min) to allow the coating to reach equilibrium with the
ﬂaskheadspace. Kwasdeterminedueingthefollowing
equation (Zhang and Pawliszyn, 1993):

K = (Arvoymovr)

whereApandAo are, respectively, the peak areas from the
GCresponsetotheﬁber-coatinganddirectgasinjecﬁcn, V,
is the volume of the coating (calculated to be 8 x 10“ cm’),
and V6 is the volume of the gas injected (100 pL). Five
measurements of K were made per volatile.

Purge-and-Trup Procedure. Purge-and-trap analysis was
performed for comparison to the SPME technique. Tenax TA
(100mg,60/80mesh,Alltech)was p1acedinglasstubee(6cm
long,4mmid.)withglasewool plugeatbothcndsoftheTcnax.
Applearomavolaﬁleewerecollectedbyinsertinga'l‘enaxglass
tube into the gas outlet of the above-mentioned sample jars
crmtainingappleﬁ-uit. Volatileewereadsorbedontethe'l‘enax
forlOminataﬂcwrateoflOOmLmin". Thetubeswcre
desorbed using a specially-built desorption unit heated to 260
'Cfor lOminwhilebeingpurgedwithulta-apuriﬁedhelium
(99.999%) at a ﬂow rate of 36 mL min". A short 0.2 m long
x 0.63 mm id. deactivated fused silica precolumn was sueched
totheinjecﬁonportofaVarianStOOWarianAnalytical). A
specially-builtcrycfocusingunitwasusedtocoolthepre-
columntoleasthan-lOO’Cduringtheheatdesorpticn
process. After crycfocusing, the aroma compounds were then
volatilizedbyheatingthecryofowsingunitto 160 ‘Cinafew
seconds.

SeparationandDetectlcn. SPlE/GC/NFMS oprpls
Volatiles. Adsorbedvolatileeweredcsorbedﬁ'omthcﬁberinto
the HP-8890 GC as described previously. Volatiles were
separated usingacapillary column (HP-6, 6 m x 0.1 mm id.,
0.34pmcoatingthicknees). 'I‘hecarriergaewasultrapuriﬁed
helium (99.9995) at a ﬂow rate of 0.6 mL min". The
temperatureprogramwas isothermalfor 1.6 minat40°Cand
thenraisedattherateof60 ’C min“th60°C, andhcldfor
2min. 'I‘thC/MStransfcrlinctemper-aturewas220'0.
Volatile detection was performed by TOFMS using electron
impact ionization (POD-660, LECO Corp., St. Joseph, MI).
MassepecbawerecollectedatarateoftOspectr-a/sovera
range of m/z 40-300. 'Ihe ionization energy was 70 017.

Data were analysed using LEGO deconvolution software.
Identiﬁcation of volatile components was conﬁrmed by com-
parison of collected mass spectra with those of authenticated
standardsandspectraintheNaﬁonal InstimteforStandsrds
andTechnology(NIST) massspectrallibrary, SearchVeraion
1.1.

Plugs-and-ﬁup/GC of Apple Volatiles. Volatiles were
desorbedfromtheTenaxasdeeaibedaboveandseparated
usingacapillarycolumn(HP-6,46mlong x 0.26mm1d., 0.26
pm thick coating). 'Ihe temperature program was isothermal

124

Analyeis ct Flavor Coupounds Using SPME and GC/TOFMS

analnmal l lLAAlAAA

-§— Baylscetste
1.860074 {- W

lxnxlxxxj.xa

 

     
    

 

 

 

"i"'é"'i"'i'ﬁiﬁ'1
Adsorption'l'lme(min)

Hgnrel. GCIF'IDresponseforamixtmeofbutylaceteteand

hexylacetateasaﬂ'ectedbyadsorptionﬁmeonaSPMEﬁber.

 

r
I
f
i
I

”-1
U

 

‘.4E401 migiiii.ninrr.1.mai.i.m
1
Hexylscetate
128*0'F {-

 

-§-Bulylsceus - 4.

Y fr r

1 .EOO'l-l
8.064%-

0.“-

FIO Response

4.0905. ,

c)

2.0E+M~

 

 

0.084-(Xlr.T
0 2

:7

I'I

3's1b1'21'41's1'52'02'22'4
AdsorptionTlme(mln)

Figurel. GC/FHJresponseforbutylaoetateandhexyl

acetateinacomplexmixtureofapplevolatilesasaﬁectsdby

adsorptiontimsonaSPMEﬁber.

for6minat40‘Candthsnraised8‘Cmin"t0260'Cand
heldfor16min. Detecn'onwasbyF'IDat300'C.

RESULTS

Sampling. SPME Saturation ﬁnetics. 'lhe equili-
bration times for butyl acetate and hexyl acetate
standards using the PDMS coating were approximately
2and6min,respecn'vely(Figure 1). Thetimesrequired
forbutylandhexylacetatetoreachequilibriumwhen
inacomplexmixturefrom ripeapplefruitwereap-
proximately4and12min,respectively(Figure2).
Othercompoundsinthemixturehavingrapidadsorp-
tion kinetics included 2-methylbutyl acetate and l-bu-
tanol (Figure 3). However, hexyl 2-methyl butanoete
anda-farneeenedidnotreachequilibriuminﬂmin
(Figure4). TheGCresponseasshowninFigure3for
2omethylhutyl acetate in the complex mixture declined
astheadsorptiontimeincreasedbeyond 12min. The
nearly linearly increasing response for most compounds
appearedtobeintheﬁrst 1—4min.

81’“ Response Linearity. The response ofSPME
ﬁber to volatile concentration was linear for the three
compounds tested after 6-min adsorption (Figure 6). For
TOFMS detecn'on at 10x background noise, the con-
centrations of butyl acetate, ethyl 2-methyl butanoate,
and hexyl acetate extrapolated to 0.008, 0.0012, and
0. 004 uL L‘ ,reepectively.

SPME Partition Coefficients. The PDMS coating/air
partition coeﬁcients of SPME ﬁbers for tested aroma
impact apple volatiles ranged 33-fold from 1-butanol

l
' E 3.0E+07~

J. Am’c. FoodChem. Vol. 45. No. 5. 1997 1am

.1_ajxl 1 L41 1 11111 1-1

-§- 2mm
—§— 1am

 

1.65406-

1.ZE+06« ,-

8.0E+05~

FID Response

4.06‘054 )-

 

 

0.08m1ﬁjﬁ F ' T‘j *T V l ' Y ‘ l V I ' Y ' I ' l
o 2 4 s 01012141618202224
MeorptlonTlme(min)

 

    
 

Figures. GC/Fﬂ)responseof20methylbutylacstateand
bumun‘mmixtmeofapplevolatilesasaﬁectedby
adsorpﬁonﬁmeonaBPMEﬁher.
5.0907 A1414+Apeehhiihiud
{who-Fm 1
“)an {—Hsayl-Z-msmylm

9 2.050074

ll.

1 .OE4-O7-J

 

 

 

o 2 i'8'i'1b'1'2'1'4'1's'1'e'io'52'2'4
Adsorptionﬂmemln)
Figured. GCIFIDresponseofa-farneseneandhexylz-
methylbutanoateinacomplexmixtureofapplevolatileeas
aﬁ‘ectedbyadsorpﬁantimeonaSPMEﬁber.

 

“£006 A A A 1 A A A ‘ A A A 1 A A k
6 Mac“ .
5.0E+06- I EﬂM-z-mslnytbinnoats

literals“ ’

 

 

 

 

1"'iﬁ'§"
VolatileConcentraﬂonmpm)

Figu-e6. Eﬁ’ectofvolatile concentrationonthe

linearityofGC/Fmresponseforbutylacetete,ethyl2-

laiutanoateégaghexylacetateadsorbedontoaSPMEﬁber
minat .

with 1.7 x 10‘ to 6-methyl-6-hepeten-2-one with 6.6 x
10‘ (Table 1). The dominant inﬂuence appears to be
the number of carbons in these linear molecules.
SPME Response Variability. With the exception of
butyl acetate, the calculated standard errors in these
testsresultedincoeﬁcientsofvariationrangingfrom
2to 10%. Butyl acetatevariedheyondthisrange inthe
linearity tests at concentrations above 3 pL L“.
Separation. SPME/GO of Apple Volatiles. Using
rapid chromatography techniques, apple volatiles con-

125

m4 J. Agic. Food Chem, Vol. 45. No. 5. 1997

Table 1. Experimental Partition Coefﬁcient (log K)
ValuesbetweenCoatlngandGasPhaseofSome
Volatiles in Apples

 

 

compound partition coeff log K
butyl acetate 2.7 x 10‘ 4.44
ethyl 2omethy1butanoate 3.36 x 105 6.61
hexyl acetate 2.3 x 10‘ 6.36
l-butanol 1.7 x 10‘ 4.21
l-hexanol 2.1 x 10‘ 6.32

6-methy1-6-hepten-2-one 6.66 x 10‘ 6.76

Tablet VolatileCompoundslnGoldenDelidousApple
Fruit. SampledwlthBPMEG’DMS, 100pm)and
Identiﬁedbymms

 

retention time (min)

 

 

peak volatile compd SPME purge-and-trap
1 pentane 1.082
2 acetone 1.132
3 l-butanol 1.322
4 propyl acetate 1.494
6 propyl propanoate 1.624
6 butyl acetate 1.619 13.46
7 ethyl 2-methylbutanoate 1.639 13.80
8 2-methylbutyl acetate 1.763 16.80
9 propyl butanoate 1.802
10 butyl propyrate 1.932
11 pentyl acetate 2. 006
12 butyl 2-methylbutanoate 2.106
13 butyl butanoate 2.164
14 hexyl acetate 2.396 21.34
16 butyl-24nethylbutanoate 2.437
16 pentyl butanoate 2.460
17 hexyl propyrate 2.492
18 propyl 2-methyl-2-butenoate 2.660
19 hexyl 2-methylpropyrate 2.866
20 not identiﬁed 2.893
21 hexyl butanoate 3.183
22 butyl hexanoate 3.202
23 4-methoayallylbensene 3.223
24 hexyl 2-methylbutanoate 3.264 32.76
26 2-methylbutyl hexanoate 3.396
26 hexyl pentanoate 3.633
27 2-methy1propyl 2—methyl- 3.764
butanoate
23 hexyl haxanoate 4.002 38.82
29 u-farnesane 4.463 41.76

taining1-16carbonscouldbee1utedinabout2min
(Table2,Figure6a). Inall,29compoundsotherthan
CO. were detected and identiﬁed in Golden Delicious
apple ﬁ'uit. Using the SPME, dominant volatiles by
response were in ' order as follows: u-farne-
sane, hexyl 2-methylbutanoate, hexyl hexanoate, hexyl
butanoate, butyl hexanoate, hexyl acetate, and 2-
methylbutyl acetate. There was a tendency for the
responsetoincreasewithlarger,morelipophilicand
nonpolar compounds. Propyl acetate, butyl acetate,
hexyl acetate, ethyl 2-methylbutanoate, 2-methylbutyl
acetate, hexyl butanoate, butyl hexanoate, and hexyl
hexanoate are dominant volatile compounds in apples
(Paillard, 1991). a-Farnesene and 4-methoxysllyl-
benzene were also identiﬁed.

Purge-and-Tmp/ GC of Apple Volatiles. The chroma-
tography of the purge-and-trap sampled volatiles took
inexcessof40min,whichwasroughly 10timeslonger
than for the rapid chromatography techniques used with
the SPME device. Five of the compounds separated
were identiﬁed by comparison of retention times to those
of authenticated compounds (Figure 6b).

Detection. SPME/GC/TOFMS of Apple Volatiles.
Detection and quantiﬁcation of apple volatiles from a
complexmatrixviaTOFMS usingTCC didnotrequire
baseline resolution. The rapid spectral acquisition rate

Songstal.

of40 spectra/s ofthe TOFMS permitted 40-80 spectra
to be collected over the typical 1-2-s peak widths
(Figure 7). Co-eluting compounds such as butyl hex-
anoate and hexyl butanoate diﬁ'ering by at least one m/z
could be identiﬁed and quantiﬁed. For instance, butyl
hexanoate and hexyl butanoate were successfully de-
convoluted using m/z 117 and m/z 89, respectively,
although elution times diﬁ'ered by only 0.2 3. Since
TOFMS generates unskewed spectra consisting of true
fragmentation relationships, complete spectral charac-
terization is readily accomplished for each analyte.
These compounds were quantiﬁed using the unique ions
found in the full spectra, which were identiﬁed either
in the NIST library or generated from authenticated
standards. The total time invested per analysis for
SPME/'lCC/TOFMS was approximately 10 min.

Purge-and-Trap IGC/FID of Apple Volatiles. Detec-
tion and quantiﬁcation of apple volatiles from a complex
matrixusinganFIDrequiredbaseline resolution. This
was obtained using a longer column and slower chro-
matographic techniques (Fisure 6b). The total time
invested per analysis for purge-and-trapltraditional GC/
FID was between 100 and 120 min.

DISCUSSION

Investigations ofaromasynthesisinbiological sys-
temsoftenrequireananalyticalsystemthatiscapeble
of both high throughput and adaptability to static and
ﬂow-through measurements of various volumes. To
date, much of the time inveetrnent in ﬂavor analyses
has been for sample collection and preparation. Sam-
pling methods for extracting ﬂavor compounds from
horticultural products include direct headspace (Neu-
beller and Bochloh, 1982), liquid-liquid extraction
(Weurman, 1969), supercritical extraction (Bundschuh
et aL, 1986), and dynamic headspace with adsorbents
such as charcoal (Streif, 1981) or Tenax (Dirinck et al.,
1989). However, the disadvantages of these methods
are the lengthy analysis time, which has signiﬁcantly
limited sample throughput and added to experimental
complexity.

Ideally, analytical methods for ﬂavor research on
hortimltural produce should be fast, inexpensive, sol-
ventlees, relatively independent of the instrument de-
sign, and amenable to automation. They should also
beapplicabletogaseousandliquidsampleeandhavea
largelineardynamicrangewhileretainingexcellent
detectionlimits. Furthermore,theyshouldbeableto
function as a screening technique or be used in the
quantitative analysis of selected volatile compounds or
classes of compounds tor monitoring physiological proc-
esses that can be directly related to biochemical changes.
SPME meets these criteria (Louch et al., 1992; Zhang
and Pawlisxyn, 1993; Buchhols and Pawlissyn, 1994).

The primary factors aﬁecting the linear range and
sensitivityofdetectionusingSPMEaretheﬁber’s
staticnaryphaseandthepropertiesoftheGCdetector.
Inourresearch,8PMEexhibitedgoodlinearityof
responseforvolatilesranginginconcentrationfromppb
to ppm with negligible eﬁ'ects from matrix varian'ons
(i.e., water vapor). Volatile concentrations in horticul-
tural produce vary from 0.01 to 10 ppm (Paillard, 1990).
Dominant volatiles like butyl acetate and hexyl acetate
can reach the 0.6-1 ppm level during fruit ripening
(Paillard, 1990),andhexanalintomatoescanbeashigh
as 3.1 ppm (Buttery et aL, 1989). This suggests that
SPMEsamplingshouldbesensitiveenoughtobeused
for We investigations of impact volatiles. For

126

Analysis 01 Flavor Compounds Using SPME and sorrows

 

 

 

J. Agric. Food Chem. Vol.45, No. 5:109? 1m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a _ 14 212224 29 SPME
.. (Zmlnsampllng)
33 1
C
o a
Q-
(0
o a:
m .1
m at
.2 -
E ‘ L
O
a: 1
' r l T ' 1 r 1 I T l
0 1 2 3 4 5
Time(min)
b 6 s 14
_ 29 T
g (waistliwlo)
O
Q
(I)
0)
a: ‘ 2e
0
.5 d
(0
7) _ i711 l. a
a: .r.,.'..,ﬁ .1.-.,...,
O 10 20 3O 40 50
Time(min)
Plgure6. QuaﬁuﬁvemmpaﬁmndchmmmgnmbrSWFMBMmﬁﬁondpurn-md-mpwhmehudw
anslysiswithGC/F'ID.
am +++‘*4‘ volatilesandtheirrespectivediﬂ’usioncoeﬁcients.
11.05405 —RT'° ZhengandPawliszyn(1993)reportedthatBTEX(ben—

O mm0m117)’

RslstlvsRssponss

 

 

 

ﬁentimesdiﬁ'erbyapproximatelyOZs.

8PME,thecapacityoftheﬁberisafunctionofthe
stationary-phase volume and adsorption coefﬁcient for
thevariousvolatiles. ThePDMScoatingusedinthis
shady favored adsorption of larger, more lipophilic
compounds. However, adsorption can be tailored to ﬁt
theneedsofaparticularresearchthrust. Forinstance,
itismosteﬁcienttousethickﬁlmsforcompoundswith
low partition coeﬁcients (Zhang and Pawlisxyn, 1993);
inourcase,thinﬁlmsareadequateforanalyteswith
high partition coeﬁcients such as 6-methyl-6-hepten-
2-one and u-farnesene. To date, however, there are not
adequate data regarding partition coeﬁcients of SPME
coatings for fruit volatiles.

The equilibration time ofthe SPME ﬁber plays an
important role in quantitative sampling and is aﬂ‘ected
bytheKvalueoftheSPMEcoatingforindividual

 

sane, toluene, ethyl benzene, and xylene isomers) re-
quire just 2-6 min to reach equilibration. In our
experiments using the mixed standards and apple
samples, 6-8 min were required for most volatiles.
Some compounds like hexyl 2-methylbutanoate and
a-farneeenetakemorethan 24mintoreachequilibri-
um. These observations suggest that the kinetic be-
havior of the adsorption process, which diﬁ'ers with the
nature ofeachanalyte,isafactorthatmustbeaceom-
modated in any experimental protocol. Some com-
pounds, such as butyl acetate and 2-methylbutyl ace-
tate, exhibited mauix eﬂ‘ects and are negatively inﬂu-
enced by other materials, which continue to increase in
concentran'on (e.g., a-farnesene and hexyl 2-methyl-
butanoate). This strongly suggests that a competition
forelimitedadsorptioncapacityexistsmakingit
advisable to avoid approaching a saturation conditiar
oftheﬁbercoatingmaterial. Sincetheresponsevaria—
timinsamplingisrelaﬁvelylow,the8PMEcanbeused
quantitatively if adsorption time is precisely maintained
and the adsorption has not approached satnration.
'I’rvo minutes were required for sampling via SPME,
generating a volatile proﬁle that is qualitatively com-
parable with data using purge-and-trap
this and earlier reports (Dirincket al., 1977; Kakiuchi,
1986; Song and Bangerth, 1993). However, it should
be pointed out that there may be marked quantitative
diﬂ'erences in the same compounds between SPME and
purgeand-trap sampling. For example, the higher
proportion of u-farnesene relative to other volatiles for
SPME as compared to purge-and-trap reported here
(Table 2) andelsewhere (KakiuchietaL, 1986; Songand
Bangerth, 1993) is most likely due to discrimination 1m
thebasisofpartitioncoeﬁcientsandalsoadsorption

127

not J. Apia Food Chem. Vol. 45. No. 5. 1997

kinetics. Calibration of the SPME ﬁber will be required
for accurate quantiﬁcation.

The rapid sampling of apple volatiles by SPME is
complemented by the features of the TOG/TOFMS
employed, which enabled analysis times to be reduced
from hours to minutes. While a narrow bore column
(0.1 mm i.d.) has been used to improve GC resolution
and speed (Sacks and Akard, 1994), co-elution was
unavoidable. However, the high spectral acquisition
rate of the TOF mass spectrometer and the data quality
is suﬁcient to permit quantitation of co-eluting com-
pounds with elution times differing by as little as 0.2 s.

SPME sampling is solventless and much faster than
purge-and-trap sampling. With the combination of fast
GC and the time-of-ﬂight mass spectrometer, the quan-
titative and qualitative analysis time can be shortened
at least 10-fold. The system should be appropriate for
real-time analysis of time-course studies on volatile
biosynthesis by horticultural produce. It is independent
of sample size, which opens the new possibilities for
biochemical investigations using small (1-2 mL) vials
of plant material.

CONCLUSIONS

The SPME system was found to be a convenient and
appropriate sampling technology for rapid qualitative
and quantitative analysis of volatile ﬁom horticultural
produce. Sampling time can be reduced approximately
60-fold relative to conventional purge-and—trap. The
primary limitation of the SPME system is that samples
cannot be conveniently stored.

Time-compressed chromatography was successfully
accomplished using a narrow bore (0.1 mm i.d.), short
(6-m) column, and rapid temperature ramp rate (60 °C/
min). Chromatographic time was reduced 10-fold rela-
ﬁve to conventional methods with good resolution.

Use ofTOFMS was found to be ofsuﬁcient sensitivity
and speed to permit rapid and eﬁcient detection of
apple aroma volatiles. The speed of the detector, based
on time array detection, makes it well-suited for ap-
plication to rapid GC separations with the resultant
time compression of peaks. Importantly, the rapid
spectral generation rates and unskewed nature of
spectra obtained ambled analytical resolution of chro-
matographically unseparated compounds. Because of
the independence from chromatographically induced
fragmentation pattern skewing, the resulting spectra
were ‘classic’ in nature and easily library searchable.
Additional speed is possible, since the mass spectrom-
eter was operating at approximately 10% of its maxi-
mum rate, raising the possibility of having detection
times of as little as 12 s for complex mixtures.

The analysis system described here appears to be
highly appropriate for aroma analyses. The speed of
the system overcomes the primary limitation of SPME
ﬁber (i.e., sample storage) and permits collection of
pertinent biological data in real time. Typical daily
sampling throughput was in the range of 10—20 samples
and could be increased to over 100. The factor now
limiting total analytical speed is the data handling (i.e.,
identiﬁcation, quantiﬁcation, and reporting of detected
compounds).

ABBREVIATIONS USED

FID, ﬂame ionization detector; m/z, mass to charge
ratio; PDMS, polydimethylsiloxane; RTIC, reconstructed
total ion current; SPME, solid—phase microextraction;

Song at al.
'I‘CC, time compressed chromatography; TOFMS, time.
of-ﬂight mass spectrometry.
LITERATURE CITED

Arthur, C. L.; Pawlisxyn, J. Solid-phase Microextradon with
'lhermalDesorptionusinthsedSilicaOpticalF'ibsraAnaL
Chem. 1990, 62, 2146-2148.

Arthur, C. L.; Killam, L. M.; Buchhols, K. D.; Pawliszyn, J.;
Berg, J. R. Automation and Optimization of Solid-phase
Microextracﬁon. AnaL Chem. 1992, 64, 1960-1”.

Buchholz, K. D.; Pawliszyn, J. Optimization of Solid-phase
Microextraction Conditions for Determination of Phenols.
Anal. Chem. 1994, 66, 160-167.

Bundschuh, E.; Tylla, M.; Baumsnn, G.; Giersdmer, K.
Gewinnung von natuerlichen Amman aus Reststoﬁ‘en der
Lebensmittelproduktion mit Hilfe der COrHochdruckex-
traktion. LebensmaWiss. Technol. 1986, 19, 493-496.

Buttery, R. G.; Ling, L. C.; Guadagni, D. G. Volatilitiee of
Aldehydes, Ketones, and Esters in Dilute Water Solution.
J. Agric. Food Chem. 1969, 17, 386-389.

Buttery, R. G.; Teranishi, R.; Flath, A.; Ling, L. C. Fresh
tomato volatiles: Composition and sensory studies. In
Flavor Chemistry: Trends and Developments; Teranishi, R,
Buttery, R. G., Shahidi, F., Eda; ACS Symposium Series
388; American Chemical Society: Washington, DC, 1989.

Dirinc'k, P. Schreyen, L, Schamp, N. Aroma Quality Evalu-
ation of Tomatoes Apples, and Strawberries. J. Agric. Food
Chem. 1977, 25, 769-762.

Dirinck, P.; De Pooter, H.; Schamp, N. Aroma Development
in Ripening Fruits. In Flavor Chemistry: Trends and
Developments; ACS Symposium Series 388; Teranishi, R.,
Buttery, R. G., Shahidi, F., Eda; American Chemical
Society: Washington. DC, 1989; pp 23-34.

Gardner, B. D.; Johnson, J. A.; Artaev, V. B.; Sparkman, O.
D.; Holland, J. F. Rapid Sampling for Rapid Analysis: A
New Method for VOCs in Drinking Water. Proceedings of
the 43th ASMS on Mass Spectrometry andAllied
Topics, Atlanta, GA, May 21-26, 1996.

Kakiuchi, N.; Moriguchi, S.; Fukuda, H.; lchimura, N.; Kato,
Y.; Banba, Y. Composition of Volatile Compounds of Apple
Fruits in Relation to Cultivate. J. Jpn. Soc. Her-tic. Sci.
1986, 56, 280-289.

Louch. D.; Motlagh, 8.; Pawlisxyn, J. Liquid-Coated Fused
Silica Fibers. Anal. Chem. 1992, 64, 1187 -1199.

Mattheis, J. P.; Fellman, J. K.; Chen, P. M.; Patterson, M. E.
Changes in Headspace Volatiles During Physiological De
velopment of Bisbee Delicious Apple Fruit. J. Agric. Food
Chem. 1991, 39, 1902-1906.

Neubeller, J .; Buchloh, G. Aromen und andere Fruchtin-
haltsstoffe verschiedener Erdbeereorten. Erwerbsobstbau
1978, 20, 189-192.

Nielsen, F.; Olsen, E.; Fredenslund. A. Henry's Law Constants
and Inﬁnite Dilution Activity Coeﬁcients for Volatile
Organic Compounds in Water by a Validated Batch Air
Stripping Method. Environ. Sci. Technol. 1994, 28, 2133-
2138.

Paillard, N. M. M. TheFlavoroprples, PearsandQuinces.
InFood Flavors. Part C: TheFlavoroth-uits; Morton, I.
D., MaCleod, A. J ., Eda; Elsevier Science Publishers B.V.:
Amsterdam, 1990; pp 1-42.

Robbins, G. A.; Wang, 8.; Stuart, J. D. Using the Static
Headspace Method to Detennine Henry’s Law Constants.
Anal. Chem. 19%, 66, 3113-3118.

Sacks, R.; Akard, M. High-speed GC analysis of VOCs:
Tunable Selectivity and Column Selection Environ. Sci.
Technol. 1994, 28, 428—433.

Song, J .; Bangerth, F. The Production and Changes of Volatile
Compounds from Apple Fruits with Different Fruit Matu-
rities. Acta Hortic. 1993, 368, 160-169.

Steﬂ‘en, A.; Pawliszyn, J. Analysis of Flavor Volatiles using
Headspace Solid-phase Microextraction. J. Agric Food
Chem. 1996, 44, 2187-2193.

Streif, J. Vereinfachte Methods sur echnellen gaschromatog-
raphischen Besﬁmmung von Fluechtigen Aromastoﬁ'en.
Gartenbauwissenschaﬁ 1981, 46, 72—76.

128

Analysis of Flavor Compounds Using SPME and GC/TOFMS

Werkhoﬂ’, P.; Bretschneider, W. Dynamic Headspace Gas
Chromatography: Concentration of Volatile Components
after Thermal Desorption by Intermediate Cryofocusing in
a Cold 'Ih'ap. J. Chromatogr. 1037. 405, 87-98.

Weurman, C. Isolation and Concentration of Volatiles in Food
Odor Research. J. Agric Food Chem. 1989, 17, 370-384.

Yang, K.; Peppard, T. Solid-phase Microextraction for Flavor
Analysis. J. Agric. Food Chem. 1994, 42, 1926- 1930.

Zhang, Z. Y.; Pawliszyn, J. Headspace Solid-phase Micro-
extraction. Anal. Chem. 1993, 66, 1843-1862.

J. Agric. Food Chem, Vol. 45. No. 5, 1997 1007

Received for review October 28, 1996. Accepwd Febniary 21,
1997.‘ Mention of a trademark and proprietary product does
not imply endorsement of the products named or criticism of
similar ones not named.

JF’9608229

 

' Abstract published in Advance ACS Abstracts, April
15, 1997.

Appendix B

129

// 05-16—95 TOFSORC3.CPP

/* program to calculate ideal space-focusing acceleration field,
modified from tofsourc.cpp to account for different x widths. */

#include <stdio.h>
#include <math.h>

// array size
#define MAXARRAY 251

// "constants" to change for various runs

#define MASS 56

#define CHARGE 1

#define DISTANCE 153 // drift length in cm
#define TIME 20.778 // time-of-flight in usec

// constants
#define CONVERSION_CONSTANT 1.017941
#define TOLERANCE .00000000005

main()
{
int umss=MASS, // mass in u
Q=CHARGE. // charge
no_of_fields=MAXARRAY-1,
i, j, k, m, n, m_old;

f10atd=DISTANCE, // distance D in cm
x[MAXARRAY], // array size
C=CONVERSION_CONSTANT; // conversion to non SI units
double E[MAXARRAY], // electric

sumej, sumek, sumEn, sumei,
E_oldhigh, E_oldlow,
U0, totalx, totalV;

long double t;

for (i=0; i<MAXARRAY; ++i)

{

E[i]=0; // initialize all E[i]
x[i]=0.01; // initialize all x[i]
}
E[O]=1000; // seed 1st E
x[1]=.648; // set 2nd x

// initialize variables to zero
m=0;

m_old=m;

E_oldlow=0;

E_oldhigh=0;

U0=0;

tota1x=0;

tota1V=O;

130

// print header info
printf("mass=%d d=%f x[1]=%f x[2]=%f T=%f\n E[1]=%f\n",
mass, d, xll], XIZ]. TIME, Elli);

for (m=1; m<=no_of_fields; ++m) // set m=1 to adjust all fields

{

if (m_old!=m) // do if lst time through for this m

{
Elml=E[m-1];
E_oldhigh=0;
E_oldlow=0;
m_01d=m;

}

sumEn=0;
for (n=1; n<=m; ++n)

{
sumej=O; // initialize sum xEj
for (j=n; j<=m; ++j)

{
}

sumek=0; // initialize sum xEk
for (k=n+1; k<=m; ++k)

{
}

sumEn += (sqrt(U0+q*sumej)-sqrt(U0+q*sumek))/E[n]; // sum En
} // end for n loop

sumej += x[j1*E[j];

sumek += xlk]*E[k];

sumei=0; // initialize sum Ei
for (i=1; i<=m; ++i)

{
}

// find the actual time, t
t=C*((d*sqrt(2*mass)/(2*sqrt(q*sumei)))+(sqrt(2*mass)/q)*sumEn);

sumei += xli]*E[i];

// compare actual time, t, with target time, TIME
if ((t>TIME) && (t-TIME)>TOLERANCE)

E_oldlow=E[m];
if (E_oldhigh==0)

{
Elm] += 0.1*E[m];
}
else
E[m]=(E[m]+E_oldhigh)/2;
}

_-m;

131

if ((t>TIME) && (t-TIME)<=TOLERANCE)

{

totalx=totalx+x[m];
totalV=tota1V+(E[m]*x[m]);
printf("%f %f \n", totalx, totalV);

if ((t<TIME) && (TIME-t)>TOLERANCE)
E_oldhigh=E[m];
E[m]=(E[m]+E_oldlow)/2;
_-m;

if ((t<TIME) && (TIME-t)<=TOLERANCE)
totalx=totalx+x[m];
tota1V=tota1V+(E[m]*x[m]);
printf("%f %f \n", totalx, totalV);

} // end for m loop

return(1);
} // end fcn

132

NONLINEAR.mcd Program for modeling a multi-gridded ion source.

Given the ion source parameters, the arrival time at the detector is calculated
for ions beginning at the acceleration region edge. The program then
calculates the electric ﬁeld and electrode voltages required for elements in the
ionization region. (Written in Mathcad ver. 4.0)

Step 1: Identify ion source parameters
Distance between ionization region
mass m t: 56 electrodes beginning at acceleration
region electrode (cm) .

chare ::
9 q ' Ll:=0.124 L4:=0.302 L7:=0.302

Unit COWBFSiO" constant c := 1.0 1 7941 L2 := 0303 L5 := 0.303 L8 := 0.302
acceleration length (cm) d .20 648 L3 1:0.302 L6 120.302 L9 :=0.259

distance to detector (cm) D I: 153 Ionization region center point

' between electrodes (cm).
voltage applied to __ ._ ._ '_
acceleration region (volts) v0 .- 1600 $1 .—0.062 54 .-0.151 s7 .—0.151
522:0.151 sS I=O.l51 583:0.151
Initial energy (eV)
Uot=0 532:0.151 56 ::0,151 s9 1:0.130
. . . . . V0 3
Calculation of acceleration region electnc ﬁeld EO .= —d— E0 = 2.469-10

Total length of ionization region Ll + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 = 2.499

Seed values for electric field calculations El 2: 80 E4 2: 8.4 E7 := 8.4
EZI=IO E5':=8.4 E8 I=8.4
531:8.3 E6 2:8.4 E91=8.4

Step 2: Calculate ion arrival time (in nsec) at detector for ions traversing only
the acceleration region.

tD :ZC.€I—,\/«/—2—:% td ;:C.£gg.( iuo+q'd'EO—JU0)
2- v0 ‘1'

tD=20.603 td=0.175

t2=tD+td t=20.778

133

Step 3: Calculate the electric ﬁelds required for each ionization region where
the axial location of the ionization region is between the electrodes at s(x). .

Given

D iUo+q'sl-El— (U0 [ﬁo+q-(sl-El+d-EO)— ’Uo+q-sl-El]
me. +

2m- +

(2-«/VO+sl-El) q'E‘ “'50

El I=Find(El) V1 I=L1‘EI + V0

 

 

E1 = 327.759 V1 = 1.641‘103

Given

r J—
U +q-sZ-EZ— ’U
tIC-«lZ-m~ D 0 0

+
<2-«[VO+sZ-E2+Ll-El) q'EZ

[x/UML q-(sZ-EZ+ L1-E1)—JUO+ q-sZ-EZ]
+ qu

JU0+ q-(sZ-E2+ Ll-El +d-E0) - JUo+q-(52-EZ+ Ll-El)
L+ “‘30

 

 

 

 

 

 

 

 

521: Find(E2) V2 i=L2-EZ+ L1-E1+ V0

132 = 158.807 V;2 = l.689°103

Given

/u +q-53-E3- [u
t-C- 2'm- D + 0 0

(2-Jv0+s3-53+ L2-EZ+Ll-El) “'53
[JUOi-q-(SB-EZH- L2-E2) — JUo+q-s3-E3]
+ q-EZ ..
JUO+ q-(s3-B3 + L2-E2+ Ll-El) — Ju0+ q-(s3-E3 + L2-E2)
+ q-El
E+ q-(s3-E3 + L2-EZ+ LI 121 + d-EO)

 

 

 

 

 

 

 

 

+ —JU0+q-(s3-E3+ L2-E2+ Ll-El)
+
i q-EO

E3 ::Find(E3) v, :zL3-E3 + Lz-Ez'+ Ll-El + v0
3

 

 

 

E3: 127.518 V3=1.727°10

134

Given

D
tIC- 2-m-

 

 

2-«/V0 + s4-E4+ L3-E3 + L2-E2+ Ll-El

/U0+q-s4-E4— /U0 [JUO+q-(s4-E4+L3-E3)— iUo+q-s4-E4]
+ +

 

 

 

 

 

 

 

q-E4 q-E3
JUO + q-(s4-E4 + L3-E3 + L2'E2) JU0+ q-(s4-E4+ L3-E3 + L2'EZ + [.1 -El)
+ -JUO+ q-(s4-E4+ L3-E3) + -JUO+ q-(s4-E4+ L3-E3 + L2-E2)

+ (1'52 + q-El

 

JU0+ q'(s4-E4+ L3~EB + L2-EZ+ Ll-El + d-EO)

 

+ -Ju0+ q-(s4-E4+ L3-E3 + L2-E2+ Ll-El)
+

 

 

L q-EO
E4 I=Find(E4) VA I=L4~E4+ L3-E3 + L2-E2 + L1 151 + V0

E4 = 114.243 V4 = l.762°103

Given

D
tBC- 2°m-

 

 

2~\[V0 + sS-ES + L4-E4 ~l— L3-E3 + L2-E2 + Ll-El

’Uoi-q-sS-ES- ’Uo [JUo+q-(55-ES+L4-E4) — i'Uo+q-s5-E5]
+ +
q ES

 

 

eel

 

 

 

 

 

q-E4
r
JUO + q°(SS-E5 + L4-E4 + L3-E3) JUO + q.(ss -E5 + L4-E4 + L3-E3 + L2-E2)
+ -JU0 + q-(sS-ES + L4-E4) + -JUO + q-(sS -E5 + L4'E4 + L3-E3)
+ . +
q-E3 q-EZ

 

JU0+ q-(sS-ES + L4-E4+ L3-E3 + L2-EZ+ Ll-El)

 

 

+ -JUO + q-( sS-ES + L4-E4 + L3~E3 + L2-E2)

 

q-El
JUO + q-( sS-ES + L4‘E4 + L3-E3 + L2-E2 + Ll E] + d-EO)

 

 

+ -Ju0+ q-(ss-ES + L4-E4+ L3-E3 + L252 + Ll-El)
+
L q'EO

 

 

ES I=Find(E5) V‘ IZLS-ES + L4-E4+ L3-E3 + L2-E2+ L1-El+ V0

ES = 105.454 5 = 1.794'103

Given

135'

q

 

 

+

 

(2 wfs6-E61— LS-ES + L4-E4+ L3-E3 + LZ'EZ + L1°E1+ V0)

 

't‘
q-E6

 

«[110 + q-(so-E6 + LS-ES + L4-E4)

 

 

 

[Juo'l‘ q~(s6-E6+ LS-ES) — [110+ q-s6-E6] + -JUO+ q-(so-Eo+ 1.5-ES)
+

+
q-ES q-E4

JU0+ q-( s6'E6+ LS-ES + L4-E4+ L3-E3)

 

 

+ -JUO+ q-(s6-E6+ 1.5-E5 + L4-E4)

 

. q-E3
JU0+ q-( $6136 + L5~E5 + L4~E4 + L3-E3 + 1.252)

 

 

1+ -4111“ q-(s6-E6 + LS-ES + L4-E4+ L3-EB)

 

_ q-EZ
Juo+ q-(s6-E6 + LS-ES + L4-E4+ L3-133 + L2-E2 + L1 .131)

 

 

+ -JUO+ q-(s6-E6 + LS-ES + L4-E4+ L3-E3 + L2-E2)

 

_ q-El
JU0+ q-(s6-Eo + LS-ES + L4-E4+ L3-E3 + L2-EZ+ L1-El+ d-EO)

 

 

 

+ -JUo+q-(s6-E6+ LS-ES + L4-E4+ 1.3133 + L2-E2+ Ll-El)

 

 

q-EO

E6 3: Find( E6) V‘ 2: 1.6-E6 + LS-ES + L4'E4+ L3'E3 + [.2132 + L1 -E1 + V0

3

E6 = 99.364 V6 = 1.824°10

136

Given

D

 

tIC- 2-m-

 

s7- E7+L6 E6+L5 E5+L4 E4+ L3 E3+L2~ E2+L1- E1+V0

:[Uo +q-s7 E7— {U0 +[,JUO +q--(s7 E7+L6E6)— iUo +q-s7 E7]

 

 

 

 

 

 

 

q- -E7 q E6
JUO+ q-( s7-E7+ L6-E6 + LS-ES) JUO+ q-(s7-E7 + 1.6-E6 + LS-ES + L4-E4)
+ -JU0+ q-(s7-E7 + L64E6) + -JU0 + q-(s7-E7+ 1.6-E6 + 1.585)
+ +
q-ES q-E4

 

JUO + q-( s7-E7 + L6-E6 + 1.5-ES + L4-E4 + L3-E3)

 

+ -Juo+ q-(s7-E7 + L656 + L555 + L4-E4)

 

- q-E3
JUO + q-(s7-E7 + L656 + L555 + L4-E4 + L353 + L252)

 

 

+ -Juo + q-(s7-E7 + L656+ L555 + L4-E4 + L353)

 

. q-EZ
JUO+ q-(s7-E7+‘L6.E6+ LS-ES + L4‘E4+ L3-E3 + L2~E2 + L1 ~E1)

 

 

 

+ -Juo + q-(s7-E7 + L656 + LS-ES + L454 + L353 + L252)

 

. q-Ei
JUoi- q-(s7-E7+ L656 + L555 + L454 + L3-E3 + L252 1» L151 + d-EO)

 

 

 

+ -JUO+ q-(s7-E7+ L6-E6+ [.5135 + L4~E4 + L3-E3 + LZ-EZ + Ll-El)

 

 

q'EO

57 I=Find( 57) v, z: L757 + L6-E6 + 1.555 + L454 + L353 + L2-E2 + L151 + v0
57 = 94.799 v7 = 1.852103

Given

th-

2-m-

 

1

137

D

 

 

2-4/58-138 + L7-E7 + L6-E6 + LS-ES + L4-E4+ L3-E3 + LZ-EZ + Ll £1 + V0

iUo+q-s8-E8 — fl}: [JUO+q-(s8-E8+ 1.7-E7) — ﬁlo+q-s8-E8]
+ +

q-E8 q~E7
JUO + q-( s8-ES + L7~E7 + L6-E6) JUO + q'(s8-E8 + L7-E7 + L6-E6 + LS-ES)

 

 

 

 

 

 

 

+ '«jU0+ q-( 53.53 + L7'E7) + -JUO + q-( s8-E8 + L7'E7 + 1.6-E6)
+

+

 

 

q56 q-ES
JUO + q-(s8-E8 + L7-E7 + L656 1» LS-ES + L4-E4)

 

 

+ ,JUO+ q.(58-E8 + L7-E7+ L656 + L555)
+

 

r q-E4
JUO + q-( $8138 + L75E7 + 1.6-E6 + LS-ES + L4-E4 + L3-E3)

 

 

+ -JU0 + q-( s8-E8 + L7-E7 + 1.6-E6 + L5-E5 + L4-E4)

 

_ q-E3
«[Uo + q-( s8-E8 + L7-E7 + L6-E6 + LS-ES + L4-E4+ L3-E3 + L2-E2)

 

 

L+ -JU0 + q~(s8-E8 + 1.7-E7 + 1.6-E6 + LS'ES + L4-E4 + 1.353)

 

. q-E2
JUO+ q-(s8-E8 + L7-E7+ L6'E6 + LS-ES + L4-E4+ Lil-E3 + L2-E2+ Ll -El)

 

 

 

+ —JU0 + q-( s8-E8 + L7-E7 + L6-E6 + L5'E5 + L4-E4+ L3-E3 + L2-E2)

 

. q'El
JUO + q-(sB-ES + L7-E7 + L6-E6 + LS-ES + L4-E4 + L3-E3 + L2-E2 + Ll E] + d-EO)

 

 

 

+ -JU0+ q‘(58-E8+ 1.7-E7 + L6-E6 + 1.5-ES + Lil-134+ L3-E3 + L2-E2+ Ll-El)
+ L.

 

q-EO

E8 Z=Find( E8) V°1=L8-E8 + L7°E7+ L6-E6+ 1.5-ES + L4°E4+ L3'E3 + L2-E2 + L1-El-1- V0
E8=91.205 V8: 1.88'10

3

138

Given

D
tIIC- 2-m-

 

 

s9 E9+L8- --E8+L7 --E7+L6 E6+L5 ES+L4 E4+L3- --E3+L2 E2+Ll El +VO

:iUo +q--s9 E9— /U0 +0[JU +q(s9 E9+L8- E8)— iUo +q-s9 E9]

 

 

 

 

 

 

 

 

. q 59 q 58
JUO + q-(s9-E9 + L8-ES + L7-E7) JU0+ q-(s9-E9 + L8-E8 + L7-E7 + L6-E6)
+ —JU0+ q-(s9-E9+ L8-E8) + —JU0+ q-(s9-E9 + L8-E8 + L7-E7)

+ : - q-E7 + q-E6

 

E) + q-( s9'E9 + L8'E8 + L7-E7 + L6~E6 + LS-ES)

 

 

+ -JU0+ q-(s9'E9 + L8'E8 + L7-E7 + L6-E6)

 

q-ES
JUO +q --(s9 E9+ L8 E8+L7 E7+ L6 E6+L5- E5+L4- -E4)..

 

 

+ -JUO + q-( s9-E9 + L8'E8 + L7-E7+ L6-E6 + 1.5-ES)

 

. q-E4
JUO + q-( s9-E9 + L8OE8 + L7-E7 + L6-E6 + LS-ES + L4'E4 + L3-E3)

 

 

+ -JUO + q-(s9'E9 + L8-E8 + L7-E7+ 1.6-E6 + LS-ES + L4-E4)

 

. q-E3
JUO + q-( s9-E9+ L8-E8 + L7-E7 + L6-E6 + LS-ES + L4'E4 + L3-E3 + L2-E2)

 

 

 

+ —JU0 + q-( s9-E9 + L8-E8 + L7'E7+ L6-E6 + LS'ES + L4-E4 + L3-E3)

 

qEZ
ﬁo +q(-s9-E9+L8 E8+L7 E7+L6-~E6+L5 E5+L4 E4+L3- E3+L2--E2+L1E..l)

 

 

+ ’ﬁo + q'( s9-E9 + L8-E8 + L7'E7 + L6-E6 + LS'ES + L4-E4 + L3-E3 + L2-E2)

 

q-El

U0 + q- s9°E9 + L8-E8 + 1.7-57+ 1.6-E6 + L5-E5
+ L4-E4+ L3-E3 + L2'E2+ L1-E1+ d-EO

+ -fio + q-(s959 + L858 + L757 + L656 )

 

 

+ LS-ES + L4-E4 + L3-E3 + L2-E2 + Ll 5E!
L q'EO

 

 

 

59 :: Find( 59) Va z: L959 + L858 + L757 + L656 + L555 + L4-E4+ L353 + L252 + L1 51 + W

E9 = 88.356 V9 = 1.903'103

Summary of required voltages

v. = 1.641-103 v, =1.762-103 _, =1.852-103
v, =1.689-103 v, =1.794-103 v9 =1.88-103
v = 1.727-103 v6 = 1.824-103 v9 = 1903-103

3

139

Step 4: Calculation of actual ﬂight time for ions beginning at any axial location
within the ionization region. The output is graphical.

$1 I=0.000,0.001.. Ll

iU +q-sl-El — /U i
f(sl)I=C- 2-m' D + 0 o...
(2-4/v0+s151> 9'51
[JUo+q'(sl-El+d-EO)— /U0+q-sl-El]
+

q-EO

 

 

 

 

 

s2 3= 0.000, 0.001 .. L2

8(s2) :05. D + JU°+q'sz'Ez"/U—°
(2.4/v0+s2-52+L151) “'52
[JUo+q-(s2~E2+Ll-El)— /U0+q-s2-Ez]
+ q-El
[JUOJrq-(SZ‘EZJrLl-El+d-E0)~JU0+q-(s2-E2+Ll-El)]
1+ q-Eo

 

 

 

 

 

 

 

 

 

 

s3 1: 0.000, 0.001 .. L3

p

’u +q-s3-E3— ’U
h(s3)I=C- 2-m- D + ° °...

(2,-Jv0+ s3-E3+ L2-E2+ Ll-El) “'3
LJU0+ q-(s3-E3 + L2'E2) — ’Uo+ q-s3-E3]
+ _ (”32
1JU0+ q-(s353 + L2-E2+ L151) — JUOJF q-(s3-E3 + L2-E2)]
r q51
H/qur q-(s3-E3 + L252+ L151 +d-EO) — [110+ q-(s3-E3 + L2-E2+ LIEU]
L q-EO

 

 

 

 

 

 

 

 

 

 

 

 

140
s4 :=0.000,0.001.. L4

D {U +q-s4-E4— ’U
i(s4) I=C- 2m- 0 o

+

 

 

 

 

 

 

 

 

 

<2-Jv0+s454+L353+L252+L151> q“
ﬁlo+q-(s4-E4+L3-E3)— ’UO+q-s4-E4
+ (1.53
[ﬂow-(5454+ L353+ L252) - JU0+ q-(s4-E4+ L3-E3)]
+ F qEZ
JUO+ q'(s4-E4+ L3-E3+ L2-E2+ Ll-El) — JUO+ q~(s4-E4+ L3-E3 + L2-E2)]
+ -

 

- q-El
JUO + q-( s4-E4+ L3-E3 + L2-EZ + Ll-El + d-EO)

 

 

 

+ -Juo+ q-(s4-E4+ L3-E3+ L2-E2+ Ll-El)

 

 

1 (150

s5 2: 0.000,0.001.. L5

j(s5)f=C' 2-m- D

 

 

Z-JVO + sS-ES + L4-E4 + L3-E3 + L2-E2 + Ll -E1

/U0+q-s5-E5- luo [JUo+q-(s5-E5+L4-E4)— /U0+q-s5-E5]
+ ...

q-ES q-E4
JUO + q-( sS-ES + L4-E4+ L3-E3)

 

+

 

 

 

 

+ -JUO + q-(sS-ES + L4-E4)
_ q'E3
JUO + q-(sS-ES + L4-E4+ L3-E3 + L2-EZ)

+

 

 

 

1+ '4on + q.(s5-55 + L4-E4+ L3-E3)

 

. qiEZ
JUO+ q-(sS-ES + L4-E4+ L3-E3 + L2-E2 + L1 ~El)

 

 

+ Juo + q-(s5-E5 + L4-E4+ L353 + L2-E2)

 

_ q-El
«[Uo + q-(sS-ES + L4-E4+ L3-E3 + L2-EZ+ Ll-E1+ d-EO)

 

 

 

1+ -Juo+ q-(sS-ES + 1.4-54+ L353 + 1,252+ L151)

 

 

 

q-EO

 

141

s6 1: 0.000,0.001.. L6

. D .
k(s6)i=C~\/2-m-

2-JVO+ $6136 + LS-ES + L4-E4 + L3-E3 + L2-E2 + Ll -El

’Uo+ q-s6-E6 - JU—o JU0+ q-(s6-E6+ LS-ES) - ’Uo+ q-s6-E6
+ . q-E6 + q-ES
JU0+ q-(s6'E6 + LS-ES + L4-E4)

 

 

 

 

 

 

+-JUo+q-(s6oE6+ LS-ES)

L

. q-E4

JU0+ q-(so-E6+ LS-ES + L4-E4+ L3-E3)

 

 

 

 

+ -JUO + q-(s6-E6 + LS-ES + L4-E4)

 

. q-E3
JU0+ q'(s6-Eo + LS-ES + L4-E4+ L3-E3 + L2-E2)

 

 

+ -JUo+q-(s6-E6+ L5-E5 + L4-E4+ L3-E3)

 

P q-E2
JU0+ q-(s6-E6 + L54E5 + L4-E4+ L3-E3 + L2-EZ + Ll oEl)

 

 

 

+ -JUO + q-(s6-E6 + LS-ES + L4'E4+ L3-E3 + L2-E2)

 

- q-Ei
J11” q-(s6-E6 + L555 + L454 + L353 + L252 + L151 + 450)

 

 

 

L+ -JU0+ q'(s6°E6 + L555 + L4-E4+ L353 + L2-EZ+ L151)

 

 

q-EO

 

142

s7 1: 0.000, 0.001.. L7

D

 

[(87) :0 2m-

 

s7 E7+L6 E6+LS ES+L4 E4+L3 E3+L2E2+ L1E1+V0

j0/U4—qs7o—E7 iUo +[f+q~--O(S7E7+L6E6)—’U +q-s7E7]

q E7 q- -E6
JUO + q-( s7'E7+ L6-E6+ LS-ES)

 

 

 

 

 

+ -JU0 + q-(s7-E7 + L6-E6)
q-ES
JUO + q-(s7-E7 + L6-EG+ LS-ES + L4-E4)

 

 

 

+ -Juo + q-(s757 + L6-E6 + L555)

 

qE4
JUO +q ~(s7- E7+ L6 E6+ L5 E5+L4 E4+L3 E3).

 

 

+ -JU0 + q-(s7-E7 + L6-E6 + L5-E5 + L4~E4)

 

r q-E3
JUO + q-( s7-E7 + L6-E6+ LS-ES + L4-E4 + L3-E3 + L2-E2)

 

 

 

+ -,\/Uo+ q-(s7-E7+ L6-E6+ L5-E5 + L4-E4 + L3-E3)

 

qEZ
JUO +q (s7 E7+L6 E6+L5 E5+L4- -E4+L3 E3+L2 E2+Ll- E1)-

 

 

 

+ -JU0+ q-(s7-E7 +. L6-E6 + LS-ES + L4-E4 + L3oE3 + LZ-BZ)

 

q-El
JUO+ q-( s7-E7 + L6-E6+ LS'ES + L4-E4+ L3-E3 -1— L2-E2 + Ll-El + d-EO)

 

 

+ W/Uo+ q-(s7oE7+ L6-E6+ 1.5-ES + L4'E4+ L3-E3 + L2-E2 + Ll-El)

 

 

q-EO

 

143

58 i: 0.000, 0.001.. L8

n(s8) i=C- 2-m'

 

l

D

 

 

Z'JSS-E8 + L7E7 + L6-E6 + LS-ES + L4-E4+ L3~E3 + L2~E2+ L1-E1+ V0

/U0+q-88E8- ’Uo [JUo+q-(88-E8+L7E7)- {UO+q-s8-E8]
+ + ...

. q-E8 qE7
4E0 + q-(58 -E8 + L7E7 + L6-E6)

 

 

 

 

L+ -JUO + q~(s8-E8 + L7-E7)
_ q-E6
JU0+ q-(sS-E8 + L7E7 + 1.6-E6 + LS-ES)

 

 

 

 

+ -JUO + q-( s8 E8 + L7E7 + L6-E6)

 

- q'ES
Juo+ q-(s858 + L757 + L656 + 1.555 + L4-E4)

 

 

 

+ -Juo + q-(sS-ES + L7-E7+ L656 + L555)

 

. qE4
JU0+ q-(s8E8 + L7E7+ L6-E6 + LSES + L4-E4 + L3E3)

 

 

+ -JU0 + q-( s8-E8 +- L7E7 + L6-E6 + LSES + L4E4)

 

. q-E3
JU0+ q-(s8E8 + L7E7 + L6-E6 + LS-ES + L4-E4 + L3E3 + L2E2)

 

 

+ -JU0 + q-(sS-ES + L7-E7+ L6-E6 + LS-ES + L4-E4 + L3-E3)

 

_ q-E2
JUO+ q-(s8-E8 + L7E7+ L6-E6+ LS-ES + L4E4+ L3-E3 + L2-E2 + Ll El)

 

 

 

, _ J00 + “,3 .53 + L757 + L6-B6 + 1.555 + L454 + L353 + L2-E2)

 

 

q-El
Uo+ q- s8-E8 + L7E7 + L6-E6 + LS-ES
+ L4E4+ L3-E3 + L2-E2 + L1-E1+ d-EO

 

+ - Uo+ q- s8-E8 + L7-E7+ L6-E6 + LS-ES
+ L4-E4 + L3-E3 + L2-E2 + Ll -El

q-EO

+

 

59 i: 0.000, 0.001..

p(s9) 2:05:

 

144

L9
D

 

 

2-4/59-E9 + L8-E8 + L7E7+ L6-E6 + L5-E5 + L4-E4 + L3-E3 + L2-E2 + L1-E1+ V0

 

{Uo+q-s9-E9— [U0 [JU0+q-(s9-E9+L8E8)- [Uoi—q-s9-E9]
+ +

. q-E9 q-E8
JUO + q'(s9-E9 + L8E8 + L7E7)

 

 

 

+ -JUO+ q-(s9-E9 + L8-E8)
q-E7
JUO + q'(s9-E9 + L8-E8 + L7-E7+ L6-E6)

 

 

 

1+ "JUN q-(s959 + L8-E8+ L757)

 

r q-E6
JUO + q-(s9-E9 + L8E8 + L7E7+ 1.6-E6 + LS-ES)

 

 

 

+ -JUO + q-(s9-E9 + L8-ES + L7E7 + L6-E6)

 

qES
JUO+ q'(s9-E9 + L8~E8 + L7-E7 + L6-E6 + LS-ES + L4-E4)

 

 

+ -JUO + q-(s9-E9 + L8-E8 + L7E7 + L6'E6 + LS-ES)

 

- q-E4
/uo + q-(s9-E9 + L858 + L7-E7 + L656 + L555 + L454 + L353)

 

 

+ '4/Uo+ q-(s9-E9 + L8-E8 + L757 + L656 + L555 + L4-E4)

 

_ q-E3
JUO + q -( s9E9 + L8E8 + L7E7 + L6E6 + LS-ES + L4-E4 + L3-E3 + L2-E2)

 

 

+ -JUO+ q-(s9-E9 + L8-E8 + L7E7 + L6-E6+ LS-ES + L4-E4+ L3-E3)

 

q-E2

Uo + q- s9-E9 + L8-E8 + L7-E7 + L6-E6 + LS-ES
+ L4-E4 + L3'E3 + L2-EZ+ L1 '51

+ - JU" + q-(s9-E9 + L858 + L757 + L656 )

 

 

+ LS-ES + L4-E4+ L3-E3 + L2-E2
. q-El
U0 + q- s9-E9 + L8-E8 + L7E7 + L6-E6 + LS-ES
+ L4-E4+ L3-E3 + L2-EZ 1- LI El + d-EO

+ — uO + q- 5959+ L8-E8 + L757+ L6-E6+ L555
l + L454 + L353 + L252 + L151

q~E0

 

 

 

 

 

--.al

I
1
l .
l

14,

 

145

x12=L1 x2i=xl+L2 x3f=x2+L3 x43=x3+L4 x51=x4+L5
x61=x5+L6 x73=x6+L7 x81=x7+L8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20.8 —
its!)
20.78 —'
8(52)
h( 53)
i( 54)
Km 20.76 -
1((56)
l(s7)
n(s8)
20.74 —
p( s9)
20.72 “
20.7

 

 

sl ,52+xl ,s3+x2,s4+x3,s$+x4,s6+x5,s7+x6,s8+x7,s9+x8

146

LINEAR.mcd Program for modeling two-stage linear-acceleration ion source
parameters. Given the ion source parameters, the arrival time at the detector is
calculated for ions located axially along the ionization region. The data is
plotted and saved to a ﬁle. (Written in Mathcad ver. 4.0)

Step 1: Identify ion source parameters

mass m 2: 56
charge q t: 1
Unit conversion constant C 1: 1.017941
acceleration length (cm) d 3: 0.648

distance to detector (cm)

D :: 153
voltage applied to ._ __
acceleration region (volts) V0 -- 1629 TOtal VOIIage VI -— 1887
Initial energy (eV)
U0 1: 0 Ionization length (cm) s := 2.5
. . . . . V0 3
Calculatlon of acceleration region electric ﬁeld E0 .=_ BO = 2.514-10
d
51 2:91:23 51:1031

S

Step 2: Calculate ion arrival time (in usec) at detector for ions traversing only
the acceleration region.

t toll—ili;— tdi= E [U +q-d~E0— U
D (24:0) q'E°( o J?)

tD=20.419 td=0.l73 ti=tD+td t=20.592

Step 3: Calculate the electric ﬁelds required for each ionization region where
the axial location of the ionization region is between the electrodes at s(x).

 

 

 

 

 

 

 

 

 

 

 

 

S 1: 1,2.. 250
JU +q'0.0l-s-El — ’U
15:20 2-m- D + 0 0
<2 Jv0+0.01-s-51) 9'51
U -0.01--51 d-EO _J ., ..
21.008 e +4] 0+Q( s + ) U0+q OOl sEl
20.928 - / I .- _ q-EO
ts 20.848 »— 4. ,
20.767 -/ 41
Output file = D.\F|LES\MCAD\19513

20.687

100

 

Appendix C

147

ON THE MASS-INDEPENDENCE OF NON-LINEAR
ION ACCELERATION

The theory describing non-linear ion acceleration incorporates, as a fundamental
tenent, the expectation that space focusing is mass independent. While the data used to
validate the concepts of non-linear ion acceleration and complete space focusing are
derived from the temporal characteristics of a single m/z, the theory is expanded to cover
the entire m/z range. The justiﬁcation for this is based on an evaluation of the analytical
equations describing ion motion, previously derived in Chapter 4. According to Equation
4-8, the time-of-ﬂight for any ion is proportional to the square root of the m/z times the
electric ﬁeld function. The time will change with the m/z, but the electric ﬁeld remains
constant. Likewise, as multiple linear electric ﬁelds are added, as in Equation 4-20, the
time-of-ﬂight remains proportional to the square root of the m/z. A second, and more
direct argument for mass independence comes from an evaluation of Equation 4-14,
where it is clearly seen that the distance at which two ions, beginning from different axial
locations within the ion source, become coincident, is independent of mass. While
Equation 4-14 is derived from the Wiley/McLaren two linear ﬁeld model, the same mass-
indepcndent relationship holds when a series of linear ﬁelds is incorporated in the model.
A third argument comes from an evaluation the potential proﬁle required to achieve
complete space focusing, derived using the LINEAR program. This computational
program veriﬁes that, for a given instrumental conﬁguration, the potential proﬁle
required to provide complete space focusing for a low m/z (m/z 50) is identical to the

potential proﬁle required to achieve the same effect with a relatively high m/z (m/z 5000).

   

MICHIGAN $1an UNIV. LIBRQRIES
lllllllllllWilllllllllllllllHllllllllllllllllllllllllllllll
31293016826285