.11.?! :2? .5411:

 

 

.
itflu. ,
)

H... x. L

w HHnJuxnzmmUQ
tron”: #9?
. .rIAHMMMGw

- . an“;
r.

. f. «.
. 1 nihwvéha

3‘ I. \E
1‘ ‘09,”,

11% p
, #13: .
,ztr... ‘

-o. .3: l?.;nfiaw.ulm.v»mm
WHHJJMAI... ”1‘... NAHFI .I

 

. . ‘ . . . 1.1.04.4.
, .. M will. 1'»
, . . I.“ . t... Ix». I
. J: V , . . . ‘ . vii, ,
.l . . ‘ . , ‘ .
07;. In“... ‘

 

 

THgsfib

05%)

MICHIGANS

IIIIIIIIIIIII III/II

Wm."

II IIZIIIIII III/IIIIII

This is to certify that the

thesis entitled

A Study of the Gas Phase Ion-Molecule
Chemistry of Glycerol

presented by

Guy Kirby Smith

has been accepted towards fulfillment
of the requirements for

___SLEI§_dI"Ia egree in Mira/.0 I i

60L, MM

 

Major professor

Date /2' Z 0‘75—

0-7639 MS U i: an Affirmative Action/Equal Opportunity Institution

 

 

 

PLACE It RETURN BOX to remove thle checkout tron your record.
TO AVOID FINES return on or More dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Nflrmetlve Action/Ewe! Opportunity lnetltuion

1

A STUDY OF THE GAS PHASE
ION-MOLECULE CHEMISTRY
OF GLYCEROL

BY GUY KIRBY SMITH

A THESIS

Submitted to
Michigan State Universiy
in partial fulfillment of the requirement
for the degree of

MASTER OF SCIENCE

Department of Chemistry
1995

 

ABSTRACT

A STUDY OF THE GAS PHASE ION-MOLECULE
CHEMISTRY OF GLYCEROL

By

Guy K. Smith

In an effort to understand the origin and formation of the background ion detected
in Fast Atom Bombardment mass spectrometry, a study of the gas phase ion—molecule
chemistry of glycerol was completed. The ion-molecule reactions of both the positive
and negative ions of glycerol with neutral glycerol were determined. The studies were
conducted using the technique of Ion Cyclotron Resonance mass spectrometry.

The existence of gas phase glycerol dimers was proven. The evidence for even
larger gas phase glycerol clusters is examined. The dominant ion-molecule reactions
found for the fragment ions of glycerol are the formation of a fragment ion- glycerol
cluster. Some of the fragment ion—glycerol clusters cluster with an additional glycerol.
The studies of the negative ions of glycerol confirm these results.

Several ion-molecule reactions for both the positive ions and negative ions of
glycerol were detected. The fragmentation mechanisms of both the positive ions and the

negative ions of glycerol are proposed.

DEDICATION

This Thesis is dedicated to my mother, Annabelle Cherry Smith, for only her
continual support has made this long journey possible. She gave me the courage and

ambition to persevere.

iii

ACKNOWLEDGMENTS

I would like to extend a very special thanks to Dr. John Allison for all his
guidance, time and especially his patience. The experience that I have gained in the
Allison group is a useful resource that I will use to further my academic and professional
careers.

I would also like to thank Dr. Ledford, Dr. Pinnivia, Dr. Jackson and Dr. Stille for
serving on my committee. Their comments were most helpful.

I would like to thank Marty Rabb, Ron Haas, Scott Sanderson, and Tom Clark for
their invaluable assistance in completing this research. I would also like to thank Russell
Geyer, Dick Minke and Sam Jackson for making this research possible.

The Ion Cyclotron Resonance Mass Spectrometer, "an0", is still working.

iv

TABLE OF CONTENTS

List of Tables
List of Figures

Chapter One: Fast Atom Bombardment
Introduction
The Mechanisms of Ion Formation in FAB
FAB Studies of Glycerol

References

Chapter Two: The Results of the Positive Ion
Studies of Glycerol

Proof of the Existence of Glycerol Dimers
Evidence for the Existence of Gas Phase Glycerol Clusters
The Source of the Glycerol Clusters

The Origin, Structure and Chemistry of the Fragment Ions
of Glycerol

The High Mass Ions of Glycerol
Proton Transfers from the Fragment Ions
Conclusions

References

Chapter Three: The Studies of the Negative Ions
of Glycerol

The Formation of the Negative Ions
The Negative Ion ICR Mass Spectrum of Glycerol

A Comparison of the Negative Ion Mass Spectra of
Glycerol and Ethylene Glycol

Page #
\lii

viii

15
25

27
32
33
36

38
47
55
57
59

61
61
63

67

The High Mass Anions of Glycerol

The Evidence for the Existence of Glycerol Clusters

A Comparison of FAB and EI Mass Spectra of Glycerol
Mass-to-Charge Ratio Verification

References

Chapter Four: The Experimental Section
The Theory of Ion Cyclotron Resonance Spectrometry

Introduction
Ion Motion in an ICR
Ions in Resonance
Double Resonance Experiments
The Marginal Oscillator
Chemicals used in the Experiments

References

Chapter Five: The Appendix
Detailed Description of the ICR components
The Procedure for a Typical ICR Experiment with Glycerol
Positive Ion Spectra

Negative Ion Spectra

vi

71
73
74
75
76

77
77
77
78
82
83
85
88
89

90
9o
94
94
95

LIST OF TABLES

Chapter One: The Introduction

Table 1.1: Molecular Weights of Compounds Produced from
Glycerol by Ar Atom Bombardment

Chapter Two: The Results of the Positive Ion
Studies of Glycerol

Table 2.1: The Relative Intensities of the Low Mass Positive
Ions of Glycerol

Table 2.2: The Positive Ions of Glycerol and Glycerol-d5

Table 2.3: Relative Intensities of the ICR Spectra of Glycerol

Table 2.4: The High Mass Ions in the ICR Mass Spectrum of
Glycerol

Chapter Three: The Studies of the Negative Ions
of Glycerol

Table 3.1: The Compositions of the Fragment Anions of
Glycerol

vii

Page #

18

29

3O
36
53

63

LIST OF FIGURES

Chapter One: Fast Atom Bombardment

Figure 1.1- The Sputtering of Glycerol by Fast Ar Atom
Bombardment

Figure 1.2- A Schematic Diagram of the Fast Atom Gun on
the JEOL HX-l 10

Figure 1.3- The FAB Mass Spectrum of Tryglycince
(MW of 189) Dissolved in a Glycerol Matrix

Figure 1.4- The Glycerol-Vacuum Interface

Figure 1.5- a) The Gas Phase FAB Spectrum of Glycerol
b) The EI Spectrum of Glycerol at 70 eV
0) The Partial FAB Spectrum of Glycerol

Figure 1.6- A Comparison of the OH“ Negative Chemical
Ionization Mass Spectra of Irradiated and
Unirradiated Glycerol

Figure 1.7- The Neutral Compounds Formed by the Coupling

of Glycerol Derived Free Radicals

Figure 1.8- The FAB Spectra of Glycerol Taken at 1 Minute
Intervals Until All of the Glycerol is Consumed

Page #

11

17

2O

23

Chapter Two: The Results of the Positive Ion Studies of Glycerol

Figure 2.1- A Comparison of the ICR Mass Specrrum with an

El Mass Spectrum of Glycerol

Figure 2.2- The Proton Transferred Structures of the
Methanol Dimer: Computer drawings of
the calculated equilibrium geometries of

the ionized methanol dimers at MP2/6-3IG”

Figure 2.3- A Comparison of the ICR Mass Spectrum with the
Glycerol Sample in a Capillary Tube vs. Glycerol

Evaporated from the ICR Cell Walls

Figure 2.4- The Fragmentation Pathways of the Protonated
Glycerol Molecule

Figure 2.5- The Dehydration of Protonated Glycerol

Figure 2.6- The Possible Structures of the C3H702+ Ion

viii

28

34

37

39

40

41

Figure 2.7- The Possible Structures of the C3H50+ Ion
Figure 2.8- The Structures of the C2H50+ Ion

Figure 2.9- The Structures of the C2H3O+ Ion
Figure 2.10- The Fragmentation of Glycerol by Alpha Cleavage

Figure 2.11- The Formation of the C2H30+ Ion from the
C2H502+ Ion

Figure 2.12— The Formation Mechanisms for the m/z = 62
and 74 Ions

Figure 2.13— Two Formation Mechanisms for the m/z = 44 Ion
Figure 2.14- The High Mass Ions of Glycerol
Figure 2.15- The Relative Intensity of the Cluster Ions of

Methanol vs. n, the number of Methanols in the
Cluster

Chapter Three: The Negative Ion Studies of Glycerol
Figure 3.1- The Low Mass Anions of Glycerol

Figure 3.2- Proposed Structures for the (C3H503)' Ions

Figure 3.3- The Proposed Concerted Loss Fragmentation
Mechanism

Figure 3.4- The Proposed Fromation Mehcanism for the
m/z = 47 Anion

Figure 3.5- The Formation Mechanism of the m/z = 45 Ion
Figure 3.6- The High Mass Anions of Glycerol

Chapter 4: The Experimental Section

Figure 4.1- The Path of a Positive Ion in a Uniform Magnetic
Field

Figure 4.2- The ICR Cell

Figure 4.3- The Cycloidal Path of a Drifting Ion in the ICR Cell
Figure 4.4- Ion in Resonance in the ICR Cell

Figure 4.5- A Typical ICR Double Resonance Spectrum

ix

43
45

46
48

49

50

51

52

58

66
68

69

70
72

78

8O
81
83
84

Figure 4.6- The Equivalent Circuit of the LC Tank Circuit
used in the Marginal Oscillator Detector

Figure 4.7- Block Diagram of the Marginal Oscillator

Figure 4.8- The Block Diagram of the Ion Cyclotron
Resonance Spectrometer at Michigan State
University

Chapter Five: The Appendix

Figure 5.1- The Sample Inlet to the ICR

85

86
87

94

Chapter One:

Fast Atom Bombardment

Introduction

Sputtering is a phenomenon that can be used to generate gas phase ions from
solids. Sputtering was first reported by Grove in 1853.1 Grove observed that gaseous
ions eroded the cathode in a gas discharge source. J. J. Thomson observed sputtering in
1910.2 Thomson experimented with "Kanalstrahalen" (an ion beam) striking a metal
plate. Secondary rays (ions and particles) were emitted in all directions. Only a small
fraction of the secondary rays were charged. The sputtering phenomenon can be simply
described. When a solid is bombarded by high energy (2-10 kilo electron volts (KeV))
particles some of the solid is vaporized.3 A small fraction (approximately 1%) of the
vaporized material appears is in the form of positive and negative ions.4

When sputtering is mated with analysis of the secondary ions by mass
spectrometry, the method of analysis is called Secondary Ion Mass Spectrometry (SIMS).
The SIMS technique is used to analyze the surfaces of solids. SIMS can also be used to
analyze non-volatile organic compounds. The organic analyte is applied as a monolayer
to a metal target.5 The monolayer is bombarded with fast ions and a mass spectrum is
obtained. Since the sample exists as a monolayer, only approximately 1 x 1012
molecules can be deposited in the area bombarded by the ion beam.5 The small number
of molecules in the target area results in a short sample lifetime. Sample lifetimes as

short as several seconds have been reported.3 Reducing the current of the bombarding

2

ions improves the sample lifetime but a reduction in ion intensity results. The application
of SIMS to solid organic analytes is called Organic or Molecular SIMS.

The sputtering phenomenon depends on the transfer of momentum from the
impinging particle to the atoms of the solid. The impact of the fast ion causes a collision
cascade which desorbs atoms and ions from the solid analyte. See Figure 1.1. The beam
of fast argon atoms (Ar) strikes the glycerol (G) matrix ejecting neutral glycerol
molecules, protonated glycerol molecules (G+H)+ and fragment ions from glycerol
(GF‘I'). The sputtering yield is the number of molecules sputtered per incident particle.
The sputtering yield generally increases with the energy and mass of the bombarding
atoms or ions.4 The sputtering yield is also a function of the angle of incidence.4 Since
the energy and angle of incidence of the bombarding particles can be controlled, a.
focused beam of particles can be used to probe the surface composition of a solid.

A fast neutral atom impinging on a solid analyte can also sputter ions into the gas
phase.4 The advantage to using fast atoms instead of fast ions is that the possibility of
charging of the solid analyte is eliminated. Charging of the analyte can interfere with the
collection of the sputtered ions by distorting the electric fields in the source of a sector
mass spectrometer. Fast atoms are also not slowed or deflected by the electric fields in
the source of any mass spectrometer.4

During Organic SIMS experiments, Barber et a1.3 observed that low vapor
pressure oils and liquids (pumping fluids, Apiezon oils, Santovac 5, Convalex 10 and
siloxanes) gave SIMS spectra which lasted for hours. These low vapor pressure liquids
were common contaminants detected in Organic SIMS.3 These observations led to the
idea of using low vapor pressure viscous fluids as solvents in solutions of nonvolatile
organic molecules. The solutions were found to give long lasting and stable SIMS
spectra.3

Barber et al.3 dissolved the analytes in glycerol. These glycerol-analyte solutions

were applied to a metal target and bombarded with fast atoms.3 The fast atoms deposit

 

 

 

 

 

 

 

Fast Ar Atom Bombardment

Figure 1.1: The Sputtering of Glycerol by

4

their energy initially in the glycerol matrix since the impacting atom is much more likely
to strike glycerol molecules that analyte molecules. The energy is then transferred to the
dissolved analyte by collision between the glycerol and the analyte molecules. The ions
generated from the analyte-glycerol solution include protonated analyte molecules and
some fragment ions from the analyte. These ions can be detected for a much longer time
than the ions from a pure analyte monolayer. In this document, fast atom bombardment
(FAB) is defined as the bombardment of a liquid sample by fast atoms. Glycerol is the
most commonly used matrix in fast atom bombardment.

Glycerol has several properties that make it an excellent FAB matrix. The ability
of glycerol to dissolve many polar organic analytes is its first useful property. If the
analyte is not dissolved in the FAB matrix, few analyte ions are observed.6 The second
useful property of glycerol is its low vapor pressure. Glycerol does not evaporate rapidly
in a vacuum. The vapor pressure of glycerol at room temperature is approximately one
millitorr.7 The high viscosity of glycerol helps prevent the glycerol solution from
draining from the metal target. Glycerol has a dielectric constant of 42.5 Debyes.8 This
dielectric constant is large enough to completely dissolve 1:1 electrolytes and gives
glycerol similar acid-base chemistry to that of water.9 This allows the pH of the
glycerol-analyte solution to be manipulated to enhance FAB's sensitivity for certain
analytes. The last important property of glycerol is its low reactivity with the analyte
molecules. This property assures that the FAB mass spectra contains mostly ions from

the glycerol and the analyte with as few ions from reaction products as possible.
The Mechanisms of Ion Formation in FAB
Figure 1.2 is a diagram of a typical fast atom bombardment source.6 The fast

atoms exit the FAB gun with a kinetic energy of 2 to 10 KeV and impact the glycerol and

analyte solution that is applied to the stainless steel target. The heavier noble gases

 
 
  
    
  
  
  
  

loe Source nun; & Filament (3050 v. Xe 10-5 torr)
Xe (8) + e' (my) —§L> Xe“ + 26

Path ofthe MW(°-50V&0V)

Fast Atoms

Colllsloe Chamber (0 v. Xe 10-4 torr)
. CE
3‘“ (3 keV) + Xe (3. thermal) —">
Xe (-3 keV) 1’ Xe“ (thermal)

ExitNonk&XeGsIrdet(-JMV.X¢10‘3lorr)
CE = Charge Exchange

Feet Xe Atom: (6 keV)

 

 

I I FAB Ion Source
Figure l. 2: A Schemr rc Diagram 0 the Fast Atom Gun on the JEOL HX-llO

(actually a SIMION- model drawn. to scale)

6

(argon and xenon) are the fast atoms typically used. The heavier the bombarding atom,
the higher the momentum of the atom becomes. The sputtering yield increases as the
momentum of the bombarding atom becomes larger. The fast atom beam impacts the
target at an angle of approximately 64 degrees from the normal.10 This angle allows
more of the secondary ions generated in the source to be detected by the mass
spectrometer.

A typical liquid SIMS (FAB) spectrum of a low molecular weight analyte is
shown in Figure 1.3.6 The fragment ions of the triglycine analyte are detected at m/z =
30, 76, 87, 115, 133 and 145. The protonated triglycine molecule (triglycine has a MW
of 189) is detected at m/z = 190. The molecular ion of triglycine is not very abundant.
Ions from the glycerol matrix are detected at nearly every m/z value. The more intense
ions from the glycerol matrix are the fragment ions at m/z = 45, 57 and 75, the protonated
glycerol molecule(m/z = 93) and the proton bound glycerol dimer (m/z = 185).

The basic question to be answered for fast atom bombardment is "Where and
how are the secondary ions detected in the FAB experiments formed?" The wide
variety of analytes and matrices that can be used in FAB tends to confuse the answer to
the question. The exact mechanisms of the formation of the secondary ions is not known.
In many FAB experiments, the secondary ions may be formed by several distinct
mechanisms.

Figure 1.4 shows the glycerol solution on a metal target being bombarded by the
fast atom beam. The fast atoms pass through a gaseous layer above the solution before
impacting the glycerol-analyte mixture. This gaseous layer consists of glycerol
molecules evaporating from the surface of the solution as well as glycerol molecules
desorbed into the gas phase by the impact of the fast atoms. The solution is a mixture of
the glycerol matrix and the analyte and is much thicker than a monolayer.9 The solution

layer may be more than 100 angstroms thick.9 The matrix and analyte molecules are

 

 

 

 

.8.
Z . a: o: n:
o 1 01 I
— . .
u «o. w =
m (OT—1.0+2T—r
< I
o 8. w... or.“
_ A
5 .
p 8.
0 r
a A
”A r E
- Nan
n L
v. . 3

Or

a

 

 

.3

G
3 ma

 

 

 

gr

Ewe...“ run .25 Eva 38m mama—.5: a. elm—woman 93$ 3. 33

 

8 Eu
33:1. 39:1
3 cu 3 335+.
O
n N 2: M.“ o o:
N
\ 1
52 /n:~ ,z& =\
O
c we.
a»
3
.3 5m
€3.27.
Eu
.8 ac

Emma—<2. m: » Q6626. Zen—ax

 

 

.8
“3+3?

Nan N u
woo

 

Region \ v z ’ @

 

Phase

 

Figure 1.4: The Glycerol-Vacuum Interface

9
desorbed into the gas phase by the impact and the heating of the solution as the energy of

the fast atom is dissipated. The ions formed are detected by the mass spectrometer.

It is generally accepted that three distinct types of molecular ions can be produced
by particle bombardment of molecules at interfaces.11 The first type of molecular ions
result from direct desorption of precharged compounds. The second type of molecular
ions are the even electron ions formed by the clustering of the analyte molecules with
inorganic cations and anions. The last type of molecular ions are radical cations and
anions. In addition to the molecular ions formed in FAB, fragment ions of the analyte
and the matrix are also generated.

The ions generated by FAB could be formed by one or by a combination of up to
four basic mechanisms. The first possible mechanism is the ionization of gas phase
molecules directly by the impact of the fast atoms. The fast atoms exit the gun and first
pass through a gaseous region above the glycerol-analyte solution as shown in Figure 1.4.
Fast atoms with energies of up to 10 KeV possess sufficient energy to ionize and even
fragment neutral molecules. The ionization step is shown in Equation 1. Morelas12

measured the cross section for the ionization of glycerol by 10 KeV xenon atoms to be 20

A2.

Xe(fast) + G ---> Xe + 6“ + e' (1)

The large cross section for the ionization of glycerol leads to the conclusion that gas
phase ionization is surprisingly efficient9. However, the ionization of gas phase neutral
molecules is not the dominant mechanism of ion formation in FAB.

The vapor pressure of glycerol at 25 0C is approximately one millitorr. This
pressure is 100 to 1000 times the pressure of the FAB source. In addition to the
evaporating glycerol more glycerol is added to the gaseous layer by the impact of the fast

atoms. Each fast atom that impacts the solution is estimated to add 1000 glycerol and

10

analyte molecules to the gas phase. 13 The gas pressure above the solution of a typical
Open FAB source has been estimated to be 10'5 to 10'4 torr.9 Despite the efficient
ionization of the gas phase molecules it is probable that most of the ions detected in FAB
are not formed in the gas phase. The gas phase glycerol has approximately 3.2 x 1012
molecules/m3 versus the 8.25 x 1021 molecules/cm3 of the liquid phase. 14 The high
energy fast atoms are also efficient in ionizing molecules in the liquid phase.9 Assuming
the ionization efficiencies are equal in the gas phase and the liquid phase a fast atom
would have to travel through 10 meters of the gas phase to have an equal number of
collisions as occur in a 100 Angstrom path in the liquid phase. Experimental evidence for
the gas phase ionization not being the dominant mechanism in FAB comes from
comparing the FAB spectra of gaseous glycerol to the FAB spectra generated by
bombarding neat glycerol. Bojeson et a1.15 ionized gas phase glycerol using a fast atom
beam that did not strike any liquid glycerol. The FAB spectrum of gas phase glycerol is
shown in Figure 1.5.15

The gas phase FAB mass spectrum of glycerol has the base peak representing the
m/z = 61 ion. The gas phase FAB spectrum of glycerol resembles its electron impact
mass spectrum. The m/z = 61 ion has low relative intensity in the FAB spectrum of
liquid glycerol. The m/z = 93 (G+H)+ ion makes up the base peak in the FAB spectra of
liquid glycerol and ions consisting of (Gn+H)+ can be detected for n values as large as 15
or more.3 The peak representing the m/z = 93 ion has a low relative intensity in the gas
phase FAB spectrum of glycerol. Most of the ions formed in FAB do not originate from
direct gas phase ionization.9

A second possible mechanism is that the ions exist in the glycerol and analyte
solution and are merely desorbed into the gas phase by the impact of the fast atoms. In
this mechanism the fast atoms merely supply the energy needed to desorb the ions into

the gas phase. Evidence in support of this mechanism is found in the FAB spectra of

11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a)
100 ,0
'7. i
.
50- r
61
xe+ ..
I 1.3 I 1W" -
1 “431 -Jr‘t _, A - a S- 1 t - lull—r.
50 100 ml:
(b)
100- ‘6‘ t
v. ‘
q ‘3 I-
507 ‘0 h-
. ’~ .
15 l I‘ :
<04") .
I “ l ‘ 1 r
50 100 mlz
(C, +
100 M" .
'4 93 l-
v. ‘ ‘
50-
7s
‘ 1.5 l I
'21 h" . .
v 4 ‘ I v v i v T r
50 100 ml:

Figure 1.5: a) The Gas Phase FAB Spectrum of Glycerol
b) The EI Spectrum of Glycerol at 70 eV
c) The Partial FAB Spectrum of Glycerol

12

salt-glycerol mixtures. In the case of sodium chloride, sodium ions are detected in
positive ion FAB spectra and chloride anions are detected in negative ion FAB spectra. 16

Glycerol has a dielectric constant of 42.5 Debyes. The dielectric constant of
glycerol is high enough to completely dissociate 1:1 electrolytes.8 Sodium chloride is
completely dissociated in the glycerol matrix. If the sodium chloride is completely
dissociated, then the sodium cations and chloride anions that are detected in the FAB
experiment are not formed by the fast atoms impacting solid sodium chloride. The ions
are desorbed from the solution by the impact of the fast atoms.

The addition of acids to the glycerol matrix is a common method to improve the
quality of the FAB mass spectra of polypeptides. The basic principle is to increase the
concentration of protonated polypeptides in solution so that more protonated molecules
(M+H)+ would be desorbed by the fast atoms' impacts. However the increase in the
relative intensity of the protonated polypeptide ions, (M+H)+, detected in the FAB
spectra is often small. 17

If the preexisting ions in the solution are desorbed directly into the gas phase in
FAB, the addition of a salt or an acid to the matrix should increase the total ion current of
the mass spectra. Sunner et al.18 found that the addition of salts and HCl to the glycerol
matrix did not increase the total ion current. The salts and HCl added did change the
relative intensities of the ions such as protonated glycerol (G+H)"’ and sodiated glycerol
(G+Na)"' that are detected. Sunner attributes the lack of an increase in the total ion
current to three effects. 18 The ion/ion recombination in the gas phase may limit the total
number of ions reaching the mass spectrometer. As the number of ions desorbed
increases the ion/ ion recombination may become more efficient. A second effect is
caused by the solvation shell of glycerol molecules around the charged species. The
solvation shell reduces the concentration of the charged species near the solution/vacuum
interface. This redistributes the analyte ions deeper into the matrix. The third effect is

the increase in the matrix's surface tension as the glycerol evaporates. Kriger et a].19

13

demonstrated the increased surface tension causes an efficient transport of ions away
from the area of the solution being irradiated.

The detection of protonated or sodiated molecules in the FAB spectrum of an acid
or salt containing solution does not prove that the ions are desorbed from the solution.
The FAB spectrum of NaCl solutions contains peaks representing sodium cations and
chloride anions. These sodium cations could form sodiated molecules by gas phase ion-
molecule reactions with the neutral molecules.

Preexisting ions may be readily desorbed for analytes such as salts or organic
acids. However, the FAB spectrum of glycerol contains lower mass ions with m/z values
of 31, 45, 61 and 75, in addition to the protonated glycerol molecule (m/z =93).15 The
lower mass ions detected in the FAB spectrum of glycerol are not ions that are known to
exist in liquid glycerol. Some of the low mass ions detected in the FAB spectrum of
glycerol are fragment ions detected in the EI mass spectrum of glycerol. Most of the ions
detected in the FAB spectra of glycerol are not preexisting ions directly desorbed from
the liquid glycerol.

The third possible mechanism in FAB ionization is that the ions are formed in and
then desorbed from the liquid phase by the impact of the fast atoms. The impact of the
fast atom on the solution causes a "collision cascade" with desorbs many neutral
molecules into the gas phase. The impact of the fast atoms form the ions directly in the
collision cascade where the ions are immediately desorbed.

When a fast atom strikes the glycerol, the momentum is transferred from the fast
atom to the glycerol molecules. A series of violent collisions between the fast atom and
the glycerol molecules along the impact track form a collision cascade. The collision
cascade is shown in Figure 1.1. The momentum transfer is so great that up to 1000
neutral glycerol molecules are ejected into the gas phase per fast atom impact.20 The
vast majority of glycerol molecules ejected are intact and neutral. However, ions and

fragment ions of glycerol are also formed in the collision cascade.

14
Most of the ions initially formed in FAB are formed in the liquid phase due to the

high number density of liquid glycerol (8.25 x 1021 molecules/cm.3). The ionization
cross sections for gas phase and liquid phase glycerol by fast atoms are not significantly
different.9 Fast atoms, with kinetic energies of up to 10 KeV, are capable of ionizing
neutrals in condensed phases.

Ions formed in the glycerol liquid are desorbed into the "selvedge" region above
the surface. Ion-molecule reactions in the gaseous "selvedge" region modify the ions
detected in the FAB spectra.

The nature of the interface between the glycerol and the vacuum is not clear. The
initial collision cascade releases neutral molecules into the gas phase creating a localized
high pressure gas region called the "selvedge" region.9 The density change in the
selvedge region could be very sharp as in conventional liquid/gas interfaces or could be a
more gradual change. One model of the FAB selvedge is that the ions are initially formed
and desorbed in the collision cascade. These ions are both the fragment ions and the
molecular ions. The glycerol molecules are then desorbed more gradually by a near
thermal process of evaporation. The selvedge region cools rapidly due to the energy
required to break the hydrogen bonds between the glycerol molecules before the glycerol
molecules can evaporate. Heat is also conducted away into the matrix. The ions that are
already desorbed form clusters with the gas phase neutral molecules. The extent of gas
phase ion-molecule chemistry that occurs in the selvedge region remains to be
determined.

Experimental support for the occurrence of ion-molecule reactions in FAB is that
some analytes with gas phase proton affinities lower than the matrix do not have a
protonated molecule in their positive ion FAB spectrum. A good example is the lack of a
protonated molecule, (M+H)+, in the FAB spectra of monosaccarides in diethanolamine
and triethanolamine matrices.21 This effect can be so pronounced as to eliminate the

protonated molecule of an only slightly less basic analyte in a glycerol matrix. An

15

example of this is the FAB spectrum of a cyclic monosaccharide in glycerol. The
protonated molecule of the cyclic monosaccharide (M+H)+ cannot be detected even
though the gas phase basicities of glycerol and the cyclic monosaccaride are very
similar.22 The gas phase basicities, not the liquid phase basicities, determines how easily
each of the compounds in the mixture are detected as a protonated molecule.

If gas phase ion-molecule reactions are the source of the ions detected in FAB, the
fragment ions should protonate the neutral molecules. The dominant positive ions
detected in FAB are the protonated glycerol and protonated analyte molecules. Some of
the fragment ions must have a lower proton affinity than the glycerol and so react to form
the (G+H)+ ion. The analytes which can be detected as a protonated molecule (M+H)+
will have proton affinities that are greater than that of the fragment ions. These ideas
explain why some analytes with proton affinities lower than that of glycerol have a strong
protonated molecule in their positive ion FAB spectra.

The fast atom bombardment mass spectrometry of some crown ethers determined
that the relative intensity of the peak representing the (M+H)+ ion depends on the source
pressure.23 The (M+H)"' ions pressure dependence is a property of an ion formed in an
ion-molecule reaction. The experimental evidence is strong for ion-molecule reactions

modifying the ions detected in FAB.

FAB Studies of Glycerol

F. H. Field performed a FAB experiment in which pure glycerol was bombarded
with 5 KeV positive Ar fast atoms until the glycerol sample was completely consumed.24
As the time of irradiation increases, the (Gn+H)+ ions disappear and a new set of ions is
detected. After all of the glycerol was consumed, a semisolid mass with crystalline

material immersed in a clear liquid remained on the probe tip. Field explained his results

16

as a combination of glycerol evaporation which is accelerated by the fast atom
bombardment and a change in the chemical composition of the glycerol.

The irradiated glycerol was dissolved in water and a Negative Ion Chemical
Ionization (NICI) study using OH' ions as the reagent ions was performed.24 A
comparison of the negative chemical ionization spectra of unirradiated glycerol and the
irradiated glycerol is shown in Figure 1.6.24 The much larger number of ions detected in
NICI spectra of the irradiated glycerol indicates the presence of many compounds in
addition to glycerol. Field compared the NICI mass spectra of unirradiated glycerol and
irradiated glycerol to identify the new compounds. The anions with m/z values of 89,
111,121, 127,151,167, 181, 197, 211, 213 and 243 are assigned by Field to be (M-H)‘
anions of new compounds. The NICI of most organic compounds (using OH' as the
reagent ion) forms (M-H)’ anions. Field lists in Table 1.1 the molecular weights of the
proposed compounds formed by the Ar fast atom bombardment of glycerol. A typical
FAB mass spectra of glycerol was compared with ions the NICI mass spectra. The FAB
mass spectra of compounds in a protic matrix contains peaks for the (M+H)+ ions. Table
1.1 has a list of the (M-tH)+ ions detected in the FAB mass spectra of glycerol. These
ions lend support to be view that fast atom bombardment produces new chemical species.
Field attributes the formation to the new compounds in glycerol to radiation induced
reactions which differ from both conventional radiation chemistry and hot atom

chemistry.24

% Base Intensity

17

 

 

 

 

 

100 183

801
z. I
'53
5 ‘°‘
'5 I
E «sol
a? 91

111
ZO'I ...
" r 3 203
I 89 :3 15 1

 

0 4 A r ' ' l
20 60 100 140 180 220 260 300
Mass (m/z)

OH‘ NICI Mass Spectum of Unirradiated Glycerol

 

 

 

 

 

 

 

100
121
1
so.
‘ 91 183
60.
q
401
5E
l
213
20- 39 5 a 243
I V-l f
OI I I I'va'uj'jfiwv'j
20 60 100 140 180 220 260 300 340
Mass(m/z)

OH' NICI Mass Spectum of Irradiated Glycerol
(11.8 Min. irradiation)

Figure 1.6: A Comparison of the OH“ Negative Chemical
Ionization Mass Spectrum of Irradiated and
Unirradiated Glycerol

18

Table 1.1: Molecular Weights of Compounds Produced from Glycerol by Ar Atom
Bombardment

OH' NICI - '

 

The FAB spectrum of pure glycerol contains ions at nearly every mass up to and
beyond 2000 Daltons.25 These "background" ions may originate from products of
radiation chemistry occurring in the matrix. The impact of very high energy (10 MeV)
particles and ions with solids is known to ionize, fragment and to produce free radicals
from the solid. Gross et a1.25 postulate that fast atoms with energies of up to 10 KeV are
energetic enough to cause reactions in the glycerol matrix similar to those observed in
solids in the radiation chemistry of high energy particles and ions.

Gross et a1.25 studied the origin and structure of these glycerol-derived
"background" ions using tandem mass spectrometry. Gross et a1. group the "background"
ions into four categories by the ions' origin. These categories are: (1) matrix-derived free
radical coupling products, (2) thermal decomposition products of the coupling products,
(3) small neutral molecules formed by addition of or loss of a H atom by a free radical,
and (4) aggregates of glycerol with the preceding 3 categories of products. The ions

detected include protonated (positive) ions or deprotonated (negative) ions formed from

the neutral products of the radiation chemistry.

19
During fast atom bombardment, free radicals are formed from glycerol molecules.

The large amount of momentum transferred to the radicals in their formation prevents
their immediate recombination.25 These free radicals couple with other free radicals to
form new covalent molecules (See Figure 1.7). The newly-formed covalent molecules
may lose neutral groups such as H20, H2CO or CH3OH. Formaldehyde has been
detected in the FAB spectrum of glycerol.26 The loss of water, formaldehyde and
methanol from heavier free radical coupling products generate molecules of intermediate
molecular weights. The protonation or deprotonation of the newly formed molecules can
cause unimolecular rearrangements and condensation reactions. The protonated coupling
products can also undergo condensation reactions with glycerol or can form proton bound
clusters. The deprotonated (negative) ions cluster with glycerol and other polar
molecules also. Gross et a1.25 postulate that the most abundant "background" ions
detected in the FAB spectrum of glycerol are protonated or deprotonated free radical
coupling products.

The radiation chemistry hypothesis can also explain the detection of background
ions with masses as large as 2000 Daltons. The free radicals couple to form higher
molecular weight molecules. These less volatile heavier molecules accumulate in the
matrix over time. When a subsequent fast atom impact occurs, the heavier molecules
may form radicals and couple to form even heavier molecules. It is also possible for the
heavier molecules to be ionized, desorbed and detected.

The ions with intermediate masses (100 to 180 Daltons) detected in the FAB
spectrum of glycerol are a challenge to the free radical coupling hypothesis. The free
radicals formed from glycerol appear to be carbon centered.25 Carbon centered free
radicals are free radicals that have the site of the unpaired electron on a carbon atom. The
carbon centered free radicals of glycerol have masses of 31, 61, 75 and 91 Daltons. The
simple combinations of the possible carbon centered free radicals form neutrals with

atomic Weights of 106, 122, 136, 152, 166 and 182. The neutrals become ions by

20

crew

(lawn H2014 emu
——cu,oa -CH,OH iron
HzOH 14201-1 cuzou

G1“ : 675‘631 G122 :691-631 6122 : Q1-GB1

HO—(ln-l, CH20H
H—C— HCH20H HOCH; -—OH
H: HQOH

31303 Grs'Gor G152 36511391
Ho—cu-r2 cram HO—CHz CH20H
:i—iflfl H—é—OH
H, “2'0” PIC—£142 éHz'OH

Gm; : 6.165 6102 3 691691

Figure 1.7: The Neutral Compounds Formed By
The Coupling of Glycerol Derived Free Radicals.
The free radicals bonded to form each coupling
product are listed under each structure.

21

protonation or deprotonation. These intermediate mass ions are dehydrated with the
possibility of subsequent water, formaldehyde and methanol loses. Gross et a1.25
postulated that the neutral losses are the source of the intermediate mass ions of between
100 and 180 Daltons. There are 24 positive ions of low abundance between 100 and 180
Daltons that are of unexplained origin using Gross et al.'s hypothesis. Though of low
relative intensity in the mass spectra, these 24 ions are of similar relative intensity of
many of the ions explained by Gross et al.'s hypothesis. The best example of an ion of
unexplained origin is the m/z = 143 positive ion. This ion could be formed from the

free radicals coupling products of atomic weights of 152, 166 and 182. The loss of five
H2 molecules from the protonated 6152 molecule would be needed to form the m/z =
143 ion. The most plausible route from the protonated 0166 molecule to the m/z = 143
ion is the loss of water followed by the loss of three H2 molecules. The protonated 6132
molecule would need to lose two water molecules and two H2 molecules to form the m/z
= 143 ion. All of these possible formation routes from the free radical coupling products
to the m/z = 143 ion are ruled out by the reported compositon of the m/z = 143 ion. The
m/z = 143 ion has a composition of C7H1103. Gross et a1.25 did not publish a
fragmentation mechanism for the formation of the m/z = 143 ion. In addition, the
positive ion with a 185 Daltons mass has the highest relative intensity in the mass spectra
between 100 and 200 Daltons.25 This ion (GZ+H)+ is not a protonated free radical
coupling product.

A possible source of the (G2+H)+ ion is ion-molecule reactions in the selvedge
region. The fast atom's impact can fragment and ionize the glycerol molecules. The ions
formed and desorbed into the gas phase can undergo ion-molecule reactions with gas
phase glycerol. Since the ion-molecule chemistry of glycerol is not known it is not
possible to conclude how many of the "background" ions detected by Gross et al.25 are

actually ion-molecule reaction products. The role of ion-molecule chemistry in the

22

formation of the ions detected in the FAB studies of glycerol was investigated by Allison
et al.2'7

Allison et 211.27 performed a study of the time dependence of the relative
abundance of the glycerol ions generated in FAB. The FAB spectra of pure glycerol were
recorded once each minute until all of the glycerol had evaporated. In this experiment it
is possible to vary the extent that gas phase ion-molecule reactions change the ions that
have desorbed from the glycerol. When the FAB beam is first turned on, the number of
glycerol molecules in the selvedge region is at a minimum. And as the glycerol
evaporates to dryness, the number of glycerol molecules in the selvedge region again
decreases. These events reduce the number of ion-molecule reactions that occur.

Allison et al.27 suggest that low mass ions (with masses of 93 Daltons or less) are
the primary positive ions produced by the fast atom bombardment of glycerol. These ions
are desorbed into the high pressure selvedge region where the ions react with the neutral
molecules. The low mass fragment ions of glycerol protonate the glycerol and analer
molecules in the selvedge region. The protonated molecules are then detected by the
mass spectrometer. When the fast atom bombardment commences, the protonated
glycerol molecules have a high relative abundance (See Scan 1 of Figure 1.8). As the
bombardment continues, the density of the gas molecules in the selvedge region
increases. The probability of ion-molecule reactions between the protonated molecules of
the glycerol and the neutral molecules increases. Further collisions between the
protonated glycerol molecules and neutral glycerol form proton bound cluster ions such
as m/z = 185, 277 and 369 (See scan 10 of Figure 1.8). As the glycerol depletes, the
pressure in the selvedge region falls and the primary ions relative intensity increases (See
scan 30 of Figure 1.8).

In FAB, negative ions are formed from glycerol by proton loss. The (G-H)‘ anion

(m/z = 91) can cluster with neutral glycerol molecules to form the anions (26-H)',

23

 

 

 

 

 

 

 

 

 

an as file!
1
10
b 30
E 1
5
g 40
46
52
J

 

 

 

 

m/z

Figure 1.8: The FAB Spectra of Glycerol Taken at 1 Minute
Intervals Until All of the Glycerol is Consumed

24
(3G-H)‘ and (4G-H)'.27 No dehydration products of the glycerol anions were observed.
Potential ion-molecule products was reported at m/z = 181 and 113.27

The low mass fragment anions of glycerol were observed after a long period of
irradiation. These reagent anions have m/z values of 17, 25, 43, 45, 59, 71 and 91. The
anions with mass to charge ratios of 71, 59, 45 and 25 Daltons do not appear in the FAB
spectra taken at short irradiation times. These ions could be either the primary reagent
anions of negative ion FAB or anions of radiation products.

It is impossible to study the ion-molecule chemistry of glycerol directly in a FAB
experiment. Several researchers have alluded to gas phase ion-molecule reactions
occurring during fast atom bombardment.23t27 These postulated ion-molecule reactions
will involve both the fragment ions of glycerol and the glycerol neutrals. No complete
studies of the ion-molecule chemistry of glycerol are known. The objective of this
research is to determine the gas phase ion-molecule reactions of the positive and negative
ions of glycerol with neutral glycerol molecules. This research can improve the
understanding of the formation of the analyte and background ions detected in FAB.

The ion-molecule chemistry of glycerol was studied using an Ion Cyclotron
Resonance (ICR) spectrometer. The ions were generated from gas phase glycerol by
electron impact ionization. Both the positive and negative ions of glycerol can be
generated by varying the electron energy. Ion cyclotron resonance mass spectrometers
are designed to study gas phase ion-molecule reactions. The ion-molecule chemistry of
glycerol can be studied up to the ICR's practical upper mass limit of 300 Daltons. This
mass range is sufficient to determine the reactivity of the fragment ions of glycerol and to
investigate the proposed clustering behavior of glycerol molecules with the ions derived

form glycerol.

1)
2)
3)

4)
5)
6)
7)

8)

9)
10)

11)
12)
13)
14)
15)
16)
17)

18)
19)

25

References

Grove, W. R. Philos. Mag. 1853, 5, 203
Thomson, J. J. Philos. Mag. 1910, 20, 752

Barber, M.; Brodoli, R. 8.; Elliot, G. J .; Sedgwick, R. D.; Tyler, A. N. Anal.
Chem. 1982, 54, 645A

Mc Neal, C. J. Anal. Chem. 1982, 54, 43A
Scheifers, S. M.; Hollar, R. C.; Busch, K. L.; Cooks, R. G. Am. Lab. 1982, 19
Rouse, J. Ph. D. Thesis, Michigan State University, 1993

R. C. Weast (ed.), Handbook of Chemistry and Physics, 55th Edition, CRC Press,
Boca Rotan, 1975

R. C. Weast (ed.), Handbook of Chemistry and Physics, 68th Edition, CRC Press,
Boca Rotan, 1983

Sunner, J. OrgMass Spectrom. 1993, 28, 805

Perle, J. Desorption Mass Spectromety: Are SIMS and FAB the Same?, ACS
Symposium Series, 1985, 126

Pachuta, S. J .; Cooks, R. G. Chem. Rev. 1987, 87, 647

Morales, A. Ph. D. 'Ihesis, Edmonton, Alberta, 1990

Sunner, J. A.; Kulatunga, R.; Kebarle, P. Anal. Chem. 1986, 58, 1312
Calculation by author

Bojeson, G.; Moller, M. Int. J. Mass Spec. Ion Proc. 1986, 68, 239

Kathleen Noon, Private Communication

3) Malomi, A.; Marino, G.; Miloni, A. Biomed. Environ. Mass Spectrom. 1986,
63, 477 b) Gegarly, A.; Van den Heuvel, H.; Claeys, M., Spectrosc. Int. 1989, 7,
133 c) Allmaier, G.; Schmidt, E. R.; Spectrosc. Int. 1985, 4, 297

Sunner, J. A.; Kulatunga, R.; Kebarle, P. Anal. Chem. 1986, 58, 2009

Kriger, M. 8.; Cook, K. D.; Short, R. T.; Todd, P. J. Anal. Chem. 1992, 64, 3052

20)

21)

22)

23)

24)
25)
26)

27)

26

a) Wong, 8. S.; Rbllegen, F. W. Nucl. Instrum. Methods Phys. Res. 1986, 314,
436 b) Jiang, L. F.; Barofsky, E.; Barofsky, D. F., Proceedings of the 36th ASMS
Conference on Mass Spectrometry and Allied Topics, San Francisco, California,
1988, 1209 c) Wong, S. S.; Rbllegen, F. W.; Manz, 1.; Przybyiski, M. Biomed.
1:49:35: Spgecétrom. 1985, 12, 43 (1) Todd, P. J .; Groenwold, G. S. Anal. Chem.

, 5 , 95

a) Puzo, G.; Prome, J. C. Org. Mass Spectrom. 1984, 19, 448-51 b) Harada, K.;
Suziki, M.; Kimbara, H. Org. Mass Spectrom. 1982, I 7, 386

a) Tonduer, Y.; Clifford, A.; Deluca, L. M. Org. Mass Spectrom. 1985, 20, 157
b) Cooks, R. G.; Kruger, T. L. J. Am. Chem. Soc. 1977, 99, 1279

Curcuruto, 0.; Traldi, P.; Moneti, G.; Corda, L.; Podda, G. Org. Mass Spectrom.
1991, 26, 713

Field, F. H. J. Phys. Chem. 1982, 86, 5115
Gross, M. L.; Caldwell, K. A. J. Am. Soc. Mass Spectrom. 1994, 5, 72

Lehmann, W. D.; Kessler, M.; Kbnig; W. A. Biomed. Mass Spectrom. 1984, I I,
217

Allison, J .; Szekely, G. In press

Chapter Two:

The Results of the Positive Ion Studies of Glycerol

A published electron impact ionization mass spectrum of glycerol reports peaks
representing ions with masses of 31, 43, 44, 45 and 61 Daltons.l This published EI mass
spectrum is typical of the other EI mass spectra of glycerol found in the literature. 1’23
When the mass spectrum of glycerol was taken using the Ion Cyclotron Resonance (ICR)
spectrometer a different spectrum was obtained. The ICR detected additional ions with
m/z values of 57, 75, 93, 117 and 119. The m/z = 93 ion is the protonated glycerol
molecule (G+H)+ which should be formed by ion-molecule reactions. The appearance of
the two spectra are radically different. In the published spectrum the base peak is m/z =
61. In the ICR spectrum the base peak represents the (G+H)‘*' ion. Table 2.1 gives the
relative intensities of the ions detected in both spectra. Figure 2.1 shows both mass
spectra.2

Table 2.1 shows major differences in the relative intensities of the all of the ions
with masses from 15 Daltons to 93 Daltons. The most important differences are that the
m/z = 93 (G+H)+ ions and the m/z = 75 ((G+H)"' - 18) ions in the ICR mass spectrum do
not exist in the EI mass spectrum. Also the m/z = 57 ion is much more abundant in the
ICR mass spectrum. The ions of m/z = 117 and 119 are ion-molecule product ions
formed in the ICR and are not reported in the published EI mass spectra.

The double resonance ICR studies of the low mass fragment ions of glycerol
proved that all of the ions with masses of 75 Daltons or less were not products of ion-

molecule reactions. The (Ga-H)+ ion could be an ion-molecule product since the ion's

27

28

Alt-93

 

 

 

El Spectrum of Glycerol

 

1200
1000 - [

800?
“I
400.
200%

0-1 :r 4%

o 20 40 60 80100120140
m/zValues

Figure 2.1: A Comparison of the ICR Mass Spectrum of Glycerol
with the El Mass Spectrum of Glycerol

Relative Mulder

 

 

 

29
mass is greater than that of glycerol. The ICR double resonance studies of the (G+H)+
ion did not find any ionic precursors when the glycerol sample was first introduced into a
clean ICR spectrometer. These results are unexpected. Alcohols do undergo ion-
molecule reactions to form the (ROH+H)+ ion.4 The fragment ions of glycerol are
expected to have lower proton affinities than glycerol and should protonate the glycerol

monomers.

Table 2.1: The Relative Intensities of the Low Mass Positive Ions of Glycerol

Ion in the El Mass Spectra Ion in the ICR Mass Spectra

 

Glycerol also forms a protonated glycerol dimer (G2+H)+ with a mass-to-charge
ratio of 185. The (G+H)+ ion is the principal source of the (G2+H)+ ion. Between 51%
and 66% fewer of the m/z = 185 ions are detected when the (G+H)"' ion is ejected.
Several of the lower mass fragment ions (m/z = 43, 44 and 75) also have weak double

l'CSOIlflIlCC I'CSpOIlSCS.

30

The large reduction in the relative intensity of the (024-11)+ ion when the (G+H)+

ion is ejected is proof that the double resonance experiments work under these

experimental conditions. The (G2+H)'*' ion can be formed from the (G+H)+ ion by one

of three possible mechanisms. Equations 1 to 3 show the mechanisms.

(G+H)+ + Gn => (Gm+1+H)+ + 611.111 (1)
(G+H)+ + Gn => (Gn+H)+ + G (2)
(G+H)+ + 0,, => (Gn+1+H)+ + hv (3)

Equation 1 shows a classic ion-molecule mechanism where an ion collides with a
neutral forming a new ion and a new neutral. A proton transfer to a glycerol dimer is
shown in Equation 2. The last possible mechanism is the formation of a (G2+H)+ ion by
clustering a (G+H)+ ion and glycerol and stabilizing the complex by emitting a photon.5

An ICR study of deuterium labeled glycerol (glycerol-d5) was undertaken to
determine the elemental composition and structure of the low mass fragment ions. A

50:50 mixture of glycerol-d5 and glycerol was studied. The ions detected in the 50:50

mixture are presented in Table 2.2.

Table 2.2: The Positive Ions of Glycerol and Glycerol-d5

CH3 CD2HO+
C2H C2D3O+
C2H40+ C2D3H
C2H C2D3H20+

C2D4HO+

C3D50+
H5C202+ H2C2D302+

C3H7Oz+ C3D5H202+
+ +

C6H1 C6D5H1
1

 

3 l
The glycerol-d5 that was used has the deuterons attached to the carbon backbone

of the glycerol.6 The compositions of the fragment ions of the glycerol-ds can be
deduced by comparing the changes in mass between the ions from glycerol and the ions
from glycerol-d5. The double resonance studies of the fragment ions and the protonated
glycerol ions (both deuterated and undeuterated) were completed.

The evidence for the (G+H)"' ion (rn/z=93) being a product of ion-molecule
reactions is weak. When the glycerol sample is first inserted into an unexposed ICR, no
ionic precursors to the (G+H)+ ion are detected. Only after the ICR has been exposed to
and coated with glycerol are ionic precursors to the (G+H)+ ion detected.

When a layer of glycerol had been deposited on the stainless steel walls of the
ICR spectrometer, the ionic precursors of the (G+H)"’ ions have weak double resonance
responses. Higher power double resonance experiments, the m/z = 61 ion and the m/z =
43, 44 and 45 ion cluster are observed to be precursors ions to the (G+H)+ ions. The
double resonance response of the m/z = 43, 44 and 45 ion cluster is strongest for the m/z
= 44 ion but could not be resolved for the individual ions. Only 5% fewer (G+H)+ ions
are detected when the m/z = 43, 44 and 45 ions are ejected. The ejection of the m/z = 61
ion reduces by 10% the number of (G-I-H)+ ions detected. The double resonance study of
the protonated glycerol-d5 ion (G-d5+H)+ in the 50:50 mixture experiment confirms that
C2H40+ (m/z = 44) and the m/z = 61 ions transfer protons to glycerol molecules to a
slight extent. The (G-d5+H)+ ion has as ionic precursors the C2H40+, C2,l)3HO+ (m/z
= 47), m/z = 61 and m/z = 64 ions in the 50:50 mixture experiment. The m/z = 47 ion
(C2D3OH)+ is the companion ion to m/z = 44 (C2H40)+ ion. Likewise H2C2,D3Oz+
(m/z = 64) is the companion ion to the H5C202+ ion (m/z = 61). The deuterated ions

have similar reactivity to the undeuterated ions.

32

Proof of the Existence of Glycerol Dimers

In the 50:50 mixture of glycerol and glycerol-d5, three protonated glycerol dimers
are detected. These ions are (Gz+H)+ (m/z = 185), (G+G-d5+H)+ (m/z = 190) and
((G-d5)2+H)+ (m/z = 195). A double resonance study of the precursors of the three
protonated dimers proved that glycerol dimers are being protonated.

The ICR double resonance studies of the three proton bound glycerol dimers
determined that both (G+H)+ and (G-ds+H)+ are ionic precursors of all three of the
protonated dimers. The (G2+H)+ ion could be formed by the (G+H)"' ion reacting with a
neutral glycerol or a glycerol containing cluster as shown in Equations 1 and 3. The
(G+H)+ ion or the (G-d5+H)+ ion could form the (G2+H)+ ion by transferring a proton to
a glycerol dimer. Proton transfer is the only explanation for the (G-d5+H)+ being a
precursor to the (G2+H)"’ ion.

The (G‘IrG-ds-t-H)+ ion, (m/z =190), could be formed by either the (G+H)+ ion or
the (G-d5+H)+ ion combining with either glycerol or glycerol-ds or reacting with clusters
containing these monomers. A proton transfer to a glycerol-glycerol-ds dimer could also
explain the double resonance results. None of the reactions shown in Equations 1 to 3
could be ruled out as a source of the m/z = 190 ion.

The (G-t-H)+ ion has a mass of 93 Daltons and can only form an ion with a mass
of 195 Daltons ((G-d5)2+H)+ by transferring a proton to a (glycerol-ds)2 dimer. This
proton transfer explains how the lower mass fragment ions of glycerol and the
(G—d5+H)+ (m/z = 98) ion can be ionic precursors to the m/z = 185 ion and how the
(G-d5-t-H)‘+ ion (m/z = 98) can be an ionic precursor to the (G2+H)+ ion.

The evidence for the existence of glycerol dimers and the lack of detectable ion-
molecule reactions forming the (G+H)+ ion leads to a logical question: Do glycerol

clusters exist in the gas phase?

33

Evidence for the Existence of Gas Phase Glycerol Clusters

The formation of protonated molecules by electron impact ionization of a neutral
clusters has been observed for hydrogen bonded clusters. For example, the H30+ ion is
formed by the ionization of (HzO)2.7 Theoretical calculations suggest that the (I-120)2+
ion has a proton transferred (HzOH+---OH) structure and is represented as a complex
between a oxonium ion and an OH radical (HzOH‘i'n-OH).8 The (I-IzO)2+ cluster
dissociates into (H30)+ and an OH radical.

In the EI mass spectra of acetone clusters9, the ions (CH3+, CH3CO+ and
CH3COCH3+) that are also detected in the EI mass spectrum of acetone are detected. In
addition, the protonated and unprotonated acetone cluster ions of the types
(CH3COCH3)nH"' and (CH3COCH3)n‘" are present. Shukla et a1. studied the ions
formed by electron impact ionization of neutral clusters of primary, secondary and
tertiary alcohols.10 The aliphatic alcohols formed the neutral clusters by adiabatic
expansion of alcohol vapor through a supersonic jet.10 When the alcohol clusters are
ionized, protonated alcohol cluster ions of the type ((ROH)n+H)+ are detected.
Repeating the experiment with deuterated alcohols of the type ROD, the composition of
the protonated molecular clusters formed is ((ROD)n+D)"'.10 This result proves that the
hydroxyl hydrogen atoms are the source of protons in the formation of the protonated
clusters. The exception to this pattern is methanol. CH3OD clusters form both
((CH3OD)n+D)+ and ((CH3OD)n+H)+ ions.11 Alcohol clusters exist as proton transfer
complexes which dissociate into protonated molecules and alkoxide radicals when
ionized. This dissociation is shown in Equation 4.10

(R'CH20H)2 + e' => (R'CHzO) + (R'CH20H)H+ + 2e‘ (4)

Methanol is unique in that it also can form proton transfer complexes using both the
protons of the methyl group and the hydroxyl protons. 11 The transition states for the

hydroxyl proton transfer and the methyl proton trnansfers are shown in Figure 2.2.11

34

(a)

(b)

 

Figure 2.2: The Proton Transfered Structures
of the Methanol Dimer: Computer
drawings of the calculated equilibrium
geometries of the ionized methanol
dimers at MP2/6-31G**

35

N 0 molecular ions ((ROH)n+) are detected in the mass spectra of aliphatic alcohol
clusters. Earlier studies have shown that the formation of ((ROH)n+H)+ ions is always
energetically more favorable than the formation of (ROH)n+ ions.1 1’12 The EI mass
spectra of aliphatic alcohol clusters have all of the normal fragment ions detected in the
mass spectra of the corresponding alcohol monomers. 10

If the glycerol molecules exist as clusters in the ICR, the ICR mass spectrum
should have similar features to those reported for the EI mass spectra of alcohol clusters.
Peaks within the EI mass spectrum of alcohol clusters represent a series of protonated
cluster ions of the type ((ROH)n+H)"’, where n can have values of up to 4 or 5 or more.
The ICR mass spectrum of glycerol has a series of ions of the type ((G)n+H)+ with n
values of 1, 2 and 3. These ions have m/z values of 93, 185 and 277. The alcohol
clusters do not form (ROH)n+ ions. The EI and ICR mass spectra both lack a glycerol
molecular ion. This is consistent with the lack of a molecular ion in the EI mass spectra
of similar alcohols such as 2-propanol, ethylene glycol and sec-butyl alcohol.1 The ions
of m/z = 31, 43, 44, 45, 57 and 61 are the normal fragment ions of glycerol found in the
published EI spectra. These fragment ions are also found in the ICR spectrum of
glycerol. All of these observations support the possibility that gas phase glycerol clusters
exist in the ICR.

The double resonance data from the mixed glycerol (do and d5) experiments
proves that gas phase glycerol dimers exist in the ICR. Evidence for the existence of
larger glycerol clusters in the ICR comes from the double resonance study of the m/z =
277 ion (Gg+H)+. This ion, a protonated glycerol trimer, has the m/z = 185 ion as its
only ionic precursor. The double resonance response to the ejection of the (G2+H)+ ion
is not very strong. Only a 12.5% reduction in the number of m/z = 277 ions is observed.
The protonated glycerol trimer may be formed by the three reactions shown in Equations
1 to 3. The weak double resonance response of the (G3+ H)+ ion may be due to its

primary source being the ionization of G4 clusters.

36

The Source of the Glycerol Clusters

The ICR spectra of glycerol taken at similar pressures have large variations in the
relative intensities of the ions. Figure 2.3 shows a pair of glycerol spectra taken at similar
pressures. The spectra obtained vary depending on how the glycerol is vaporized The
normal procedure is to insert a direct probe into the inlet arm of the ICR. The glycerol is
contained in a capillary tube. Both the inlet arm and the direct probe are heated. The
glycerol spectra taken when the glycerol is evaporated from inside the capillary tube all
have the (G+H)+ ion as the base peak. Figure 2.3A shows a typical spectrum.

After a series of experiments, enough glycerol coated the walls of the ICR cell
and the source inlet to provide a moderate (7.0 x 10"6 torr) pressure of glycerol without
inserting the capillary tube. When the glycerol is evaporated from the glycerol layer
inside the ICR the base peak in the spectra is the C2H502+ (m/z = 61) ion. The (G+H)"'
ion has a very low relative intensity in these spectra. An example spectrum is shown as

Figure 2.3B. Table 2.3 gives the relative intensities of the ions.

Table 2.3: Relative Intensities of the ICR Spectra of Glycerol
Figure 2.3A (5.6 x 10-6 Torr) Figure 2.313 (7.0 x 106 Torr)

 

37

 

8
u
i
8
I
'5’ 1 r:
3‘ u
or -.
'3 R
- ' 2.9:: -9.
”a 0-0 : ,_ n—
“..‘17 '2 . -" 37%
E? e = .
5161on 9.3kilo0auss

its“

 

 

Figure 2.3: A Comparison of the ICR Mass Spectrum of Glycerol
Evaporated from a Capillary Tube with the ICR Mass Spectrum
of Glycerol Evaporated from the ICR Cell Walls

38
The data in Table 2.3 suggests that when the glycerol is evaporated from the walls

of the ICR most of the glycerol in the gas phase is in monomeric form. When the
capillary tube is also a source of glycerol, the most abundant ion detected is (G+H)+. The
(G+H)+ ion is principally an E1 fragment ion of a glycerol cluster. The source of the
glycerol clusters is the liquid glycerol in the capillary tube or the high pressure region
above the glycerol liquid.

The Origin, Structure and Chemistry of the Fragment Ions of Glycerol

Chhabil Dass studied the structure of the fragment ions of protonated glycerol
using a Collision Inducted Dissociation (CID)-mass analyzed kinetic energy (MIKES)
experiment. 13 In CID, ions are accelerated and collide with monoatomic gas molecules.
The collision fragments the heavier ions into smaller ions and neutrals. Dass analyzed the
fragment ions using a sector mass spectrometer in the mass analyzed kinetic energy
(MIKES) mode. CID-MIKES allows the experimenter to observe the unimolecular and
metastable fragmentation of a selected ion. Figure 2.4 shows their proposal for the
fragmentation pathways of the protonated glycerol molecule. 13

The fragmentation of the protonated glycerol molecule is shown in Figures 2.4
and 2.5. The m/z = 75 ion (C3H702)+ ion and the m/z = 57 ion (C3H50)+ ion are
formed by successive dehydrations of the protonated glycerol molecule. Several possible
structures for the m/z = 75 ion are shown in Figure 2.6. The only ion-molecule reaction
of the m/z = 75 (C3H702+) ion is the transfer of a proton to a glycerol dimer to form the

(G2+H)+ ion The proton transfer is shown in Equation 5.

C3H7O2+ + G2 => (G2+H)+ + C3H602 (5)

39

 

 

 

 

 

 

CH2OH _CH3OH CHOH
CHOH +i” CH20H
I m/z=6l
CHQOH "CH20
mlz=93 \
+CH20H
l _ H20 I'll/2:31
mlz=29 —H O
2
+CH2 —> m/z=57/
I - H20
CHOH
l -CH3OH V
CHQOH ’ III/2:43
Ill/2:75
-CH20H ‘__H
- CH2O m
_H 1 _l+
CH2=CH-+OH2 > CH2=CHOH
III/2:45 m/z=44

Figure 2.4: The Fragmentation Pathways of the
Protonated Glycerol Molecule

am a
acnnln
HI In

0 ._.
H
H|n_elo

H

a

o H

4 OICIH

+m _
Cola
H _ H
Hlfllo

H

3

9

__

/z

m

N H
n _
+ C IR
_
NW C IIH
__
C
/
O
H\ H
7
5
__
/z
m

Figure 2.5: The Dehydration of Protonated

Glycerol

41

a) b) H

‘i‘ ’1‘ 0—" '*
/
0—C—C—C/ \ / \ /
H/ h I \ C—C H
H H / \
/° 6’
H H/ \H
C) H d) H
1'}- \O
H
\ /O\ /H J:
c—c H H— —H
H/ \ / |
/ H +
H 0—H \ / \ /
C—C
/
e) H H H
|
H—f—H
+
0
H H
\C/ \C/
o/ \
/ H
H

Figure 2.6: Possible Structures of the C3H702+ Ion

42
The C3H50+ (m/z = 57) ion has 9 known structures. 14 The possible structures
are presented in Figure 2.7. The C3H50+ ion has been formed by a wide variety of
experimental methods. The experimental methods include the metastable decay of

C4H30+ and C3H¢50+ ions as well as EI ionization of precursors such as methyl
propionate, but-2—en—1-ol, l-peten-3-ol and l-ethylcycloproponol.13,14,15 Despite the
varied sources of the C3H50+ ion the researchers report that the C3H50+ ions are a
mixture of two isomers, the acylium (CH3CH2C=O+) ion and the hydroxycarbenium
(CI-Iz=CHCI-I=0H+) ion. The acylium and the hydroxycarbenium isomers of the
C3H50+ ion have been previously reported to be the most stable structures. 14

The C3H50+ ion may undergo a condensation reaction with glycerol to form the

m/z = 131 ion as shown in Equation 6.
C3H50+ + C3H303 => C6H1103+ + H20 (6)

The double resonance response of the m/z = 131 (C6H1103)+ ion is very weak so it is
possible that the (C6H1103)+ ion is a fragment ion of a glycerol dimer. A second
possible explanation is the low abundance of the isomer of the C3H50+ ion which reacts
with glycerol resulting in the formation of the (C6H1103)+ ion. The C3H50+ ion does
not protonate glycerol or glycerol dimers.

The m/z = 57 ion (C3H50+) reacts with glycerol in an ion-molecule reaction with
the loss of formaldehyde to form the C5H903+ ion . This reaction with glycerol to form
the m/z = 119 ion and formaldehyde is shown in Equation 7.

C3H50+ + C3H303 => C5H903+ + CH2O (7)

One of the known structures of the C3H50+ ion is protonated acrolein. The
proton affinity of acrolein is 811 kilo Joules/mole.16 Despite the proton affinity of the

acrolein being less than the proton affinity of glycerol, and a relative intensity of 164, the

43

a) H H D H

\ / H
H\ /C= + H—CE—C—C—OE
=C l
H \H H H
b) H H g) “\C/H+
H—(IS—C—CEO+ H [\(li
| \
III H ,C—C —H
H
+ /H
C) H = h) H
\C—C/ C\ |
H/ — \ H H— (IL--H
H
[0+
C:\C
H/ \
d) H H
l
0+ i)
/ \ H
‘H \/\\ H
l C_C
H H’ \C—H
I
e) ill H
0+
H
\/ \ )1
,C—C::C\
H H

Figure 2.7: The Possible Structures of
the C3H50+ Ion

44

C3H50+ ion does not protonate glycerol or the glycerol dimers. The C3H50+ ion does
not have a protonated acrolein structure in this experiment.

The C2H50+ ion (m/z = 45) is formed by formaldehyde loss by the C3H702+
ion. This ion has four stable structures. 13 The structures are shown in Figure 2.8. The
structures are (CH3C=O+H2), (+CH20CH3), (+CH2CH2OH), and a protonated ring
structure. The C2H50+ ion does not react with glycerol.

The C2H4O+ ion (m/z = 44) formed by the E1 ionization of glycerol has a vinyl
alcohol structure (CHz=CHOH)+.l3 The abundant C2H40+ ion is formed by several
fragmentation mechanisms. The C2H40+ ion is generated by a two step fragmentation
from the molecular ion of glycerol. The C2H4O+ ion also is generated by alpha cleavage
reaction of the C3H702+ ion and by a H radical loss by the m/z = 45 ion. 13 The
C2H40+ ion protonates glycerol and its dimers. The neutral product formed by the
proton loss is acetaldehyde.

Bass13 and R. G. Cooks et a1.17 studied the possible structures of the C2H3O+
(m/z = 43) ion. The five possible structures are shown in Figure 2.9. Both Dass and R.
G. Cooks et al. report that the C2H3O+ ion generated by the BI ionization of glycerol is a
mixture of two structures. The structures are shown in Figure 2.9 as the acetyl ion(a) and
the oxiranyl ion (b). The C2H30+ ion forms the C5H9O3+ (m/z = 117) ion by a
condensation reaction with glycerol as shown in Equation 8.

C2H30+ + C3H303 => C5H903+ + H20 (8)

The C2H3O+ ion has only a slight tendency to transfer a proton to a glycerol
dimer. This supports Dass and R. G. Cooks et al.‘s conclusion that the C2H30+ ion is

not a protonated ketene (b) or protonated alcohol (e) structure. The lack of a significant

proton transfer also supports the conclusion that the C2H30+ ion does not isomerize

upon collision.

45

a)HH b)H H

& +
H—C— =O—H H—C— o—C+

1'. l l

c) d) E]

H H +
\C=C/ H—CQC—H
./ \zls_.. l l

H

Figure 2.8: The Structures of the C2H50+ Ion

46

a) d)

H H
| + l +
H-(IE—C EO H—$— O=C
H H
b) e)
H
+/
>C=C=O+\ H—C "=" c—O\
H H H

Figure 2.9: The Structures of the C2H3O+ Ion

47

The electron impact ionization of glycerol yields important ions with m/z values
of 31, 43, 44, 61, 62 and 74. The dominant ion fragmentation mechanism in the electron
impact (EI) ionization of alcohols is the alpha cleavage reaction. 10 The alpha cleavage
reaction of the glycerol molecule forms the abundant m/z = 61 ion and the much less
abundant m/z = 31 ion detected in the EI mass spectrum of glycerol. Figure 2.10 shows
the alpha cleavage reaction. The loss of water molecules is also an important mechanism
in ion formation. 10 The m/z = 43 ion is formed by alpha cleavage of glycerol with the
dehydration of the resulting m/z = 61 ion (See Figure 2.11).

The abundant m/z = 44 ion is formed by two fragmentation steps from the
molecular ion of glycerol. The first step, shown in Figure 2.12, is the loss of either water
or formaldehyde to form the m/z = 74 or m/z = 62 ions respectively. The m/z = 74 ion
then loses formaldehyde and the m/z = 62 ion loses water to form the m/z = 44 ion.
Figure 2.13 is a diagram of the mechanisms of the 2nd step in the formation of the m/z =
44 ion.

The High Mass Ions of Glycerol

The high mass ions in the ICR spectrum of glycerol are ions with mass-to-charge
ratios greater than 92. Table 2.4 is a list of the ions detected in a typical high mass ICR
spectra (Figure 2.14). The relative intensities of the m/z = 241 and 277 ions were

calculated using another higher mass ICR spectra that has peaks representing the
(G2+H)+, (G3-l-H)+ and the m/z = 241 ions.

48

-+
H H OH

 

I | |
m/Z = 92 H __ C_C _ H
OH OH H
Alpha / \lpha
Cleavage Cleavage
Reaction Reaction
H
H :I: E +0: m/z - 31
_ l— I >C— H
OH OH H
i i“ i‘
ii
on n on
Alpha Alpha
Cleavage Cleavage
Reaction Reaction
H H
| +O/ H
H _ C' \\ I m/z = 61
l /C— C— H
OH 3)
H H

Figure 2.10: The Fragmentation of Glycerol
by Alpha Cleavage

49

 

HK' /
+ / '0
Out =
L\\\II m/z 61
C—C— H
/ l
H
H
8 H20
v
H
+ =C=C/ m/z:43
/ \
H H

Figure 2.11: The Formation of the C2H30+ Ion
from the C2H5Oz+ Ion

50

62 and 74 Ions

Figure 2.12: The Formation Mechanisms
of the m/z

51

l C—H
. / | C |
C l H l H
I H H0 / 74
mz=
HO H20
m/z=62
CHzO
H +
| C—H
.C/l
I H
HO
m/z=44

Figure 2.13: Two Formation Mechanisms
for the m/z = 44 Ion

52

m/z = 185

 

131,133,135

m/z
m/z
[z = 229

z=149

= 151

z=117
m/z=119

 

 

75 RG - ' 16. 5 RG

Figure 2.14: The High Mass Ions of Glycerol

53

Table 2.4: The High Mass Ions in the ICR Mass
Spectrum of Glycerol

ve
Corrected for Mass

 

The double resonance studies of the high mass ions of glycerol proved that all of
the ions except m/z = 131, 133, 135 and 151 are products of ion-molecule reactions.

Alcohol clusters form protonated alcohol molecules when ionized The formation
of the protonated alcohol molecules occurs as shown in Equation 9.10

(R'CH20H)2 + e' => (R'CH20) + (R'CH20H)H+ + 2e' (9)
In the case of glycerol the R' group is (HOH2CCHOH). The 50:50 mixture study of
glycerol and deuterated glycerol showed that the proton transferred to form the (G+H)+
ion did not come from glycerol's carbon skeleton since the ICR mass spectra did not have
a peak representing the (G+D)+ ion (m/z = 94). The transfer of the hydroxyl protons of
glycerol is the same transfer observed for all of the other alcohols except methanol. 10
The glycerol dimer does not form many fragment ions upon ionization. The fragment
ions detected at m/z = 93, 131, 133, 135 and 151 are from the glycerol dimer or the
protonated glycerol dimer.

The protonated clusters of methanol, ethanol and l-propanol undergo a

condensation reaction shown in Equation 10.10, 13

(ROH)nH+ => R20H+ + H20 (10)

54
Ethanol can also form R20H+ by an ion-molecule reaction with protonated ethanol. No
ions formed by the condensation reaction (Equation 10) of the protonated or unprotonated
glycerol dimers were detected.

The fragmentation patterns of other alcohol dimers show that fragment ions are
formed by the alpha cleavage reaction of one of the alcohol molecules in the dimer. The
fragment ion of the alpha cleavage reaction remains bound to the other alcohol molecule.
Examples of this type of cluster ion are the fragment ions of l-proponal, l-butanol and l-
pentanol. These alcohols form ions of the type (ROH)nCHzOH"'.10 Secondary alcohols
such as 2-propanol and 2-butanol form ions of the type (ROH)nCI-I3CHOH"'.lo The
cluster ion formed by t-butyl alcohol is (ROH)n(CI-l3)2COH"’.10 The product ions of the
alpha cleavage reaction of glycerol are m/z = 31 (CH20H+) and m/z = 61
(I-lOCHzCHOH+). The cluster ions of a glycerol molecule bound to the CH20H+ or the
HOCH2CHOH+ fragment ions would have m/z values of 123 or 153. No ions with these
m/z values were detected. The only possible cluster ion of a known glycerol fragment ion
and glycerol is the m/z = 135 ion. The m/z = 135 ion could be a G 4» m/z = 43 ion.

The lack of fragment ions that are complexed with glycerol molecules may be due
to the temperature of the glycerol. In the other alcohol cluster experiments, the alcohol
vapor was adiabatically expanded and cooled. In the glycerol experiments, the liquid
glycerol is heated to vaporize enough glycerol to maintain a moderate pressure. The
heated glycerol dimer or cluster may fall apart upon ionization since its vibrational modes
have more energy than the ground state.

The ion-molecule reactions that form the C5H903+ (m/z = 117) ion and the
C5H1103+ (m/z = 119) ion have been discussed previously. The ion-molecule reactions
discussed in this sections are the formation of higher mass glycerol containing cluster

ions.

55
In the double resonance experiments, no ionic precursors were detected for the
m/z = 133 and m/z = 135 ions. The low relative intensities of these ions may explain why
no ionic precursors were detected.
The m/z = 57 ion (C3H50+) reacts with glycerol clusters to form the m/z = 149

ion (C6H1304+). The C3H50+ ion could combine with a glycerol molecule and emit an

infra-red photon.5
C3H50+ + (C3H803)n =---> C6H1304+ + (C3H803)n-1 (11)
C3H50+ + (C3H303)n => C6H1304+ + hv (12)

The low relative intensity of the m/z = 151 ion may explain its lack of a detectable
ionic precursor. The double resonance study of the m/z = 185 and 277 ions has been
discussed previously. The m/z = 241 ion is a cluster formed by the reaction of the m/z =
149 ion and glycerol clusters. Equations 13 and 14 show the possible reactions.

C6H1304+ + C3H303 => C9H2107+ + hv (13)

C6H1304+ + (C3H803)n => C9H2107+ + (C3H803)n-1 (14)

The double resonance response is weak. Only 10% fewer of the m/z = 241 ions

are detected when the m/z = 149 ions are ejected.

Proton Transfers from the Fragment Ions

No ion-molecule reactions with glycerol were observed for the m/z = 31 ion. The
m/z = 31 ion (CH30+) can be viewed as a protonated formaldehyde molecule. The gas
phase proton affinity of formaldehyde is 718 kilo Joules/mole.16 The gas phase proton
affinity of glycerol is 874 Kilo joules/mole. 16 The protonated formaldehyde should be
able to transfer a proton to the glycerol, Equation 15.
(HzCO)H+ + G => (G+H)+ + H2CO (15)

The proton transfer, equation 15, is exothermic by 156 kilo Joules/mole.16

56

The lack of a detectable proton transfer by the m/z = 31 ion may be due to the ion

having a H3C-O+ structure. This ion is 139 kilo Joules/mole higher in energy than the
protonated formaldehyde. 19 The H3CO+ ion should be able to transfer a proton to the
glycerol. The reaction is exothermic by 17 kilo Joules/mole. 16 The m/z = 31 ion
(CH30"') does not protonate glycerol. The low relative intensity (57) of the 0130* ion
and the weak double resonance responses found for the (G+H)+ ion may explain why this
was not detected. The companion ion m/z = 33 (CD2HO+) did not protonate either
glycerol or d5 glycerol in the 50:50 mixture experiment. Double resonance studies of the
protonated glycerol dimer show that (CI-130+) does not protonate glycerol dimers.

The m/z = 43 ion has an empirical formula of C2H30. If the C2H30+ ion
transferred a proton to glycerol, the neutral molecule would have a composition of
CszO. The proposed neuual is a ketene (HzC=C=O). The proton affinity of the ketene
is 825.5 kilo Joules/mole.l6 No proton transfer to glycerol monomers was detected
Proton transfer to the glycerol dimer occurs. The C2H3O+ ion undergoes a condensation
reaction with glycerol.

The m/z = 44 ion has a empirical formula of C2H40 and protonates glycerol and
glycerol dimers. The proton affinity of the CH3CO neutral formed upon proton transfer
has been estimated to be between 657 and 828 kilo J oules/mole. 16 The proton transfer
reaction is exothermic by 46 to 217 kilo Joules/mole.

The C3H5O+ ion (m/z = 57) does not transfer a proton to a glycerol molecule or
to a glycerol dimer. The proton affiniy of the acylium ion (CH3CH2C=O+) is 834 kilo
Joules/mole.16 The other stable structure of the C3H50+ ion is the hydroxycarbenium
(CH2=CHCH=OH+). The hydroxycarbenium (CH2=CHCH=OH+) ion has a proton
affinity of 811 kilo Joules/mole. 16

The m/z = 61 ion has a proposed structure of (HOCH2-CHO)H+. The m/z = 61
ion has a proton affinity of less than 874 kilo Joules/mole since it can transfer a proton to

glycerol. The m/z= 61 ion does not protonate the glycerol dimer. A possible explanation

57
of this fact is that the irradiating oscillator's frequency is almost exactly 3 times the
frequency of the marginal oscillator. The signal detected by the marginal oscillator
breaks up at this harmonic frequency. The signal breakup could obscure a weak double

resonance response.

The m/z = 75 ion has a empirical formula of C3H70z. A proposed structure of
this ion is (HOCH2C(OH)=CI-12)H+. This structure is a protonated alcohol. The only
ion-molecule reaction observed for the m/z = 75 (C3H70z+) ion is the transfer of a
proton to a glycerol dimer to form the (62+H)+ ion. The proton affinity of the alcohol is
less than the proton affinity of the glycerol dimer. The m/z = 75 ion is formed by the
dehydration of the (G+H)+ ion.

Conclusions

The principal source of the protonated glycerol molecules is the fragmentation of
glycerol dimers upon EI ionization. The glycerol dimer exists as a proton transfer
complex with a structure (HOCHzCHOHCHzOHz+ ----- OH2CCHOHCH20H'). The
heated glycerol dimers fall apart upon ionization yielding only minor fragment ions with
masses above 93 Daltons. Evidence for the existence of the dimer is the proton transfer
from the (G+H)+ ion to form the ((G-d§)2+H)+ ion. The proton transfer from the (G-
d5+H)+ ion to a glycerol dimer forms the (Gz+H)+ ion.

The weak double resonance response of the (63+ H)+ ion may be due to its
primary source being the ionization of G4 clusters. A more probable explanation of the
glycerol data is that the (G3+H)+ is the product of successive additions of glycerol to the
protonated glycerol molecule. A fraction of the protonated glycerol molecules react to
form protonated dimers. Some of the protonated dimer reacts further to form a
protonated trimer. The successive additions can explain the monotonic decrease of

relative intensity of the (G+H)+, (Gz+H)+, and (G3+H)+ ions. Setting the relative

58
intensity of the (G+H)+ ion as 1000, the ((32+H)+ ion has a relative intensity of 610 and
the (G3+H)+ ion has a relative intensity of 20. The data for the relative intensity of
ionized methanol clusters does not show either a rapid or monotonic decrease in ion
intensity with increasing cluster size. See Figure 2.15.11 The weak double resonance

response observed for the (03+ H)+ ion is due to the very low relative intensity of the ion

and not to the existence of G4 neutral clusters.

 

 

'1 l2 ’3 4 5 sfiqnw

 

 

 

 

 

Ion Intensity (Absolute Units)

 

 

LA 1414 L A 14; A A_LA n A HLA

20 4o 60 so loo 120 140 160180200 220
Mass (m/z)

 

Figure 2.15: The Relative Intensity of the Cluster Ions of Methanol
vs. n, the number of Methanols (M) in the Cluster

59

The ions of glycerol undergo few ion—molecule reactions except for the addition

to glycerol. A condensation reaction between glycerol and the abundant C2H30+ ion

occurs. The C3H50+ ion reacts with glycerol to form formaldehyde and the C8H1103+

ion.

1)
2)

3)
4)

5)
6)

7)
8)
9)
10)
11)

12)

13)
14)

References

Stenhagen, E., Abrahamsson, S.; Atlas of Mass Spectral Data, Volume I, John
Wiley & Sons, 1969

a) Stenhagen, E., Abrahamsson, S.; Regisu'y of Mass Spectral Data, Volume 1,
John Wiley & Sons, 1970

Bojeson, G., M?ller, M., Int. J. Mass Spectrom. Ion Processes, 1986, 68, 239

McLafferty, F. W., Interpretation of Mass Spectra, 3rd edition, University
Science Books, 1980

Beauchamp, J. L., Caseiro, M. C.; J. Amer. Chem. Soc., 1972, 94, 2638

The 1,1, 2, 3, 3d 5 glycerol was obtained from Combridge Isotope Laboratories.
The deuterium content was reported as 98%. The deuterons are reported to be
attached to the carbon atoms.

Fricke, J., Jackson, W. M., Fite, W. L., J. Chem. Phys., 1972, 57, 580
Shinohara, H., Nishi, N., Washida, N., J. Chem. Phys., 1986, 84, 5561

Stace, A. J ., Shukla, A. K., J. Phys. Chem, 1982, 86, 865

Stace, A. J ., Shukla, A. K., J. Phys. Chem, 1988, 92, 2579

Lee, S. Y., Shin, D. N., Cho, S. G., Jung, K., Jung, K. W., J. Mass Spectrom.,
1995, 30, 969

Rollgen, F. W., Beckey, H. D. 2., Naturforsch., A: Phys., Phys. Chem.,
Kosmophys., 1974, 29A, 230

C. Dass, Org. Mass Spectrom., 1994, 29, 475

F. W. McLafferty, J. Wachs, C. Koppel, P. P. Dymerski, F. M. Bockhoff, Adv.
Mass Spectrom., 1978, 7, 231

15)

l6)

17)
18)
19)

60

a) Zwinselman, J. J., Harrison, A. (J., Org. Mass Spectrom., 1984, I9, 573

b) Bouchoux, G., Org. Mass Spectrom., 1985, 20, 154, c) Bouchard, G.,
Hoppilliard, Y., Flammang, R., Maquestiau, A., Meyrant, P., Org. Mass
Spectrom., 1983, 18, 340, (1) Hudson, C. E., McAdoo, D. J., Org. Mass
Spectrom., 1982, I7, 366, e) Turecek, F., Hanus, V., G umann, T., Int. J. Mass
Spec. Ion Proc., 1986, 69, 217

Lias, S. (3., Bartmess, J. E., Liebman, J. E., Holmes, J. L., Levin, R. D., Marland,
W. G.; J. Phys. Chem. Ref. Data, 1988, 17, Suppl. 1

Eberlin, M., Majumdar, T. K., Cooks, R. G., J. Am. Chem. Soc., 1992, 114, 2884
Stace, A. J., Shukla, A. K., J. Amer. Chem. Soc., 1982, 104, 5314-18

Allison, J ., Szekely, G.; In press

Chapter Three:

The Studies of the Negative Ions of Glycerol

The Formation of the Negative Ions

The negative ions of glycerol are formed by bombarding the vaporized glycerol
with low energy electrons. The electrons were emitted by the hot rhenium filament and
accelerated by a 1.00 to 8.00 volt difference between the filament voltage and the
napping plate voltages. The voltages of the trapping plates were negative to prevent the
negative ions formed from being neutralized. The voltages of the drift plates were
adjusted to cause the negative ions to drift to the analyzer region.

Negative ions are formed by the interaction of electrons with neutral molecules.
Three different classes of mechanisms for negative ion formation are known},2 For
electrons of very low kinetic energies (near 0 eV), the electron attaches itself to the
molecule. This mechanism is called nondissociative electron capture. The resulting
anion may emit a photon to stabilize the electron-molecule complex. Equation 1 shows
nondissociative capture of an electron by a molecule (AB).1

AB + e' => AB' (1)
Equation 2 is the simple radiative attachment of an electron to the AB molecule.2
AB + e' => AB' + hv (2)

As the kinetic energy of the electrons increases, the mechanism of negative ion
formation changes. The higher energy electron is captured by the molecule and the
resulting anion fragments. This mechanism is called dissociative electron capture.

Dissociative electron capture occurs for electrons with kinetic energies above 0 eV to

61

62

approximately 15 eV.1 Equations 3 and 4 show the dissociative electron capture of an

electron by the molecule AB.
AB + e' => A + B' (3)
AB + e‘ => A' + B- (4)

As the kinetic energy of the electron increases above approximately 10 eV, the
electron is no longer captured by the molecule. The impact of the high energy electron
merely excites the molecule AB to an electronic level that leads to the production of a
positive ion and a negative ion. This mechanism is called ion pair production. Equations
5 and 6 show a high energy electron making an ion pair from the molecule AB.

AB+e'=>A'+B++e' (5)
AB+e'=>A++B' +e' (6)

The electron energies used in the ICR experiments were large enough so that only
the dissociative electron capture and the ion pair production mechanisms formed the
anions. The experimental evidence to support this finding is that no molecular ions of
glycerol or of a glycerol cluster were detected. The attachment of an electron to a
glycerol molecule could occur at several different sites. Attachment of the electron to
one of the sigma bonds (C-C, OH, 00, 0-H) in the glycerol molecule would place the
electron in a sigma anti bonding orbital and lead to the formation of a radical and an
anion. The oxygen atoms are the most electronegative atoms in the glycerol molecule.
The oxygen atom is the preferred site for electron attachment in the glycerol molecule.
The electron cannot attach itself to the filled lone pairs of the oxygen atoms. If the
electron attached to the C-0 bond the glycerol anion could fragment into an OH radicals
and an m/z = 75 anion. No m/z =75 or 17 (OH') anions were detected. If the electron
attaches to the O-H bond, either a alkoxide radical and a hydride anion or a H atom (the
radical part) and an alkoxide anion can form. The lower mass limit used in the ICR

experiments is to high for a hydride anion to be detected. Only the alkoxide anion of

63
glycerol (m/z = 91) was detected in the ICR mass spectrum of glycerol. Other alcohols

such as ethanol, proponal, isopropanol and t-butyl alcohol also form alkoxide anions},4
The Negative Ion ICR Mass Spectrum of Glycerol
The fragment anions detected by the ICR are recorded in the negative ion ICR
mass spectrum of glycerol. These fragment anions have m/z values of 91, 89, 88, 87, 72

and 47 Daltons. Table 3.1 shows the possible compositions of these anions.

Table 3.1: The Compositions of the Fragment Anions of Glycerol

 

 

 

 

 

 

 

 

 

 

 

 

Mass in Daltons Possible Elemental Compositions

91 C3H703'
89 C3H503'
88 C3H403'
87 C3H303'
72 C203’

72 C3H402'
47 C2H70‘
47 CH302'

 

Figure 3.1 shows the low mass ICR spectrum of the anions of glycerol. The m/z
= 91 ion, (G-H)‘, is an alkoxide anion of glycerol. The loss of 2 to 4 more hydrogen
atoms yields the m/z = 89, 88 and 87 anions. The loss of H2 from alkoxide anions has
been reported by R. N. Hayes et al.3 The alkoxide anions EtO', PrO', and i-PrO' were

observed to eliminate Hz to form enolate anions. The enolate anions have a

mam—=6 wk" .28 has. 738 >533 a. $5.83—

72
89

m/z:91

 

m/z
m/z

 

I 0.50 KG
m/z = 47

m/z = 128

- 9.50 KG

m/z = 139

65
R-CH2=CHO' structure.3 The proposed structure for the m/z = 89 anion is an enolate
structure. The m/z = 89 anion, (M-3)‘, may arise from the pyrolysis of glycerol by the
thoriated iridium ion gauge filament or the rhenium electron beam filament. Caldwell et
al.5 reported the loss of H2 by primary and secondary alcohols due to the pyrolysis of the
alcohols by the hot filaments in their ICR. Most of the pyrolysis occurred at the ion
gauge filament since it had a much larger surface area than the surface area of the
electron beam filament. The pyrolyzed alcohols were deprotonated by OH' Negative Ion
Chemical Ionization (NICI) and formed (M-3)‘ anions with enolate structures.
Deuterium labeled alcohols revealed that the (M-3)‘ anion is formed by the loss of one
hydroxyl proton and one a and one ' hydrogen atom. The hydrogen atom and proton
losses reported by Caldwell et a].5 support the stability of the proposed enolate structure
for the m/z = 89 anion of glycerol. The m/z = 87 anions may have an enolate structure.
The possible structures of the m/z = 89, 88 and 87 anions are shown in Figure 3.2.

The m/z = 72 anion, (M-20)’, has two possible compositions. The possible
(C203)' anion has not been reported by other researchers. The (C3H402)' anion was
detected by Gross et al.6 in the study of the anions of glycerol generated by fast atom
bombardment. Tandem mass spectrometry was used to identify the m/z = 72 anion as
(G-H-H20-H)'.6

The base peak in the low mass ICR spectrum of glycerol is the m/z = 47 anion,
(M-45)’. This ion has two possible compositions of (CH302)' or (C2H70)'. The
(C2H70)' anions' structure is CH3CH2-OH2‘. This proposed structure requires the 0
atom's atomic orbitals to have a dsp3 configuration. The (CH302)' anion has a
H0-CH2-0' structure. The m/z = 47 anion can be formed form the (G-H)‘ ion by the
loss of C02 or by the loss of C2H40. The loss of C2H4O could occurs as the loss of
C2H40 or the loss of H20 and acetylene in a concerted fashion. The negative ion ICR
specuum (Figure 3.1) does not have peaks corresponding to the (G-H-H20)‘ or (G-H-
C2H2)‘ anions that would be detected if the losses of H20 and acetylene were not

0H 0'
[\{\M/
HO—C/ \H
l
H
on o
'O—C—%—C<
III H H
OH -
o
o \. c/
/ \

H

v

H

m/z = 89

m/z=88

m/z:87

Figure 3.2: Proposed Structures for the (C3H503)',
(C3H403)', and (C3H303)' Ions

67
concerted. Figure 3.3 shows a proposed concerted mechanism for the loss of H20 and
acetylene. The proposed mechanism for the loss of C2H40 by the (G-H)‘ anion is shown
in Figure 3.4.

A Comparison of the Negative Ion Mass Spectra
of Glycerol and Ethylene Glycol

Lloyd et al.7 studied the ionization reactions and the hydrogen-deuterium
exchange reactions in mono- and dialcohol alkoxide anions. Ethylene glycol was first
ionized using Negative Ion Chemical Ionization with 0'- as the reagent ion. The
"fragment" ions detected are m/z = 58 (C2H202)' and m/z = 45 (CH02)'. The m/z = 58
anion, (M-4), is formed in an ion-molecule reaction. The 0'- ion reacts with ethylene
glycol removing two hydrogen atoms to form water. Two more hydrogen atom are lost
by the ethylene glycol to form H2. The m/z = 45 anion is formed by an ion-molecule
reaction between the O‘- ion and ethylene glycol. The mechanism proposed by Lloyd et
al.‘7 for the ion-molecule reaction is shown in Figure 3.5. The ethylene glycol molecule
is cleaved at the CC bond and the 0'- ion adds to the one half of the molecule. The
resulting m/z = 47 anion (CH302') loses H2 to form the m/z = 45 anion. The fact that
the m/z = 47 anion is not detected in the NICI mass spectra of glycerol does not support
the idea that the m/z = 47 anion is a stable anion. However, the H2 loss by the m/z = 47
anion may be due to the ion being formed in an ion-molecule reaction. The m/z = 47
anion may need to release the excess energy from its' formation by eliminating H2 to
from the stable m/z = 45 anion.

Lloyd et al.7 also ionized ethylene glycol using Negative Ion Chemical Ionization
with CH as the reagent ion. The DH ion NICI mass specu'um has peaks representing
ions with m/z values of 61, 59, 58, 45, and 43 Daltons. Lloyd et a1:7 attribute the source

the m/z = 58 anions and some of the m/z = 45 anions to the previously described

68

 

 

r 0“ r I: r“ r
H—C—(f—C—H > H—C KC C—H
I | N I |
OH H 0' 0 H 0'

 

m/z = 91
\ H20

 

 

 

>C:C/ ¢ \Cr—C/ \HKD
C C
n/ \H u/ \H
V H
H
H Q—C—H + HO—CI—H
\ _ / I \ I
CTC. 0H 0'

/
H K,
CZHZ m/z = 47

1, 2 migration

Figure 3.3: The Proposed Concerted Loss
Fragmentation Mechanism

69

| l
H—C—C—(l3—H H—(IAL-C—O—C— H
I
OH Ill 0H 0H ['1 1'1
/C = C\
on H
i‘
HO —(l3—- H

0 .

m/z = 47

Figure 3.4: A Proposed Formation Mechanism for
the m/z = 47 anion

70

 

 

 

V
O _
a /O
\O- \O
m/z=45 m/z=45

Figure 3.5: The Formation Mechanism of the

m/z = 45 Ion in 0' Chemical
Ionization

71

ion-molecule reactions with uaces of the 0'- ion in the NICI source. The OH‘ reagent
ion can also form the m/z = 45 anion by an ion-molecule reaction analogous to that
shown in Figure 3.5. The m/z = 61 anion is an alkoxide anion formed by proton
absu'action. The m/z = 61 anion can lose H2 to form the m/z = 59 anion. The m/z = 61
anion can also lose H20 to form the m/z = 43 anion.

Neither the OH' or the 0'- Negative Ion Chemical Ionization mass spectra of
ethylene glycol are similar to the negative ion EI mass spectrum of glycerol. The 0" ion
NICI spectrum lacks the (M-1)' and (M-3)‘ anions detected with the other two ionization
methods. In addition, the (M-4)‘ ion is of low abundance in the EI mass spectrum of
glycerol. The m/z = 58 (M-4)‘ anion is the most abundant fragment anion in the 0‘- ion
NICI mass spectrum. The EI mass spectrum of glycerol lacks anions analogous to the
m/z = 45 (M-l7)‘ and m/z = 43 (M-19)‘ anions detected by the NICI of ethylene glycol.
The only ion detected in this region of the EI mass spectrum of glycerol is the m/z = 72
ion (G-H-H20-H)‘. There is no corresponding m/z = 42 ion in the NICI mass spectra of
ethylene glycol. The m/z = 47 anion (CH302)' of glycerol is a proposed intermediate in
the NICI fragmentation mechanisms but is not detected. The comparison of the NICI
mass spectra of ethylene glycol with the EI mass spectra of glycerol did not reveal any
common fragmentation mechanisms or any unusually stable ion structures. In
conclusion, the NICI mass specu'a of ethylene glycol have few similarities to the EI mass
spectrum of glycerol due to the dominant ionization mechanism in NICI being ion-

molecule reactions.

The High Mass Anions of Glycerol

The higher mass negative ions detected in the negative ion ICR spectrum of
glycerol are ion-molecule reaction products and fragment ions of glycerol clusters.

Figure 3.6 shows the ICR mass spectrum of the high mass anions of glycerol. These

72

m/z=47

 

 

5..
W
=
H
(D
m/z = 128 :
“mm/Z : 139 :3
5"
m/z = 149 (D
E
m/z = 165 (I;
m/z = 171
m/z = 173 z
m/z = 183 “a”
m/z = 190 >
E.
G
5
m
m/z = 220 9.,
m/z = 231 Q
¢<
('5
(b
'8
9..
m/z = 277

m/z=282

73
anions have m/z values of 128, 139, 149, 165, 183, 190, 220, 231, 275 and 282. Double
resonance studies of these anions determined that the m/z = 128 and m/z = 149 anions are

not products of ion-molecule reactions.

The Evidence for the Existence of Glycerol Clusters

The existence of clusters in the gas phase is demonstrated by the detection of
anions of 128 Daltons and 149 Daltons that are not products of ion-molecule reactions.
These ions are fragment anions of clustered glycerols which are most likely dimers.
Further evidence for the existence of gas phase glycerol clusters are the ion-molecule
reactions that form the m/z = 165 and m/z = 190 anions. The m/z = 165 anion could be
formed by the m/z = 72 anion reacting with a glycerol dimer and the loss of a neutral or
neutrals of mass 91. The m/z = 165 anion is 93 Daltons heavier than its only ionic
precursor, the m/z = 72 anion. The ionic precursor to the m/z = 190 anion is the m/z = 47
anion. The only way the m/z = 47 anion can form the m/z = 190 anion in one step is to
react with a glycerol dimer or an even larger glycerol cluster.

The glycerol derived anions undergo ion-molecule reactions to cluster with
neuu'al glycerol molecules or glycerol clusters. The anion- glycerol clusters stabilize by
emitting a photon or ejecting neutral glycerol molecules. Equations 7 and 8 show the
ion-molecule reactions for an arbitrary anion (A').

A' + Gn => (G+A)' + Gn-1 (7)
A' + G => (G+A)' + hv (8)

Both the m/z = 47 and m/z = 91 anions form glycerol adducts. The resulting
adducts are the m/z = 139 anion and the m/z = 183 anion (G2-H)‘. These anions, m/z =
139 and (G2-H)‘, cluster with glycerol to form the m/z = 231 and m/z = 275 (G3-H)‘

anions.

 

74
The fragment anion m/z = 128 complexes with glycerol to from the m/z = 220

anion. The m/z = 190 ion complexes with glycerol to form the m/z = 282 anion.

A Comparison of FAB and EI Mass Spectra of Glycerol

The EI mass spectrum of the negative ions of glycerol has few similarities to the
negative ion FAB mass spectrum. The principal low mass fragment ion, m/z = 47, was
not detected by Gross et al.6 Only the m/z = 91 and 89 anions are generated by both
ionization methods. Table 3.2 is a comparison of the composition of the positive ions
generated by El with the negative ions generated by FAB. Some of the fragment ions
generated by El have a similar sequence of m/z values as the anions generated in FAB.
The m/z =91 (G-H)' is the negative ion analog to the (G+H)"' ion. The difference is that
the negative ions all have two less hydrogen atoms than their positive ion counterparts.
The fragmentation mechanisms of the anions generated in FAB may be similar to the
fragmentation mechanisms of the positive ions generated by electron impact ionization.

In the high mass ICR spectrum of glycerol, six anions with m/z values of 128,
139, 190, 220, 231 and 282 Daltons are detected but are not reported in the FAB mass
spectrum. Anions of the type (Gn-H)' were detected in both the negative ion FAB mass
spectrum and the ICR mass spectrum of glycerol. In conclusion, most of the ions formed
by FAB are formed by different mechanisms than the anions generated by the El

ionization of glycerol.

75

Table 3.2: A Comparison of the Positive Fragment Ions of Glycerol Generated in El
with the Negative Fragment Ions of Glycerol Generated in FAB

Positive Ions Generated in El

Negative Ions Generated in FAB

 

 

 

 

 

 

 

 

 

 

Mass in Daltons Compositon Mass in Daltons Composition

93 (G+H)+ 91 (G-H)‘
89 (G-3H)‘

75 C3H702+ 73 C3H502'
71 C3H302‘

61 C2H502’r 59 C2H302;__
58 CZHZOZ'

45 C2H50+ 43 C2H30'

43 C2H30+ 41 C2H0'

 

 

 

 

 

Mass-to-Charge Ratio Verification

The masses of the anions of glycerol were confirmed by mixed carbon

tetrachloride/ glycerol and carbon tetrabromide/ glycerol studies. When a halocarbon such

as carbon tetrachloride is bombarded by low energy electrons, chloride anions are formed

by dissociative electron capture. Equation 9 shows the reaction to form the chloride

anions.

CC14 + e‘ => CC13- + Cl'

(9)

These chloride anions are the only ions detected in the negative ion ICR mass spectrum

of carbon tetrachloride. The mass spectrum of carbon tetrabromide is analogous with

only bromide anions being formed. The chloride and bromide anions were used as

calibration anions to determine the masses of the fragment anions of glycerol.

76
The chloride and bromide anions complex with gas phase glycerol to form (GCl)°

and (GBr)' anions. No (G2Cl)' or (G2Br)' anions were detected. Equations 10 and 11
show the possible ion-molecule reactions for the chloride anion.

C1' + Gn => G-Cl' + 011.1 (10)

C1' + 6,1 => G—Cl' + hv (11)
The masses of the complexes were confirmed by the double resonance studies of the
halide-glycerol adducts. For example, the maximum suppression of the glycerol-
chloride35 anion occurred with a frequency ratio of 0.276 or 35/127. These ion-molecule
reactions were used to produce mass marker anions to confirm the masses of the higher

mass anions of glycerol.

References

1) Bowie, J. H. Mass Spectrom. Rev. 1984, 3, 161

2) Melton, Charles E., Principles of Mass Spectrometry and Negative Ions, Marcel
Dekker, Inc., 1970

3) Hayes, R. N.; Sheldon, J. C.; Bowie, J. H.; Lewis, D. E. J. Chem. Soc., Chem.
Comm. 1984, 1431

4) Brauman, J. 1.; Tumas, W.; Foster, R. F.; Pellerite, M. J. J. Am. Chem. Soc.
1983, 105, 7464

5) Caldwell, G.; Bartrness, J. E. Int. J. Mass Spectrom. Ion Phy. 1981, 40, 269
6) Gross, M. L.; Caldwell, K. A. J. Am. Soc. Mass Spectrom.. 1994, 5, 72
7) Lloyd, J. R.; Agosta, W. 0; Field, F. H. J. Org. Chem. 1980, 45 , 3483

Chapter Four:

The Experimental Section

The Theory of Ion Cyclotron Resonance Spectrometry

Introduction

Ion cyclouon resonance (ICR) mass spectrometry is a technique well suited to the
study of thermal energy ion-molecule reactions. Ion cyclotron resonance spectrometers
trap ions in crossed electric and magnetic fields. The trapped ions collide with gas phase
molecules and may undergo ion-molecule reactions. The ICR mass spectrometer detects
the product ions of the ion-molecule reactions and can be used to determine which of the
trapped ions reacted.

Ion trapping in an ICR is based on the phenomenon that a charged particle with a
velocity component perpendicular to an uniform magnetic field travels in a circular orbit.
The size of the orbit of the charged particle depends on its mass-to-charge ratio, its
velocity, and the strength of the magnetic field. This phenomenon was utilized to
determine the mass-to-charge ratio of a proton. Hipple, Sommer and Thomas built a
specialized mass spectrometer called an omegatron to determine the mass-to-charge ratio
of the proton in 1949.1 The omegatron was the forerunner to the modern ICR mass
spectrometer.

The first conventional ICR spectrometer was marketed by Varian and Associates
in 1966. An ion cyclotron resonance specuometer can trap ions for several milliseconds
in crossed electric and magnetic fields.2 The ions react by colliding with gas phase

molecules to produce product ions. The ICR technique of double resonance allows the

77

78

experimenter to determine which ions are the precursors to the product ions. Ion

cyclou'on resonance mass spectromeu'y has been extensively reviewed in the literature},3

Ion Motion in an ICR

Ion cyclotron resonance mass spectromeu'y is based on the behavior of a charged
particle in crossed electric(E) and magnetic(B) fields. A charged particle with a velocity
perpendicular to a uniform magnetic field(B) is confined to a circular orbit with a
cyclotron frequency, (0c, in the plane normal to B. The particle is subject to a force that
is perpendicular to both the velocity of the particle and the magnetic field. See Figure
4.1.

 

Figure 4.1: The Path of a Positive Ion in a Uniform Magnetic Field4

Equation 1 shows how the magnitude of the centrifugal force on the particle is

proportional to the charge(q), the perpendicular velocity(v) and the magnetic field.

F=qu (1)

79

The cenuifugal force on an object in a circular orbit is given by F = mv2/r. The
mass of the object is m and r is the radius of the orbit. Substitution into Equation 1 yields

Equation 2.

m2=qu (2)
r
The cyclotron frequency(in radians/sec), (0c, is the ratio of ion velocity divided by
the orbit radius(v/r). Substitution and rearrangement of Equation 2 yields Equation 3.

wc=x=ali (3)
I" m

The ICR at MSU has a fixed frequency detector and scans the magnetic field
strength to detect various m/z values. At the correct magnetic field value, the cyclotron
frequency, toe, of some ion and the detector's frequency are the same. This condition is
called resonance. Equation 4 illustrates the relationship between the mass-to-charge ratio

and the magnetic field.

B zm (4)
(I

Figure 4.2 shows the ICR cell.

80

DC Voltage ant if signal from marginal oscillator
I i +DC to -DC Square Wave

www— ---/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% Drift
Plates
1:22: /l:=::::: ,l:———'—-
/ _ — —‘ .
/ 0 (I / ; Trapping
/ Plates
'—-f A J I
Source Analyzer l Collector
Region Region Region
+DC Voltage
Electron Collector Plate

Figure 4. 2: The ICR Cell

The ICR cell has three sections. The source is the section where the ions are
generated. A beam of electrons is emitted by a rhenium filament and focused into a beam
by the magnetic field. The filament is held at a large negative voltage relative to the
ground of the ICR cell. This electric field accelerates the electrons before they pass
through the source. The ions are generated by electron impact ionization. The trapping
plates in the source are held at a dc voltage. The combined electric fields of the four
source plates create a potential well to contain the ions generated. After being formed in
the source region of the ICR cell, a perpendicular set of electric and magnetic fields
superimposes a drift velocity on the orbiting ion. Figure 4.3 shows the cycloid path of a
drifting ion in an ICR cell.

81

 

 

Drift
Plates

Trapping
Plates

Figure 4.3: The Cycloidal Path of a Drifting Ion
in the ICR Cell

 
 

 

 

 

 

 

Source Analyzer Collector

The drift velocity(vd) is calculated using Equation 5.

Vd = E/B (5)

The ions drift through the analyzer section of the ICR cell and then into the collector
region.

The ions are generated in the source region of the cell and drift out of the source
region continuously. This mode of ICR operation is called the drift mode. The time the
drifting ion takes to traverse the cell can be determined.5

The analyzer section is where the ions are detected. The ICR's detector is a
marginal oscillator. The ions are detected in the analyzer section by measuring the power
absorbed by the ions in resonance with the detector. The ac signal from the marginal
oscillator is added to the dc voltage applied to a drift plate in the analyzer section.

The trapping plates in the analyzer section have different voltages applied to each
plate. The bottom trapping plate is held at a dc voltage. A 20 Hz square wave is applied

to the top trapping plate. The voltage of the square wave varies from the -dc to the dc

82
voltage applied to the bottom trapping plate. The purpose of the square wave is to

improve the signal-to-noise ratio by signal modulation.
The last section of the ICR cell is the collector section where the ions are
neutralized by a pair of collector plates. The collector plate current is measured to

determine the number of ions traversing the cell and is a great help in tuning the ICR.

Ions in Resonance

The basis of ion detection in ICR is the phenomenon of resonance. If an orbiting
ion is irradiated with an rf field that has the same frequency as the ion's cyclotron
frequency, the ion absorbs energy from the rf field. The resonant ion's kinetic energy and
velocity increase. Figure 4.4 shows the path of a resonant ion. The marginal oscillator
detects the ions by measuring the amount of energy absorbed by the ions in resonance.
The amount of energy a resonant ion can absorb is limited by the dimensions of the ICR
cell. As the ion absorbs power, the radius of the ion's orbit must increase in order to
maintain the same cyclotron frequency. When the ion strikes the walls of the ICR cell it
is neutralized. The amount of energy absorbed by resonant ions is proportional to their

number and m/z value.

83
rfsignalfrommemarginaloscillatortodriftplate

/l° r; / /
4; // j w:

Analyzer Collector

 

 

 

 

Figure 4.4: Ion in Resonance in the ICR cell

Double Resonance Experiments

Irradiating a resonant ion until it is neutralized is called ion ejection. Ion ejection
is the basis of the double resonance technique for determining the precursors of the
product ions. The double resonance technique uses two rf fields to place ions in
resonance. The product ion is monitored by the fixed frequency of the marginal
oscillator. The frequency of the second rf field is varied, ejecting the potential precursor
ions sequentially according to their resonant frequencies. When the precursor ion is
ejected, fewer product ions are formed to be detected by the marginal oscillator. The
double resonance spectrum shows a decrease in the relative intensity of the product ion
when a precursor ion is ejected.

Figure 4.5 is a typical spectrum from a double resonance experiment. Figure 4.5
is a plot of the number of ions detected(the y-axis) vs. the frequency ratio(the x axis).
This frequency ratio is the ratio of the frequency of the irradiating rf voltage divided by
the frequency of the rf voltage of the marginal oscillator. When the irradiating oscillator
is at the frequency that ejects the precursor ions, the number of product ions detected is

reduced. This causes the downward spike seen in Figure 4.5. The frequency ratio at the

84

minimum of the spike corresponds to the ratio of the m/z values of the precursor ion
divided by the product ion. In Figure 4.5, the precursor ion is (G + H)+ (m/z = 93) and
the product ion is (G2+ H)+ (m/z = 185). The ratio of the m/z values is 93/185 = 0.5040.
This is the frequency ratio measured at the minimum of the double resonance spectrum in

Figure 4.5.

Figure 4.5: A typical ICR Double Resonance Starting Frequency Ratio
Spectrum. The m/z = 185 ion has the m/z = 93 0,1300
ion as its precursor.

 

 

 

 

 

, 05040 Number
Ending of Ions
Frequency Detected
Ratio

0.”)

Ratio of the frequency of m; m 'at_ing rf voltage

frequency of the rf voltage of the marginal oscillator

85
The typical double resonance experiment starts by ejecting ions with m/z values

of 18% of the m/z value of the product ion and sequentially ejects heavier ions up to

precursor ions with m/z values of 90% of that of the product ion.

The Marginal Oscillator

The ions in the ICR cell are detected by a marginal oscillator. To detect the ions,
the marginal oscillator creates a rf field of a known radio frequency into the cell. The
resonant ions absorb energy from the rf field and the marginal oscillator measures the
amount of energy absorbed. The equivalent circuit of the marginal oscillator is shown in

Figure 4.6.6

 

Current
Source

8 _C_Z_ell fi_
é R L F C) Detector
E c

 

 

 

 

 

 

 

 

 

 

 

 

Figure 4.6: The equivalent circuit of the
LC tank circuit used in the Marginal
Oscillator Detector

The constant current generator I drives a parallel LC tank circuit at its resonant

frequency 030. The plates of the ICR cell form part of the capacitance of the LC circuit.
When the ions in the ICR cell are in resonance with the LC circuit a voltage difference is

generated across the detector(D). The voltage difference(AV) is given by equation 6,5

86
AV .-. (QA/Vo) WC)“2

(6)

where A is the power absorbed by the ions, V0 is the voltage of the tank circuit when no

ions are present and Q is the quality factor of the circuit. Equation 7 is the definition of

Q for a parallel LC circuit.5

Q = R/(ooL = R(C/L)l/2

(7)

A block diagram of the marginal oscillator detector used in the ICR is shown in

Figure 4.7.5 When ions are in resonance, the demodulator outputs a 20 Hz square wave

signal that is observed on the oscilloscope.

 

IDemodulator giutnal

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

rf output

I— 1" I

l Step and Three I Complementary I

| Current Variable Stage T— Emitter I

Drive Attenuator Limiter Follower

l l I

I l |

L ________ _ _ _ _ I l
-l

I I l l

I Tank Circuit I I I

_..l_ I J10 OF I _ 20 OF I FET I Cascade I

ICR I; —: |_ :E—l— Preamp I Gain Control |
Cell 2 2 100 £337 Freq C I

l T Adjust T I l l

l_ _ _ _ _ _ _* ______ | ______ _J

Figure 4.7: Block Diagram of the Marginal Oscillator

Block Diagram of the ICR

87
Figure 4.8 is the block diagram of the ICR at Michigan State University.

 

 
  

 

  
 

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

t’l arglnal M agnet
Oscillator Cell — '
, N
Dual Trace
Oscilliscope
Errission Magnet Power
Square Wave Selector Box Current Supply 2
Cenerator and BM M Controller I
Sweep
I Voltage I Control I

 

 

 

 

 

 

 

Lock-in . Smply
‘ Amplifier Recorder I

Figure 4.8: The Block Diagram of the Ion Cyclotron Resonance
Spectrometer at Michigan State University

The ICR has a cell located in a t-shaped vacuum chamber between the poles of the
electromagnet. The vacuum chamber is connected to a right angle isolation valve by a 4
inch stainless steel line. A refrigerated trap and a Varian 4 inch diffusion pump is used to
pump the chamber down to 5.0 x 10'7 torr. The normal operating pressure of the ICR is
5.0 to 15.0 x 10'6 torr. The pressures are monitored using a Bayer-Alpert ionization
gauge and a Veeco Instruments model 1000/1002 gauge controller.

A leak valve and a sample vial is attached to the bottom of the vacuum chamber.
This leak valve was used to leak chlorobenzene and bromobenzene into the ICR to
calibrate the mass scale of the ICR in the negative ion mode. The chlorobenzene and
bromobenzene are a source of chloride and bromide anions in the negative ion ICR

experiments.

88

bromobenzene are a source of chloride and bromide anions in the negative ion ICR
experiments.

The m/z values of the positive ions of glycerol were determined by mixing
naphthalene with the glycerol. The positive ions of the naphthalene were used as mass
markers to calibrate the mass range of the ICR.

The glycerol sample is placed in a capillary tube and spun with a cenuifuge. The
capillary tube is inserted into the ICR on a direct insertion probe. The probe tip was
heated using a TRIAC controller. There was no regulation of the temperature of the
probe tip. The TRIAC conu'oller was set to a fixed value determined by the pressure of
the glycerol desired in the ICR. The inlet tube to the ICR cell was heated to prevent

condensation of the glycerol.

Chemicals Used in the Experiments

The glycerol was reagent grade and manufactured by Matheson, Coleman and
Bell. The 1,1, 2, 3, 3-ds glycerol was obtained from Cambridge Isotope Laboratories.
The deuterium content was reported as 98%.

Naphthalene from Mallinckrodt, Inc. was used as a mass marker to calibrate the
mass range of the ICR in the positive ion experiments. Chlorobenzene and
bromobenzene from Fisher Scientific was used to calibrate the mass range of the ICR in

the negative ion experiments.

REFERENCES

1) Sommer, H.; Thomas, H. A.; Hipple, J. A. Phy. Rev. 1951, 82, 697

2) Lehman, T. A.; Bursey, M. M. [on Cyclotron Resonance Spectrometry, Wiley,
New York, 1976

3) Budzikiewicz, H. Mass Spectrom. Rev. 1987, 6, 329

4)

5)
6)

89

Halliday, D.; Resnick, R. Fundamentals of Physics, 3rd Edition, John Wiley &
Sons, 1988, 694

McMahon, T. B.; Beauchamp, J. L. Rev. Sci. Instrum. 1971, 42, 1632
Warrnick, A.; Anders, L. R.; Sharp, T. E. Rev. Sci. Instrum. 1974, 45 , 929

Chapter Five:

The Appendix

Detailed Description of the ICR Components

This appendix is a short description of the electronic components that make up the
ICR at Michigan State University. The procedure to obtain a typical spectrum is also

given.

W The ICR uses a Varian V7300 12" electromagnet. The elecu'omagnet
has a normal operating range from 0 to 23 kilo Gauss. The magnet is controlled by a
Varian FIELDIAL magnet controller which senses the magnetic field using a Hall probe.
The magnet was calibrated and its true field(y) can be calculated from the field dial
setting(x) by y = 1.2572 x + 0.005.

W The V 7X00/E-7X00 regulated Magnet Power Supply is
designed for spectrometers requiring a stable magnetic field. The power supply can
power the electromagnet to field strengths of up to 23 kilo Gauss. The magnet and the
power supply are water cooled by a Haskris heat exchanger which exchanges heat with
by tap water. The normal operating temperature of the magnet and the power supply is

between 16 and 27 degrees centigrade.

91

W The emission current controller was designed and built by
Marty Rabb. The controller regulates either the filament current(filament mode) or the

current of the ionizing elecuons passing through the ICR cell(feedback mode). The
emission current controller maintains a negative bias on the filament with reference to the
ICR cell ground. The bias voltage is supplied by a Hewlett-Packard 6299A DC power

supply through a BNC connector on the emission current controller.

W A Keithley model 480 picoammeter is used to measure the
ionizing current passing through the ICR cell. The picoammeter converts the current into

a DC voltage used by the emission current controller to regulate the ionizing current.

W The voltages applied to the eight plates of the ICR cell are
controlled from a panel on the front of the ICR. The four drift plates have dc voltages
from -3 to 3 volts supplied from the control panel. The ICR cell source trapping plates
are both have the same dc voltage of -10 to 10 volts selected from the control panel. The
analyzer section trapping plates have one trapping plate regulated to a dc voltage. The
other trapping plate is pulsed with a 20 Hz square wave with an amplitude that varies

from the -dc to the dc voltage applied to the other trapping plate.

Link-23mg The "little box" mounts on the 20 pin feedthrough and makes the
connections from the control panel to the ICR cell. The rf voltage used to do double
resonance experiments is added to one or both of the top drift plate dc voltages through
the little box. The filament on/off switch and connections for filament power and for the
ionizing current measurement are on the front of the little box. The little box supplies the
dc voltage to the marginal oscillator. This dc voltage is added to the rf signal applied to

the analyzer section's bottom drift plate. A ribbon cable from the little box to a selector

92
box on the control panel allows the signals actually applied to the ICR plates to be

monitored.

Wang A Tektronix DM 502A autoranging digital multimeter is used to
monitor the signals being sent to the plates of the ICR cell. The plate to be monitored is
selected by a set of eight switches in a selector box. The selected plate voltage signals

can also be displayed on the oscilloscope.

W A Tektronix CFG 250 was used as a rf voltage generator to supply
the variable frequency rf voltage used in double resonance experiments. The CFG 250
has a voltage controlled frequency output. The frequency is swept by applying a dc
voltage from the x-axis of the chart recorder to the voltage cont1olled gain port of the

CFG 250. The irradiating rf voltage's amplitude is adjustable.

W A Tektronix DC 503 Universal Counter is used to determine the
frequency of the rf signal generated by the marginal oscillator. The DC 503 is also used
to determine the ratio of the frequencies of the rf signal of the marginal oscillator to the rf

voltage of the double resonance oscillator.

W The ICR's detector is a marginal oscillator. The marginal oscillator
detector was built by Marty Rabb. The frequency and the amplitude of the sine wave in
the marginal oscillator is variable. The normal frequency is 138 kilo Hertz. Lower
frequencies are used to detect higher mass ions. The marginal oscillator is attached to the

top drift analyzer plate through a BNC connector on the 6 inch flange.

93
W A Tektronix FG 504 function generator is used to generate the 20

Hz square wave that is used to pulse the trapping plates and to provide the reference

signal to the lock-in amplifier.

W An EG&G Princeton Applied Research model 128A lock-in
amplifier is used to amplify and improve the signal to noise ratio of the signal from the
marginal oscillator. The lock-in amplifier uses the same 20 Hz square wave as the
trapping plate circuits as a reference signal. The 128A subtracts the signal it receives
from the marginal oscillator at the top of the square wave from signal the 128A receives
from the marginal oscillator at the bottom of the square wave. The difference in signal
intensities is the signal due to the detection of ions. The output of the lock-in amplifier

drives the y-axis of the xy chart recorder.

W A Soltec xy recorder is used to plot the results of the ICR experiments.

Oscilloscope A Tektronix model 455 dual trace oscilloscope is used to display the
output of the marginal oscillator. The amplitude of the irradiating rf voltage can also be
determined. The voltages and wave forms acrually reaching the plates of the ICR cell can

also be observed.

We: The Keithley model 610C electrometer is used to monitor the
current of the ions that drift through the ICR cell. The electrometer is connected to four
ion collector plates at the bottom of the ICR cell through a BNC connector on the 6 inch

flange.

Ionizationfiaugeflonnnller A Veeco RG 1000 ionization gauge controller with a

Bayer~Alpert ionization gauge is used to monitor the pressure in the ICR.

94
The Procedure for a Typical ICR Experiment with Glycerol

Positive Ion Spectra
The glycerol sample in a capillary tube is placed on a direct probe and inserted
into the heated inlet arm to the ICR cell. Figure 5.1 shows the sample probe and the ICR

inlet arm. The TRIAC conuoller is turned on and adjusted until the desired pressure of

glycerol is achieved.

To Vacuum Pump

Heated Chamber

To ICR
Cell

 

Glycerol in
Capillary Tube

Figure 5.1: The Sample Inlet to the ICR

The filament is turned on and the ionizing current begins to flow. The magnet is
scanned to the m/z value of one of the more intense ions and the plate voltages are
adjusted to give a large peak height. The filament voltage varies from -50 to -70 volts for
positive ions and is adjusted so as to yield the greatest number of ions. The amplitude of
the rf signal of the marginal oscillator is adjusted to as small a value as possible to give
the best resolution. The amplitude of the rf signal is typically 50 millivolts.

The magnet is set to the starting field strength(normally 0.5 or 7.5 kG) and is
scanned from low field to high field(ending at 9.5 or 16.5 kG). A typical scan rate is 1

kG per minute. The ICR mass spectrum is recorded on the chart recorder.

95
The higher mass ions above 220 Daltons are detected by reducing the frequency

of the marginal oscillator from its normal value of 138 kilo Hertz to approximately 100
kilo Hertz. The resolution of the ICR degrades so that unit resolution for ions with ni/z
values greater than 300 Daltons is not always possible.

To perform a double resonance experiment the magnetic field is first set to the
value corresponding to the product ion being investigated. The sweep conuols of the
magnet are turned off and the % sweep dial is turned to its highest setting. The CFG 250
is tuned on with normal settings of 450 millivolts and 1 mega Hertz for the rf voltage.
After the frequency ratio of the rf voltage from the CFG 250 vs. the rf voltage of the
marginal oscillator is measured by the DC 503, the frequency ratio is adjusted to a
starting ratio of 0.18. Then the double resonance scan is started. The frequency ratio of
the rf voltages of the CFG 250 and the marginal oscillator typically varies between 0.18
and 0.90. The precursor ions that have m/z values between 18% and 90% of the product
ion's m/z value are detected. When a precursor ion is ejected by the double resonance rf
voltage the intensity of the product ion is reduced creating a downward spike on the plot.

The starting frequency ratio can be reduced to values as low as 0.10 if needed.

Negative Ion Spectra

The negative ions of glycerol were formed by resonant and dissociative electron
capture. The electrons were emitted by the hot rhenium filament and accelerated by a
1.00 to 8.00 volt difference between the filament voltage and the trapping plate voltages.
The voltages of the trapping plates were negative to prevent the negative ions formed
from being neutralized. The voltages of the drift plates were adjusted to drift the negative
ions to the analyzer region.

At the low acceleration voltages used the elecu'on beam is difficult to control. In

the normal negative ion spectra the filament current is set to a constant value. The low

96
energy electrons are either captured by the glycerol or spread out, hitting the trapping

plates. The electron current detected is normally less than 1 nanoamp. All of the other
tuning, spectra collection and double resonance experiments are the same as for the

positive ions.