g
.DMW...” .
.. kw

 

- ..
a .3 .1
.-...u?&.m

5.5..

 

2!»?:
3.5.31

 

 

.. 4‘

'¥

.1. V. v)“ beir... DALI:

1‘30: I... K.)

 

 

iv

A

rHESiS
7.

MIICH IGAN STAT TEU

11111121!!! WI!!!'(WUHlW.l}”ll'IUUHHII

01413 8881

LIBRARY

Michigan State
University

ill)

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Spectroscopic Studies
of Mass-Selected,
Matrix-Isolated Cations

presented by

Jerry T. Godbout

has been accepted towards fulfillment
of the requirements for

 

Ph .D . degree in Chemistry

16:? ~' ’

‘ x. . J 3 .’.‘ 4. '
(“A Major drgfes/sor

June 26, 1996
Date

 

MS U is an Aflirman've Action/Equal Opportunity Institution 0-12771

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

  

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSUJe An Affirmative Action/Equal Opportunity Inetltwon
mount

 

SPECTROSCOPIC STUDIES OF MASS-SELECTED, MATRIX ISOLATED
CATIONS

By

Jerry T. Godbout

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1 996

ABSTRACT

SPECTROSCOPIC STUDIES OF MASS-SELECTED, MATRIX ISOLATED
CATIONS

By

Jerry T. Godbout

The elucidation of the structures of polyatomic ions has been a daunting task over
the years. Attempts at structure determination have generally employed one of two
experimental methodologies: molecular spectroscopy or mass spectrometry. While
molecular spectroscopy yields structural information directly, the high reactivity of ionic
species can make direct spectroscopic measurements exceedingly difficult. Whereas the
sensitivity of mass spectrometric methods facilitates the detection of ions, structural
information is indirectly and incompletely obtained at best.

We have merged mass spectrometric techniques with matrix isolation spectroscopy
to allow the vibrational spectra of polyatomic ions trapped in rare gas matrices to be
observed. In this method, ions are made in a conventional electron impact ionization
source and mass selected using a quadrupole mass filter. The mass-selected ion beam is
then guided to a 4 K cryogenic window where it is co-condensed with excess neon. The
trapped ions are then analyzed using conventional Fourier Transform Infrared (FTIR)
spectroscopy.

Initial feasibility studies were performed using small cations with known IR
spectra. The cations were observed without the concurrent deposition of anions, although

negative charge carriers were indirectly detected. Further experiments with trapped beams

of C02+ generated from C0; have allowed the issue of charge neutrality to be addressed.
The addition of CO; to the system results in the production of C02“ and (C02)2'. The
addition of C02 also seems to enhance the concentration of cations stabilized in the matrix
by a factor of ~5. The origin of these anions was found to be neutral C02 molecules
adsorbed onto various cold copper surfaces surrounding the matrix window. Issues such
as fragmentation of the mass-selected cations the use of visible emission as an alternate

detection method are also discussed.

Dedicated to Mom, Dad, and Joe
and in loving memory of

Grandma St. Pierre

iv

ACKNOWLEDGEMENTS

Although this dissertation bears the name of only one author, the work described
in it was made possible by the combined efforts of many people. First and foremost, I
would like to thank Professors George Leroi and John Allison for their never-ending
support, always-accessible guidance and countless suggestions when things weren’t
working. I should also thank them for their patience when I didn’t follow those
suggestions, especially when they were right after all. Working for two advisors can be a
mixed blessing, but I think that the benefits outweigh the disadvantage of having to
appease two people. Besides the obvious advantage of being able to draw from two fairly
distinct scientific backgrounds, their also distinct views on life in general gave me a
hopefirlly more balanced perspective. Interaction with and support from the other
members of the Leroi and Allison groups was also invaluable.

I once heard someone (J. P. Maier I believe) describing experiments similar to ours
as “conceptually simple, but a technical nightmare”. On that note the people who kept us
up and running need to be thanked. The machine shop stafl‘ (Russ Geyer, Dick Menke,
and Sam Jackson) always amazed me with their ability to translate what I asked for into
what I actually needed, and their handling of my jobs that always seemed to be dire
emergencies (well, they were to me). The electronics support fi'om Ron Haas and Scott
Sanderson constantly kept the electrocution hazard we called and instrument from going

up in flames, or at least diagnosed the cause and replaced the damaged components when

it did. I am sorry to say that I also need to thank the staff of the glass shop (Scott

Bancroff, Manfred Langer, and Keki Mistry). I make this acknowledgment with

reservation not because of the quality of their work (which was superb and should be

considered art) but because needing their help usually meant that I had just broken
something.

Finally, as far as science goes, I need one more. This work would not have been
possible without the dedicated efl’orts of my labmate, Elvis Presley (currently assuming the
identity of Tom Halasinski). Along with his many scientific contributions (such as the ion
source) his unique presence made all those 30 hour experiments almost tolerable, and
always interesting.

Science isn’t everything though, and on that note:

0 Dan: We never seemed to reduce Michigan’s fish and pheasant population as much
as we wanted to, but oh well.

0 Paul and Deena: Now whose going to feed me?

0 Matt: For putting up with me (I guess it’s not my lab anymore), helping with flood
prevention/recovery and most notably for your assistance in our brief stint in
aerodynamic engineering.

0 Phil: Ifall Republicans were like you, maybe, just maybe.

0 Wanda: Sometimes I still wish things had been different (one way or the other), but
I’m glad we’re still fiiends and I wish you the best.

0 Jefi‘ and Claudia: Support, advice, and fiiendship early on, even during Jeff’s “plenty

fiiendly” period.

Michelle, Don, and Donny: Home life worth leaving the lab for.
Kerry and Rob: Legendary Christmas parties.

Overall, I’d have to say that grad school was OK, but it was stressful at times. For

helping relieve that stress, I’d like to thank (in no particular order):

Any of the “Happy Hour” regulars not already mentioned (Michelle, Chris, Sara,
Nocera, Gerry).

The programmers at id Software, and Netscape Communications Corp., without whom
I would have graduated months earlier.

The gang at Dagwood’s (Brenda, Tee, and Randy) especially for their assistance
during the actual writing of this dissertation.

The Michigan DNR, the game they manage, and the Remington Arms Company.

vii

TABLE OF CONTENTS
List of Figures

Chapter 1: Introduction
I. The Spectroscopy of Polyatomic Ions
A Early Spectroscopic Studies of Polyatomic Ions
B. Laser-Based Gas-Phase Techniques
C. Matrix Isolation Studies
11. Matrix Isolation Spectroscopy of Polyatomic Ions
A. Ion Generation Techniques
VUV Photolysis
Windowless Discharge
Ion Molecule Reactions
Chemical Ionization
Pulsed Glow Discharge
Electron Bombardment
Mass Selected Deposition
riteria for Matrix Isolation
Inertness
Rigidity
Purity
Transparency
Thermal Properties
6. Volatility
C. Trapping Ions in Matrices
1. Deposition Rate
2. Matrix to Guest (MzG) Ratio
3. Counter Ions
D. Matrix Effects on Ionic Spectra
1. Number of Peaks
2. Frequency Shift
3. Spectral Distortion
4. Vibrational Relaxation
List of References

MFPPPONP‘MPP’NP

Chapter 2: Instrumental
1. Ion Source, Mass Filter, and Ion Optics
II. Vacuum System
A Vacuum Chamber
B. Pumping System

viii

><

MWMNI-i

ll
12
12
13
13
14
15
16
16
16
17
18
18
18
18
19
20
21
22
22
23
24
24
26

29
29
30
32
33

C. Vacuum Interlock
D. Pumping Procedures
E. Faraday Plate Modifications
III. Cryostat
A Refi'igerator
B. Temperature Control
C. Radiation Shield Modifications
D. Electrical Isolation of the Sample Window Holder
E. Sample Windows
IV. Matrix Gas Line
V. FTIR System
A FTIR Spectrometer
B. Interfacing of the Spectrometer to the UHV Chamber
VI. “Allions” Measurements
VII. Counter Ion Measurements
VIII. Matrix Emission Studies
List of References

Chapter 3: Preliminary Experiments: The Infrared Detection of
Mass-Selected, Matrix Isolated CF;+
I. Background
11. Experimental
A. Choice of Test Cation
B. Instrumental Parameters
III. Results and Discussion
A FTIR Results
B. Indirect Detection of Counterions
List of References

Chapter 4: The Formation of Counterions
I. Background
11. Experimental
III. Results and Discussion
A Spectroscopic Detection of Counterions
B. Indirect Detection of Counterions
C. Evaluation of Counterion Generation Mechanisms
D. Direct Observation of Anion Generation
E. Discussion
IV. Conclusions
List of References

Chapter 5: Visible Emission From Warming Neon Matrices or
It Seemed Like a Good Idea at First
I. Background
11. Experimental

ix

35
35
37
37
39
43
45
49
50
50
S4
54
55
58
60
61
68

69

69
70
70
71

72
78
84

85
85
87
91
91
95
100
105
113
115
119

121

121
125

III. Results and Discussion
A. Total Emission Studies
B. Dispersed Emission Studies
1. Ar”
2. CF;
3. C82”
4. OCS'+
5. CH3+
C. Discussion
IV. Conclusions
List of References

Chapter 6: Conclusions and Future Work
1. Conclusions
11. Instrumental Modifications
A FTIR System Improvements
1. Gain-Ranging Amplifier System
2. Electronic Spectroscopy of Matrix-Isolated Species

127
127
130
130
133
139
145
149
153
155
157

159
159
160
160
160
161

1.1

1.2

2.1

2.2
2.3

2.4

2.5

2.6

2.7

2.8

2.9

2.10

2.11

2.12

2.13

LIST OF FIGURES

A generic matrix-isolation spectroscopy experiment.

Schematic representation of a mass-selection matrix-isolation
experiment.

Schematic top view of mass-selected cation source based on modified
Finnigan 3200 quadrupole mass spectrometer.

UHV pumping system
Faraday plate assembly

Heliplex piping schematic. Pressures and temperatures shown are with
the system operating at minimum temperature.

Typical Heliplex" cooldown. Temperature is measured at the fixed
third-stage sensor.

Adaptation of the Heliplex0 cryostat to the UHV chamber.

Attachment of temperature control diodes to cryostat and optical ring
sample holder.

Modified and original radiation shield assemblies. Grid shown on
spectroscopic window was not present for all experiments.

Cross-sectional view of spectroscopic window holder modified for
electrical isolation.

Vacuum manifold used for matrix gas preparation and introduction into
UHV chamber.

Instrument for FTIR spectroscopy of mass-selected, matrix-isolated
cations showing UHV chamber, ion source, and FTIR spectrometer.

Schematic description of “Allions” experiments. Components marked
optional were not used in all experiments.

Schematic illustration of anion measurement experiments.

6

9

31

34
38

41

42

44

46

48

51

52

57

59

62

2.14

2.15

2.16

3.1

3.2

3.3

3.4

3.5

3.6

3.7

4.1

4.2

Schematic diagram of instrumentation used to collect total emission
from a warming matrix.

Schematic diagram of original instrumentation used to collect dispersed
emission from a warming matrix. The effect of undispersed light is
shown.

Schematic diagram of instrumentation used to collect dispersed
emission from a warnring matrix.

Antisymmetric stretch (v3) of mass-selected, matrix-isolated CF;+ in
neon.

FTIR absorption spectrum in the 4800—900-cm'l region of a pure neon
matrix after 11 h of deposition of CF;+ generated from CF3C1.

FTIR absorption spectrum in the 1700—900-cm'l range of a pure argon
matrix after deposition of CF31 fiom CF3C1.

Infrared absorption spectra in the 1700—1640 cm" region of neon
matrices after 11 hours of deposition of: a) m/z 69 (CF32 20 nA)
generated from electron impact of CEO; b) m/z 120 (no ion current)
while subjecting CEO to electron impact; and c) m/z 50/51
(CFf/CFgI-I”, 15 nA) generated from electron impact of CF3H.

FTIR absorption spectrum in the 970—920-cm'l range of a pure argon
matrix after deposition of CF 3+ generated from CF3C1.

Transient current observed at the Faraday plate located ~l cm from the
substrate during warming of a neon matrix after a 25 hour deposition of
20 nA of CF3+ generated from electron impact of CEO. The initial
temperature of the matrix was ~45 K The final matrix temperature of
~20 K is reached within 10 seconds.

Transient current observed at the Faraday plate located ~l cm from the
substrate during warming of a argon matrix after a 25 hour deposition
of 20 nA of CF 3+ generated from electron impact of CF3Cl, followed by
2 hour bombardment by 40 nA of 10 eV electrons The final
temperature of the matrix was ~40 K.

Illustration of proposed condensed phase mechanism of counterion
generation in mass-selection, matrix-isolation experiments.

Illustration of proposed gas phase mechanism of counterion formation
in mass-selected, matrix-isolated cation experiments.

xii

63

65

67

73

74

76

77

80

81

83

88

89

4.3
4.4

4.5

4.6

4.7

4.8

4.9

4.10

4.11

Illustration of proposed surface bombardment mechanisms of counter
ion formation in mass-selected, matrix-isolated cation experiments.

Infiared absorption spectrum of a neon matrix in the 1750—1350 cm'1
region following a 20 h deposition of 40 nA of C02".

Infrared absorption spectrum of a 100021 Ne:CO; matrix in the 1750—
1350 cm'1 region following a 13 h deposition of 20 nA of CF3+ formed
by electron impact of CEO.

Currents observed on the Faraday plate from a warming neon matrix
after a 27 h deposition of 40 M of C02”. The cryostat heater is turned
on a t = 0 8. At t = 30 s, the temperature was ~40 K. The potential on
the Faraday plate during the collection of the signal was -70 V with the
window holder grounded.

SIMION plots showing the potential surface formed between the
Faraday plate and the CsI window. Equipotential lines, separated in
value by 10 V, are shown. The CsI window is represented by a dashed
line in both plots. The grounded copper grid has been added to part b.
The Faraday plate has a pctential of -70 V, and all other surfaces are at
a potential of 0 V in both plots.

Currents obServed from a 1700:l Ne:C02 matrix after a 20 h deposition
of 30 M of C-IH7+ generated by electron impact ionization of toluene.
Heating conditions are the same as Figure 4.6. The potential of the
Faraday plate during collection of the signals was +70 V with all other
elements held at O V.

Infrared absorption spectrum of a 150021 Ne2C02 matrix in the 1300—
1200 cm'1 region following a 22 h deposition 35 nA of 130 eV Can”
ions (top trace) and similar amounts of 50 eV CaFa” ions (bottom
trace). In both cases, the CaFa" ions were generated from electron
impact of Can.

Infiared absorption spectrum of a 1500:1 NCIC02 matrix in the 1450—
1400 cm'1 region showing C02°+ formed when a copper grid was used
to decelerate incoming CaFa” cations. Depositions conditions are
identical to those represented by the bottom trace of Figure 4.9.

Direct measurement of anion generation, Configuration 1. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

xiii

90

92

96

98

99

102

104

106

4.12

4.13

4.14

4.15

5.1

5.2

5.3

5.4

5.5

5.6

5.7

Direct measurement of anion generation, Configuration 2. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

Direct measurement of anion generation, Configuration 3. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

Direct measurement of anion generation, Configuration 4. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

Illustration of the proposed processes occurring during deposition of a
100 eV positive ion beam and neon on the CsI sample window and the
metallic radiation shield.

The temperature of the cryostat window holder as a firnction of time as
the cryostat is turned off. The cryostat was turned off approximately
10 s after data collection was started. This temperature profile was
obtained concurrently with the emission data shown in Figure 5.3.

Total emission 1 observed during warming of a 100021 Ne2C02 matrix
after the deposition of €ng generated fiom CF3Cl.

Total emission observed during warming of a 1000:1 Ne2C02 matrix
after the deposition of CS2°+ generated from C82. This matrix was
subject to several short annealing cycles before being warmed
irreversibly.

Dispersed emission observed during warming of a 100021 Ne2C02
matrix after the deposition of Ar” generated from Ar. This single
spectrum is representative of the emission observed throughout the
entire warming.

Contour plot showing temporal profile of the dispersed emission
observed during the warming of a 1000:] Ne2C02 matrix after the
deposition of Ar'+ generated fi'om Ar.

Tentative assignments of Argon emission.

Contour plot showing temporal profile of the dispersed emission in the
350-650 nm range observed during the warming of a 100021 Ne2C02
matrix after the deposition of CF 3+ generated fi'om CF3C1.

xiv

107

108

109

116

126

128

129

131

132

134

135

5.8

5.9

5.10

5.11

5.12

5.13

5.14

5.15

5.16

5.17

5.18

Contour plot showing temporal profile of the dispersed emission in the
450-750 nm range observed during the warming of a 1000:1 Ne2C02
matrix after the deposition of CF; generated from CF3Cl.

Dispersed emission observed fi'om a warming 100021 Ne2C02 matrix
after deposition of CF; generated fi'om CF3CI. Spectra shown are
representative of the emission observed during the entire 5 nrinute
warrnup. The blue trace was taken from the data set shown in Figure
5.7 and scaled 2x. The red trace was taken fi'om the data set shown in
Figure 5.8 and not scaled

Visible CF3° emission reported by Washida et. a1. Spectrum reproduced
fi'om reference 13. Bottom trace (e) is the discharge lamp reference.

Visible CF; emission reported by Koda. Spectrum reproduced from
reference 15.

Contour plot showing temporal profile of the dispersed emission
observed during the warming of a 100021 Ne2C02 matrix afier the
deposition of CS2" generated fi'om CSz.

Dispersed emission observed during warming of a 100021 Ne2C02
matrix after the deposition of CSz'+ generated from C82. This single
spectrum is representative of the emission observed throughout the
entire warming.

Visible CS emission reported by Coxon and co-workers. Spectrum
reproduced fiom reference 18.

Contour plot showing the temporal profile of the dispersed emission
observed during the warming of a 100021 Ne2C02 matrix after the
deposition of OCS" generated from OCS.

Dispersed emission observed during warming of 3 100021 Ne2C02
matrix after the deposition of OCS'+ generated fi'om OCS. This single
spectrum is representative of the emission observed throughout the
entire warming.

Visible CO emission shown by Herzberg. Spectrum reproduced from
reference 22.

Contour plot showing temporal profile of the dispersed emission
observed during the warming of a 1000:1 NezC02 matrix alter the
deposition of CH3+ generated from CHaBr.

136

137

138

140

141

142

144

146

147

148

150

5.19

5.20

5.21

6.1

Dispersed emission observed during warming of a 100021 NezC02
matrix after the deposition of CH; generated from CH3Br. This
spectrum was taken approximately 60 s after the cryostat was turned
OE, and is representative of the early emission when the 63 7-nm band is
dominant.

Dispersed emission observed during warming of a 100021 Ne:COz
matrix alter the deposition of CH; generated from CH3BI'. This
spectrum was taken approximately 100 s after the cryostat was turned
ofi and is representative of the late emission when the 575-nm band is
dominant.

Portion of the visible CH2 emission reported by Garcia-Moreno and
Moore. Spectrum reproduced fi'om reference 25.

Schematic diagram of proposed modifications to accommodate both
FTIR and WW is absorption and lurrrinescence measurements of matrix
isolated species.

151

152

154

162

Chapter 1

Introduction

Polyatomic ions have long been of interest to chemists. They are thought to be
quite common in interstellar space1 and are also thought to play an important role in many
atmospheric processes.2 The most common technique used for the terrestrial study of
polyatomic ions is mass spectrometry. Mass spectrometric studies can provide a wealth of
kinetic and thermodynamic data, but generally provide structural information only
indirectly. A classic example is the question of the structure of the ions with the formula
C2H30+ (m/z = 43) formed from many organic molecules. The two most stable predicted

geometries for these ions are the following:

‘1 + 1 +
H\ _ H... /H
,_,.C-C=O ('C=C:O
HH- H
a b

Structure a is thought to be the more stable,3 and is generally referred to as the acetyl
cation. Structure b, referred to as oxygen-protonated ketene, is thought to be less stable
by ~42 kcal/mole.4 To determine if different precursors would generate C21130+ ions
having different structures, Weber and Levsen5 performed a series of tandem mass
spectrometric experiments on a wide variety of compounds that produce a fragment ion
with m/z 43. They determined that the m/z 43 fragment ion generated from electron

impact of glycerol had a chemistry different than that of the m/z 43 ion generated from

2

other carbonyl-containing compounds such as acetone. This difierence in chemical
behavior implied that the ions had two different structures, but it did not, however,
determine their geometries. From mechanisms based on condensed phase organic
chemistry, Weber and Levson postulated that the ion made from acetone was the acetyl
cation, while the m/z 43 ion from glycerol was probably oxygen protonated ketene. This
postulation is still, however, not without it’s critics.6
I. The Spectroscopy of Polyatomic Ions

Differentiating and assigning structures a and b would be trivial if the vibrational
spectrum of each were available. The IR spectrum of structure a should exhibit a
prominent absorption in a wavenumber region indicative of a carbonyl group, while the
spectrum of structure b would have a distinct absorption in a wavenumber region
associated with the stretching of its hydroxyl group. Molecular spectroscopy, however,
has not been routinely applied to polyatomic ions for a number of reasons. The foremost
difficulty arises fi'om the reactivity of these ionic species. Their high reactivity does not
allow them to be stable in typical confining media such as solutions. Even in the
exceptional cases where these ions can be stabilized, the confining medium typically causes
severe perturbations of the guest species. The problem of reactivity is generally
circumvented by doing studies in the gas phase. In the case of charged species, however,
some other problems arise. .Due to Coulombic repulsion, the theoretical “space charge
limit” of ions in the gas phase is about 106 ions/cm3, which corresponds to femtomolar
concentrations', far below the detection limits of routine vibrational spectroscopic

methods such as Fourier Transform Infrared (FTIR) or Raman spectroscopy. Despite

3
these experimental challenges, polyatomic ions have not completely eluded
spectroscopists.
A. Early Spectroscopic Studies of Polyatomic Ions

Beginning in the late 19605 and early 19705, the electronic absorption and emission
spectra of several small polyatomic ions in electrical discharges were observed.8 Ion
concentrations higher than the space charge limit can be achieved in discharges due to the
presence of both positively and negatively charged species. Unfortunately, these
experiments had some serious limitations. First, the discharge ionization is non-selective,
and can produce several ionic and neutral species‘. As a result, the identity of the
absorbing and/or emitting species is not always clear. Second, the ions are formed in
highly excited states, which can make their spectra quite complicated and difiicult to
assign.8 Third, the harsh conditions of the dischargeiallow only relatively small species to
be observed.

B. Laser-Based Gas-Phase Techniques

Spectroscopic observation of polyatomic ions can be facilitated by addressing one
of two major problems: namely increasing the sensitivity of the experiment, or finding a
way to achieve ion concentrations high enough for standard spectroscopic techniques.
The development of lasers in the 19705 and 19805 made polyatomic ion spectroscopy
more feasible by means of the first method. In 1976, Wing et al. reported the first infi'ared
absorption of a gas-phase, polyatomic, ionic species.9 In these experiments, the rotational-
vibrational transitions of I-ID", formed in a discharge, were driven by a C0 laser, and
absorption was indirectly detected by monitoring a change in charge exchange cross-

section with a neutral target gas. Because of it’s high sensitivity, indirect detection

4
became the basis of several other successful spectroscopic techniques such as
photofi'agmentationlo and photo-detachmentll spectroscopy.

In 1979, Saykally et a1. obtained the first direct infrared absorption spectrum of
I-IBF in a plasma, using newly developed laser magnetic resonance (LMR) spectroscopy. ‘2
In 1980, Oka and co-workers developed the highly successful method of IR difference
frequency spectroscopy and reported the IR spectrum of H33.” While these techniques
were sufficiently sensitive to be applied to ions, they still suffered from the lack of ion-
selective detection. ‘As a result, ion signals were often obscured by absorptions from the
much larger number of neutral species present in the discharge. Perhaps the most notable
breakthrough in gas phase ion spectroscopy came in 1983 when Saykally and co-workers
developed velocity modulation spectroscopy.14 In this experiment, a single wavelength IR
laser is passed coaxially through a mass-selected ion beam. By changing the velocity of
the beam, different ro-vibrational transitions can be tuned into resonance with the laser by
the Doppler shift. This provides a means of selectively exciting and detecting ions, even
though a large number of neutrals may be present. Other high-sensitivity, high-resolution
laser techniques which are variants of velocity modulation were soon developed.m6
These methods have two distinct advantages over the original discharge experiments.
First, most are inherently mass selective, so the identity, or at least the mass to charge
ratio, of the ion being studied is known. Second, the high resolution of these techniques
allows rotational transitions to be observed, which can provide very accurate structural
information about the species being studied.

Nonetheless, these techniques still have some shortcomings. One drawback is that

even with the combinations of several velocities and lasers, not all spectral regions can be

5
covered, although this limitation is being rapidly overcome by advances in laser
technology. Generally, only small portions of the IR spectrum (~O.25 cm") can be
scanned at a time, making coverage of the entire IR spectrum a very tedious and time
consuming task. This problem is further exacerbated because the positions of the
vibrational transitions are often unknown, so that a great deal of experimental time is spent
scanning spectral regions where there are none. These studies are also limited to smaller
ions for two reasons. First, these experiments are also dependent on discharge sources for
ion production.” Second, most of the structural information is extracted from rotational
analysis, which becomes more difficult as the size of the species increases.
C. Matrix Isolation Studies

In the early 19705, however, some researchers were successful in obtaining
vibrational spectra of polyatomic ions by addressing the second problem associated with
ion spectroscopy, namely by devising a way to accumulate sufficient concentrations of
ions for observation by more routine spectroscopic techniques such as FTIR and Raman
spectroscopy. The solution to this problem involved finding a medium sufiiciently inert to
constrain these ions without interacting, or reacting with them. Matrix isolation provided
this medium. This technique, developed initially in 1954 by Pimentel and co-workers,
involves stabilizing the reactive species by trapping them in a rigid, inert matrix, usually
argon or neon.18 Figure 1.1 shows generic matrix isolation methodology. Matrix isolation
was first shown to be quite useful in the study of highly reactive neutral species;19 it has
since been proven to be equally applicable to ionic species. The first report of a matrix
isolated ion was by Kasai and in 1970.20 He generated Na+ in an Ar matrix from the

photochemical reaction:

Reactive Species

    
    
 

Cryostat \,

 

Spectroscopic
Window

Figure 1.1: A generic matrix-isolation spectroscopy experiment

Na+HI+hv—>Na’ +1“ +11

In this system, an electron from the Na atom is photodetached and scavenged by the HI
molecule. Because both positive and negative ions are present in the matrix, they are
theoretically able to exist in concentrations higher than the space charge limit to allow
spectroscopic observation. These ions, however, are assumed to be physically trapped in
individual matrix “sites” and therefore do not perturb one another to any appreciable
extent. As a result, the spectra obtained are very similar to those of the isolated gas-phase
species.21 Since then matrix isolated ions have been studied with. a variety of
spectroscopic methods, including electronic absorption,22 fluorescence,23 FTIR,24
Raman,25 ESIL26 and magnetic circular dichroism (MCD).27

A wide variety of methods have been used to generate ions for matrix isolation
studies, although photolysis of neutral precursors has been the most popular and
successfirl. The common factor of most of the ionization techniques used to date is that
the ionization takes place either in situ or immediately prior to co-condensation with the
matrix gas.28 Since these methods can generate relatively large numbers of ions,
traditional spectroscopic techniques and equipment may be used, making these
experiments much simpler than the laser-based studies. Being able to use common
spectroscopic techniques has made matrix isolation a very usefirl method of determining
the structures of polyatomic ions.

Most of the matrix isolation studies to date, unfortunately, are not without their
drawbacks. The most serious problem generally involves the ionization process. In
general, ionization techniques such as UV photolysis are non-selective, and give rise to

many ionic and neutral products. As a result, the identity of the species being studied is

8
not always known,28 resulting in the same problems that hindered the early discharge
experiments.

The goal of the work described in this dissertation was to develop a methodology
for structural analysis of polyatomic species. The approach used to achieve this goal
involves coupling mass spectrometric technology with matrix isolation spectroscopy. This
experiment and its principal instrumental'components is depicted schematically in Figure
1.2. In such an experiment, ions would be generated in a remote source, mass selected
using a quadrupole mass filter, and directed towards a growing matrix. It is believed that
this methodology will combine the advantages of mass spectrometric selection with
current spectroscopic technologies (matrix and gas phase), while also eliminating their
drawbacks. First and foremost, direct structural analysis is possible with the application of
molecular spectroscopy, particularly FTIR spectroscopy. Through the use of as mass
filter, the identity (more specifically, the m/z value) of the ion being studied would now be
known, making the task of spectral assignment much easier. This technique can be
thought of as having ion-selective detection, provided that conformation is obtained by the
proper control experiments This method was first developed by J. P. Maier and his group
in 1989, who used it to measure the absorption spectra a wide variety of cations.29 The
technique was further extended to laser induced fluorescencew(LIF) by Sabo et al, who
reported the emission spectrum of mass-selected, matrix-isolated C82”.30
11. Matrix Isolation Spectroscopy of Polyatomic Ions

It has been said that the microscopic “molecular” properties of a species can only
be deduced fiom gas phase studies, since perturbations due to intermolecular forces can be

very significant in the liquid and solid phases.“ The more reactive a species becomes, the

Ne inlet

        

positive ion
source

negative ion/
e" source

 

Figure 1.2: Schematic representation of a mass-selection matrix-isolation
experiment.

10
more problematic these intermolecular (or interionic) forces become. With very reactive
species, it is usually necessary to work at very low pressures (concentrations) and high
effective temperatures. Even if these “isolated” conditions can be achieved, many reactive
species still have very short lifetimes, thus making spectroscopic observations challenging.
Matrix isolation provides an alternative to gas-phase studies. Matrix isolation is defined as
the trapping of a reactive species in a rigid cage of a chemically inert substance (the
matrix) at low temperatures. It is imperative that all of the above criteria are met for the
species under study to be truly isolated. Cage rigidity is necessary to prevent the loss of
reactive species by diffusion. Matrix inertness prevents the loss of sample by reaction.
The low temperature contributes to cage rigidity and reduces the probability of internal
rearrangement of the reactive species. These conditions, required to successfully reduce
interactions to the point of “pseudo gas phase behavior,” severely restrict the choices of
matrix materials, and requires that a combination of several different technologies such as
cryogenics, vacuum methods, and spectroscopy, be applied.
A. Ion Generation Techniques

Several techniques have been employed to generate ions in matrices. Matrix
isolated ions can be of two kinds. The first kind is referred to as "chemically bound" ion
pairs, such as Li“02',32 where, the cation and anion are in the same matrix cage, and can
be thought of as a “chemically bound” pair. Their close proximity allows considerable
overlap of the electron density of the ions, so the species may not resemble their isolated
counterparts. The second kind of ion encountered in matrix isolation is an isolated ion,

that is a part of what is called a Coulomb ion pair. In this case, the cation and anion are in

separate matrix cages, and are therefore physically isolated from each other; however, they

l l
are still stabilized by their coulombic force fields. The following is a brief review of some
of the more successfiil ion generation and trapping schemes.
1. VUV Photolysis

This was the earliest technique, and is still the most widely used. As noted earlier,

in 1970 Kasai and co-workers20 produced Na+ ions in an argon matrix by the
photochemical reaction:

Na+Hl+hv—>Na‘ +I’ +H
Electrons and H atoms can freely migrate in an argon matrix and can cause neutralization

of cationic species. In this system, H1 is used as a trap for the electrons removed from the

Na atoms. HI undergoes dissociative electron capture and forms H and 1., which
immobilizes the electrons and prevents the neutralization of Na’. An interesting aspect of
this method is that light lower in energy than the ionization energy of the neutral precursor
will induce photoionization. This can be understood in terms of the Coulomb-pair ions.

The stabilization energy, E is calculated to be:

E = I.E.—E.A.— e

 

471’67‘
Where I .E . is the ionization energy of the electron donor, E .A . is the electron affinity of
the acceptor, a is the permittivity of the matrix, and r is the distance between the two
ions. This is an example of ions being far enough away from each other to be chemically
isolated, yet their coulomb fields still interact to stabilize the ion pair.33 UV photolysis is
usually performed in situ, although the matrix gases can be irradiated during deposition,

also. This technique has the advantage of being relatively simple to implement, but suffers

12
from one major drawback. It is non-selective and usually results in the generation of
several neutral and ionic products from the neutral precursor.“

2. Windowless Discharge

After photolysis, this is the second most popular form of ion generation. In this
technique, the matrix gas is passed through a microwave discharge before condensation.
Some of the atoms exit in a metastable state after passing through the discharge. There
are two possible channels for ionization. The first is energy transfer upon a collision with
a metastable atom. The second route is photoionization from the radiation released by the
relaxing metastable atom.” This technique has been used to study ions such as N202",36
C02”,37 NH,” 04”,” and many others. With this method, secondary photophysical
process are not as prevalent as in VUV photoionization, and it also seems to have greater
ion yields than VUV photoionization. While this technique is somewhat more selective
than VUV photolysis, there still can be several ionic products from a given precursor.

3. Ion Molecule Reactions

One of the more novel methods of ion generation involves the reactions of ions

and neutral molecules. This technique was first reported by Knight, et al., in 1984."0

Some of the systems studied were:
C0 + CO” —> C303"
N2" + CO ——> NzCO”
N2” +N2 —> N,"
In these systems, the mechanism of ion formation is most likely co-condensation of the ion

and neutral in the same matrix cage. The cage effect increases the collision frequency and

13

leads to formation of the ion-molecule pair. These experiments are very interesting in that
along with spectroscopic information, the reaction chemistry can provide some insight into
the dynamics of matrix isolated systems.

4. Chemical Ionization

In 1989 Hacaloglu and Andrews reported the use of a modified Finnigan 3200
chemical ionization source for generating ions for subsequent matrix isolation."l Argon
and naphthalene vapor were passed through the source and focused onto the matrix
window by a negatively biased (relative to the ion source) ring electrode placed near the
substrate. They reported the appearance of CmHg" in the infrared spectrum, but it was
not determined whether the ions were formed in the source, by collision with accelerated
argon ions or by the discharge conditions existing in the vicinity of the ring electrode.
This source was also used to study negative ions. A mixture of argon and C12 was passed
through the source and the ring electrode was biased to accelerate negative ions to the
window. They reported the appearance of C12" and C13' at higher source pressures.

5. Pulsed Glow Discharge

There has been debate about whether or not a sufficient number of ions can be
externally generated and deposited onto a rare gas matrix for FTIR observation. A report
by Vala and co-workers have shown that this is indeed possible.42 In their experiment,
ions are formed in a plasma generated between a stainless steel tube held at ground
potential and a wire connected to a hemispherical grid held at +3 kV. The grid accelerates
positive ions to the cryostat window, with focusing provided by a cylindrical lens. In
preliminary experiments, a mixture of argon and p-dimethoxybenzene (PDMOB) was

passed through the pulsed discharge. Many new bands appeared in the infrared spectrum

14
compared to observations in the absence of discharge. To test whether these were bands

due to the PDMOB cation, photobleaching experiments were performed. During

photobleaching, the intensity of a known electronic absorption of PDMOB+ was

monitored. Four of the new IR bands had photolytic decay characteristics which matched

that of the electronic absorption, and were assigned to PDMOB+. It is important to note.
however, that the pulsed glow discharge is still a very non-selective technique, and that
this experimental configuration provides only charge-selection, rather than mass-selection
during deposition. Of the many new bands present under discharge conditions, only four
were assignable to PDMOB‘.

6. Electron Bombardment

This technique was pioneered by Knight in 1983."3 Electrons boiled off of a
heated filament were focused onto and accelerated toward the matrix substrate. Knight
found it necessary to float the matrix substrate at a positive potential to overcome the
charge accumulation that developed. He used this method to study small ions such as
H20", CO", N113", and N2". Electron bombardment proved to produce greater ion
signals than conventional ionization techniques such as windowless discharge. This
method has the advantage of variability of the electron beam intensity and energy, but it is
still not very selective. The experiments are also difficult to set up; Knight went to great
lengths to electrically isolate the matrix target. Andrews has also reported use of this
technique.‘4 The Andrews group developed a thermionic emitter source capable of
delivering 10-200 mA of electrons to a ring electrode placed near the matrix. Continuous
bombardment of a 100021 argon2H20 matrix resulted in spectroscopically observable

amounts of OH' and Ar.,H+ ions, as well as OH radicals.

15
In recent years, this technique has perhaps been most successfully used in the Vala
laboratory. By subjecting a gaseous mixture of argon, CCla, and precursor vapor to
electron bombardment prior to condensation, they have been able to acquire the visible as
well as FTIR spectra of the radical cations of several polycyclic aromatic hydrocarbons

49
8 and tetracene. 1n

(PAHs) such as napthalene,"5 anthracene,46 pyrene,47 pentacene,4
these experiments, CO. is used as an electron scavenger to prevent neutralization of the
cations. Although this technique does generate large amounts of cations, it still may not
be a generally applicable technique. While electron bombardment resulted in the
production of the molecular ions of the compounds studied, less stable precursors may
undergo extensive fragmentation. Also, as in their pulsed glow discharge experiments,
assignment of vibrational transitions depends on a correlation with a known electronic
transition of the cation. Electronic transitions of the PAH radical cations are known, but
this is not generally the case for polyatomic ions.

7. Mass Selected Deposition

The one feature common to all of the techniques mentioned previously is that the
ionization is non-selective, albeit in varying degrees. Interpretation of the ensuing spectra
would be much simpler if only one ionic species were present in the matrix, and its identity
were known. In 1989, Maier and co-workers reported the development of the first
successfirl mass-selective ion deposition source.29 In their experiments, ions are formed in
a remote source, mass filtered by a quadrupole mass spectrometer, and directed to the
matrix substrate. As stated earlier in this chapter, the electronic absorption spectra of

several cations were measured in the 220—1200-nm region by the waveguide technique. In

1991, Sabo et al. also reported the use of mass selected deposition to obtain the laser

l6
induced fluorescence (LIP) spectrum of mass-selected C52" in argon.30 The project
described in this dissertation involves the application of mass-selected deposition to FTIR
spectroscopy and gaining a more thorough understanding of this rather unique deposition
process, particularly the maintenance of charge neutrality even though only positive ions
are being presented to the matrix.
B. Criteria for Matrix Isolation

l. Inertness

This is perhaps the most obvious and important criterion for a matrix material.
The matrix must be sufficiently inert so it will not chemically or physically perturb the
guest species. The noble gases and molecular nitrogen are generally the only substances
inert enough to trap highly reactive species, and therefore cryogenic temperatures must be
employed to obtain rigid matrices. In the case of polyatomic ions, their high electron
affinities leads to significant charge exchange interactions with xenon, krypton, and argon,
necessitating the use of neon as the matrix host material. In some cases where the ion has
an unusually low electron affinity ( < 11 eV) argon may be used.

2. Rigidity

For the guest to be trapped and kept from reacting with other species in the
matrix, the matrix must be rigid. Most matrices are considered rigid at 0.3 T ,,., where T ,,, is
the melting point of the host material. 0.37),, corresponds to 7.3 K and 25 K for neon and
argon, respectively. Being far below the boiling point of nitrogen (77 K), attaining these
temperatures requires the use of special cryogenic techniques. Liquid hydrogen (20 K)
and liquid helium (4.2 K) were the first refrigerants used. As liquid helium became more

readily available the use of liquid hydrogen was disfavored because of its explosive

17

potential. In the early 19705 mechanical refrigerators capable of maintaining 12 K were
developed and greatly expanded the field. Mechanical refrigerators capable of cooling to
3.6 K are now available, allowing neon matrices to be employed without the use of costly
liquid helium. Maintenance of low temperatures also requires high vacuum conditions,
since these cryogenic temperatures cannot be maintained at high pressures due to the
increased heat load; moreover, the cooled substrates act as cryopumps, and would trap
undesirable background gases in the matrix.

At temperatures above 0.5T... matrix is considered non-rigid, and diffusion of
trapped species occurs. In this case, reactive species will be lost due to reactions with
other guests. Between 0.3 T ,,, and 0.5T," a process known as annealing occurs. Annealing
is described as a rearrangement of the matrix at an atomic level to the most stable crystal
structure. During this process, large trapped species can cause local rearrangement to
give the best, or least perturbing cage. Annealing is often used to detect and minimize
peaks caused by trapping of a guest species in multiple matrix sites. While this technique
is commonly used with argon matrices, the relatively narrow annealing temperature range
and volatility of neon makes the annealing of its matrices quite challenging.

3. Purity

Matrix impurities may cause perturbation of the guest species, should they become
close enough to interact. Since it is essential that the guest be trapped in a uniform, inert
matrix, the matrix material must therefore be very pure. This necessitates the use of high
purity matrix material and high vacuum techniques to prevent contamination during

deposition of the host and guest species on the cooled optical substrate.

18

4. Transparency

The matrix material must be transparent to the radiation being used for
spectroscopic observation. Noble gas solids are exceptionally suitable in this respect.
Being atomic solids, they exhibit no rotational or vibrational transitions. Argon and neon
provide a spectroscopic window from the vacuum ultraviolet to approximately 100 cm'l.
where lattice vibrations become excited. The noble gases are also diamagnetic, which
makes them transparent to ESR spectroscopy, although the magnetically active nuclei of
krypton and xenon can affect the hyperfine splittings of trapped species.3 1

5. Thermal Properties

During condensation, the latent heat of fusion (L,) of the matrix material is

released. This heat must be conducted away from the surface through the matrix by the
cryostat. The matrix must have a high thermal conductivity so that this heat conduction
does not cause any local warming. Matrices experiencing local warming due to low
thermal conductivity are ofien opaque and highly scattering, and can possibly lose reactive
species through diffusion.

6. Volatility

In order to prevent loss of matrix material and the species trapped within due to
evaporation, the matrix material must not be volatile at the temperatures being used.

C. Trapping Ions in Matrices

The trapping of ions in matrices is at times more an art form than a scientific
process. Along with the proper choice of matrix material, the proper deposition
conditions must also be maintained to ensure successful trapping. Optimization of such

parameters as deposition rate and matrix to guest (M/G) ratio are critical. Also, since it

19

seems that the matrix must be approximately electrically neutral, counter ions must also be
present to maintain charge neutrality.

l. Deposition Rate

As described previously, the temperature of the matrix must be kept below during
deposition 0.37,. to prevent annealing or diffusion. The matrix gas must also condense as
fast as possible to ensure efficient trapping; is. the latent heat of fusion of the matrix gas
must be conducted away by the cryostat as quickly as possible. This requires not only the
use of a sufficiently powerful cryostat, but a slow enough deposition rate so that the latent
heat of fusion released does not cause an appreciable warming of the matrix. The power

liberated by the latent heat of fusion is given by the equation:

44(7— T.)

W

Lfn =

where

Lf= latent heat of firsion (“’Zmfl

n = deposition rate ("’015 )
11 th I d t"t (——cal )
= erma con uc wr
. y cm-s-K

T, To = surface and matrix temperature (K), respectively
w = matrix thickness
A = matrix surface area

The thickness of the matrix is given by;

20

III

u _
pA
where t is the deposition time and p is the molar density of the matrix material. The

surface temperature of the matrix can now be described by the equation:

T T Lfnzt
— 0+ zipAz

Substitution of typical parameter values indicate that deposition rates of l-18 mmol/hr

 

will result in a temperature increase of approximately one degree after several hours of
continuous deposition. The deposition time will affect the maximum usable flow rate.
When long depositions are expected, the matrix gas must be deposited very slowly to
avoid excessive heating.50

2. Matrix-to-Guest (M:G) Ratio

To ensure complete isolation, the M:G ratio must be high enough that very few
guests will have another guest as a nearest neighbor. For example, in the body centered
cubic structure of solid neon, a substitutional site (created by the removal of one neon
atom) has twelve nearest neighbors. A guest occupying that site may be considered to be
isolated if all of its nearest neighbors are neon atoms. The probability that this guest is not
isolated can be expressed as:31

P = (l — r)”

where r is the ratio of guest to matrix atoms. With a M:G ratio of 10021 11.4% of the
guests will have another guest as a nearest neighbor, leaving only 88.6% of the guest
molecules completely isolated. A M:G ratio of 100021 will provide 98.8% isolation, and a

ratio of 10 000:] will provide 99.9 percent. Due to their size, most polyatomic species

21
will occupy multiple substitutional sites (site created by the removal of more than one
neon atom). As an example, a species three times the size of a matrix atom will most
likely occupy a site created by the loss of one atom, and its 12 nearest neighbors. Such a
site creates a cage with 122 nearest neighbors. The probability of isolation in this cage is
given by:
P z (1_ r)”:

requiring ratios of 10 000:1 or greater to achieve 99% isolation.3 '

3. Counterions

It is usually assumed that an ion-containing matrix maintains at least approximate
electrical neutrality. This neutrality is made possible by the existence of counter ions in
the matrix, i.e. ions of opposite charge to the ions being studied. Since most matrix
studies involve cations, the counter ions are usually negatively charged. Many researchers
report the presence of both positively and negatively charged species in matrices but
usually no correlation can be made between the relative intensities of the two types of
ionic species.“’“ To date, the most popular explanation for the formation of these anions
is that they are formed by electron attachment to electronegative matrix impurities such as

02, C02, and OH (from H20).52 Although this concept seems quite intuitive, definitive

experimental evidence is still lacking.

While counter ions are usually assumed to exist and not considered eXplicitly,
some experiments have been performed that address the issue in some way. In one set of
experiments, Knight purposely doped his matrix with an electrophile, F2, in hopes of

53

increasing the number of cations stabilized in the matrix. Although an electron capture

product was found afier photolysis (Ff), the yield of the cation was not affected

22

appreciably by the intentional addition of the counterion. This contrasts with reports from
the Vala laboratory that the addition of CG. as an electron scavenger dramatically
improves the yield of stabilized cations.45

While many of the ionization techniques, such as photolysis and discharge
ionization, are capable of producing both positively and negatively charges species, this
raises a special concern in the case of mass-selected deposition. In this case, ions of one
type of charge only are intentionally introduced into matrix. 15 there some mechanism by
which anions can be generated, or will a separate source of negative ions be necessary to
achieve the necessary approximate electrical neutrality of the matrix?

D. Matrix Effects on Ionic Spectra

Although matrices are ideally a non-perturbing medium, the interactions between
the guest and host are sufficient to cause some spectral changes relative to the gas phase
spectrum. Matrix vibrational spectra are characterized by narrow bands relative to the gas
phase. Electronic transitions, however, are generally much broader in matrices than in the
gas phase. In general, matrices have been seen to affect the position, appearance, and
number of peaks. A brief discussion of these effects follows.

1. Number of Peaks

Since the matrix is never a perfectly ordered solid, all of the possible trapping sites
are not identical. A5 a result, the interaction forces felt by the guest ions are not uniform.
The differences in the interaction may be enough to produce distinct vibrational motions.
Such multiple site features can usually be detected by annealing. During the annealing
process, the matrix becomes more uniform, as do the trapping sites. If a certain multiplet

is due only to inequivalent matrix sites, the multiplet will become a singlet. The multiplet

23
may remain, however, if the inequivalent site is due to an interaction with an impurity in
the matrix.

2. Frequency Shift

There is in generally a shift between a matrix absorption frequency and the gas
phase absorption frequency. These shifts can be explained by the types of interactions
which also describe shifts between gas and solution phase frequencies. Solution
interactions are generally described in terms of a combination of dipole-dipole, dipole-
induced dipole, ion-ion, ion-dipole, and London dispersion forces. The strength of these
interactions depends to a large extent on the polarizability of the molecule, and of the host.
In a matrix, as in solution, the general trend is towards a lowering in energy (redshifting)
of molecular transitions.

One fimdamental difference between matrices and solutions is that the solvent cage
is much more rigid in matrices. This gives rise to an interaction generally not encountered
in solution, repulsive forces. The neglect of these forces is a serious shortcoming in the
prediction of matrix spectral shifts, as these forces may become strong enough to
overpower the attractive forces. This especially true of low energy vibrations, where the
matrix cage can severely distort the vibrational motion. The total matrix frequency shift

can be summed up by the expression"1

AV : Vmatnx _ Vgas : Avelec + Avrnd + Avdisp + Avrep
where mea refers to the matrix-isolated frequency, v80, refers to the gas-phase
fiequency. Ava“ refers to the shifts due to the permanent dipole distribution, Avmd refers

24

shifts due to induced dipole/induced dipole interactions, Arm refers the shifts due to
dispersion forces and Av"? refers to the shifts due to repulsive interactions.

As the host polarizability increases, interactions with guest species increase, and
there is a trend towards redshifted frequencies. The shift usually becomes less negative as
the frequency decreases. As the frequency of the vibration becomes even lower, the shift
may become positive.31 This is because vibrations characterized by a small force constant
are easily distorted by the matrix cage. A recent review by Jacox55 reports vibrational
matrix shifts for diatomic species to be less than 1% for neon, and less than 2% in argon,
with red shifts being more common than blue shifts. Larger than normal shifts are found
in systems which exhibit 1) hydrogen bonding within the matrix, 2) cage stabilization of
weakly bound species, 3) species with a large dipole moment, where it is assumed that
both the gas phase and matrix transitions are correctly assigned. Neutral molecules are
generally less affected than ions by this effect since they less strongly polarize the host
atoms.

3. Spectral Distortion

In general, distortion of the guest spectra by the host is negligible for neutral
species, but can be very significant for ionic species. Argon is fairly polarizable, and can
interact with trapped ions by ion induced—dipole interactions. These interactions do not
severely distort vibrational transitions, but they can markedly affect electronic spectra.56
The electronic transitions are usually red shifted and broadened. As a result, ions with an
electron affinity greater than about 10 eV experience electronic spectral distortion in
argon. Neon is much less polarizable than argon, and is therefore the matrix of choice for

ionic studies.

4. Vibrational Relaxation

In addition to the effects resulting from the perturbations of the electronic
structure of the trapped ions, fluorescence experiments are affected in another way by the
matrix. When an ion is excited to a vibrational level within an excited electronic state, two
relaxation pathways are possible. The ion can fluoresce directly from the excited
vibrational state to the ground electronic state, or it can relax to the ground vibrational
level of the excited electronic state and then fluoresce to the ground electronic state. In
the gas phase, the first pathway is commonly observed. Only large molecules having a
high density of vibrational states exhibit vibrational relaxation in the excited electronic
state. In matrices, vibrational relaxation is invariably observed, even for small molecules.
This is because the phonon modes of the matrix are an efficient channel for the dissipation
of vibrational energy.” This simplifies the spectra, but also limits the vibrational
information obtainable from matrix fluorescence experiments to the ground electronic

state of the guest.

(1)
(2)

(3)
(4)
(5)
(6)
(7)
(8)

(9)

(10)
(11)

(12)
(13)
(14)

(15)
(16)
(17)

(13)
(19)

List of References
Szczepanski, J .; Vala, M. Nature, 1993, 363, 699.

McEwan, M. J .; Phillips, L. F. Chemistry of the Atmosphere; Halsted: New York,
1975; Chapter 6.

Bouchoux, G.', Hoppilliard Y. J. Mol. Struct. 1983, 104, 365.

Based on HF calculation using 6-3 lg" basis set.

Weber, R.; Levsen, K. Org. Mass Spectrom. 1980, 15, 138

Terlouw, J. K.; Heerma, W.; Dijkstra, G. Org. Mass. Spectrom. 1980, 15, 660.
Stepnowski, R.; Allison, J. Anal. Chem. 1987, 59, 1072A.

Herzberg, G. In Molecular Ions: Spectroscopy, Structure, and C hemistry; Miller,
T. A., Bondybey, V. E, Eds; North Holland: New York, 1983; pp 1-9.

Wing, W. H; Ruff, G. A.; Lamb, W. E; Speezeski, J. J. Phys. Rev. Lett. 1976, 36,
1488. '

Carrington, A; Sarre, P. J. M01. Phys. 1977. 33, 1488.

Cosby, P. C.; Ozenne, J. B., Moseley, J. T.; Albrittion, D. L. J. Mol.
Spectrosc. 1980, 79, 203.

Saykally, R. J .; Evenson, K. M. Phys. Rev. Lett. 1979, 43, 515.
Oka, T. Phys. Rev. Lett. 1980, 45, 531.

Gudeman, C. S.; Begemann, M. H; Pfafi‘, J .; Saykally, R. J. Phys. Rev. Lett. 1983,
50, 727. '

Foster, S. C.; McKellar, A. R. W.; Sears, T. J. J. Chem. Phys. 1984, 81, 578.
Kawaguchi, K.; Hirota, E. Ann. Rev. Phys. Chem. 1985, 36, 53.

Coe, J. V.; Saykally, R. J. “Infrared Laser Spectroscopy of Molecular Ions,” in Ion
and Cluster Ion Spectroscopy and Structure; J. P. Maier, Ed. Elsevier Science
Publishers B. V., Amsterdam, 1989 Chap. 5

Whittle, E.; Dows, D. A; Pimentel, G. C. J. Chem. Phys. 1954, 22, 1943.

Perutz, R. N. Ann. Rep. Prog. Chem. 1985, 82, 157.

26

(20)
(21)
(22)
(23)
(24)

(25)

(26)

(27)

(23)

(29)
(30)

(31)

(32) ,

(33)
(34)
(35)
(36)
(3 7)
(38)
(39)

27

Kasai, P. H. Acc. Chem. Res. 1971, 4, 329.

Jacox, M. E. Chem. Phys. 1994, 189, 149.

Andrews, L.; Keelan, B. W. J. Am. Chem. Soc. 1981, 103, 99.
Bondebey, V. E.; Miller, T. A. J. Chem. Phys. 1980, 73, 3053.
Jacox, M. E.; Thompson, W. E. J. Chem. Phys. 1989, 91, 1410.

Szczepanski, J.; Personette, W.; Pellow, R.; Chandrosekhar, T. M.; Roser, D.;
Cory, M.; Zemer, M. Vala, M. J. Chem. Phys. 1992, 96, 35.

Knight, L. B.; Steadman, J. J. Chem. Phys. 1982, 77, 1750.

Rivoal, J. C.; Emampour, J. S.; Zeringue, K. J.; Vala, M. Chem. Phys. Lett. 1992,
92, 313.

Bondybey, V. E. “Time-Resolved Laser-Induced Fluorescence Studies of
Spectroscopy and Dynamics of Matrix-Isolated Molecules,” in Chemistry and
Physics of Matrix-Isolated Species; Andrews, L.; Moskovits, M. Eds; Elsevier
Science Publishers B. V.: Amsterdam, 1989, Chapter 5, p 129.

Fomey, D.; Jakobi, M.; Maier, J. P. J. Chem. Phys. 1989, 90, 600.
Sabo, M. 8.; Allison, J.; Gilbert, J. R.; Leroi. G. E.App1. Spectrosc. 1991, 45, 535.

Cradock, S.; Hinchcliffe, A. J. Matrix Isolation; Cambridge University:
Cambridge, 1975; Chapter 2.

Knight, L. B., Jr, J. Chem. Phys. 1983, 78, 6715.
Peruz, R. N. Chem. Rev. 1985, 85, 97.

Prochaska, F. T.; Andrews, L. J. Am. Chem. Soc. 1978, 100, 2102.

Jacox, M. E.; Thompson, W. E. Res. Chem. Int. 1989, 12, 33.

Jacox, M. E.; Thompson, W. E. J. Chem. Phys. 1990, 93, 7603.
Thompson, W. E.; Jacox, ME. J. Chem. Phys. 1989, 91, 1410.
Thompson, W. E.; Jacox, M. E. J. Chem. Phys. 1989, 91, 3856.

Thompson, W. E.; Jacox, M. E. J. Chem. Phys. 1989, 91, 3826.

(40)

(41)
(47-)
(43)
(44)

(45)

(46)

(47)

(43)

(49)

(50)

(51)
(52)

(53)

(54)

(55)
(56)

(57)

28

Knight, L. B. Jr.; Steadman, J .; Miller, P. K; Bowman. D. E.; J. Chem. Phys,
1984, 80, 4593.

Hacaloglu, J .; Andrews, L. Chem. Phys. Lett. 1989, 160, 274.

Szczepanski, J.; Personette, W.; Vala, M. Chem. Phys. Lett. 1991, 185, 324.
Knight, L. B. Jr. J. Chem. Phys. 1983, 78, 6715.

Suzer, S.; Andrews, L. J. Chem. Phys. 1988, 88, 916.

Szczepanski, J .; Roser, D.; Personette, W.; Eyring, M.; Pellow, R.; Vala, M. J.
Phys. Chem. 1992, 96, 7876.

Szczepanski, J .; Vala, M; Talbi, D.; Parisel, 0.; Ellinger, Y. J. Chem. Phys. 1993,
98, 4494.

Vala, M; Szczepanski, J .; Pauzat, F.; Parisel, 0.; Talbi, D.; Ellinger, Y. J. Phys.
Chem. 1994, 98, 9187.

Szczepanski, J .; Wehlburg, C.; Vala, M. Chem. Phys. Lett. 1995, 232, 221.

Szczepanski, 1.; Drawdy, J .; Wehlburg, C.; Vala, M. Chem. Phys. Lett. 1995, 245,
539.

Moskovits, M; Ozin, A. In Ctyochemistty; Moskovits, M.; Ozin, A. Eds; John
Wiley and Sons, New York, 1983, Chapter 4.

Suzer, S.; Andrews, L. J Chem. Phys. 1988, 88, 916.
Andrews, L. Ann. Rev. Phys. Chem. 1979. 30, 79.

Knight, L. B. Jr.; Earl, L.; Ligon, P. R.; Cobranchi, D. P. J. Phys. Chem. 1986, 85,
1228.

Downs, A. J .; Peuke, J. C. Molecular Spectroscopy: A Specialist Periodical
Report; Burlington House: London, 1973, 1, Chapter 9.

Jacox, M. E.; J. Mol. Spectrosc. 1985, 113, 777.
Bondybey, V. B; English, J. H. J. Chem. Phys. 1979, 71, 777.

Bondybey, V. E.; Miller, T. E. in Molecular Ions: Spectroscopy, Structure, and
Chemistry; Bondybey, V. E.; Miller, T. E., Eds; North Holland: New York, NY,
1983; Chap. 4.

Chapter 2

Instrumental

In order to better understand of the structure and chemistry of polyatomic species.
instrumentation has been developed to isolate and spectroscopically observe mass-
selected, matrix-isolated ions. In these experiments, cations are generated by electron
impact ionization, mass selected by a quadrupole mass filter, and directed to a low-
temperature window and condensed with an excess of neon. The trapped ions can then be
studied by a variety of spectroscopic techniques, with FTIR absorption and visible
emission being the focus of the work described in this dissertation. The instrumentation
necessary for spectroscopic observation of mass-selected, matrix isolated ions relies on the
implementation and coordination of several different technologies. Among these are mass
spectrometry, ion optics, vacuum generation, cryogenics, spectroscopy, data acquisition,
and signal processing. Instrumentation to apply the mass-selection technique to laser
induced fluorescence (LIF) has been previously developed in this laboratory1 and is
described in detail elsewhere.2 The instrumentation used in this research is based on that
LIF system; however, many modifications and improvements have been implemented.

1. Ion Source, Mass Filter, and Ion Optics

In any study of polyatomic ionic species, the first, and usually most critical, step is
to generate the ions of interest. The ion source in these experiments is a modified
Finnigan 3200 quadrupole mass spectrometer. Cations are generated in the original
electron impact ionization source. The quadrupole mass filter is then used to produce a

beam of mass-selected ions. The emerging beam is then focused and directed to the

29

30

cryogenic window by a series of ion optics and steering plates. This system is capable of
generating mass-selected ion currents of 20—80 nA for periods of up to 30 hours.

This ion source represents a great improvement over the previous ion sources.3
The Finnigan source produces ion beams nearly two orders of magnitude greater than the
original ion sources. The ion source and quadrupole are also differentially pumped, which
has reduced the influx of neutral species into the UHV chamber by two orders of
magnitude. The 1—700 Dalton mass range of the Finnigan source is also an improvement
over the 1—200 and 1—60 Dalton ranges of the RGA ion sources.3 A schematic of this ion
source is shown in Figure 2.1. A complete description of the modification of the mass
spectrometer and construction of the ion optical system can be found elsewhere.‘
1]. Vacuum System

The primary need for high vacuum conditions in matrix isolation experiments arises
from the need to thermally isolate the cryogenic substrate. Without this insulation, the
low temperatures necessary could not be achieved due to the excessive heat load from the

surrounding atmosphere. While pressures of l x 10'5 Torr are sufficient for thermal

insulation, certain aspects of our experiments required ultra-high vacuum (UHV)

conditions (< 1 x 10‘8 Torr). The main reason was to minimize the amount of residual

gases in the system, in order to reduce the possibility that the small absorption of a mass-
selected ion would be hidden by a large absorption due to residual gases. Another reason
was to ensure matn'x purity. As a rare gas matrix grows, the heat of fusion of the most
recently condensed layer must be transported through the previously condensed layers to
the cryostat. As the matrix thickens, this heat transport becomes less efficient. It is

possible that the matrix will become thick enough that any additional gas cannot be cooled

 

Figure 2.1:

31

Diffusion Pumps

 

 

 

 

E
L_ O
:<
ill ' m<
_|—am»
||| 0'2
A N— (3:
3
h (f) (I)
S a % 3,5 a
: E 8. r: E5
O a) CD Q.
m a) H _
(D O _
Clo (U 0 CD {C3
E; E O E1 'C
CU) __ CD CD
9 O ”27>
.H D.
g a
u“: ‘8
3
0'

Schematic top view of mass-selected cation source based on modified
Finnigan 3200 quadrupole mass spectrometer.

32
rapidly enough, and matrix growth will cease. Since long (25—30 h) deposition times were
expected, very slow deposition rates were necessary to ensure adequate cooling of the
matrix. With a small flux of matrix material on the substrate, the residual gases become a
greater fraction of the material condensing on the cryogenic substrate. If residual gases
make up 1% or more of the frozen material, at least 12 °/o of the ions will have a residual
gas molecule as a nearest neighbor, and hence will not be truly isolated. The need for a
such a pure matrix has dictated the design of the of the matrix isolation chamber and its
vacuum system. The maintenance of pressures < 1 x 10'8 Torr is not a trivial task, and
requires great care in choice of materials and construction methods.
A. Vacuum Chamber

The UHV chamber is constructed from a 8" OD. stainless steel cylinder. To
minimize the system’s leak and outgassing rate, all permanent joints and seals are either
welded or silver soldered, and all non-permanent seals are made with all-metal Conflat‘i‘i'
flanges. The top and bottom of the tube is fitted with 10 in. Conflat® flanges. The top
flange supports the cryostat, roughing line, and the residual gas analyzer (RGA). The
bottom flange is connected to a 10" electropneumatic gate valve which acts as a support
for the chamber and separates the chamber from the cryopump. The chamber is also fitted
with various ports to allow connection of various components such as the ion source,
optical windows, current measurement, N2 purge, etc. The ion source is connected via a
6” Conflat® flange and is separated by a manual 4" gate valve (MDC #GV-400M). Opticai
access to the matrix window is provided by windows mounted on 2%" conflat flanges.
Depending on the spectroscopic application, either KBr, cm, or sapphire windows were

used.

B. Pumping System

To keep the UHV chamber as clean as possible, a cryopump is used to evacuate it.
These pumps have the advantages of very high pumping speeds (~1500-2000 l/s). and low
base pressures, and provide an oil free environment. The chamber was originally
evacuated with a Leybold-Heraeus 8” cryopump (RPK 1500) driven by a RW3
compressor. This unit sustained irreparable failure and has been replaced with an APD
Cryogenics 8” cryopump (APD-8) driven by a HC-2 compressor. When the chamber is at
atmospheric pressure, it is first roughly evacuated with a direct drive mechanical pump
(Leybold—Heraeus Trivac D16A). To minimize the amount of backstreaming from the
rough pump, a molecular sieve trap (L—H 85415, 10 A sieves) is used. The mechanical
pump is separated from the chamber by a manual bellows valve (Hughes Aircrafi #HVV-
150-2). The main chamber is isolated from the cryopump by a 10” electropneumatic gate
valve (VAT 10146-UE40). The foreline valve of the cryopump is sealed with a
electropneumatic valve (L-H. 289-22-B1). Pressure in the chamber is measured with a
residual gas analyzer (RGA, Dycor M100). This device is essentially a small quadrupole
mass spectrometer with a high efiiciency ionisource. It has the advantage of not only
measuring the total pressure in the chamber, but also makes it possible to identify the
individual species present in the UHV chamber, as well as their relative amounts. The

base pressure of the experimental as measured by the RGA is ~5 x 10'9 Torr. During
matrix deposition, the pressure rises to ~ 7.5 x 10'7 Torr. A schematic of the UHV system

is shown Figure 2.2.

34

   
  
 
  
  
  
  

IN

Roughing
Valve

Residual Gas
Analyzer

Purging
Valve

 

H2 Vapor
Thermometer

 

(1111C \tli\\‘

   
   
  
   

Foreline
Valve

 

(flown W
' I ’1 Thermocouple

E Gauge

 

{—— Sievc Trap

\lcchumml Pump

Figure 2.2: UHV pumping system.

35
C. Vacuum Interlock

The UHV system was designed with an interlock system to protect various system
components in case of vacuum and/or power failure, and also to protect the system from
operator error. The interlock was originally designed to control and supply power to
every component of the mass-selected deposition system. While the protection scheme of
this system remains unchanged,2 the interlock is no longer be used to power the cryostat
and Finnigan ion source, due to the increased electrical power requirements of the newer
equipment.

D.‘ Pumping Procedures

Evacuation of the UHV chamber is a relatively simple procedure, and with the
interlock control system, quite safe. The vacuum system is similar to most cryopumped
systems, with one exception. In most cryopumped vacuum systems, the pump and the
main chamber are usually individually connected to the roughing pump. The cryopump is
generally evacuated through its foreline and the chamber through a different valve, while a
gate valve separates the cryopump and the chamber. This is usually done to facilitate
regeneration of the pump. This was not a great concern, since our UHV system operates
under a fairly low gas load, and the cryopump does not require regeneration for weeks at a
time. Instead, we roughly evacuate the cryopump through the main chamber. To prevent
excessive backstreaming of the mechanical pump oil, the chamber then is purged with
nitrogen while the pump cools to its ultimate temperature of ~14 K, and to pump it down
to no less than 50 mTorr before the gate valve is opened The standard pumpdown
procedure for the ultra high vacuum chamber, if the cryopump is at room temperature is as

follows:

IO.

11.

12.

13.

14.

36
shut off nitrogen purge
close cryopump foreline valve
open gate valve (should already be opened)

open roughing valve and evacuate chamber and cryopump to ~ 15.

mTorr with mechanical pump

close gate valve

turn on cryopump

close roughing valve

purge chamber with house nitrogen until cryopump is cool
once cryopump cools (~ 90 min), shut of nitrogen purge
open roughing valve and evacuate chamber to ~ 50 mTorr
close roughing valve

open gate valve

turn on RGA

engage interlock

If the cryopump is already cooled, the following procedure should be used:

1.

2.

shut off nitrogen purge

open roughing valve and evacuate chamber to ~ 50 mTorr
close roughing valve

open gate valve

turn on RGA

engage interlock

37
E. Faraday Plate Modifications

The UHV chamber was originally designed with a Faraday plate to measure the ion
current impringing on the matrix window and to allow re-optimization of the ion beam
during deposition. The Faraday plate is mounted on a linear motion feedthrough (MDC
LM-133-2) to allow it to be moved in and out of the ion beam. The replacement of the
DisplexO cryostat with the three-stage Heliplex® required some modifications to the
Faraday plate, which is now used to measure the ion current emanating from the matrix
upon warming as well. First, to account for the smaller window size, the plate was
reduced to 3/4" diameter. The Displex‘ID also did not require a radiation shield, whereas the
use of one is critical for the Heliplex”). To provide appropriate clearance of the radiation
shield, ~ 3/41' was cut off of the end of the feedthrough. A thin strip of stainless steel was
attached to the end of the feedthrough. This strip was bent forward to allow the Faraday
plate to clear the radiation shield. The Faraday plate is secured to the stainless steel strip
by a small nut and bolt, and electrically isolated by the use of ceramic washers. The
Faraday plate is electrically connected to an ammeter via a Teflon coated copper wire and
a BNC feedthrough mounted on a 1 K " Conflat® flange. This assembly is shown in Figure
2.3.
III. Cryostat

It has been the goal of this project to build an instrument which permits trapping
and spectroscopic observation of a variety of polyatomic ions of interest to chemists.
Since many of these ions have electron affinities close to the ionization potential of argon,
neon matrices must be used. Initial experiments in this project were performed using an

Air Products DisplexO 202 cryostat driven by an air-cooled 1R02A

38

Teflon coated BNC feedthrough

Cu MIR /

Faraday plate

    
 

M L 1 1|:- 10
' I I l I ' -

 

 

I I I I
DO

I «a;

---15

 

stainless steel strip

linear motion feedthrough

vacuum atmosphere

 

Figure 2.3: Faraday plate assembly

39
compressor. A description of this system and its adaptation to the UHV chamber can be
found elsewhere.2 This two-stage closed cycle unit was capable of reaching temperatures
as low as 16 K. This restricted studies to argon matrices, which severely limited the
number of ions that could be studied.

Even when they run at peak performance, the lowest temperature than can be
attained with such two-cycle Gifford-McMahon refrigerators is ~10 K. Since neon
matrices require a temperature of 7.2 K or less, a more efficient cooling system is
required. The two preferred methods are the use of liquid helium, or the use of a three-
stage closed cycle refrigerator. Since the high cost of liquid helium makes very long (24
hour) experiments prohibitively expensive, the DisplexGD was replaced with an APD
. Cryogenics HSAB Heliplex® closed cycle cryostat.

A. Refrigerator

The HeliplexQ cryostat achieves temperatures below the 10 K limitation of two-
stage Gifford-McMahon refrigerators by implementing a third cooling stage based on the
Joule—Thompson (J—T) effect. This cooling stage is based on the principle that a gas will
cool when undergoing free expansion. Each gas, however, has a characteristic
temperature (the J—T inversion temperature) above which the effect is inverted and the gas
will actually warm upon expansion. The J—T inversion temperature of helium is 100 K,
and the gas must be significantly below this temperature for cooling to be efficient. In the
Heliplex" refrigeration cycle, the helium refrigerant is split into two streams. One stream
undergoes the two-stage Gifford—McMahon cycle and is successively cooled to
approximately 50 K and 10 K. The other stream of helium is then cooled by these two

cold stations. This helium is now well below it’s J—T inversion temperature, and

4o
undergoes significant cooling upon free expansion. The expansion is controlled by
changing the diameter of the orifice through which the helium expands. This is
accomplished via a rotary motion feedthrough on the cryostat’s refrigeration shroud. The
Heliplex“> refrigeration cycle is shown in Figure 2.4. This three stage cryostat has a
cooling capacity of 1 Watt at 4.2 K, with a ultimate temperature of 3.6 K or lower. Figure
2.5 shows a typical cooldown to 4 K, which generally takes ~3 hours.

The HS-4B cryostat is factory modified for ultra-high vacuum service. These
modifications include:
o 6" o-ring flange replaced with a 8" rotatable Conflat® flange
0 o-ring sealed instrumentation feedthroughs replaced with equivalent conflat hardware
0 standard HMX series vacuum shroud deleted
0 standard evacuation port and valve deleted
0 the third stage was lengthened so that the sample window would be centered in the

UHV chamber windows

0 the rotary motion feedthrough controlling the Joule—Thompson expansion valve was

replaced with a comparable UHV compatible feedthrough

The replacement of the rotary motion feedthrough with a similar UHV feedthrough
changed the operating procedures of the cryostat. The numbering convention of the UHV
feedthrough is reversed from that of the standard feedthrough. On PIS-4B, the J—T valve
is firlly open when the micrometer reads 0.000 and is fully closed when the micrometer
reads 0.170. The micrometer should never be adjusted below 0.000 or above 0.170, or

permanent damage to the J—T expansion circuit my result.

l
4

Q a 89m
c2383 .1. iv

9 o: 59203

938388 83558 3 message
8893 05 5;» 8a 55%. 8522382 9;. 853on .88me oumEonom mama 53:3;

 

 

so: oomaw sew/W

2 es 59203 v
8: 89m L

 

 

 

5982 Pin

 

 

 

 

 

and 2&5
aommmano 20305800
52933 .71

 

 

 

 

 

 

V

300

N
01
O

N
O
O

.a
01
O

.3
0

Temperature (Kelvins)
O

01
O

O

42

 

 

   
   

 

 

 

J—T valve
set to 0.120
40 K
J-T valve
set to 0.045 ___’
137 K
J—T valve
set to 0.150 >L
20 K
0 50 100 150
Time (minutes)
Figure 2.5: Typical Heliplex cooldown. Temperature is

measured at the fixed third-stage sensor.

 

43

The addition of a J—T expansion circuit allows liquid helium temperatures to be
reached, but also introduces some complications into the maintenance of the system.
First, the helium gas is expanded through a very small orifice, which can be easily clogged
by the condensation of impurities such as air and water in the gas. This results in the
refrigerator being extremely sensitive to helium contamination. Second, the efficiency of
the J—T expansion cycle is also dependent on the initial temperature of the helium gas. If
the supply pressure of helium is too low, the temperature of the two Gifford—McMahon
stages will be somewhat high, which in turn raises the temperature of the helium stream
undergoing J—T expansion, resulting in less efficient cooling and a decreased refrigeration
capacity. Finally, the return pressure of the J—T circuit is dependent of the J—T valve
setting. At temperatures below 4.4 K, the return pressure is below atmospheric pressure.
Under these conditions, any leaks in this circuit will result in air leaking into the system,
and will severely degrade its performance. A more complete description of the system
maintenance can be found in the Heliplex® Refrigeration System Technical Manual.5

A new top flange for the vacuum chamber was designed to accommodate the
larger size of the three stage cryostat. The cryostat mates to the top flange through a 4”
OD tube which is welded to an 8” Conflate' flange. This flange then connects with an 8”
conflat at the bottom of the 5” shroud of the Heliplex". The upper end of this shroud is
connected to a 8” ConflatG' on the instrumentation skirt of the refrigerator. This assembly
is shown in Figure 2.6.

B. Temperature Control
Temperature control of the cryostat is maintained by two Si diode sensors (Lake

Shore Cryotronics DT-470), and a 36 Q foil heater, controlled by a Lake Shore 330

44

 

 

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

.———> ‘
3
Cryostat 1" Stage
<—-— Cryostat Vacuum Shroud
Radaition Shield for 2'“1
Conflat Seals and 3rd Cooling Stages
. , I: z / :r
(-——— Chamber "Top Flange"
L' :I
.2
(—-— UHV Chamber
1

 

Figure 2.6: Adaptation of the Heliplex cryostat to the UHV chamber.

45
autotuning temperature controller. One diode is affixed to the third stage of the cryostat
with Stycast epoxy. This places the sensor in close proximity to the cold head and the
heater, which results in more accurate temperature control. The calibrated diode is
attached to the outer side of the sample window holder, giving an accurate measurement
of the cryostat window temperature. The diode is fastened using bare copper wire strung
through three holes drilled into the widow holder. These wires are threaded through the
non-removable piece of the sample holder, which allows the sample window to be
removed and cleaned without removal of the diode. A thin layer of Apiezon N vacuum
grease is applied between the window holder and the Si diode to maximize thermal contact
between the two. This assembly is shown in Figure 2.7.
C. Radiation Shield Modifications

Modifications to the radiation shield were necessary to improve the ion collection
efficiency of the experiment. The original HMX series radiation shield, in the region of
the window was a 1.5”. diameter tube of nickel-plated copper. Optical access to the
sample holder was via two % ” holes drilled in the tube. Since these apertures were
smaller that the exposed window diameter of 3A”, a significant portion of the mass-
selected ion beam was being clipped. While an ion current of 25 nA was measured at the
Faraday plate directly in front of the radiation shield, only 8 nA was measured at the
sample window.

To make more efficient use of the ion beam, the apertures were enlarged to Z"
diameter. This, unfortunately had the adverse effect of increasing the amount of radiative
heat reaching the window to a level greater that it could dissipate. As a result, the

window could not be cooled below approximately 10 K. As noted above, the

46

To 10-pin Electrical
Feedthrough

'1

 

 

 

 

 

 

 

 

Diode and Heater \ E Cryostat 3rd Stage
Connections ' Cold Station
(—— Foil Heater
Control Diode 4 , .
Copper Wire
Matrix Optical Ring
' Sam le Holder
«Free Length u WII'IdOW p
Sensor Diode
Apiezon N

Vacuum Grease

Figure 2.7: Attachment of temerature control diodes to cryostat and optical ring
sample holder.

47
original radiation shield was blocking too much of the ion beam to be used in its
manufactured form. The low throughput was caused from the openings in the cylindrical

radiation shield being ~0.70”. from the surface of the cryogenic window. If this distance

could be shortened, an aperture smaller than %" could be used, with less of a shadowing
effect. The solution was to replace the bottom portion of the radiation shield with a new
end piece, which also had a rectangular cross-sectiOn like the window holder, and could be
attached to the remaining cylindrical radiation shield. The original and modified radiation
shields are shown in Figure 2.8. The first modified end piece was constructed from a solid
copper block and was fastened to the radiation shield with a two set screws. The
temperature of the window holder was now measured to be about 6.5 K. The ion current
measured at the window was now on the order of 60% of the incoming beam, instead of
30%..

While a temperature of 6.5 K is sufficient for Ne matrices, the difference from the
expected temperature was still concerning. The warming was originally attributed to poor
thermal contact between the end cap and the radiation shield. To improve the contact, the
new end piece was silver-soldered to the radiation shield. After this modification, the
temperature of the cryostat window actually rose to 10 K, which is unusable for neon.
Similar results were obtained with a end cap made from copper foil and attached with bare
copper wire. The cause of this warming was eventually found to be stray radiation leaking
into the radiation shield, being reflected off of the inside copper surface, and eventually
warming the sample window. This is a very common problem in cryogenic systems. The
most common solution to this problem is to blacken the reflective surface, but most paints

outgas significantly and are not UHV compatible. The inside of the shield cap was made

48

Top View

 

EW. modified

<9 original

I Window Holder

 

I Radiation Shield
1: Spectroscopic Window

Figure 2.8:

Side View

 

ll'_.||

 

m
I Cryostat

0 Copper Wire
- Grid

Modified and original radiation shield assemblies. Grid shown

on spectroscopic window was not present for all experiments.

49
non-reflective by painting it with Stycast epoxy, which is UHV compatible. Once the
shield cap was painted with the epoxy, the window temperature was ~ 4.5 K. which was
actually an improvement over the original radiation shield. The epoxy also showed no
outgassing detectable by the Dycor RGA. Because removal of the solid cap soldered onto
the radiation shield also proved to be cumbersome and wastefirl of relatively expensive
indium, the copper foil shield cap was used in all subsequent experiments.
D. Electrical Isolation of the Sample Window Holder

In the course of investigating the method by which counter-ions were being
formed in the matrix as mass-selected cations are being deposited, it became necessary to
electrically isolate the cryostat window holder from earth ground. This would prove to be
challenging, as the cryostat and the UHV chamber are both metallic, and connected to
earth ground by various electrical components such as the cryostat and cryopump valve
motors, and the electropneumatic valves. Our approach was to replace the screw and base
of the window holder with an electrically insulating material. The challenge was in finding
a material that is an electrical insulator, machinable, has a high thermal conductivity, and is
mechanically robust enough to withstand repeated cycling from room temperature to 4 K.

The requirement of machinability severely limited our choices of materials with
which we could fabricate the screw. Vespelm, a machinable polyamide polymer was used
for this purpose. The original copper screw was removed from the window holder and the
VespelTM screw was put in it’s place. The VespelTM screw is fastened to the window
holder by a small brass screw placed inside of the VespelTM screw. To fithher strengthen
this attachment, the two pieces were also glued together using Stycast epoxy. A sapphire

window (Edmund Scientific A43, 368) was modified by the MSU Scientific Glassblowing

50
Laboratory and used as a washer between the base of the window holder and the cryostat
cold head. Some of the original base material was removed to compensate for the
thickness of the sapphire window. This design is shown in Figure 2.9. The combination
of the Vespeln‘ screw and sapphire washer was successfiil in electrically isolating the
window holder, and had sufficient thermal conductivity to maintain the window at
approximately the same temperature as the original window holder. An electrical
connection to the window holder is made with 28 gauge magnet wire. This wire was
wrapped around the cryostat third stage for heat-sinking purposes and is connected to a
BNC feedthrough on the instrumentation skirt on the Heliplex“.
E. Sample Windows

The sample window is the standard HMX series optical sample holder (APD
78332A2). This ring holder accepts 25 x 1 mm windows. CsI windows (APD
254138132) were used for IR studies, and a 25 x 1 sapphire window (Royln Optics
55.6025) was used for the dispersed emission experiments. The Csl windows,
unfortunately, have a relatively short lifetime. The cryostat windows require frequent
cleaning. CsI is very soft, and as such the polishing process invariably removes some
material from the window. After 5—8 polishings, the windows become too thin to use and
must be replaced.
IV. Matrix Gas Line

The vacuum manifold used to prepare matrix gas and introduce the gas into the
UHV chamber is shown in Figure 2.10. This system consists of a glass manifold pumped
by a water-cooled 2” oil diffusion pump (Varian HS-2) backed by a direct drive

mechanical pump (Edwards 8). Low pressures are read with a standard Bayard-Alpert ion

To BNC
Feedthrough

f Brass Screw

Sapphire Spacer

‘: /Threaded Vespel Rod
:3 g g

:e— Stycast Epoxy

 

28-Gauge _._)

Magnet Wire
0 e
(D 0
M

Optical Ring
Sample Holder

 

 

Figure 2.9: Cross-sectional view of spectroscopic window holder modified for
electrical isolation.

52

     
  
    

Ion Gauge
Tube

Diffusion
Pump
Capacitance

Manometcr \lcclmnuyil

l’ump

Figure 2.10: Vacuum manifold used for matrix gas preparation and its introduction
into the UHV chamber.

53
gauge (thoriated iridium filament) controlled by a Veeco RG-830 ion gauge controller. A
DV-6M thermocouple is also placed after the diffusion pump to monitor the forepressure,
as well as the manifold’s pressure during rough evacuation. High ( > 1 mTorr) pressures
are read with a Celesco DMZ digital capacitance manometer equipped with a DP-30
transducer. The original display of the manometer consisted of 4 nixie tubes, all of which
have burned out. Since replacement of these tubes is prohibitively expensive, the pressure
is now read fi'om the output terminals of the manometer. The response function of this
output is 1 Volt/ 100 Torr and can be read with any voltmeter. The base pressure of this

system is ~ 7.5 x 10'7 Torr. The manifold is equipped with several ground glass and V4”

OD tubular glass connections sealed with Teflon stopcocks. A 5.225 liter bulb is used as
the main reservoir. The volume of the manifold is 0.801 liters, and the volume of the
connected manifold and bulb is 6.026 liters.

The manifold is connected to the UHV chamber by V4” OD c0pper tubing. The
coupling between the glass and copper tubing consists of two stainless steel Cajon Ultra—
Torr fittings with a V4” diameter flexible stainless steel bellows tubing in between. This
type of coupling should always be used when connecting rigid tubing to glass to avoid
breaking the glass tubulation. The matrix gas transfer line is connected to the UHV
chamber by a Cajon Ultra-Torr fitting on the inlet of a Vacuum Generators MD7 high-
precision leak valve.

High purity Ne (AGA) and Ar (Matheson), both 99.9995%, were used as received
as matrix gases. In some experiments, C02 (Aldrich, 99.8+%) was added to the matrix

gas. Matrix gas mixtures were made according to standard manometric techniques.

54
V FTIR System
A. FTIR Spectrometer

The spectrometer originally used for these experiments was a Bomem DA3 FTIR
spectrometer. This high resolution, high sensitivity, research grade spectrometer was
ideally suited for this research in terms of sensitivity and flexibility. Its interfacing to the
UHV chamber is described in detail elsewhere.2 It did however, have one serious design
flaw, which resulted in the great Christmas Eve flood of 1991. The source cooling shroud
began to leak, introducing approximately 100 gallons of H20 into the spectrometer.
promptly dissolving roughly 25% of the KBr beamsplitter. Contact with the ensuing KBr
solution also destroyed virtually all of the optics and electronics in the spectrometer bench.
The bench was a total wash (pun very much intended).

The Bomem spectrometer was replaced with a Nicolet 520P FTIR spectrometer.
This is a dedicated mid-IR instrument. The instrument’s source is a closed-cycle, water-
cooled SiC GlobarTM which is useful over the range of 250—10,000 cm". The heart of the
optical system is an air bearing Michelson interferometer equipped with a Ge/KBr
beamsplitter. The specified spectral range of this beamsplitter is 7400—350 cm". A HeNe
laser is used for position referencing of the moving mirror. The 520? bench has two
optical paths. The standard path focuses the IR beam in the bench’s sample compartment
and into the internal detector compartment. The spectrometer can alternatively be
configured to direct a collimated beam through a port on the side of the bench to a remote
location. A computer controlled mirror inside the bench is used to select the optical path.
For the experiments reported here, the external beam is focused through the UHV

chamber by a series transfer Optics originally used in the Bomem spectrometer (the only

55
ones that survived). This series of optics consists of two mirrors, M1 (90° off-axis
paraboloid, 6” effective focal length, 3” aperture) and M2 (90° off-axis ellipsoid. 2" and
22" focal lengths, 1.25" aperture). Afier the IR beam passes through the UHV chamber it
is focused onto a remote detector. Two detectors were used in these experiments. For
high sensitivity, a narrow-band mercury-cadmium-telluride (MCTA) detector (Nicolet
840-111900) was used. Unfortunately, this detector’s increased sensitivity comes at a cost
of spectral range and is usefiil only between 4800 and 720 cm". Spectral data below
720 cm'1 were obtained with a wide—band mercury-cadmium-telluride (MCTB) detector
(Nicolet #840-11200) was employed. As received fi'om Nicolet, the range of this detector
is 6,000500 cm". The original MCT crystal was damaged, and replaced with a different
crystal (Belov Technologies). This MCT crystal is somewhat more sensitive that the
original and its peak sensitivity is near 500 cm'1 which better complements the MCTA
detector. The detectors are placed in an aluminum housing mounted on a table equipped
with roller bearings for height and side adjustment of the detector. All aspects of
spectrometer bench control, data acquisition, and data processing are controlled by an
IBM compatible microcomputer, using Nicolet's PC/IR software (version 3.2).
B. Interfacing of the Spectrometer to the UI-IV Chamber:

The Nicolet FTIR spectrometer is placed on a table constructed so that the
external IR beam is at the same height as the spectroscopic windows on the UHV
chamber. Since the IR beam is difficult to visualize, alignment of the various components
of the system was performed using the HeNe beam, which follows the same optical path
of the interferometer, whenever possible. When necessary, the IR beam was visualized

with an IR viewer borrowed from the MSU LASER Laboratory. The transfer optics

56

salvaged from the Bomem system are attached on a shelf which is screwed into the
baseplate of the spectrometer bench. After the mirrors are properly aligned, the bench and
mirrors are placed close (~ 3—5 mm) to the window on the UHV chamber, and aligned so
that the HeNe beam passes through the center of the two windows and the intervening
matrix substrate of the chamber. The detector housing is then place relatively close to the
window on the opposite (exit) side of the UHV chamber. Fine alignment is accomplished
with the help of the ALIGN function of the PC/IR software. This fiinction triggers data
collection and displays the interferogram as well as its peak intensity as it is being
collected. The height and position of the detector are adjusted until this signal is at a
maximum. When the system is correctly aligned, the throughput should be 20 000 counts
with the small aperture, MCTA, gain of 2, and the CsI window in the cryostat. The
external optical system is quite efficient, achieving 50% of the throughput achieved in the
bench. Figure 2.11 illustrates this optical system, as well as the UHV chamber and mass-
selective ion source.

The windows originally used on the UHV chamber were made of Can mounted
on 2%" Conflat'D flanges (Harshaw 8960-1 -CaF2). These windows were chosen due to the
favorable vacuum properties of Can, namely high rigidity and low permeation rate, and
because its optical properties are suitable for both IR and LIF spectroscopy.
Unfortunately, the Can do not transmit light below 900 cm". To extend the spectral
range of the system, these windows were later replaced with KBr windows (Harshaw
8960-1-HBr), also mounted on 2%" Conflat‘” flanges. Although the KBr windows have a
slightly higher leak rate, the UHV system is quite capable of compensating for the

increased load, and no difference in seen in the matrix background spectra. To remove

Figure 2.11:

57

    
        
  

IR transfer optics

linear motion

3/feedthrough

gate valve

ion optics

 

 

 

Ne Sprayer

 

DC octopole
Csl window

Faraday plate quadrupole
(retracted)

ion source

diffusion pumps

Instrument for FTIR spectroscopy of mass—selected, matrix isolated
cations showing UHV chamber, ion source, and FTIR spectrometer.

' 58
atmospheric gases from the optical path, and to protect the KBr windows, the spaces
between the spectrometer, the UHV chamber, and detector housing are sealed with
polyethylene and purged with dry nitrogen supplied by bleed-off from the MSU Chemistry
Department liquid nitrogen tank.
VI. “Allions” Measurements

While initial FTIR experiments showed that mass-selected cations could be trapped
in sufficient quantities for FTIR detection, in some cases the identity of the anionic species
that must be present to maintain charge neutrality was still unknown. The “allions”
experiments were attempts to at least indirectly detect the presence of anionic species.
The idea behind these experiments was to measure the transient current produced by ions
striking the Faraday plate as the neon matrix is evaporated. A typical experimental
configuration is shown in Figure 2.12.

The ion detector used was the Faraday plate, which had been installed in the
chamber to measure and monitor the incoming mass-selected ion beam. Allions current
was measured and converted to a voltage using a Keithley 480 picoammeter. The
transient signal was then recorded with a Keithley DAS-8 data acquisition board and
processed using Keithley Easyest LX 1.0 software. Data were typically collected in the
strip chart mode at 50 Hz. The Keithley data acquisition system was later replaced with a
National Instruments AT-MIO-16X data acquisition board, controlled with National
Instruments LabVIEW 3.1 for Windows. In the LabVIEW environment, the programs
developed for instrumental control are referred to as Virtual Instruments (Vls). Data were
usually collected using the allionsllyi VI. This is a modification of a basic signal

acquisition VI supplied with the LabVIEW software package. The “wiring diagram” for

59

Cu heat shield

lll|-- window 5\ /\!!l\
grid

(opfionaD

 

 

 

 

Faraday plate

 

\ battery

(opfionaD

 

 

ADC

 

 

 

 

 

 

 

Figure 2.12: Schematic description of "Allions" experiments. Components
marked optional were not used in all experiments.

60
the allions VI can be found in the file allionl l.vi. The ADC was typically set to a :10
Volt input range and sampled at 10 Hz.

In the original allions experiments, all components were held at ground potential.
Positive and negative ions alike can strike the Faraday plate, allowing all ions to be
sampled (hence the name). Since the ions were not influenced by any directing fields,
however, cation/anion recombination was the dominant process, and currents representing
only a small fraction of the deposited ions were observed. To improve collection
efficiency, either the window/window holder or the Faraday plate was held at some
potential other than ground. Attempts to float these components with a power supply
were unsuccessfiil owing to fluctuations in the power supply that caused erratic currents
larger than those due to the ion signal. To alleviate these problems, the Faraday plate or
window holder was floated by placing a battery (either 70 or 300 Volts) in-line between
the picoammeter and the component to be floated. Use of the battery introduced an offset
due to leakage current in the BNC cables; however, it was constant and did not interfere
with the ion current measurements.

VII. Counter Ion Generation Measurements

Once the identity of the counter ions was determined (see Chapter 4), experiments
were undertaken to determine their origin. The goal of these experiments was to measure
an anion current as a function of the incoming cation beam. The ion current was measured
simultaneously at the Faraday plate in its retracted position and at a copper plate
substituted for the CsI window. The ion currents were measured and voltage converted
by two Keithley 480 picoammeters. The voltage signals were then filtered with lO-I-Iz

low-pass RC filters, digitized by a National Instruments AT-MIO-16X data acquisition

61
board and processed using LabVIEW software. In some cases it was necessary to float
the copper plate that replaced the C51 window at 300 V in order to deflect the incoming
cation beam. This was done by placing a 300 V battery (Eveready 493) between the
copper plate and the picoammeter. This experimental configuration is shown in Figure
2.13.
VIII. Matrix Emission Studies.

It is believed that most ions undergo charge recombination neutralization once the
matrix is warmed beyond annealing temperatures. Since this recombination most likely
results in highly excited neutral species, emission is expected; detection of emission and
has been suggested as a test for ion-containing matrices.6 This emission was first visually
detected in a CHgi/COZ' containing matrix. This matrix exhibited an eerie blue-white glow
for several minutes after the cryostat had been turned off.

The total emission accompanying the warming of several ion-containing neon
matrices was measured using a photomultiplier tube (Hamamatsu R928) placed in front of
a window on the UHV chamber. These signals were generally quite intense, and focusing
was not required The experimental configuration is shown in Figure 2.14. Anode
potentials of 800-1000V were applied to the PMT using a HP 6110A DC power supply.
The signal current was measured as a voltage across a 1 M0 resistor with a National
Instruments AT-MIO-16X data acquisition board and manipulated with LabVIEW 3.1
software. Data were acquired using the matrix emission V1 (emile.vi)

Wheras these experiments demonstrated that ion-containing matrices do emit upon
warming, they did little to identify the species responsible for the emission. The next step

was to disperse the emission from the warming matrix. This was accomplished using a

62

Cu heat shield

/\

E- 300 v I
Faraday plate

 

_

 

 

 

 
 
 

 

 

filter

 

 

 

 

 

 

Figure 2.13: Schematic illustration of anion measurement experiments.

63

 

Matrix \Vrndow > H

UHV Chamber —-> A

 

 

PMT

 

Figure 2.14: Schematic diagram of instrumentation used to collect total
emission from a warming matrix

u:
it
dc
be

lit

Oil

64
UV/VIS spectrometer system borrowed from the MSU LASER Laboratory. The system
consists of a 0.22 meter monochromator (SPEX 1681) with a 300 grove/mm grating
blazed at 4‘” order. The detector is a 1.024 element intensified photodiode (Princeton
Instruments TRY-1024 N/RB). Data collection and processing is controlled with a IBM
compatible microcomputer using Princeton Instruments’ NSMA software.

During the dispersed emission experiments, some problems were encountered with
undispersed light. Initially, the optical setup shown in Figure 2.15 was used. This
configuration was chosen because it involved removing only the FTIR detector, and not
the FTIR bench, making it more compatible with other experiments. However, it had
some serious problems. The distance between the cryostat window and the vacuum
chamber is ~ 8”. The first focusing optic can be at best placed just outside this vacuum
chamber window, resulting in a low collection efficiency. Also, due to the reflective
nature of the chamber interior, a great deal of scattered light was also collected. While this
scattered light wasgemission from the matrix, it had a wide angular distribution, and could
not be focused to a point. As a result, a large amount of undispersed light reached the
detector, resulting in intense, but unfortunately useless signals.

Several measures were taken to remove the undispersed light. First, the collection
optics were moved closer to the matrix window by collecting from the FTIR side of the
UHV chamber. Second, an aluminum “light pipe” was added. This pipe extends from the
cryostat radiation shield to the UHV chamber window and blocks out any light that has
been reflected from the chamber walls. In the new configuration, the light must pass
through the cryostat window. The CsI window was replaced with a sapphire window,

owing to its more favorable UVNisible transmission properties. The emission was now

65

 

Matrix Wlndow

(— UHV Chamber

 

f/4 lens

 

 

 

Undispersed Light

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.15: Schematic diagram of original instrumentation used to collect
dispersed emission from a warming matrix. The effect of
undispersed light is shown.

66
focused into the monochromator using two lenses (f/7, f/8) instead of just one. This
configuration is shown in figure 2.16. It was successfiil in eliminating the undispersed

emission and also resulted in enhanced collection efficiency.

 

 

 

 

 

 

Z

 

 

 

 

 

 

 

 

 

 

8H
f/4 lens > <.._.'> --
4 ..
f/7 lens > d. > --
‘ .- 8”
Light Pipe ~ I I
Matrix Window > r—r ~—

 

UHV Chamber ——->

Figure 2.16' Schematic diagram ofinstrumentation used to collect dispersed
emission from a warming matrix

(1)
(2)
(3)
(4)
(5)
(6)

List of References
Sabo, M. 5.; Allison, 1.; Gilbert, J. R.; Leroi, G. E. Appl. Spec. 1991, 45, 535.
Sabo, M. S. Ph.D. Dissertation, Michigan State University, 1991.
Gilbert, J. R. Ph.D. Dissertation, Michigan State University, 1990.
Halasinski, T. M. Ph.D. Dissertation, Michigan State University, 1996.
Heliplexo system manual, APD cryogenics.

Knight, L. B. Jr. Personal communication, 1992.

68

Chapter 3
Preliminary Experiments: The Infrared Detection

of Mass-Selected, Matrix-Isolated CF;

I. Background

The ultimate goal of this project is to couple mass-selective ion sources that
produce suficient numbers of both positively and negatively charged species with the
appropriate matrix-isolation hardware to allow the determination of ionic structures by
vibrational spectroscopy. We initially believed that mass-selected cation and anion
sources would both be necessary to maintain charge neutrality in the rare-gas matrix.
\Vrthout the presence of counter charges, the matrix will charge to an electrostatic
potential equal to the kinetic energy of the incoming mass-selected ions. Once this
potential is reached, the incoming ions will be deflected away from the matrix, resulting in
termination of the deposition process. Under our experimental conditions, this is
predicted to occur in ~ 3 ms,1 and corresponds to a deposition of ~ 0.5 femtomoles of
ions, roughly seven orders of magnitude below the detection limits of FTIR absorption
spectroscopy.

Although the need for dual ion sources seems essential, previous applications of
mass-selected deposition by the groups of both Maier2 and Leroi3 were successful using
only mass-selected cations. Although these studies used the more sensitive spectroscopic
probes of electronic absorption and laser-induced fluorescence, respectively, they too
should have failed without the presence of counter charges for the reasons outlined above.

The success of these experiments suggested that anions were indeed being generated by

69

70
some mechanism in these experiments. These results prompted us to begin preliminary
experiments using only mass-selected cations upon completion of the modified Finnigan
positive ion source.
II. Experimental
A. Choice of Test Cation

The first task in these preliminary studies was to choose a cation to be used to
investigate the performance of the new ion source and matrix isolation system. The
trifluoromethyl cation (CF3I) was selected as our test ion, for a variety of reasons.
Spectroscopically, many aspects of CF 3" made it a very appealing choice. Its infrared
active transitions in rare gas matrices were already known, facilitating definitive spectral

"5 These transitions, typical of vibrations involving carbon-

identification of the species.
fluorine bonds, are also quite intense, thus improving the sensitivity of the experiment.
The expected frequency of the antisymmetric stretch (v3) CF3”, ~ 1670 cm”, was above
the 900 cm'1 cutoff imposed by the Can windows on the UHV chamber, and also lies near
the maximum sensitivity of the MCTA detector. CF;+ also has a relatively low (for
cations) electron affinity (9.2 eV)‘, allowing observation in both argon and neon matrices
without significant spectral perturbation due to charge transfer.

Along with favorable spectroscopic properties, CF3+ also had several attractive
mass spectrometric properties. Relatively high currents of this cation can be generated by
electron impact of several gaseous precursors, such as CF 3Br, CF3Cl, CF, and CF 3H.
C133" (m/z 69), is the most abundant ion in the 70 eV EI mass spectra of all four

precursors. The large mass separation of the halogenated fragment ions allows the beam

intensity to be increased by decreasing the resolution of the quadrupole, without

71
introducing ionic species other than CF;+ into the experiment. The only exception was
possible “interference” of CFzH+ in an investigation of €in when CF3H was used as the
precursor. The decision to use CF 3” was cinched when we discovered that we already had
a large tank of CF;C1 in our laboratory.
B. Instrumental Parameters

Beams of CF3+ ions were generated by electron impact of several precursors using
the modified Finnigan 3200 quadrupole mass spectrometer that is described in detail
elsewhere.1 Typical CF; currents, measured at the sample window with a Faraday plate

in the experiments reported, ranged fi'om 1.5-2.5 x 10" amperes. This rate corresponds to

an ion flux of 0.6-0.9 nanomoles per hour. The beam of mass-selected ions was co-
deposited with excess neon or argon onto the sample window (CsI or sapphire) held at ~ 5
K. High-purity neon (AGA) and argon (Matheson), both 99.9995%, were used as the
matrix gases. The flow rates of the matrix gas through the sprayer located in fi'ont of the
window was ~ 0.5 mmole/h in all experiments. All halocarbon gases (CF3C1, CF3BT, CF4,
CF 3H) were used without further purification. For the experiments described in this
chapter, IR radiation was transmitted through the UHV chamber using Can windows.
FTIR data were recorded by averaging either 256 or 1024 spectra between 4800 and 900
cm”, taken at l-cm’l resolution. The data were background corrected by ratioing with a
spectrum collected prior to matrix deposition.

The movable Faraday plate used to measure the mass-selected ion current at the
sample window was also used to collect and measure the current of both positively and
negatively charged species as they were released from the matrix upon warming. The

matrix was warmed by using the foil resistance heater supplied with the cryostat, and

72
controlled by a Lake Shore 330 temperature controller. The heater was set at the HIGH
power level and a setpoint of 40 K. The current from the Faraday plate was measured and
converted to a voltage signal using a Keithley 480 picoammeter, digitized with a Keithley
DAS-8 data acquisition board controlled with Keithley Easyest LX 1.0 software. Unless
otherwise noted, data were collected in the strip chart mode at 50 Hz.
III. Results and Discussion
A. FTIR Results

Co.deposition of 15-25 nA beams of €ng fiom CF3C1, CF3Br, CF, or CF3H and
neon for up to 25 hours onto the cryogenic substrate resulted in the appearance of two
new infiared absorptions, at 1670 and 1651 cm". Only these bands, shown following
generation of €ng fi'om CEO in Figure 3.1, among absorptions due to matrix-isolated
H20, CO; and precursor molecules, have intensities that correlate with the integrated ion
current. As shown in Figure 3.2, the most prominent of these background absorptions are
those of the precursor, CF3C1, and of C02. CD; which is always present as a background
gas. After 11 hours of deposition, the absorbances of the most intense of these transitions,
v, of CEO at 1104 cm”,7 and v3 of CO; ' were ~ 0.20. We assign the 1670- and 1651-
cm'1 peaks to the antisymmetric stretch (v3) of mass-selected, matrix-isolated CF3+ trapped
in difi‘erent matrix environments. Although 19 cm’1 is unusually large for a site splitting,
this assignment has been supported by other researchers.’

The positions of the 1670- and 1651-cm“l bands are consistent with previous’ and
later’ assignments of v; of CF; matrix-isolated in neon. In the neon matrix work by Jacox
and co--workers,9 split absorptions at 1670, 1664 and 1651 cm'1 were assigned to v3 of

CF3+. While we did not generally observe the feature at 1664 cm", our work and that of

73

 

 

(1001IKLL h

 

 

W

Absorbance

 

J l

 

 

1680 1660 1640

Wavenumbers (cm-1)

1700

Figure 3.1: Antisymmetric stretch (v,) of mass-selected, matrix-isolated CF ,i in neon.

74

 

 

 

 

 

 

 

 

0.05 A.U. co,
0)
g cr=,cr
(U
E
O
(D
.D
<
H20 H20
1 M K
5000 4000 3000 2000 1000

Wavenumbers (cm")

Figure 3.2: FTIR absorption spectrum in the 4800—900—cm'l region of a pure neon
matrix after 11 h of deposition of CF; generated from CF,Cl.

75
Jacox indicated that the relative intensities of these absorptions are somewhat sensitive to
the composition of the matrix. m The antisymmetric stretch of €ng also has been reported
in the range 1663-1667 cm'1 in argon matrices."”"12 When we deposited mass-selected
€ng generated fi'om CEO in an argon matrix at 5 K, an absorption at 1667 cm'1 with an
apparent shoulder at 1665 cm'1 was observed; see Figure 3.3.

Although the fiequencies of the new absorptions that we observed are consistent
with those of CF33, there was still the possibility that these absorptions were due to an
exotic neutral species created during operation of the ion source. Control experiments
demonstrate that the 1670- and 1651-cm'l absorptions are not due to neutral species
generated during the ionization process, and are indeed due to mass-selected, matrix-
isolated CF3+. The results of these experiments are summarized in Figure 3.4. Trace a
shows the IR absorption between 1700 and 1640 cm“1 of a neon matrix after 11 h of
deposition of CF; (m/z 69, 20 nA) generated from CF3CI. When the mass spectrometer
was set to select m/z 120, but all other experimental parameters were unchanged, no ions
were transmitted for deposition in the matrix, yet the flux of neutrals from the ion source
remained. During such experiments, the peaks at 1670 and 1651 cm'1 did not appear
(Figure 3.4, trace b), but the rest of the spectrum was unchanged. To determine if these
features were due not to mass-selected CF3", but to CF 3” formed upon collisions of the
ions with solid surfaces, gas phase species, and/or the growing matrix, mass-selected
Cle-F (m/z 51) [along with a small amount of CF;+ (m/z 50)], generated from CF3H, was
deposited into the neon matrix. As illustrated in trace c of Figure 3 .4, the new peaks were
again absent. It should be noted that bands attributable to CFgH’ or CF 2* were not

observed, but based on previous assignmentsll the CFZIF absorptions would have been

76

 

10.01 A.U.

 

 

 

 

 

 

G)

O

C

.23 mo

0 r I

3
<2 CF!

1 700 1 650 1600 1550

Wavenumbers (cm")

Figure 3.3: FTIR absorption spectrum in the 1700—900-cm'l range of a pure
argon matrix after deposition of CF; from CF,Cl.

 

 

 

 

I 0.002 A.U.

a)
o
c
(U
.0
5 m/z 69
a)
in”

m/z 120

m/z 50/51

1 700 1 680 1 660 1 640

Wavenumbers (cm")

Figure 3.4: Infrared absorption spectra in the 1700 - 1640 cm’1 region of neon
matrices after 11 hours of deposition of: a) m/z 69 (CF3+, 20 nA)
generated from electron impact of CF3C1; b) m/z 120 (no ion current)
while subjecting CF3C1 to electron impact; and c) m/7. 50/51
(CF2+/CF2H+, 15 nA) generated from electron impact of CF3H.

$1113

78

obscured by background water absorptions, and the amount of €in deposited was too
small to observe in this experiment.

Ifthe CF; ions are partially neutralized during deposition, spectral features due to
CF; should be observed. An extremely weak absorption was observed in neon matrices at
1254 cm", which agrees with previous assignments5 for CF 3' in neon. The low intensity of
this feature (absorbance = 0.0007), indicates that only barely detectable quantities of CF;
are present in the matrix. Features due to other neutralization/fragmentation products,
such as CF; ’3 and CF ‘3 are not observed.

B. Indirect Detection of Counterions

In the absence of counter charges, one cannot accumulate a sufficient number of
positively-charged species in a matrix to obtain a measurable infrared absorption spectnim.
When pure neon was used a the matrix material, no absorptions were observed that could
be assigned to an anionic species. In some experiments CC14 was added to the matrix gas
(1:500) to attempt to increase cation yield, and to investigate the possibility that electrons
were the counter charges. CC14 acts as an electron scavenger by the following reaction:

CCl4 +e' —> CCl,’ +Cl’

If electrons were present and migrating through the matrix, the reaction with CCL will
immobilize them, and hopefully reduce the number of ions lost to neutralization.
Unfortunately, the products of this reaction are not directly observable by our
instrumentation at that time. Electrons, and the chloride ion, of course, have no IR
spectnim and the highest energy IR-active vibration of CC13', at 898 cm", lies just below
the low energy cutofi‘ of the Can windows of the UHV chamber. The 1670— and 1651-

cm'1 peaks appeared both with, and in the absence of, CC14 in roughly equal intensity.

79
This result was consistent with our spectroscopic findings that neutralization was not
occurring on a large scale. It also indicated the counter charge in these experiments was
probably not a free electron.

When argon was employed as the host, we also observed new FTIR bands at 938
and 933 cm", shown in Figure 3.5, when Cng was deposited. These bands correspond
closely to those previously (and tentatively) assigned to CF3C1' in argon by Prochaska and
Andrews.‘ Presumably this species is formed by electron attachment to CF 3C1, which is
present in the matrix in abundance. The origin of the electrons, however, is not clear.
They may be extracted fi'om the grounded window holder, or they could be ejected from
the various grounded components of the cryostat upon impact of the ion beam and
subsequently deposited into the matrix.“ Since corresponding peaks are not readily
observed in our experiments with neon matrices, this may indicate that the mechanism of
counter-ion formation is matrix dependent.

By an as yet unknown mechanism, electrical neutrality was being maintained.
Some negatively-charged species are in fact present and were indirectly detected in the
matrix, although their identity was not certain. When the Faraday plate, installed to
monitor the impinging ion flux, was placed directly in fi'ont of the cryostat window as
neon matrices containing mass-selected cations were warmed to about 20 K current was
measured. As shown in Figure 3.6, very small (picoamp) transient signals due to both
positive and negative species entrained in the vaporizing matrix were measured when the
neon matrix was warmed. For the control experiments, where the peaks attributable to
€ng were absent, no such transient ion signals were observed. These results indicate that

negatively-charged species are being produced and matrix isolated as a result of the

80

 

0.01 A.U.

 

 

CF,C|' ?

WWW WW

1 k I . l . l a l . l

 

é

Absorbance

 

 

 

 

 

970 960 950 940 930 920
Wavenumbers (cm")

Figure 3.5: FTIR absorption spectrum in the 970—920-cm’I range of a pure argon
matrix after deposition of CF,+ generated from CF,C1.

81

 

 

 

 

 

 

2?

3- 20

9‘3 0

S

0 -20 -

2

s -40 -

O.

E‘ -60

8

ts

LL I - I - I .. I
5 15 25 35

Time (seconds)

Figure 3.6: Transient current observed at the Faraday plate located ~l cm from
the substrate during warming of a neon matrix afier a 25 hour
deposition of 20 nA of CF3+ generated from electron impact of
CEO. The initial temperature of the matrix was ~4.5 K The final
matrix temperature temperature of ~20 K is reached within 10
seconds.

82
deposition of positive ions. This experiment also provides some insight into the
composition of the matrix. If the matrix were to have a homogeneous distribution of
positive and negative charge carriers, they would either recombine and be neutralized, or
their signals would cancel. The presence of transient currents of both sign may indicate
that the counter-ions are being formed in discrete time periods, rather than continuously,
suggesting some sort of layering in the matrix. When an argon matrix containing CF3+
was subjected to two hours of 40 nA, 10 eV electron bombardment (the CF 3* absorption
was not reduced after this electron bombardment), the transient signal shown in Figure 3.7
was measured as the substrate was warmed. In this case, desorbed negative charge
carriers clearly dominate. ‘ This experiment suggests that the matrix can tolerate some
charge imbalance. Perhaps during positive ion deposition, counter-ions are formed only
after the potential in the matrix is sufficiently high to activate the counter ion generation

process.

20

Faraday Plate Current (pA)
C

Figure 3.7:

83

 

 

 

 

. # l I I _I .

 

5 1 5 25 35
Time (seconds)

Transient current observed at the Faraday plate located ~1 cm
from the substrate during warming of a argon matrix after a 25
hour deposition of 20 nA of CF; generated from electron impact
of CF3C1, followed by 2 hour bombardment by 40 nA of 10 eV
electrons The final temperature of the matrix was ~40 K.

(1)
(2)
(3)
(4)
(5)

(5)
(7)

(8)
(9)
(10)

(11)
(12)
(13)
(14)

List of References
Halasinski, T. M. Ph.D. Dissertation, Michigan State University, 1996.
Forney, D.; Jakobi, M.; Maier, J. P., J. Chem. Phys. 1989, 90, 600.
Sabo, M. S.; Allison, J .; Gilbert, J. R; Leroi, G. E., Appl. Spectrosc. 1991, 45,
153881135“ F. T.; Andrews, L. J. Am. Chem. Soc. 1978, 100, 2102.

Chemistry and Physics of Matrix-Isolated Species, Andrews, L., Moskovits, M.,
Eds; Elsevier Science Publishing: Amsterdam, 1989.

Lossing, F.P., Bull. Soc. Chim. Belges, 1972, 8], 125.

Biirger, H.; Burczyk, K. ; Bielefeldt, D.; Willner, H.; Ruoff, A.; Molt, K.,
Spectrochim. Acta, 1979, 35A, 875.

Jacox, M. E.; Thompson, W. E. J. Chem. Phys. 1990, 93, 7609.
Fomey, D.; Jacox, M. E.; Irikura, K. K. 'J. Chem. Phys. 1994, 100, 8290.

Jacox, M. E., National Institute of Standards and Technology, personal
communication, 1994.

Kelsall, B. J.; Andrews, L., J. Phys. Chem. 1981, 85, 2938.
Jacox, M. E. Chem. Phys. 1968, 83, 171.
Milligan, D. E.; Jacox, M. E., J. Chem. Phys. 1968, 48, 2265.

Hagstrum, H. D.; Becker, G. E., Phys. Rev. 1967, 159, 572.

84

Chapter 4:

The Formation of Counterions

I. Background

For the past two decades, matrix isolation has been utilized for the spectroscopic
analysis of ions. A variety of cations, from simple diatomics including NO+ 1 to
polyatomic species such as the molecular ion of pentacerre2 (CnHu'i) have been
characterized by this method. In all of these studies one has to assume that overall charge
neutrality of the matrix is maintained in some way. Consider as an example the generation
of Cch” cations in a neon matrix by photoionization of matrix-isolated CgFgf
Presumably, when C6F5.+ is made some counterion is generated as well. In this case, like
many others, the counterion may be an atomic species such as F, formed by electron
capture of the precursor in the matrix, which is not detectable by ESR or vibrational
spectroscopy. Spectroscopic peaks have been assigned to negatively charged species,4
although correlations between anion and cation abundances usually cannot be made. In
such experiments, in which the matrix is maintained at temperatures below 10 K, one may
expect a variety of species to be present in the matrix that could capture electrons or in
some way form stable anions, such as H20 forming OH'.

A few laboratories have reported success in the spectroscopic analysis of cations
that have been created in conventional mass spectrometric ion sources, and mass-selected

7 The most common mass

before their deposition into a growing inert gas matrix.""
spectrometer utilized for this purpose has been the quadrupole mass filter. If one must

also deposit counterions into the matrix, then these experiments will become very

85

86

challenging. Even though mass spectrometry is a relatively mature field, the generation of
large quantities of negative ions is still a “black art”. The experiments described in
Chapter 3 of this dissertation showed that depositing a beam of CF;+ ions into a growing
Ne matrix over a period of hours resulted in the accumulation of sufficient numbers of
cations for detection using FT IR spectroscopy without the concurrent deposition of
negatively charged species. Clearly, this experiment should not have worked; in the
absence of counter charges, the accumulated positive charge on the sample window would
become limiting for continued deposition on the millisecond timescale. Although
spectroscopic evidence of counterions was lacking in those experiments, we developed a
technique whereby both positively and negatively charged species present in the matrix
could be, and were, indirectly detected on a Faraday plate, held at an electric potential of 0
V, as the matrix was warmed.’

How are counterions generated in matrices upon cation deposition? Are there
some unique features of the experimental setups, that have yielded results to date, that are
critical to success? Clearly, a more complete understanding of counter ion generation is
needed. This chapter describes the direct observation of counterions in matrices formed
during the codeposition of mass-selected cations and neon. We have discovered that the
addition of carbon dioxide to the neon matrix gas during deposition of a mass-selected ion
beam results in detectable quantities of C02". Use of CO; as a matrix dopant also
enhances positive ion accumulation efficiencies. The yield of stabilized cations increases
by roughly a factor of five when C02 is present. Experiments utilizing matrices doped

with CO; and the measurement of the current due to negatively-charged species have been

87
used to determine the mechanism that may provide the counter-charged species when
mass-selected cation beams are deposited in low temperature matrices.

We have considered three possible mechanisms, illustrated in Figures 4.1—4.3.
Mechanism 1 (Figure 4.1) is a condensed phase mechanism, in which accumulating
positive charge in the matrix creates a field which is sufficient to extract electrons from the
grounded metal window holder (with which the matrix is in contact). Thus, the primary
negative charge is an electron, which migrates through the matrix and is captured by some
molecular species capable of forming a stable anion. Mechanism 2 (Figure 4.2) is a gas
phase mechanism. The incoming ions have sufficient kinetic energies to ionize gas phase
species (exemplified by C02 in Figure 4.2) in front of the growing matrix, forming positive
and negative charge carriers. The negative charge carriers, which may be electrons or
negative ions depending on the molecules that are present in this region, are attracted
toward the positive potential of the matrix. The nascent cation is repelled by the same
field. Mechanism 3 (Figure 4.3) is a surface bombardment mechanism. Similar to
Mechanism 2, incoming ions collisionally generate gas-phase electrons or negative ions
which are subsequently attracted to the matrix. However, in this case the collisions are
with cold metal surfaces and the molecules adsorbed on them (again exemplified by CO; in

Figure 4.3).

11. Experimental
High purity neon (AGA), 99.9995%, was used as the matrix gas. Carbon dioxide,
99.8+%, was obtained from Aldrich Chemical Co. Reagent grade carbon disulfide was

obtained from Fisher Scientific Co and used as received. The matrix gas, pre-mixed with

88

Mechanism 1

Cu Heat Shield
/ (40 K) \

         

a a mass-selected cation
a electron

O—C—O neutral CO2

Figure 4.1: Illustration of proposed condensed phase mechanism of counterion
generation in mass-selection, matrix-isolation experiments.

89

Mechanism 2

Cu Heat Shield
/ (40 K) \

    

mass-selected cation

e
a secondary cation
e

electron

a O—C—O neutral CO,

Figure 4.2: Illustration of proposed gas phase mechanism of counterion
formation in mass-selected, matrix-isolated cation experiments.

Mechanism 3

Cu Heat Shield
/ (40 K) \

    

electron

a mass-selected cation

O—C—O neutral C02

0’0 ‘0 C02

Figure 4.3: Illustration of proposed surface bombardment mechanisms of counter
ion formation in mass-selected, matrix—isolated cation experiments.

91

measured amounts of CO; in some experiments, was deposited onto the sample window at
a rate of ~0.5 mmole/hr. The energies of the mass-selected cation beams were in the
range of 100—130 eV. In some of the experiments described in this chapter, a 70%
transmission copper grid was fit onto the face of the CsI sample window as shown in
Figure 2.8. Isolation of the grid and window holder makes it possible to float these
components at various potentials, both during the ion accumulation process and while
measuring the transient current of positively and negatively charged species as they are
released from the matrix upon warming.

Potentials on components used as current detectors were established by placing a
battery in-line between the metal component and a picoammeter. Electrostatic field
simulations were performed with SIMION PC/PS2 Version 500'

III. Results and Discussion
A. Spectroscopic Detection of Counterions

After successfully obscuring the infrared detection of mass-selected, matrix-
isolated CF 3+ ions generated from a variety of precursors,’ we next selected for study the
radical cation C02”, again using neon as a matrix. These experiments were successful,
with accumulated COz'+ detected at 11422 cm", which had been previously assigned to v;
for the cation by Jacox and Thompson.9 As shown in Figure 4.4, in these experiments we
also observed spectral features at 1658 and 1665 cm", assignable" to C02” (v3) and
(CO;);". This negative ion absorption had also been observed in other cation deposition
experiments where CO; was not intentionally introduced into the experiment, but was
present as a major component of the background gas, in part emanating at low levels from

ceramic elements of the vacuum system. CO; is present in greater abundance in the

92

 

 

 

H20
0 01 AU
CO,-

0)

s \

(U

.D

5

8 CO; C0; +
< \ 20.

 

1 400 1 300

l - l

1800 1700 1600 1500
Wavenumbers (cm")

 

 

Infrared absorption spectrum of a neon matrix in the 1750-1350 cm'1

Figure 4.4:
region following a 20 h deposition of 40 nA of C0,".

93
vacuum chamber during COz'+ depositions owing to the incomplete pumping of the
neutral precursor floor the ion source housing.

The C02” anions observed could have a direct lineage to the incoming cations.
That is, through a charge inversion process, some fraction of the incoming C02” could be
converted into C02”. To investigate this, we chose to repeat the experiment with CF3".
Figure 4.5 shows the spectrum observed when CF 3” is mass-selected and deposited, using
a Ne:COz matrix gas mixture of approximately 100021. Signals representing C02” are still
present, showing that a process is occurring which results in their formation, but they are
not formed directly fi'om the incoming cations.

The presence of anions in the matrix is crucial to isolate sufficient quantities of
cations for detection by vibrational spectroscopy. The use of COz-doped neon matrices
and the subsequent formation of C02” provided two advantages. First, as previously
noted, cation trapping is more efficient, allowing for shorter experiments. Second, it
provides a direct spectroscopic probe of the counter charge identity and relative
abundance, since we believe that, at the C02 levels used, a substantial fraction of the
anionic species present in the matrix is C02” or (CO:):". The growth curves for the
cation and anion signals show a clear correlation between the relative amounts of these
species.

The observation of C02" does not in itself provide information as to the
mechanism of anion formation during cation deposition. However, the determination and
elimination of mechanisms is made easier with the capability for spectroscopic detection of

the anions present. We use the spectroscopically observable absorption due to C02” as a

94

 

I 0.005 A.U. co;

 

o
‘c’

g 00:00; “2°

5 l l
o

g or;

 

 

 

 

 

 

l .. l - l - l

1700 1650 1600 1550
Wavenumbers (cm")

 

Figure 4.5: Infrared absorption spectrum of a 100021 Ne:CO2 matrix in the
1750—1350 cm'I region following a 13 h deposition of 20 nA of CF,+
formed by electron impact of CF,C1.

95
direct spectroscopic probe in our study of anion formation during mass-selected, matrix-
isolated cation experiments.
B. Indirect Detection of Counterions

Isolation of the window holder also permits its use for measurement of transient
currents from the matrix upon warming, similar to use of the Faraday plate which was
employed in the past for these experiments. New modifications to this experiment also
include the ability to float the window holder and Faraday plate at various potentials, in
order to increase ion collection efficiency and to obtain separate signals for anions and
cations by creating fields that will separate charges. Unexpected results are sometimes
obtained when ions are collected as the matrix is warmed. This is particularly true when
the Faraday plate is biased, and the window holder is held at ground. An example,
obtained when the Faraday plate was biased at -70 V with respect to the grounded
window holder, is shown in Figure 4.6. Although one would expect only positive ions to
be collected, both positive and negative charges clearly strike the Faraday plate.

Several points are relevant in considering such results. As the matrix warms, the
dynamics for released, charged species are complex. Warming may occur first at the
matrix/window interface. It is unclear how inhomogeneous heating, with a neon matrix
that rapidly vaporizes, influences the emanating ions. There are a variety of forces that act
on ions which determine their trajectories as the matrix vaporizes. The first is applied
fields, which we create to separate ions of opposite charge. However, ion/ion interactions
likely create stronger and more complex fields. The many collisions of the ions with the
desorbing neon must also be considered. Further, the distribution of anions and cations

within the matrix and across the cold window need not be homogeneous. Thus, it is not

1000

‘6'

Current (pA)
O

6

-1000

Figure 4.6:

96

 

 

 

 

 

0 5 1 0 1 5 20 25 30
Time (seconds)

Currents observed on the Faraday plate from a warming neon matrix
after a 27 h deposition of 40 nA of C0,“. The cryostat heater is
turned on a t = 0 s. At t = 30 s, the temperature was ~40 K. The

potential on the Faraday plate during the collection of the signal was
-70 V with the window holder grounded.

97

surprising that, even when a field is established to collect cations on the Faraday plate,
some anions reach the collector as well. Also, while an experiment with the window
holder at ground and the Faraday plate at -70 V may be assumed to create conditions to
collect only cations at the Faraday plate, all points across the salt window are not at a
potential of 0 V. This situation is illustrated in Figure 4.7a, which shows a SIMION plot
of the potential gradient formed between the elements that create it, the Faraday plate and
the other metallic surfaces. The window location is indicated by the dotted line in the
figure. The full -70 V is not realized by ions originating at the window, and the field is
clearly not linear between the salt window and the collector. The complex potential
surface on which desorbed ions move is not that which we hoped to form.

To create a more ideal electrostatic situation for detection and possibly sorting of
desorbed ions, we placed a metallic grid across the face of the CsI window. A SIMION
plot of the potential gradient formed, with the grounded grid present and with -70 V
applied to the Faraday plate, is shown in Figure 4%. The potential surface is now
considerably more linear between the sample window/grid and the Faraday plate. Figure
4.8 shows the currents observed on both surfaces with the Faraday plate held at a potential
of +70 V as a matrix is warmed. In this plot it is clear that negative signals dominate at
the Faraday plate and positive signals at the grid collector on the sample window,
suggesting that desorbed ion motion is influenced primarily by the applied field in this
case.

The results of all configurations of these experiments show that extensive
cation/anion recombination takes place, and that the signals measured are only a very small

fiction of the total charge present in the matrix before warming. A deposition of 20 nA

Figure 4.7:

 

98

 

 

 

 

 

 

-10v -60v

 

 

 

SIMION plots showing the potential surface formed between the
Faraday plate and the CsI window. Equipotential lines, separated in
value by 10 V, are shown. The CsI window is represented by a
dashed line in both plots. The grounded copper grid has been added
to part b. The Faraday plate has a potential of -70 V, and all other
surfaces are at a potential of 0 V in both plots.

 

 

 

 

 

 

 

 

50 '
40 Current measured
at the grid and window
holder on the
30 " sample window
20 '
A 10 '
<
c J LJL
v 0 I ”M
4—3
C
9 .
h 0 [ 'v
8
-10 '
Current measured
-20 ' at the Faraday
plate
-30 '
-40 '-
'30 L l I I I I l
0 5 10 15 20 25 30

Time (seconds)

Figure 4.8 Currents observed from a 1700:1 Ne:CO2 matrix after a 20 h
deposition of 30 nA of CH; generated by electron impact ionization
of toluene. Heating conditions are the same as Figure 4.6. The
potential of the Faraday plate during collection of the signals was +70
V with all other elements held at 0 V.

100

of cations for 5 h corresponds to 2 x 10" ions in the matrix (assuming all ions are

isolated). A 50 nA ion signal for 1 s, similar to those shown in Figure 4.8, corresponds to

2.5 x 1013 ions. Clearly, most matrix-isolated ions are neutralized or impinge on surfaces

other than those used here as collectors upon warming. Also, because the magnitudes of
the signals vary between experiments, quantitative information cannot be extracted fi'om
them. However, such measurements provide a means for quickly detecting the presence
of cationic and anionic species, of particular importance in experiments where this cannot
be done spectroscopically. The technique is useful in optimizing the experimental system,
since emitted ion currents can be obtained after < 2 hours of ion deposition, whereas
longer times are required when spectroscopic detection is the only tool available for
determining matrix composition.
C. Evaluation of Counterion Generation Mechanisms

In Mechanism 1, we considered the possibility that accumulating cations in the
matrix would create a collective electric field that would assist in the extraction of
electrons from the grounded nickel-plated copper window holder into the matrix with
which it is in intimate contact. To investigate this mechanism, we electrically isolated the
window holder, thus terminating the connection to its source of electrons (ground). If
only cations are accumulated in a growing matrix, and the impinging ions have kinetic
energies of 100 eV, the potential of the matrix cannot exceed 100 V. This will be
established within the first second of deposition. If this potential is sufficient to extract
electrons (or anionic impurities) from the grounded window holder, then isolating the
window holder from ground should prohibit the continuous deposition of cations over

periods of several hours. Experiments performed with the isolated window holder, with

101
all other parameters the same, resulted in no change in the infi'ared absorption spectrum.
Both cationic and anionic species were observed with absorption intensities equivalent to
the earlier experiments in which the window holder was not isolated. Mechanism 1 can
therefore be eliminated.

Unfortunately, it is not experimentally straightforward to distinguish between
Mechanisms 2 and 3. To determine if Mechanism 3 is operative, one may consider ways
to deposit ions without having some fraction of them bombard metal surfaces. However,
the mass-selected ion beam cannot presently be focused sufficiently well to impinge only
on the growing neon matrix and not the surrounding metallic surfaces. Isolation of the
metallic surfaces from ground may be useful; however, this changes the electric potentials
involved in the experiment due to static charging by the unfocused ion beam. Removal of
the radiation shield, another possible approach, would compromise the thermal aspects of
the experiment.

Insights regarding Mechanisms 2 and 3 were first obtained during attempts to
observe IR transitions of mass-selected, matrix-isolated CsFa”. Several new absorptions,
shown in Figure 9, were observed. These transitions, unfortunately, could be assigned to
CF, CB, and CFg, presumably formed fi'orn the fragmentation of Cst”. To attempt to
reduce this fragmentation, the energy of the ion beam was reduced to 50 eV by using the
electrically isolated metallic grid on the CsI window, as described above, and applying to it
a positive potential of +80 V during the course of the matrix deposition. All surrounding
surfaces were at ground potential. The bottom trace of Figure 9 shows that this approach
was successful in reducing the fi'agmentation of the incoming cations. Other results of this

experiments were unexpected. Along with a decrease in fragmentation products, a large

102

 

0.002 A.U. CF,
CF2

CF 130 eV
(D
O
C
to
.Q
L
O
a)
.Q
<

50 eV

 

 

 

1280 1260 1240 1220
Wavenumbers (cm")

Figure 4.9: Infrared absorption spectrum of a 150021 Ne:CO2 matrix in the
1300—1200 cm‘I region following a 22 h deposition 35 nA of 130 eV
COF,’ ions (top trace) and similar amounts of 50 eV C,F,’ ions (bottom
trace). In both cases, the QR; ions were generated from electron

impact of C,F,.

103

absorption due to COz'+ at 1422 cm'1 was observed. This absorption, shown in Figure 10,
is larger than the signal observed when C02” is intentionally deposited into the matrix.
We attribute the appearance of COz'+ to ionization of neutral CO; molecules in the matrix
by electrons that have been formed by cation bombardment of surfaces and are accelerated
towards the matrix by the potential applied to the grid. If Mechanism 2 is operative,
electrons would most probably be formed immediately in front of the growing matrix
where the density of gas phase particles is greatest. An electron formed in this area would
likely not be accelerated to a kinetic energy required to ionize other neutral species in the
matrix so efiiciently. Only in Mechanism 3 would the electrons experience the full 80 V
acceleration, allowing them to ionize neutral species in the growing matrix. Regarding
Mechanism 3, large (~3 0%) secondary electron emission from metal meshes bombarded
with anions has been recently reported. '° Such processes are of concern to mass
spectrometrists who work with negative ions, since ion/surface interactions can yield
electrons which are detected along with incoming anions, making signals artificially large.
Schlag and co-workers'0 attribute the electron origin mainly to the adsorbed residual
gases, which is consistent with the temperature dependence they observed. It seems likely
that a similar mechanism for electron emission occurs to some extent in our experiments.

Other evaluations of Mechanism 2 suggest that it is not the process by which
positive ions are created and deposited into the matrix. As stated in Chapter 3,
experiments performed with mass-selected ion beams of CB” fail to produce absorptions
due to isolated Cng ions in the matrix. Only when ion beams of CF;+ are mass-selected
and matrix-isolated are absorptions due to €ng observed. This is reasonable; even if 100

eV CFz'+ ions collided with neutral, gas phase precursor CF3C1 molecules to form Cng

104

 

Absorbance

 

0.001 A.U. 002+

 

 

 

WW

I l l l l l

 

 

1450 1440 1430 1420 141 0 1400

Figure 4.10:

Wavenumbers (cm")

Infrared absorption spectrum of a 150021 Ne:CO2 matrix in the
1450-1400 cm" region showing CO,“ formed when a copper grid was
used to decelerate incoming C,F.“ cations. Depositions conditions
are identical to those represented by the bottom trace of Figure 4.9.

105
and Cl’, the CF 2” in the matrix would not create a field that would attract gaseous, low
energy CF3+ ions toward the window. Only the ions in the mass-selected ion beam have
sufficient energy to overcome the slight positive potential of the matrix and become
isolated. Therefore, when the window grid is positively biased and we detect C02”, this
cation is not being deposited into the matrix, but is formed in the matrix upon electron
bombardment.
D. Direct Observation of Anion Generation

To firrther investigate whether Mechanism 2 or 3 is the cause of negatively
charged species being deposited into the matrix, the C31 window was replaced with a
capper plate, to measure current that may impinge on the location of the salt window in
these experiments. This copper plate can be grounded or biased while being used as a
Faraday plate. Several experimental configurations were investigated, and four are shown
as Configurations 1-4 in Figures 4.11—4.14, respectively. In each case, the experimental
configuration is shown on the left, the results of the experiments are shown in the middle,
and our interpretation of the results is shown on the right. For each pair of data traces,
the top trace shows current measured at the copper plate that was substituted for the C51
window, and the bottom trace is current measured at the retracted Faraday plate, which
lies approximately 1 inch to the right of the copper plate, outside of the radiation shield.
Each pair of traces was obtained simultaneously. A manual gate valve was opened/closed
to start/stop the incoming C82” ion beam. The ion beam was presented for two one-
minute periods (approximately 30—90 seconds and 150-210 seconds into the
measurement) during the displayed time period. Matrix gas (Ne2C02 1000:1) was present

in all experiments.

Configuration 1

 

 

 

. ISnA

I
Current

 

 

 

 

O

 

 

 

 

I: :l

 

 

 

 

 

 

 

 

 

 

0 :[05nA

30 90 150 210
Time (seconds)

 

H:

it:

Figure 4.11: Direct measurement of anion generation, Configuration 1. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

107

Configuration 2

 

I20 pA

0

Current

    

{713/ M M

IZII—EJ_I__?\

 

 

 

Current

 

 

 

 

 

 

 

30 90 150 210

Time (seconds)
£621 :—
Figure 4.12: Direct measurement of anion generation, Configuration 2. The

experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

108

Configuration 3
FI‘ITUIS nA

Time
1 1 A
“V I n

30 90 150 210
Time (seconds)

 

Current

 

 

 

 

Current

 

 

 

 

O

 

 

557-532,?

Figure 4.13: Direct measurement of anion generation, Configuration 3. The
experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

109

Configuration 4

 

 

 

 

 

 

 

I 1 "A
E
/ 5
O

T \E Vi

9
‘5
O

C t v_._ I1 nA

 

 

 

 

 

30 90 150 210

Time (seconds)
t°$(\—§
Figure 4.14 Direct measurement of anion generation, Configuration 4. The

experimental setup is indicated on the left, data in the center, and the
interpretation on the bottom.

l 10

When all surfaces are at room temperature, Configuration 1 (Figure 4.11), the
incoming cation beam has a current of approximately 30 nA, as shown in trace 4.11a. The
retracted Faraday plate, which sits to the side of the radiation shield, collects
approximately 1 nA of cation current (trace 4.11b), demonstrating that not all of the
incoming ions strike the surface on which the spectroscopically-observable matrix is
formed.

Are negative charges being formed in this experiment, when gases are present and
all surfaces are at room temperature? To test this, Configuration 2 (Figure 4.12) is used.
The copper plate is biased at a positive potential of 250 V. This design is preferable for
detecting any negative current formed for the following reason. In Configuration 1, trace
4.11a could represent the sum of positive and negative ions striking the copper plate.
With a +250 V bias on the copper plate, incoming cations are repelled and only negative
charges are collected. Trace 4.12s shows that a small negative current is formed when a
positive ion beam is admitted into the experimental chamber. We believe that the signal
shown in this trace represents negative charges which are formed by ion bombardment of
nearby metal surfaces, on which molecules are adsorbed. These negative charge carriers
are attracted to the positively-biased copper plate. Note that the fraction of incoming
positive ions collected by the Faraday plate in Configuration 2 is slightly larger than in
Configuration 1 due to the positive potential applied to the copper plate in the latter case.

If Mechanism 3 is operative, then cooling the radiation shield should have a
dramatic effect on such measurements. This situation is created using Configuration
3.(Figure 4.13) The cryostat radiation shield can be cooled to a temperature of ~40 K,

and, via a heater on the window holder block, the copper plate can be maintained at

1 1 1

approximately 225 K. Under such conditions, with Ne and C02 present, CO; should
accumulate on the radiation shield, leading to the formation of a larger negative current
upon cation bombardment. When the radiation shield is cooled, and the copper plate held
at ground, the data shown in traces 4.13a and 4.13b are collected. As in Configuration 1,
incoming cations are detected. The negative deflections which are superimposed on the
cation signal in traces 4.13s and 4.13b are due to negatively-charged species striking the
copper plate and the Faraday plate. Several aspects of the data warrant comment. First,
we have evaluated many experiments that employ Configuration 3 and the negative
deflections detected at the copper plate, such as those clearly seen in the 150—210 second
time window of trace 4.13a, correlate with identical features in the current detected at the
Faraday plate, trace 4.13b. This suggests that ion bombardment of the cold surface results
in ejection of negative charges into the gas phase. They then impinge on a variety of
grounded surfaces, and thus show a specific time correlation. Second, it is unexpected
that the negative charge generation is not continuous. This remains the most surprising
aspect of this data set. Finally, the negative current deviations are relatively small
compared to the cation signal. If this is the case, then the process of ion bombardment of
cold surfaces on which CO; is condensed is not yielding an anionic current sufficient to
balance the charge of the incoming cationic current. However, in Configuration 3 the
copper plate is held at a potential of 0 V; during the experiment it is likely that the matrix
holds a net positive charge.

To optimize our ability to measure negative charges from the cation bombardment
of the cold radiation shield, Configuration 4 is used. With a positive potential on the

mpper plate, as in Configuration 2, incoming cations are repelled from this plate and a

1 12

field is created which will collect most of the negative charges that are formed. As
expected, the grounded Faraday plate shows a simple cation signal in trace 4.14b.
However, the temporal and signal intensity aspects of the data shown in 4.14a are
surprising. Negative charge generation under these conditions is very different fi'om that
detected in Configuration 2. Negative charge is not detected promptly upon cation
bombardment, and does not cease promptly when the incoming ion beam is blocked by the
gate valve. Also, negative charge currents larger than expected are collected. At present
this is not fiilly understood. The total cationic current striking the entire radiation shield is
at most 200 nA; the observations require that more than one negative charge carrier is
generated per incoming cation. This would suggest a surprisingly efficient utility of the
cation kinetic energy. Unfortunately, we are working in an energy regime in which few
relevant ion/surface interaction experiments have been performed. Considering that
anions impinge on the biased surface at energies of ~250 eV in this configuration, it is
possible that part of the total negative current measured in 4.14a includes anions losing
two electrons each at the surface and leaving as cationic species.ll That is, the current
sensed may not be a simple measure of the primary negative charges formed. The
difference between the data in traces 4.12a and 4.14a may also suggest that part of the
current detected in 4.14a is field assisted, due to the positive potential on the copper plate.

However, the important observation from Configuration 4 is that the number of
negative charges that can be produced per second, induced by cation bombardment, is
enhanced when the radiation shield is cooled. Since we know that a portion of the ion
beam strikes this copper surface, and that C02 condenses on the surface, we present this

as strong support for Mechanism 3. It should be noted that results similar to these are

113

obtained when the Cal window is present and the window holder is used to measure ionic
currents, and that the results are independent of the identity of the mass-selected cation. It
was also observed that the absorptions due to C02” are much less intense over the course
of a positive ion deposition when argon is used as the matrix gas in place of neon.
Presumably, argon competitively adsorbs onto the cold metal surfaces, reducing the
probability for negative ion formation from adsorbed C02. While the behavior of C02 is
the focal point here since it is added to the matrix gas, other molecules such as water can
likely play the same role when there is no matrix additive introduced to assist in counter
ion formation.

Explicitly, we suggest that both electrons and anions leave the cold radiation shield
upon energetic ion bombardment. Experimental results presented above in the discussion
of Figure 4.10, where a biased grid was placed on the window, suggested that electrons

are released, consistent with the report of Schlag et al.10 It was also observed that the

absorbance due to the (cog-(cor) anion is present whenever C02” is spectroscopically
observed. Despite widely varying NezC02 ratios in the matrix, the ratio of peak intensities
of the absorbances due to the two species is maintained in the 5:1 to 8:1 (monomerrdimer)

range. This suggests that cation bombardment of the cold, COz-covered metal surfaces

produces electrons, C02” and (C02)-(C02") in relatively constant ratios, presumably
because the concentration of CO; on the surface is more constant than either in the gas
phase or the matrix, for the variety of experiments we have performed.
E. Discussion
The literature on the structure and physical characteristics of C02” and (CO;){, as

well as information available fiom the surface science literature concerning the structure of

114
C02 on metal surfaces, yields an important context for the proposed mechanism. There
are several reports confirming the creation of negatively charged species upon

""2 including the formation

bombardment of metallic surfaces with positively charged ions,
of C02".13 The temperature of the metallic surfaces surrounding the matrix window must
play an important role in Mechanism 3. The radiation shield itself is maintained at
approximately 40 K, cold enough to freeze the relatively high concentration of C02 fiom
the nearby matrix gas sprayer, but not cold enough to fieeze the neon matrix gas. C02 is
therefore selectively collected onto the metal surface.‘ A portion of the diverging positive
ion beam hits the copper radiation shield continuously during our experiments. It seems
likely that the bombardment of these surfaces with positive ions may sputter off negatively
charged 00;" ions and electrons. For some C02" ions, the time of flight to the matrix
may exceed the lifetime of the ion (20—90 microseconds), "“5 and they may dissociate into
neutral CO; and electrons. However, electrostatic field simulations using SIMION show
that negative species formed near the edge of the hole in the radiation shield reach the
positively-charged matrix in times shorter than the reported lifetime.

The fact that 002” is observed in our experiment may be a key to the mechanism
of its formation. While the C02” anion was first detected in matrices by ESR in 1961
following gamma ray irradiation of an alkali halide matrix of sodium formate,“ the nature
of its stability has been the subject of considerable discussion. It has been suggested that
isolated 00;” should not exist, unless stabilized by complexation in a matrix with a
cation" such as Nai. Several reports have concluded that CO; adsorbed on a copper

surface is a bent and negatively charged entity." The electron affinity of C02 has been

measured to be approximately -0.6 eV. 1’ This does not indicate that the anion is

115

inherently unstable, but reflects the relative stability between the neutral, linear molecule
and the bent anion. '9 C02” ions can be generated in negative ion mass spectrometry fiom
gas phase organic molecules that contain a “bent CO2 unit”.“'20 Comparisons of CO2"
with the stable, isoelectronic species NO2' have been made as well.21 Thus, while the
electron affinity of the linear neutral molecule is negative, direct CO2” generation from
surface-adsorbed CO2, where the molecule is bent and can readily accept an electron, or
electron capture of CO2 molecules in very low temperature inert gas matrices are certainly
plausible processes to yield this anion in a neon matrix.

The CO2 dimer electron affinity has been estimated to be +0.8 eV by collisional
electron transfer between a crossed molecular beam of C02 clusters and a seeded
supersonic beam of alkali-metal atoms.22 Although ab initio calculations for the electron
afiinity for the dimer show that it may still be slightly negative (-0.31 eV),23 it is apparent
that solvation of a single CO2 molecule by other molecules, such as CO2, increases its
electron amnity. The metallic surfaces in our experiments may not only assist in forming
bent CO2, which has a positive electron afiinity, but their cold temperature may also help
to form clusters of the CO2 molecules, further facilitating the formation of anions.

In our laboratory, CO2" has been detected in matrix isolation experiments
involving a variety of mass-selected cations. These ions include both even electron cations
(CIH, CF33, C2H2+, C2H2O+) and radical cations (CO2'+ CS2”, CeFe”).

IV. Conclusions

A working model, which incorporates the mechanism suggested and other aspects

of the experiment that we now realize are important, is presented in Figure 4.15.

Successive points in time are shown, representing one extreme view of the system. At

116

v,,,=0v

 

 

    

 

.e‘ e. 'b.

tzx ms me=+IOOV

 

 

 

 

 

 

.e‘ t '6. 'et

Figure 4.15: Illustration of the proposed processes occurring during deposition of a
100 eV positive ion beam and neon on the CsI sample window and the
metallic radiation shield.

117

t = 0, incoming cations are collected in the growing inert gas matrix. If the ions have
kinetic energies of 100 eV, the matrix can sustain a net positive charge of +100 V. At this
point, t = 1: ms, subsequent incoming cations are repelled fiom the matrix and no firrther
accumulation can occur. The positively-charged matrix then attracts negative charges
emanating fi'om condensed molecules on cold surfaces. The matrix continues to attract
the negative charges until the net matrix potential is 0 V (t = x + y ms). Additional
incoming cations can then be deposited and the process can continue.

This is clearly an extreme model, which would lead to a layered matrix. More
realistically, some mean potential of the matrix must be established, which lies between 0
and +100 V. The matrix, then, always has a net positive potential, attracting negative
charges. However, its potential is never sufficiently high to repel incoming cations, since
anion generation and accumulation is also occurring at a high rate. The cations, thus,
serve a variety of purposes. Not only are they the analytes in these spectroscopic
experiments, but they create a positively charged matrix, and this is required for the
accumulation of negative charges that are formed on nearby surfaces.

In some experiments, the quantity of available cations may exceed the number of
available anions, and the mean matrix potential will be high. In other experiments, where
generation of negative charges occurs with high rates, the mean matrix potential may be
close to 0 V. The mean matrix potential must have an important influence on the results
of experiments involving matrix isolation of ions. How can a beam of 100 eV CF;+
cations lead to their spectral detection, when only 3-4 eV are required to break a bond?
We propose that the mean matrix potential serves to decelerate the incoming ions, creating

a field which slows them down so that the full initial kinetic energy is not relevant to the

118
deposition process. The deposition kinetic energy is determined by the difference between
the accelerating voltage in the ion source and the mean matrix potential,
“K. E (mm) = e(Vm.2- me). This may explain why some cations are more
difficult to detect than others in such experiments. An ionic species such as C5F5” which
we can generate and deposit, but have not yet detected using vibrational spectroscopy,
may create a situation in which the mean matrix potential is close to 0 V; thus the
incoming ions maintain high kinetic energies when they encounter the surface, resulting in
collision-induced dissociation and charge transfer processes. In contrast, the experiments
involving CF;+ may lead to sufficiently high mean matrix potentials that the incoming
cations undergo relatively “soft landings”. Clearly the mean matrix potential, while

dificult to directly measure, is an important experimental consideration in the success of

such experiments.

(1)
(2)

(3)
(4)
(5)

(6)
(7)

(8)

(9)
(10)

(11)

(12)
(13)

(14)
(15)
(16)
(17)

List of References

Jacox, M. E.; Thompson,W. E. J. Chem. Phys. 1990, 93, 7609.

(a) Szczepanski, J.; Wehlburg, C.; Vala, M. Chem. Phys. Lett, 1995, 232, 221.
(b) Hudgins, D. M.; Allamandola, L. J. J. Phys. Chem. 1995, 99, 8978.

Bondybey, V. E.; Miller, T. A. J. Chem. Phys. 1980, 73, 3053.
Prochaska, F. T.; Andrews, L. J. Am. Chem. Soc. 1978, 100, 2102.

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

Maier, J. P. Mass Spectrom. Rev. 1992, 11, 119.

Knight, L. B., Jr. in: Chemistry and Physics of Matrix-Isolated Species, Andrews,
L., Moskovits, M., Eds; Elsevier Science Publishing: Amsterdam, 1989.

Dahl, D. A.; Delmore, J. E. SIMION PC/PSZ Version 5.00, Idaho National
Engineering Laboratory, 1991.

Jacox, M. E.; Thompson, W. E. J. Chem. Phys. 1989, 91, 1410.

Fraunberg, W.-D. v.; Drechsler, G.; BaBmann, C.; Boesl, U.; Schlag, E. W. Int. J.
Mass Spectrom. Ion Processes 1994, 133, 211.

Bier, M. E.; Vmcenti, M.; Cooks, R. G. Rapid Commun. Mass Spectrom. 1987, 1,
92.

Middleton, R.; Klein, J.; Fink, D. Nucl. Instr. and Meth. 1989, B43, 231.

(a) Amot, F. L.; Beckett, C. Proc. Roy. Soc. (London), 1938, A168, 103. (b)
Schubert, S.; Imke, U.; Heiland, W.; Snowden, K. J .; Reijnen, P. H.;K1eyn, A. W.
Surface Sci. 1988, 205, L793.

Cooper, C. D.; Compton, R. N. Chem. Phys. Lett. 1972, 14, 29.

Compton, R. N.; Reinhardt, P. W.; Cooper, C. D. J. Chem. Phys. 1975, 63, 3821.
Ovenall, D. W.; Whifi‘en, D. H. Mal. Phys. 1961, 4, 135.

Koppe, R; Kasai, P. H. J. Phys. Chem. 1994, 98, 11331.

119

(13)

(19)

(20)

(21)
(22)

(23)

120

(a) Linder, H.; Rupprecht, D.; Hammer, L.; Muller, K.; J. Electron Spectrosc.
Relat. Phenom. 1987, 44, 141. (b) Copperwaite, R 6.; Davies, P. R; Morris, M.
A; Roberts, M. W.; Ryder, R. A Catal. Lett. 1988, 1, 11. (c) Carley, A. F.;
Roberts, M. W.; Strutt, A. J. Catal. Lett. 1994, 29, 169.

(a) Cadez, I.; Tronc, M.; Hall, R I. J. Phys. B: Atom. Molec. Phys. 1974, 7, L132.
(11) Krauss, M.; Neumann, D. Chem. Phys. Lett. 1972, 14, 26.

Ropp, G. A; Melton, C. E. J. Am. Chem. Soc. 1958, 80, 3509.

Carmichael, 1.; Bentley, J. J. Phys. Chem. 1985, 89, 2951.

(a) Bowen, K. H.; Liesegang, G. W.; Sanders, R A; Herschbach, D. R. J. Phys.
Chem. 1983, 87, 557. (b) Quitevis, E. L.; Herschbach, D. R. J. Phys. Chem. 1989,
93, 1136.

Fleischman, S. H.; Jordan, K. D. J. Phys Chem. 1987, 91, 1300.

Chapter 5
Visible Emission From Warming Neon Matrices
or

It Seemed Like a Good Idea at First

I. Background

In some FTIR experiments, we did not observe absorptions assignable to the mass-
selected cation being deposited, even though the total integrated ion current deposited was
comparable to those of experiments in which IR transitions were observed. There are
three general, likely explanations for these negative results:

0 the ions are being neutralized in the matrix

0 the mass-selected ions undergo fiagmentation upon
collision with the matrix

0 the ions are intact in the matrix, but a relatively low
molar absorptivity makes them unobservable by our
spectrometer.

The results of the experiments in which mass-selected, matrix-isolated cations
were observed by FTIR spectroscopy indicate that neutralization in the matrix does not
occur to an appreciable extent. It seems probable that the mass-selected cations would
fiagment as a result of a 130-eV collision with the matrix substrate. The actual collision
energy, however, may be much less since it depends on the potential of the matrix window
as well as the kinetic energy of the ion beam. This potential could presumably have any
value between 0 and 130 V. Unfortunately, this effective potential cannot be measured,

and it may be different for different chemical systems. Fragmentation of the incoming

121

122

cations could be troublesome for three reasons. First, the mass-selectivity would be
violated. Then, the fragments themselves may have a low molar absorptivity, making their
detection dificult. Moreover, even if it is observable, the IR spectra of the fragments may
not be well known, making their identification difficult. If the lack of observation of the
cation is due simply to inadequate instrumental sensitivity, then instrumental modifications
can be made to improve the overall sensitivity of the experiment. Unfortunately,
’difi’erentiation between fragmentation and insufficient sensitivity requires that the
sensitivity improvements be made.

Electronic spectroscopy (absorption and emission) has been used to identify mass-
selected, matrix-isolated cations.”2 Electronic absorption will not provide ground-state
structural information directly, and emission measurements may not be fully diagnostic.
Nonetheless, these techniques do have the advantage of being 10-1000 times more
sensitive than FTIR. If the emission spectra of the mass-selected cation, and its
neutralization product and its possible fi'agments are known, emission spectroscopy might
help to detemrine if the mass-selected cations are not observed because of fragmentation,
or simply because of inadequate sensitivity.

Unfortunately, the adaptation of a traditional luminescence spectrometer to the
UHV chamber was net feasible. The was due mainly to the inability to accommodate a
light source and detector without major modifications to the chamber. We were,
however, able to perform some preliminary emission experiments using what is
presumably charge-recombination energy from the matrix-isolated cations and anions as an
excitation source. While the emitting species observed in such experiments are not the

ions originally trapped in the matrix, it was hoped that by using systems we believe we

123
understand, we could establish a relationship between the ions in the matrix and the
resulting emission. If a relationship could be established, emission spectroscopy could be
used as a sensitive, albeit indirect method to identify species present in the matrix that are
not observable by FTIR spectroscopy.

The observation of emission from a warming rare-gas matrix has been suggested as
an informal test for the presence of ions in the matrix.3 The emission, however, could
emanate fiom a variety of sources. The emission could occur as the result of the
recombination of oppositely charged species. In the gas phase, the recombination creates
one or more highly excited neutral species, as demonstrated below:

C0,” + C0,” —-> CO,’ + CO,’
The neutral species may then relax to their ground states by the emission of light. In this
example, both products are shown in an excited electronic state, but in many cases only
one of the products may be formed in an excited electronic state. Rather than simple

electron transfer, the cation and anion may also undergo some chemistry, for example:

CO,” +C02" —> C0+O2 +CO
—2 C03 +CO

The energy deposited into the neutral species is the difference between the electron
amnity of the cation and of the anion. Since the cations we have studied to date have
electron affinities of 8.9 eV or greater, the charge recombination process can deposit a
significant amount energy into the products. Because the amount of energy involved is
more than enough to break a chemical bond, the recombination products often release
internal energy by breaking one or more bonds. This is fi'equently observed, ‘ for example,

in the recombination of a polyatomic cation with an electron, such as:

124

OCS+ + e’ -> C0' + 8‘
Again all, or some, of the neutral products may be formed in an excited electronic state.
Other possible sources of emission are excited neutral species formed from the
recombination of neutral radical species that may be present in the matrix, such as:

CH; + CH; —) C2H5°
Excited species may also be formed from radical-molecule or ion-molecule reactions
between the wide variety of species present in the matrix. The species available for such
reactions include the ubiquitous CO2 and H20, as well as the ion precursor and possibly
some of its neutral fragments.

Certain aspects our system merit elaboration. One concern is the identity of the
negative charge carrier. The matrix hosts used in these experiments were a mixture of Ne
and CO2. In these matrices, we believe that a large number of the negative charge carriers
are present are in the form of CO2“ or one of its aggregates.’ The electron affinity of CO2'
is negative, yet the anion has a relatively long lifetime.‘ It is not clear if this species will
remain intact once the matrix is no longer rigid. It is possible that above the annealing
temperature, the predominant negative charge carrier could be an electron rather than
CO2“. Electrons are considered to be relatively mobile in rare-gas matrices.’ If they are
the dominant negative charge carrier, then the conditions under which recombination
occurs are not clear. Ifelectrons are flee to migrate before the matrix surrounding them is
completely vaporized and larger species are still trapped, then the recombination will
occur under condensed-phase conditions. While fi'agmentation may be expected in gas
phase neutralization, it should be less likely under inert, condensed-phase conditions. If

the ions recombine in the matrix, the matrix cage effect will suppress fragmentation, and

125

the excess energy will be effectively dissipated by numerous collisions with the host atoms
and molecules. If the recombination occurs in the gas phase, then fragmentation is
expected to occur readily.
II. Experimental

The apparatus used to observe the emission fi'om warming matrices is described in
Chapter 2. High purity neon (AGA), 99.9995% premixed with C02 (Aldrich, 99.8+%) in
an ~1000:1 ratio was used as the matrix gas. The matrix deposition rate was ~0.5
mmol/h. Compounds used as precursors for the mass selected cations [Ar (AGA), CF3C1
(Matheson), OCS (Matheson), CS2 (Fisher) and CH3Br (Aldrich)] were all used as
received. The energy of the mass-selected cation beams was 130 eV. Deposition times
depended on the intensity of the emission fi'om the different ionic species and varied from
5 to 25 hours. The matrix was warmed by turning ofi’ the cryostat after stopping the
deposition. The time-dependence of the window temperature is shown in Figure 5.1. The
apparatus used to collect the emission from warming matrices is described in detail in
Chapter 2, Section VII. All optical measurements were made with the lab as darkened as
possible. During the collection of the total emission, the PMT was generally operated at a
potential of 800—1000 V. Data were collected at 10 Hz for 300 seconds. In the dispersed
emission experiments, the monochromator was used with slit widths fi'om 1.5 to 3.0 mm,
depending on the emission intensity of the particular system. The 300 groove/mm grating
used in this study had a relative linear dispersion (RLD) of 14.8 nm/mm. This resulted in a
slit-width limited spectral resolution3 of 22 and 44 nm (FWHM), respectively. Spectra

were collected by integrating over 5 3 intervals, for a total of 300 s.

126

 

.s .s
«F O)
1 l—

.a
N
I

 

Temperature (Kelvins)
8

 

. - - l

- I .

 

Figure 5.1:

0 50 1 00 150 200 250 300

Time (seconds)

The temperature of the cryostat window holder as a function of time as
the cryostat is turned off. The cryostat was turned off approximately
10 s after data collection was started. This temperature profile was
obtained concurrently with the emission data shown in Figure 5.3

127
III Results and Discussion
A. Total Emission Studies

The first attempt to observe emission from a warming ion-containing matrix was
canied out by simply visualizing the matrix as the cryostat was turned 03’. After a 24 h
deposition of CHf, generated fi'om CH3BI', the matrix exhibited a blue-white emission as
it warmed. This emission was plainly visible 1 m away from the outside of the UHV
chamber, and lasted for several minutes.

Figure 5.2 shows the undispersed emission fi'om a 100021 Ne:CO2 matrix
containing CF; generated fiom CEO. The emission fiom a 100021 Ne:CO2 matrix
containing CS2" generated from CS2 is shown Figure 5.3. This matrix had been subjected
to several short annealing cycles before being warmed beyond 14 K.

These annealings are believed to be responsible for the qualitative differences
between the two spectra. While most of the matrix gas and mass-selected ions deposited
are frozen onto the cryostat window, some may end up on metallic window holder. Due
to its higher thermal conductivity, the window holder heats faster than the CsI window.
The first sharp signal in Figure 5.2 is attributed to the ions isolated on the window holder,
while the second broad signal is due to the ions on the CsI window. In the CS2°+
experiment, the ions isolated on the surface of the window holder were most likely lost
during the repeated annealing cycles. The difference between the times of the two
emission maxima is most likely a result of timing inconsistencies in the experiment and the

CS2" deposition being longer, resulting in a larger matrix that takes longer to warm.

128

 

 

 

A — emission
:2 — blank
C
D
2‘
(U
h
:1:
.Q
h
<
v
>
:2:
U)
C
a)
4—1
5
I - l - I - l - l - l - I

 

 

 

0 50 100 150 200 250 300
Time (seconds)

Figure 5.2: Total emission observed during warming of a 1000:1 Ne:CO2 matrix
after the deposition of CF,’ generated from CF,C1.

129

 

 

 

 

 

32‘

"E . .
3 — emrssron
a. — blank
3

E

S,

3‘

'5

c

.9

E

0 50 100 150 200 250 300

Time (seconds)

Figure 5.3: Total emission observed during warming of 2 100021 Ne:CO2 matrix
after the deposition of CS,“ generated from CS,. This matrix was
subject to several short annealing cycles before being warmed
irreversibly.

130
B. Dispersed Emission Studies

While the total emission experiments helped quantify and provide some temporal
information about the emission fiom the warming ion-containing matrices, they could
provide no insight as to the identity of the emissive species. It was hoped that the species
responsible for the luminescence could be identified by dispersing the emission. Owing to
experimental limitations, however, interpretations of the results of these measurements are
highly speculative. First, the low resolution of the spectra make vibrational or rotational
analysis impossible. The limited spectral range of the photodiode array (3 50—800 nm) also
limits the amount of information that can be obtained. The question of the conditions
under which recombination occurs further complicates assignment of the emission.

1. Ar”

The warming of a 1000:1 Ne1C02 matrix after depositing Ar” generated from Ar
resulted in the emission shown in Figure 5.4. The temporal profile of the emission is
shown in Figure 5.5. The intensity of the spectra varied with time, but the same bands,
centered roughly at 589, 550, 532, and 494 nm were observed, indicating that the same
species is responsible for the emission throughout the warming period. I

Ar'+ was chosen to simplify analysis of the experiment, since at least the
complication of fragmentation is obviously not possible in this case. Using Ar”, however,
does add some complications. Its relatively high electron afiinity (15.75 eV) 9 will allow
charge exchange chemistry with CO2 and H20 which is not possible with the other ions
studied. This chemistry could lead to the formation of CO2+ and H20. The emission
should presumably emanate from an excited Ar atom, CO2 or H2O. The emission

observed can be assigned to transitions originating fiom the 3p’8p (J= 1) and 3p55d (J=2)

131

 

 

Intensity (Arbitrary Units)

 

 

 

350 400 450 500 550 600 650
Wavelength (nm)

Figure 5.4: Dispersed emission observed during warming of a 1000:] Ne:CO2
matrix after the deposition of Ar” generated from Ar. This single
spectrum is representative of the emission observed throughout the
ntire warming.

132

 

- high intensity
600 -
-
-
- low intensity

550 -

500 -

Wavelength (nm)

450 -

400 -

 

50 100 1 50 200 250
Time (seconds)

 

 

 

Figure 5.5: Contour plot showing temporal profile of the dispersed emission
observed during the warming of a 100021 NezCO2 matrix after the
deposition of Ar" generated from Ar.

133
electronic configurations of Ar.’ Although the spectral bands are quite broad for atomic
transitions, their width (~20 nm FWHM) is what is expected from the 1.5 mm slit width
used in this experiment. Both of the upper electronic configurations have the same core
configuration (3p’) as Ar'+ and Al for each transition is :H, following the general rules for
atomic transitions. 1° These assignments are demonstrated graphically in Figure 5.6.

2. CF;

When a 1000:l Ne2CO2 matrix containing CF; generated from CF3C1 was
warmed, emission fi'om approximately 400 to 700 nm was observed. Figures 5.7 and 5.8
show the temporal evolution of the emission over two partially overlapping wavelength
ranges. Figure 5.9 shows individual spectra from each wavelength range in greater detail.
The temporal profile of this emission is similar to the total emission results shown in
Figure 5.2. The spectra collected vary only in intensity, indicating that the same species is
responsible forthe emission throughout the warming process.

The assignment of this emission (and of all the polyatomic ions) unfortunately is
not as straightforward as that for the visible luminescence observed in the Ar'+ experiment.
The emission observed during the recombination of CF 3” is consistent with previous
assignments for both CF; 11.12.13.14 and CF2.ls Washida et al."'m3’“ observed emission in
the 400—700-nm range during vacuum UV photolysis of several trifluoromethyl
compounds. A sample of the emission collected during these studies is reproduced in
Figure 5.10. It should be noted the authors reported that the position and width of this
band were dependent of several experimental conditions such as the photolysis
wavelength, pressure, and the presence of a buffer gas. Based on experimental

resultsu'u’l3 and theoretical calculations," they assigned this emission as originating from

Energy (cm")

134

 

 

 

 

 

 

 

 

 

 

 

 

 

127 000+ Ar ionization limit (Ar" 3p‘)
124 311 Sunset). J=1
122 086 3p’5d. J=2
% 532 nm
> 494 nm
> 589nm
> 550 Wt
107 054 3p54p. J=0
105 660 3135413, J=2
105 500 Jr- 3p54p, J=3
104 102 ._J 3p54p. i=1

 

0 '— Ground State of Ar 3p“

Figure 5.6: Tentative assignments of Argon emission.

600

550

500

Wavelength (nm)

400

Figure 5.7:

135

 

 

 

.-
.. '

igh intensity -

low intensity

 

 

l I I L l

 

100 150 200 250
Time (Seconds)

Contour plot showing temporal profile of the dispersed emission in
the 350—650 nm range observed during the warming of a 1000:1
Ne:CO2 matrix after the deposition of CF,’ generated from CF,C1.

136

 

700 -

"650—

600

Wavelength (nm

 

- high intensity

' :5
.5; g -
- low intensity

 

 

1 1 l . I .

 

Figure 5.8:

50 100 150 200 250
Time (Seconds)

Contour plot showing temporal profile of the dispersed emission in
the 450—750 nm range observed during the warming of a 1000:1
Ne:CO2 matrix after the deposition of CF,’ generated from CF,C1.

N
O
O
O

1500

1000

500

Intensity (Arbitrary Units)

0

Figure 5.9:

137

 

 

 

 

400 450 500 550 600 650 700
Wavelength (nm)

Dispersed emission observed from a warming 1000:l Ne2CO2 matrix
after deposition of CF; generated from CF,C1. Spectra shown are
representative of the emission observed during the entire 5 minute
warmup. The blue trace was taken from the data set shown in Figure
5.7 and scaled 2x. The red trace was taken from the data set shown in
Figure 5.8 and not scaled

138

v

 

400 500 530 fi WP v 890
(o) CF30 ,Ar
(b) CF38! , H
(C) CF3C| , H
(d) CF38! , O
Q) CF30 , Kf (COle

 
 
 
 

 

(0)

J11.“
$0.1th

AL

400 500 W 700
Wave length (nrn )

Intensity

 

 

 

Figure 5.10: Visible CF,’ emission reported by Washida et. al. Spectrum repoduced
from reference 13. Bottom trace (e) is the discharge lamp reference.

139

one of two CF 3' Rydberg states approximately 8 eV above the ground state (E ' and A!)
and ending in another Rydberg state (A,’) approximately 6 eV above the ground state.

When CF2,+ is neutralized to form CF3°, the radical is formed with as much as 8.9 eV of
excess energy. Given the typical C-F bond strength of 5.5 eV,“ it could be possible to

form a CF2 fragment with as much as 3.4 eV (27 424 cm") of excess energy. This is

sufficient to access a low lying triplet state (E’B,, 19 824 cm”) of CF2.ls The

phosphorescence from this state to the ground state (1714,) has also been reported15 in the

450—700-nm range. The emission spectrum observed by Koda is reproduced in Figure
5.11. Unfortunately, a definitive assignment of the matrix emission cannot be made with
the data currently available. CE" and CF2. could be differentiated by their UV emission
spectra, where these two compounds have distinctly different emission bands.

3. C82”

The warming of a 100021 Ne:CO2 matrix after deposition of CS2°+ generated from
CS2 resulted in a relatively complex emission band from 400—700 nm. A typical temporal
profile of the emission is shown in Figure 5.12, and an individual spectrum is shown in
greater detail in Figure 5.13. This emission, however, was not very reproducible. The
energies of the admittedly ill-defined features were somewhat consistent, but their relative
intensities were not. FTIR experiments with CS2°+ generated fi'om CS2 indicated that after
several hours of deposition the cold copper surfaces seemed to become “fouled” and
exhibit severely decreased anion production. This could possibly be the result of a 0,8,

polymer forming on the copper surface. These emission experiments may be even more

140

 

 

 

l

 

 

 

 

sdofi'f 550""600"' 650
wave length (nm)

Figure 5.11: Visible CF, emission reported by Koda. Spectrum repoduced
from reference 15.

12% :_,_-—_—,_.;9 1.;; fl
- high intensity
A v
E -
5 u
.C - low intensity
4.:
C)
C
9
OJ
>
g
I l I 1 I l

141

 

 

 

50 100 150 200 250
Time (Seconds)

 

Figure 5.12 Contour plot showing temporal profile of the dispersed emission
observed during the warming of 2 100021 Ne:CO2 matrix after the
deposition of CS,” generated from CS,.

142

 

2% :5

Intensity (Arbitrary Units)
N
§

 

 

450 500 550 600
Wavelength (nm)

500 ‘

 

400

Figure 5.13: Dispersed emission observed during warming of a 1000:] Ne:CO2
matrix after the deposition of CS,” generated from C8,. This single

spectrum is representative of the emission observed throughout the
entire warming.

143
sensitive to the experimental parameters, and perhaps they should be repeated under more
reproducible conditions.

The C82” emission experiments can be interpreted in similar fashion to the Cng
results. When CSz’+ is neutralized, the C82 neutral is formed with as much as 10.1 eV of
excess energy. ‘7 This is more than sufficient to break one of the GS bonds, which only
requires 4.5 eV." The resulting CS fragment could then be formed with as much as 5.6
eV (45 170 cm") of excess energy. The emission observed correlates roughly with that
previously assigned to transitions within a triplet manifold of CS,” reproduced in Figure
5. 14. The lowest of these states lies approximately 27 700 cm'1 above the ground state of
CS. The emission observed would therefore have to originate from energy levels from

43 100 to 50 400 cm’1 above the ground state. While these could be assigned to

vibrationally excited CS in either the (7 3A or the E3 2‘ state, '9 population of the upper
levels would require more energy than can be attained from neutralization alone. The
extra energy needed, however, is not very large (5200 cm", 0.64 eV) and could possibly
be supplied by the energy of the collision between the ions if they experience some
acceleration towards each other or from some electric field before recombining.
Another possibility is that the emitting species is 82, formed from the reaction:
C8;+ + negative charge carrier' —) C + S2
This reaction is exothermic by approximately 9.9 eV.20 This, however, is more than twice

the bond dissociation energy of 82, 4.4 eV.21 If the internal energy of the 8; molecule is

quenched rapidly enough by the matrix, emission from it could be observed. The lowest

21

energy symmetry-allowed transition of S; is the 3322; +2732; transition. The energy

PM

 

‘ur- _

Ari’Fi) . SCCI,“ ?

'._._...... ‘_:.'.

 

 

 

 

Relative Intensity
I

F

 

t LJl 14 JLJ_LJ ll
114glittsbtituggu 60 65 7O

Wavelength (A x 10‘)

 

 

Figure 5.14: Visible CS emission reported by Coxon and co-workers. Spectrum
reproduced from reference 18.

145

difference between these two states, however, is 31 835 cm",21 which is too large for the
emission observed.

This observed emission could also be assigned to transitions of C82. The energy of
the observed emission corresponds reasonably with the energy difference between the

3‘82 (45 950 cm") and 53A, (26187 cm") levels." If these Rydberg states are

responsible for the emission, then the relaxation of the neutral, excited C82. by collisions
with the matrix would need to be very efficient to prevent fragmentation.

4. OCS°+

The warming of a 1000:] Ne:CO; matrix after deposition of OCS'+ generated from
OCS resulted in very intense emission in the 400—600—nm range. The emission intensity
varies with time as the matrix is warmed as shown in Figure 5.15. The same spectmm is
observed throughout the warming, varying only in intensity. A single spectrum, showing
the spectral distribution in greater detail, is reproduced in Figure 5.16.

The emission from the OCS'+ experiments is equally ambiguous. The broad feature
centered at ~420 nm could be. assigned to the well known 3'2 41'1"! transition of CO.
The spectrum observed from a gas phase CO sample is reproduced in Figure 5.17.22 This
CO emission has been previously observed in OCS passed through a RF discharge.4
Bezuk et al. found that the CO fragments were not formed directly in the F 2 state, but
were excited into that state by collisions with electrons in the plasma. In our experiment,
if the only energy available is the electron affinity of OCS", then the CO fragment can be
made with at most 7.5 eV (60 200 cm”), after 3.7 eV (29 844 cm") is used to break the
C-S bond. This is not sufficient energy to access even the 2' H state. The remaining

energy (26 750 cm", 3.3 eV), could, however, be supplied by the some other source, such

146

 

   

GOOL _ - high intensity
‘ S =
F -
- low intensity
A 550—

 

Wavelength (nm

 

l 1 l | t

l
o 50 100 150 200 250
Time (Seconds)

Figure 5.15 Contour plot showing the temporal profile of the dispersed emission
observed during the warming of a 1000:] Ne:CO2 matrix after the
deposition of OCS" generated from OCS.

147

 

% %

Intensity (Arbitrary Units)

 

 

J L 4 l

 

Figure 5.16:

400 A 450 ‘ 500 550 600
Wavelength (nm)

Dispersed emission observed during warming of a 1000:] Ne:CO2
matrix after the deposition of OCS” generated from OCS. This single
spectrum is representative of the emission observed throughout the
entire warming.

148

Comet-Tail Bands (co+)

 

 

 

 

 

High-Pressure Bands (0,) fielder-Johnson Bands (00+)
fi— _I
-. , a rj'lf'lfi __’_“_~__ .. :FL. ....J ”r
(a) ».-?;"' $1,” .. "7 :“ ‘ - 1 4-.
.4 y r". :3. 5...: ll-.. \ 'i‘
J 1 H Bonds
23 «an, 3"!
o ‘53 3’,
M c-s w
v v n

 

Figure 5.17: Visible CO emission shown by Herzberg. Spectrum reproduced
from reference 22.

149
as collision with the negatively charged species (such as an electron) as in the study of
Bezuk et al.4

If it is assumed that the excited OCS. molecule is quenched rapidly enough to
prevent dissociation , then this emission could also be attributed to a transition from the
5 state to the Z state.23 The energy of this transition (v00) is approximately 24 600 cm",
in reasonable agreement with the observed emission.

5. CH;“

The warming of a 1000:] Ne:COz matrix after deposition of CH;+ generated fi'om
CH3BI' resulted in the observation of two bands, centered at approximately 575 and 637
nm. The temporal profile is shown in Figure 5.18. The results from this ion system were
unique in that the spectral distribution of the emission did change with time. Figure 5. 19
shows an emission at early time, when the 63 7-nm feature is prominent. At later times, as
shown in Figure 5.20, the 575-nm feature is dominant.

The emission from the CH3+ depositions was by far the most puzzling because the
recorded spectra did not exhibit the simple time variance observed for the previous four
ions. Spectra recorded at different times are distinctly different. The temporal evolution
of the 63 7-nm feature is similar to that observed for the other experiments. The 575-nm
feature, however, reaches maximum iintensity at a later time. If fragmentation of the CH
(4.9 eV, 39 523 cm") bond occurs, the resulting CH2 fragment can be formed with as

much as 40 653 cm“1 of excess energy. This would allow the formation of methylene in
the E" A, state. Both bands observed could be assigned to the E'A, +513, transition of

CH2,” with varying degrees of vibrational excitation. The bands also coincide with

150

 

' _ 1 -high intensity
700 - -
' -
A . — - low intensity
E
E5, 650 -
E.
m I
0:)
F) 600 -
>
“5 r
E
550 -

 

 

 

500=.|.|.I.I.1.
50 100 150 200 250

Time (Seconds)
Figure 5.18 Contour plot showing temporal profile of the dispersed emission

observed during the warming of a 1000:1 Ne:CO2 matrix after the
deposition of CH,’ generated from CH3Br.

151

 

i

as
§

5;
5

3:
8

.2:
§

 

.5

 

Intensity (Arbitrary Units)

§

§

 

 

 

500 550 600 650 700 750
Wavelength (nm)

Figure 5.19: Dispersed emission observed during warming of a 1000:] Ne:CO2
matrix after the deposition of CH; generated from CH,Br. This
spectrum was taken approximately 60 s after the cryostat was turned
off, and is representative of the early emission when the 63 7-nm band is
dominant.

235

1400

1200

Intensity (Arbitrary Units)

§§

a:
8

Figure 5.20:

152

 

 

I - - l J

- l 4

 

 

500 550 600 650 700 750
Wavelength (nm)

Dispersed emission observed during warming of a 1000:] Ne:CO,
matrix after the deposition of CH,+ generated from CH,Br. This
spectrum was taken approximately 100 s after the cryostat was turned

off, and is representative of the late emission when the 575-nm band is
dominant.

153

assignments made to various ro-vibrational transitions within the 3 IBl —>5 'A, manifold,”
although the extensive band structure observed by Garcia-Moreno and Moore, a small
portion of which is reproduced in Figure 5.21,” was not observed here. The different
temporal profiles of the two bands suggest that they not be assigned to the same electronic
manifold. If fi'agmentation does not occur, both transitions, assuming some vibrational
excitation, could be assigned to the 52.4; or C’E" 632/1, transition of CH3.“

An alternate explanation is that the 637-nm emission is coming from CH, which
could be produced by an H2 elimination such as:

CH; + negative charge carrier’ -) CH + H2

This reaction is exothermic by ~5.2 eV.20 While this is more greater than the dissociation
energy of CH (3 .47 eV)”, the molecule could be kept intact by the matrix cage. The
emission observed could be assigned to the Ezz‘orZZA ->X”I'I transition of CH?” if
the molecule is vibrationally excited. Perhaps the early emission is from CH3, or CH
recombining in the matrix while the emission at later time is due to gas-phase

recombination which results in the fiagmentation of CH3" to CH2°.

C. Discussion
Although definitive assignments of the emission observed fi'om warming ion-
containing matrices were not possible, a few general conclusions, can be made. First,
because the emission observed in each of the five sets of experiments was different, it is
clearly dependent in some way on the mass-selected cation which is deposited into the

matrix. The relationship of the observed emission to the identity of the cation is not

154

 

 

 

 

 

 

, 936
.3 734
5
.33
3 318
35 836
O
x 818 l 9‘ 3
fr a i A A - 4 W A A; L
589 592 595 $98 601 604
Fluorescence Wavelength (nm)
Figure 5.21: Portion of the visible CH, emission reported by Garcia-Moreno and

Moore. Spectrum repoduced from reference 25.

155

certain. Second, if the observed emission is indeed due to the neutralization of the matrix-
isolated, mass-selected cations, the most probable assignments take place while the matrix
is still somewhat rigid and can suppress fi'agmentation. Lastly, in all but one case (CHf),
the spectral distribution of the emission did not exhibit significant temporal change; that is,
the same spectrum was always observed, only the intensity changed with time. This
suggests that in most cases the same species or set of species is responsible for emission
throughout the warming process.
IV. Conclusions

The results of these luminescence experiments from warming matrices that contain
trapped ions show that this technique will not be generally usefiil as an indirect method for
determining the composition of ion-containing matrices. Some of the problems are due to
experimental limitations. The total emission experiments performed with the PMT
detector indicate that a relatively large amount of light was being produced. In some
cases, however, relatively low light levels were observed when the PDA/monochromator
system was used. Much of the emission from those systems may be occurring in the UV.
Since. the spectral range of the PDA detector is limited to the visible wavelengths, this
potential information is being lost. Use of a PMT as a UV detector is not feasible, owing
to the transient nature of the emission. Due to the low light levels, relatively large slit
widths must be used in the dispersed emission experiments, which severely degrades the
resolution of the spectra that are obtained.

Even if improvements were made in the instrumentation, such experiments would
be of only limited value as a more sensitive detection method, as proposed in the

introduction of this chapter. The crucial limitation of the methodology is that the ions in

156

the matrix, which we seek to characterize, are not the species that produce the emission.
Another fundamental question that remains pertains to the conditions under which
recombination occurs. The total emission experiments performed during annealing
demonstrate that emission can be observed without significant loss of the matrix or its
contents as seen by FTIR29 This result indicates that if recombination is in fact
responsible for the emission, it occurs in the matrix presumably as diffiision of the trapped
species is permitted, rather than in the vapor phase. This question might be answered by
repeating the experiment for the dispersed emission with the C02°+ ion. If the
recombination is a gas phase process, a CO fragment is expected. In that case, the
emission should be similar to that observed for OCS“.

Although the recombination/emission experiment could be improved, clearly the
most effective use of electronic spectroscopy (absorption or luminescence), would involve
observations on the ionic species while they are still trapped in the matrix. This would,
however, require significant modifications to the instrument. These modifications are

outlined in the following chapter.

(1)
(2)
(3)
(4)
(5)

(6)

(7)
(3)

(9)

(10)
(11)
(12)
(13)

(14)

(15)
(16)
(17)

List of References

Fomey, D.; Jakobi, M.; Maier J. P. J. Chem. Phys. 1989, 90, 600.
Sabo, M. 8.; Allison, J.; Gilbert, J. R.; Leroi, G. E. Appl. Spectrosc. 1991, 45, 535.
Knight, L. b. J. Personal communication, 1992.

Bezuk, S. J.; Miller, L. L.; Platzner, I. J. Phys. Chem. 1983, 87, 131.

Godbout, J. T.; Halasinski, T. M.; Leroi, G. E.; Allison, J. J. Phys. Chem. 1996,
100, 2892.

(a) Cooper, C. D.; Compton, R N.; Chem. Phys. Lett. 1972, 14, 29.

(b) Compton, R N.; Reinhardt, P. W.; Cooper, C. D. J. Chem. Phys. 1975, 63,
3821.

Kasai, P. H. Acc. Chem. Res. 1971, 4, 329.

Ingle, J. D. Jr.; Crouch, S R. Spectrochemical Analysis; Prentice Hall: Englewood
Cliffs, New Jersey; p 71.

Moore, C. E. Atomic Energy Levels Volume I; NBS circular 469, 1949, pp
2] 1-226. -

Kuhn, H. G. Atomic Spectra; Academic: New York, NY, 1969; Chapter 1.
Suto, M.; Washida, N. J. Chem. Phys. 1983, 78, 1007.
Suto, M.; Washida, N. J. Chem. Phys. 1983, 78, 1012.

Suto, M.; Washida, N.; Akimoto, H.; Nakamura, M. J. Chem. Phys. 1983, 78,
1019. -

Washida, N.; Suto, M.; Shigero, N.; Nagashima, U.; Morokuma, K. J. Chem.
Phys. 1983, 78, 1025.

Kodo, S. Chem. Phys. Lett. 1978, 55, 353.
Hildenbrand, D. L. Chem. Phys. Lett. 1975, 32, 523.

Herzberg, G. Molecular Spectra and Moleculw Structure Volume III - Electronic
Spectra and Electronic Structure of Polyatomic Molecules; Krieger: Malabar, F1,
1991 ; pp 600—60 1 .

157

(13)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(23)

(29)

158

Taylor, G. W.; Setser, D. W.; Coxon, J. A; J. Mol. Spectrosc. 1972, 44, 108.
Field, R W.; Bergeman, T. H. J. Chem. Phys. 1970, 54, 2936.

Thermochemical data taken fiom: Lias, S. G.; Bartmess, J. E.; Liebmann, J. F.;
Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref Data. 1988, 17,
Suppl. 1.

Herzberg, G. Molecular Spectra and Molecular Structure Volume I - Spectra of
Diatomic Molecules; D. Van Nostrand: Princeton, NJ, 1950; p 566.

Herzberg, G. Molecular Spectra and Molecular Structure Volume I - Spectra of
Diatomic Molecules; D. Van Nostrand: Princeton, NJ, 1950; p 3 5.

Herzberg, G. Molecular Spectra and Molecular Structure Volume III - Electronic
Spectra and Electronic Structure of Polyatomic Molecules; Krieger: Malabar, F1,
1991; pp 599-601.

Herzberg, G. Molecular Spectra and Molecular Structure Volume III - Electronic
Spectra and Electronic Structure of Polyatomic Molecules; Krieger: Malabar, F1,
1991; pp 583-584. ‘

Garcia-Moreno, 1.; Moore, C. B. J. Chem. Phys. 1993, 99, 6429.

Herzberg, G. Molecular Spectra and Molecular Structure Volume III - Electronic
Spectra and Electronic Structure of Polyatomic Molecules; Krieger: Malabar, F1,
199]; p 609.

Herzberg, G. Molecular Spectra and Molecular Structure Volume I -Spectra of
Diatomic Molecules; D. Van Nostrand: Princeton, NJ, 1950; p 518.

Castillejo, M.; Martin, M.; de Nalda, R; Oujja, M. Chem. Phys. Lett. 1995, 23 7,
368.

Halasinski, T. M.; Godbout, J. T.; Allison, J.; Leroi, G. E. J. Phys. Chem. 1996,
accepted.

Chapter 6

Conclusions and Suggestions for Future Work

1. Conclusions

The ultimate goal of the research described in this dissertation was to obtain the
vibrational spectra of mass-selected, matrix-isolated cations. This goal was reached: the
FTIR spectra of several mass-selected, matrix-isolated cations such as CF3“,l C02”,2 and
C82" 3 generated in an electron-impact ionization source have been observed. Another
goal of this project was also to gain an understanding of the mechanism by which charge
neutrality is maintained during the deposition of cations. We have addressed this issue and
discovered that negatively charged particles are produced from the bombardment of cold
metallic surfaces, and the molecules adsorbed on them. The addition of CO; to the system
was found to result in the production of large quantities of C02“ and (CO:);".’2

While this research was successfirl, not all issues have been resolved, and several
questions remain. One interesting finding. of this work was the observation of
fi'agmentation products fi'om some cations, such as C.;F¢;°+ and OCS'+ but not from others,
such as CF 3+ and C82”. The fragmentation was first thought to result from collisions of
the relatively high (130 eV) kinetic energy ion beam with the growing matrix. These
fragments may, however, be formed fi'om the neutral molecules fiozen onto parts of the
cryostat assembly. Bombardment of the molecules on these surfaces by the mass-selected
cation beam can fragment and vaporize the adsorbed molecules.

Cations with low molar absorptivities still provide a challenge. The cations

observed to date have had relatively high molar absorptivities. Ions with medium or low

159

160
absorbances still cannot generally be observed in a reasonable amount of time (1 filament
lifetime, 25—50 hours), owing to a combination of low beam intensity and instrumental
sensitivity.
11. Instrumental Modifications
A. FI'IR System Improvements

1. Gain-Ranging Amplifier System

The implementation of a gain-ranging amplifier in the FTIR detection system will
reduce the background noise of the system by at least an order of magnitude, facilitating
the observation of very weak absorptions. While this type of signal processing was used
in the Bomem DA3 spectrometer, it is not present in the Nicolet 520P spectrometer,
which replaced the Bomem. The hardware necessary to implement gain-ranging
amplification was purchased with funds from a Chemistry Department Equipment
Committee grant. The electronic circuitry necessary to measure the FTIR detector output
and subject it to variable-gain amplification has been designed and installed. The circuit
diagrams of the components are shown in Appendices X-X. The hardware is firlly
functional, but some LabVIEW programming is required to make the system operational.
This programming involves data storage and manipulation. The raw data collected across
the separate gain channels must stored as they are collected. Once the data collection is
over the information fi'om the separate gain channels must be processed and used to
“splice” one final interferogram together. The interferogram will then be transformed to a
absorption spectrum. This processing can be accomplished with the built-in data handling
and math firnctions of LabVIEW. If more extensive data manipulation is required, C

subroutines can be implemented into the LabVIEW VI.

161

B. Electronic Spectroscopy of Matrix-Isolated Species

The ability to observe the electronic as well as the FTIR spectra of the species
trapped in the rare gas matrix would be very advantageous. Although electronic
spectroscopy does not yield structural information as directly as vibrational spectroscopy,
its higher sensitivity makes the technique usefiil in experiments when simple detection of
ions is the goal, rather than structural analysis. The higher sensitivity of electronic
spectroscopy (both absorption and emission) will facilitate the detection of species present
in amounts to small too observe by FT IR. Since electronic spectroscopy is more sensitive,
the electronic transitions of ions have been studied and documented much more
extensively than their vibrational spectra. Simultaneously performing both experiments
would provide the advantage of the ability to correlate electronic and vibrational
transitions, when the latter are observed. This technique has been used very successfirlly
by the Vala group to assign vibrational transitions of matrix-isolated ions.‘

The implementation of an electronic absorption/emission spectrometer that can
collect data concurrently with the FTIR system will require substantial modification to the
existing UHV and optical systems. One way to accommodate both experiments is to
modify the chamber to have two separate beam paths, as shown in Figure 6.1. The dual
path system has several advantages. The primary advantage of this configuration is that
different windows can be used for each beam path, so there will be no compromise in
accommodating the IR and UV spectra] ranges with a single window material.
Fluorescence as well absorption experiments could be performed if desired. If absorption
experiments are performed (IR as well as UV/VIS), they will have the benefit of an

efi‘ective doubling of the sample pathlength.

162

F—‘W

FTIR
Spectrometer

 

 

 

 

  
  

 

 

U
UVNis .9; 1')
Source 3. 5
2
Rotatable |
Matrix Window .
r
UV/Vls
Detector

 

Figure 6.1: Schematic diagram of proposed modifications to accommodate
both FTIR and UV/Vis absorption and luminescence
measurements of matrix isolated species.

163

There are, however, some drawbacks to this system. Since the cryostat will need
to be rotated into different positions for deposition, and FTIR and UVN is spectroscopy,
the Conflato flanges used in the cryostat’s current vacuum housing will need to be
replaced with a dual concentric o-ring system. It may also be a challenge to find a mirror
material with sufficient reflectivity from the mid-IR to UV. If the proposed geometry is
used, the Faraday plate's linear motion feedthrough (LMF) and electrical feedthrough will
need to be moved to accommodate the new optical port. An alternative to this is to use

fiber optics (if UHF-compatible are available) to simplify the WMS optical path.

(1)

(2)

(3)

(4)

List of References

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

Godbout, J. T.; Halasinski, T. M.; Leroi, G. E.; Allison, J. J. Phys. Chem. 1996,
100, 2892.

Halasinski, T. M.; Godbout, J. T.; Allison, J.; Leroi, G. E. J. Phys. Chem.
submitted.

(a) Szczepanski, J.; Roser, D.; Personette, W.; Eyring, M.; Pellow, R.; Vala, M. J.
Phys. Chem. 1992, 96, 7876. (b) Szczepanski, J.; Vala, M; Talbi, D.; Parisel, O.;
Ellinger, Y. J. Chem. Phys. 1993,98, 4494. (c) Vala, M; Szczepanski, J .; Pauzat,
F.; Parisel, O.; Talbi, D.; Ellinger, Y. J. Phys. Chem. 1994, 98, 9187. (d)
Szczepanski, J.; Wehlburg, C.; Vala, M. Chem. Phys. Lett. 1995, 232, 221. (e)
Szczepanski, J.; Drawdy, J.; Wehlburg, C.; Vala, M. Chem. Phys. Lett. 1995,
245,539.

164

 

"illllllllllllllll“