Fourier transform infrared studies of guest-host interactions in ice

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
FOURIER TRANSFORM INFRARED STUDIES OF GUEST-HOST
INTERACTIONS IN ICE

by
George Kumi
____________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

August 2007

Copyright 2007 George Kumi
ii
Dedication

To the Gardiner and Kumi clans
iii
Acknowledgements

The doctoral degree is inherently an individual honor, and yet it is oftentimes a
culmination of efforts and commitments by many to one person. It has taken a
village to guide, support, and inspire me in my academic endeavors, and in life. It is
a village of people too numerous to name, and no measure of thanks is
commensurate with the gratitude I will always feel for their help. That said, it is
nothing short of an honor to single out a few who have been instrumental in helping
me at USC.
I am grateful to my advisors, Hanna Reisler and Curt Wittig. It has been a
pleasure to interact with such insightful and ebullient mentors, and I will continue to
seek their counsel for as long as they will permit it. In particular, I am indebted to
them for taking a chance on me when all I had to offer was enthusiasm and a promise
to work diligently.
A postdoc, Wousik Kim, and a senior graduate student, Minda Suchan,
welcomed me to the project and made me feel useful. Indeed, for a while I was
Minda's shadow, learning from someone who embraced readily the responsibilities
of teaching a novice. Samantha Hawkins (a.k.a. Karate Sam) and Sergey Malyk were
delightful partners who ensured laboratory discussions were always engaging and
not always about science. I consider myself fortunate to have met all these
individuals. I would like to express my thanks to Oscar Rebolledo-Mayoral who
joined the project just as I was leaving it and willingly helped to prepare this
iv
dissertation. I am grateful also to Sun Lee, my first office mate and now 'adopted'
sister, who tolerated my various eccentricities with a smile and contentiously tried to
ensure that I did not skip lunch.
My time at the Seaver Science Center was spent in an atmosphere that
engendered friendships and collaborations among the numerous chemistry groups
situated at this center. Thus, I must extend my sincere thanks to Professors
Bradforth, Koel, Krylov, Mak, Reisler, Villesov and Wittig for fostering this
environment. I have enjoyed getting to know many of the individuals who were
members of their groups during my graduate studies, and I appreciate the support
and assistance these individuals provided. Thanks also to all the support staff at the
Chemistry department, especially to Valerie Childress, Michele Dea, Robert Martin,
Heather Meunier-Connor, Jennifer Torres and Yuki Yabuta.
Last, but not least, I thank my family for their unwavering encouragement and
support during this venture; I am privileged to have had such support throughout all
my undertakings.
v
Table of Contents

Dedication ii
Acknowledgements iii
List of Figures vii
Abstract xiii
Chapter 1: Introduction 1
1.1 Wading Through Water 1
1.2 The Low Temperature Ice Phases: ASW, CI and HI 3
1.3 Icy Mixtures 10
1.4 Thesis Overview 16
1.5 Chapter 1 References 18

Chapter 2: Fourier Transform Infrared Spectroscopy and Temperature
Programmed Desorption 23
2.1 Absorption Spectroscopy 23
2.2 Fourier Transform Infrared Spectroscopy 26
2.3 The Michelson Interferometer 27
2.4 The Fourier Transform 32
2.5 Practical Aspects of FTIR Spectroscopy 33
2.6 Temperature Programmed Desorption 42
2.7 Chapter 2 References 46

Chapter 3: Experimental Apparatus 48
3.1 Introduction 48
3.2 The UHV System 48
3.3 The Top Tier: the FTIR Chamber 51
3.4 FTIR Optics and Set-up 52
3.5 The Surface Manipulator 56
3.6 The Substrate 57
3.7 The Sample Holder 59
3.8 The Sample Holder: the Second Design 62
3.9 Sample Preparation 65
3.10 Chapter 3 References 67

vi
Chapter 4: Transport and Guest-Host Interactions in Ice 68
4.1 Introduction 68
4.2 Experimental Details 71
4.3 Results: Interactions of CO
2
with ASW 74
4.4 Results: N
2
O Interactions with ASW 82
4.5 Discussion 90
4.6 Summary 98
4.7 Chapter 4 References 101

Chapter 5: The Nature of Trapping Sites in Ice 103
5.1 Introduction 103
5.2 Experimental 106
5.3 Results 107
5.4 Discussion 111
5.5 Epilogue 113
5.6 Chapter 5 References 115

Chapter 6: Future Experiments 117
6.1 Amorphous Materials 117
6.2 ASW: Transport and the Amorphous-Crystalline Interface 118
6.3 Experimental Strategy 122
6.4 Experimental Details: the Current Configuration 127
6.5 Experimental Details: the Modifications 129
6.6 Summary 131
6.7 Chapter 6 References 133

Bibliography 135
vii
List of Figures

Figure 2.1 Radiation (shown in blue) from a broadband IR source is split
by a beamsplitter (red) and directed toward two mirrors. The
radiation reflected from these mirrors recombines at the beam
splitter. The radiation reflected back to the source during this
recombination has been omitted for clarity. The moving mirror
is used to change the path difference between the two beams. 27
Figure 2.2 The intensity at the detector oscillates as a result of
interference. This is shown in plots of ) 2 cos( 1 x vΔ + π versus
phase difference for plane waves of light with (a) one
wavelength, 2000 cm
-1
, (b) two wavelengths, 2000 and 2200
cm
-1
, with the same irradiance, and (c) two wavelengths, 2000
and 2200 cm
-1
, of different irradiance (the 2000 cm
-1
wave has
0.6 the irradiance of the 2200 cm
-1
wave). 30
Figure 2.3 The Fourier transform of a boxcar function (shown in (a)) is
the sinc function (shown in (b)), and this function has several
side lobes adjacent to its main feature. (c) The spectrum of a
monochromatic line takes on a sinc(x) lineshape as a result of
the inability to measure the interference signal over
∞ ≤ Δ ≤ ∞ − x . The width of the feature obtained is shown. 35
Figure 2.4 A triangular apodization function (shown in (a)) Fourier
transforms to yield a sinc
2
function (shown in (b)), which has
side lobes that are small compared to the main feature.
Although there is significant overlap when two sinc
2
functions
are separated by the distance shown in (c), these two functions
(blue) are still distinguishable in the combined trace (red). At
this separation, two sinc functions would display no overlap in
their main feature. 37
Figure 3.1 A schematic of the UHV Chamber. The numerous ports on
tiers 1 and 2 have been omitted for clarity. The entire UHV
system is evacuated by a turbomolecular pump located on the
second tier. 50
Figure 3.2 A schematic of the FTIR chamber. The axis system shown
relates the orientation of this view to the view depicted in
Figure 1. 52
viii
Figure 3.3 A schematic showing the IR beam path from the source to the
detector. All pertinent components outside the UHV chamber
and the spectrometer are placed in purge boxes (shown in
green). 53
Figure 3.4 The peak-to-peak absorbance routinely obtained in the ~2700
cm
-1
region with the experimental configuration described in
this chapter. 56
Figure 3.5 The surface manipulator that is attached to the UHV system.
The stainless steel tube of the manipulator and the copper
piece attached to this tube are shown explicitly in the inset. 58
Figure 3.6 A drawing of the surface holder used to attach the surface to
the manipulator. The two main copper parts of the holder are
labeled A and B. One of the two copper plates that are used to
sandwich the surface is shown. 60
Figure 3.7 A drawing of sample holder 2; the copper foil is attached to
the two main parts of the holder (A and B) using a screw
(inserted into the tapped hole labeled C). 63
Figure 3.8 The homemade heater is a wire coil with alumina sleeves. The
sleeves insulate the wire and ensure current travels throughout
the entire length of the wire. 65
Figure 4.1 (a) Varying amounts of CO
2
deposited at 90 K onto ASW
films of approximately constant thickness (∼65 layers). CO
2

exposure times (in minutes) at constant pressure (4 × 10
-8

Torr) were: (i) 1.5 (ii) 3 (iii) 6. The broad water feature at
~3250 cm
-1
provides estimates of relative film thickness. The
horizontal scale is expanded to emphasize (b) the CO
2
and (c)
13
CO
2
ν
3
region. 75
Figure 4.2 The same amount of CO
2
was deposited at 90 K onto ASW
films of varying thickness. The films were exposed to 4 × 10
-8

Torr CO
2
for 3 min. H
2
O exposure times at 5 × 10
-8
Torr were:
(i) 2 (ii) 4 (iii) 8 and (iv) 16 min, corresponding to 10, 20, 40
and 80 layers, respectively. Entries (a) and (b) show spectral
regions for CO
2
and H
2
O, respectively. 76

ix
Figure 4.3 Spectra of the CO
2
that remained after depositing at 90 K
equal amounts of CO
2
(3 min at 4 × 10
-8
Torr) onto ASW
films of varying thickness followed by raising the temperature.
The temperature was raised to 105 K where it was held for 15-
20 minutes before each trace was obtained. Entries (i) – (iii)
correspond to ASW films of thickness 20, 40, and 80 layers,
respectively. Panels (a) and (b) show the same traces on
different horizontal scales. 78
Figure 4.4 Spectra of the CO
2
that remained after depositing different
amounts of CO
2
onto ASW films of constant thickness (~65
layers) and then increasing the temperature to 105 K. The
spectra were recorded at 105 K, 15-20 min after the reaching
this temperature. The CO
2
exposure times at 4 × 10
-8
Torr
were (i) 1.5 (ii) 3 and (iii) 6 min. See Figure 4.1 for the spectra
recorded following CO
2
deposition at 90 K. 79
Figure 4.5 CO
2
(4 × 10
-8
Torr, 4 min) was deposited onto a 90 K ASW
film of 40 layers. The temperature was then raised to 105 K
for 15 min, at which time a spectrum was recorded. The
figure depicts the evolution of the 105 K spectra as the
temperature was increased in steps to the temperatures shown
and held at these temperatures for the duration of a scan (∼8
min). The spectra were recorded immediately after each
temperature increase and at the temperatures indicated. Entries
(a) and (b) show the spectral region for CO
2
and H
2
O,
respectively. The broad H
2
O feature (~3250 cm
-1
) changes
upon annealing past 165 K because of the ASW-to-cubic
phase transition. 80
Figure 4.6 Different amounts of CO
2
on an ASW film (~1000 layers).
The amount of CO
2
was varied via the sequence: (i) no
deposition (ii) 4 × 10
-8
Torr for 2 min and (iii) 1 × 10
-7
Torr
for 16 min. Spectra shown were recorded immediately after
each deposition. The inset shows the d–OH feature is red
shifted as a consequence of CO
2
deposition; namely the
feature at ~3700 cm
-1
moves to ~3650 cm
-1
. 82
Figure 4.7 (a) Varying amounts of N
2
O were deposited onto 90 K ASW
films. The broad H
2
O feature (~3250 cm
-1
) provides an
estimate of relative film thickness, which were approximately
equal to 70 layers. The N
2
O exposure times at 4 × 10
-8
Torr
were: (i) 2 (ii) 4 and (iii) 8 min. (b) The horizontal scale is
expanded to emphasize the N
2
O ν
3
region. 83
x
Figure 4.8 (a) ASW films of different thicknesses were prepared at 90 K
by vapor depositing H
2
O at 4 × 10
-8
Torr. (a) H
2
O spectra (i) –
(iv) were recorded with deposition times of (i) 4, (ii) 8, and
(iii) 16 min, corresponding roughly to 15, 35, and 70 layers.
(b) Each film was exposed to 4 × 10
-8
Torr N
2
O for 4 min, and
the corresponding N
2
O spectra are labelled (i)-(iv). 84
Figure 4.9 ASW films at 90 K were exposed to 4 × 10
-8
Torr N
2
O for 4
min. The temperature was then raised to 105 K and held there
for 15-20 min before a spectrum was recorded. Approximate
ASW thicknesses for (i)–(iv) were 15, 35, 70, and 150 layers,
respectively. Entries (a) and (b) show H
2
O and N
2
O features,
respectively. 85
Figure 4.10 Spectra of the N
2
O that remained after depositing different
amounts of N
2
O onto ASW films of same thickness (70
layers) and then increasing the temperature to 105 K. The
temperature was then raised to 105 K and held there for 15–20
min before a spectrum was recorded. N
2
O exposure times at 4
× 10
-8
Torr for (i)–(iii) were 2, 4, and 8 min, respectively. 86
Figure 4.11 N
2
O (4 × 10
-8
Torr for 4 min) was deposited onto a 90 K ASW
film of 40 layers. The temperature was then raised to 105 K
for 15 min, at which time a spectrum was recorded. The figure
depicts the evolution of the 105 K spectra as the temperature
was increased in steps to the temperatures shown and held at
these temperatures for the duration of a scan (∼8 min). The
spectra were recorded immediately after each temperature
increase and at the temperatures indicated. Three scans (~5
min in duration) were recorded at 165 K; the time interval
between reaching 165 K and commencing each of these scans
is specified for three relevant spectra. 87
Figure 4.12 Spectra were recorded after sandwiching N
2
O between two
ASW films at 90 K, then raising the temperature to 105 K, and
keeping it there for 15 min. The amount of deposited N
2
O and
the total number of water layers (~80) is the same for all
spectra. The ratios of bottom layer to top layer thickness are:
(i) 80:0, (ii) 60:20, (iii) 40:40, and (iv) 20:60. 88

xi
Figure 4.13 Spectra were recorded for different amounts of N
2
O deposited
onto an ASW film of ~1000 layers at 90 K. The amount of
N
2
O was increased via the sequence: (i) no deposition, (ii) 4 ×
10
-8
Torr for 4 min, (iii) 1 × 10
-7
Torr for 3 min, and (iv) 2 ×
10
-7
Torr for 7 min. Spectra were recorded immediately
following each deposition. The inset shows the d–OH feature
is red shifted as a consequence of N
2
O deposition. 89
Figure 4.14 90 K ASW films of approximately the same thickness were
exposed to 4 × 10
-8
Torr of CO
2
for the same duration (3 min).
After CO
2
deposition, the temperature was increased to 105 K
and held there while a spectrum was recorded. Traces are the
spectra recorded for: (i) a film that was not annealed prior to
CO
2
exposure and (ii) a film that was annealed for 15–20 min
at 120 K and then re-cooled to 90 K prior to CO
2
exposure.
The inset shows these spectra on an expanded horizontal scale
with the corresponding absorbance magnification. 97
Figure 5.1 A TPD trace of a CO
2
film prepared at 90 K on MgO(100).
The TPD experiment was performed by heating the surface at
~1 K/s and monitoring a mass-to-charge ratio (m/e) of 44.
MgO exposure to CO
2
was carried out at a pressure 4 × 10
-8

Torr for 3 min. 107
Figure 5.2 CO
2
was deposited

(4 × 10
-8
Torr for 3 min) onto an ASW film
(5 × 10
-8
Torr for 7 min) ~80 layers thick. The H
2
O desorption
was monitored by measuring m/e 18. Panels (a) and (b) show
the TPD traces of CO
2
and H
2
O respectively. The TPD trace
of H
2
O was scaled for clarity. 109
Figure 5.3 TPD traces of CO
2
(4 × 10
-8
Torr for 3 min) deposited onto
ASW films (~80 layers). The temperature to which each
sample was annealed prior to recording FTIR spectrum is
indicated. As the traces show, the CO
2
TPD feature at 185 K
was not affected by the annealing to 105 K. 110
Figure 5.4 Spectra (p-polarized) of ASW films (~80 layers) exposed to
CO
2
(4 × 10
-8
Torr for 3 min). Each sample was annealed to
the temperature indicated and then re-cooled to 90 K. The
broad H
2
O feature (~3250 cm
-1
) changes upon annealing past
165 K because of the ASW-to-cubic phase transition. The
inset shows the expanded scale of the CO
2
ν
3
band absorbance
region with the corresponding magnification factor. 110
xii
Figure 6.1 An ASW film with domains of crystalline ice (shown in red)
may possess cracks or voids (shown in blue) at the
amorphous-crystalline interface. 119
Figure 6.2 The slabs (shown in blue) represent isolated regions of ASW
on a supporting substrate. As these isolated regions approach
the phase transition temperature, there may be a significant
increase in lateral mobility (indicated by the black arrows) and
material may be transported to the surrounding regions. 121
Figure 6.3 (a) An ASW film is formed on a supporting substrate. (b) A
mesh (shown in black) is placed in front of the film, and the
film is irradiated. (c) All the ASW in the exposed areas
desorb, leaving the structure depicted in blue. To form isolated
rows of ASW with the axes of the rows parallel to the y-axis
shown, the mesh can be translated in small increments along
the x-axis. At each increment the substrate is irradiated to
desorb any exposed ASW. To form isolated areas of ASW (the
black squares), this process has to be done along both axes of
the mesh (shown as x and y). 123
xiii
Abstract

Guest-host interactions in amorphous solid water (ASW) films doped with CO
2

or N
2
O were examined experimentally. Investigations focused on exploring
molecular transport and morphology in ASW. The main diagnostics were Fourier
transform infrared (FTIR) spectroscopy and temperature programmed desorption
(TPD).
ASW films were prepared at 90 K. Dopants exposed to these films saturate all
the favorable sites within the film interior. The subsequent accumulation of dopants
occurs atop the ASW films, and results in a polycrystalline film. Infrared signatures
permit a distinction for these two cases; vibrational modes of the crystalline
overlayer, and an easily discernible peak for dopants residing within the ASW film.
Desorption experiments show ASW traps some of the dopants residing within its
bulk. As anticipated, some of these trapped species are released during the ASW-to-
cubic ice transition at ~160 K. Annealing the 90 K ASW films to 120 K prior to
dopant exposure lowers the film's capacity to include dopants within its bulk. No
substantial differences were observed for CO
2
and N
2
O. ASW dangling OH bonds
displayed a red shift of roughly 50 cm
-1
upon interaction with these dopants.
Sandwich structures were prepared at 90 K, i.e., ASW-dopant-ASW. At 105 K,
some of the dopant molecules sandwiched in between the ASW escape by diffusion
through the top ASW layer, and the remaining molecules appear to be distributed
within the ASW film interior.
xiv
The factors mediating the ratio of dopants released during the phase transition to
dopants that co-desorb with crystalline (cubic) ice were studied. The infrared
signatures for these two sets of molecules suggest their local environments are
similar. Future experiments using small molecules to explore molecular transport at
complex interfaces are outlined.

1

Chapter 1: Introduction
1.1 Wading Through Water
In all of science, there is arguably no substance that has been investigated more
times than water. On one hand, these efforts are indicative of a complex system that
has been difficult to characterize using the available experimental and theoretical
tools. On the other hand, these endeavors reflect the privileged status held by this
critical life component. In spite of the progress that has been made in understanding
the properties of water, there is plenty that remains unresolved.
Why has water been so difficult to characterize? Water is the richest one-
component system known, with numerous phases each possessing distinct properties
[1]. The predominant intramolecular interaction in pure water is hydrogen bonding
[1, 2] and this is a challenging type of interaction to model in any system.
Furthermore, because of its superior solvating capabilities, ensuring low levels of
impurities in water is easier said than done. These impurities can lead to erroneous
conclusions about experimental results. The example of "polywater" comes readily
to mind; several years of research were required to conclude that this was not another
phase of liquid water but simply liquid water containing small quantities of various
impurities [3].
Everyday processes involving water are often rather complex. For example, it is
well-known that water freezes to ice upon cooling, and yet there are very few
molecular dynamics simulations illuminating the details of this phenomenon [4]. In
2
fact, the inordinate number of water studies clouds the reality that there is no
"universally consistent fundamental description of water systems" [5]. The drive for
a molecular level understanding of the properties and role of water in its condensed
phase continues to fuel studies in a broad range of scientific fields. Atmospheric
chemistry, molecular biology, and interstellar science are all areas where attaining
this state of knowledge is critical to advances in basic and applied research.
Solid water, or ice as it is commonly known, has garnered special attention as an
important surface for heterogeneous reactions [6-9]. The availability of water
molecules for reaction and product stabilization appears to be a key factor mediating
some heterogeneous processes [10]. Novel insights into important stratospheric
chemical processes have been gleaned by focusing on the interaction of small (less
than 5 atoms) molecules with ice films. In particular, studies of atmospherically
important chlorine reservoir species, such as HCl and chlorine nitrate (ClONO
2
),
have shown that ice can influence the ozone cycle [6, 8]. The photochemistry of
molecules constrained in or on ice is also becoming an area of fertile research,
especially since work demonstrating that glycine, adenine, and serine form naturally
after the ultraviolet (UV) photolysis of certain ice mixtures [11].
Among vitreous solids, the amorphous form of ice appears to be a substance of
extraordinary character. Its low frequency collective vibrations, or phonons, exhibit
characteristics similar to that of phonons in a crystalline material [12] rather than the
behavior often observed in glassy materials. These results have been interpreted to
mean that amorphous water can exist as disordered state with an unusually low
3
degree of local disorder. There is also speculation that amorphous water may provide
the first clear evidence of "polyamorphism", the coexistence of two equilibrated
amorphous phases [13, 14]. In addition, there are indications that this ice phase may
serve as a model system for liquid water [15, 16] and other amorphous solid systems
in general.
Performing experiments in ultrahigh vacuum (UHV) minimizes the concentration
of unwanted species, and this reduces or eliminates many of the complexities
associated with competing processes. In particular, this type of environment is
crucial to simplifying the characterization of solid water. Only three ice phases can
be cooled enough to exhibit a vapor pressure of less than 10
-10
Torr: amorphous solid
water (ASW), cubic ice (CI) and hexagonal ice (HI) [1]. These ice forms are the only
thermodynamically stable phases in UHV, i.e., the only phases capable of existing at
these low pressures (< 10
-9
Torr). However, HI is difficult to prepare using the
convenient method of vapor deposition, and thus experimental studies in these
simplified environments generally focus on ASW and CI.

1.2 The Low Temperature Ice Phases: ASW, CI and HI
In general, ASW may be defined as a "solid phase of water that is metastable
with respect to its crystalline phase" [17]. Here, the word metastable is meant to
convey that ASW is never the most thermodynamically stable phase under any
conditions. The existence of ASW stems solely from a kinetic favoring of its
formation, over a more stable solid water form, in certain circumstances.
4
Nevertheless, ASW is the most prevalent ice phase in the universe [18]. It is found
mainly in the interstellar medium, where it is usually combined with other condensed
gases. In this environment, it is subject to UV radiation, ion winds, and cosmic rays,
all of which may induce a host of effects; namely, chemical reactions, desorption,
and lattice changes [19]. Needless to say, ASW is the solid water phase most
relevant to the astrophysical community, and it has been the subject of many reviews
[13, 14, 20].
The definition of ASW proffered above belies the vigorous debate concerning the
very nature of ASW. The first report on ASW dates back to 1935; X-ray diffraction
studies by Burton and Oliver [21] observed a diffuse pattern for water vapor
deposited onto a copper rod at temperatures less than ~150 K. Since then, some
groups have argued that different forms, or polymorphs, of amorphous ice (AI) exist.
"Amorphousness" is difficult to quantify, and thus these groups have resorted to
using certain properties, such as density, to distinguish between these AI
polymorphs. Based on X-ray diffraction studies, Narten and coworkers [22] reported
a "high density" (1.07 g/cm
3
) amorphous phase prepared by vapor deposition at 10
K. This "high density" phase irreversibly changed to a "low density" (0.94 g/cm
3
) AI
over the range of 30-80 K. The compression of HI reportedly produces an AI with a
density of 1.31 g/cm
3
, and this ice remains amorphous even after compression [23],
slowly transforming into an ice form with a density of 1.17 g/cm
3
. Whether ASW
does indeed have several polymorphs or if these varying densities stem from some
other phenomenon is still an open question.
5
One of the important properties of ASW is its porosity, especially since this
creates a relatively large available surface area for heterogeneous processes. ASW
diffraction studies, which focus on short-range atomic oxygen distributions, provide
no details about pore volumes. Therefore, the densities derived from these studies do
not account for pore volumes in the ratio of mass to volume, and they are sometimes
referred to as intrinsic densities [19]. In fact, due to its high porosity (relative to "low
density" AI), the bulk density of "high density" AI is expected to be less than the bulk
density of "low density" AI. Here, the term bulk density implies a density that takes
into account phase porosity.
The porosity of ASW has been inferred from gas adsorption experiments at low
temperatures (less than 30 K) [24, 25] and from other isothermal gas adsorption
experiments [16, 26-28]; the large adsorption areas determined were not reconcilable
with a non-porous substance. It is worth noting that gas adsorption studies provide
information on the pores connected to the vacuum but say little or nothing about the
enclosed pores. Measurements of the refractive index of ice, related to porosity by
the Lorentz-Lorenz relation, can also be used to estimate ASW porosity [29].
However, there has been little progress on determining typical ASW pore
dimensions, let alone pore distributions. Even the nature of the ASW surface remains
unresolved [30, 31].
Why is ASW porous? It has long been speculated that at low substrate
temperatures vapor deposited molecules quickly dissipate energy upon incidence,
and therefore tend to remain close to the point of landing [1]. Such simple "hit and
6
stick" models have been used to depict how the porosity of a film depends on the
incident angle of the incoming particles [25, 32]. These simulations are consistent
with experiments probing the dependence of ASW morphology on the H
2
O
molecular beam incidence angle. Nevertheless, while there are postulates on how the
energy dissipation probably occurs for ASW [19], there does not appear to be any
experimental evidence directly supporting any of these rationalizations.
ASW films, vapor deposited at low temperatures, have displayed irreversible
changes in properties upon annealing to temperatures well below (~ 30 K) any
appreciable phase change [32-34]. Many groups have observed ASW spectral
changes induced by annealing ASW deposited at low temperatures [33, 35, 36].
Hallbucker and coworkers [37] reported a heat release, presumably due to relaxation,
during the heating of ASW. This heat release was accompanied by changes in the X-
ray diffraction pattern. In addition, gas phase adsorption experiments [32, 33]
suggest a decrease in ASW porosity occurs even when ASW films are annealed to
temperatures not conducive to crystallization. Mainly because of compelling gas
phase adsorption experiments [32], these irreversible changes have been interpreted
as signaling slight molecular rearrangements in ASW that result in permanent pore
closures. It is worthwhile noting that in light of this information, the general
assumption that ASW by definition is porous may not always apply. Further
investigations are required for a more complete picture.
There are a variety of methods used to form ASW: vapor deposition onto a cold
substrate [21, 34], the rapid cooling of liquid water [16, 38, 39] and the high pressure
7
"amorphization" of crystalline ice [40-42]. Vapor deposition is by far the easiest
technique for UHV studies, and thus it is almost always the method of choice in such
endeavors. Kimmel and coworkers [32] found that the porosity of ASW films,
examined via gas adsorption experiments, were dependent upon the incident angle of
the collimated H
2
O molecular beam used to form the film. Many agree that this work
demonstrates the sensitivity of ASW characteristics to growth conditions. This
dependence suggests that the conflicting results concerning the physical properties of
ASW may stem from different growth conditions. Indeed, there are speculations that
perhaps a continuum of AI structures can be formed, each with varying bulk physical
properties dependent upon growth conditions [19].
In short, the current understanding of basic ASW properties is fragmentary. In
part, this is because ASW does not display the properties of a single, well-defined
phase, and consequently it has been a challenging ice form to characterize. There is a
need for further experimental investigations that shed light on the size range of
pores, the mechanisms responsible for the irreversible structural changes upon
annealing, and the structure of the ASW surface. From a theoretical standpoint, a
detailed understanding of the processes (e.g., hydrogen bonding) that combine to
yield the characteristic ASW infrared spectrum remains an outstanding problem.
More importantly, it is apparent that to perform meaningful comparisons between
various ASW studies, growth conditions must be specified precisely.
Like ASW, cubic ice is also a metastable ice phase. First reports about its
existence originate from electron diffraction studies by Konig in 1942 [1]. In these
8
studies, the diffuse diffraction pattern of ASW changed upon annealing into a pattern
consisting of sharp rings that suggested a cubic crystalline ice structure (similar to
that in diamond). The metastable nature of CI was deduced from the irreversible
transformation of CI to HI at temperatures greater than ~200 K [1]. Moreover, CI
produced when ASW is annealed does not revert to ASW upon cooling, and this
shows that CI is a more thermodynamically favored form of ice than AI. This phase
transition from ASW to CI has been the subject of much scrutiny, and it purportedly
occurs over the range of ~135-165 K [13, 33, 34, 43].
Unlike ASW, cubic ice displays the properties of a well-defined phase, and thus
many of its basic attributes are known. As mentioned above, its structure has been
precisely determined from diffractions studies, and density measurements [1, 2] all
generally show a density of ~0.94 g/cm
3
. Compared to ASW, CI exhibits very little
ability to trap gases, and this suggests that it is non-porous. CI is routinely prepared
in UHV by annealing ASW [34, 44] or vapor depositing H
2
O onto a substrate with a
temperature of 135-155 K [45, 46]. However, there are other documented methods of
producing CI in other environments, i.e., not low pressures [47]. The phase transition
from ASW to CI is accompanied by spectral changes, and these changes provide a
convenient indication of this process. Unfortunately, there are a limited amount of
techniques that can be used to confirm a total transformation. It is thus worth bearing
in mind that experiments with CI are susceptible to the question of phase purity and
the possibility of ASW domains within CI.
9
The solid water form obtained when liquid water is frozen at atmospheric
pressures is HI. Consequently, it is the ice form commonly referred to as "ice". Like
CI, it is crystalline and displays well-defined characteristics [1]. Several diffraction
studies have confirmed its structure, although the exact position of the hydrogen
atoms was temporarily a point of contention [1].
Forming HI via vapor deposition requires substrate temperatures ~200 K and
higher. At these temperatures, the H
2
O desorption flux is high (> 10
-5
Torr) [48], thus
and ambient pressures greater than 10
-5
Torr are needed to form a film [49]. These
pressures increase the likelihood of trapping unwanted species within the film during
sample preparation. More importantly, it is difficult to pump out systems quickly
(within ~15 minutes) after such exposures. Thus, for mostly experimental reasons
associated with its preparation, HI is rarely studied in a simplified UHV
environment. This is unfortunate because there are important processes that involve
this particular ice form. For example, polar stratospheric clouds, which play a major
role in the ozone cycle, have HI as a major constituent [29, 50, 51]. Nevertheless,
because of the similarities in the physical structure of CI and HI, many experiments
performed in low pressure environments utilize CI as a model for HI [29, 50, 52].
The current state of knowledge on low temperature ice, and also of water in
general, has been acquired over several decades by accruing an expansive body of
theory and experiments. It is prudent, if not logical, to assume that the lingering
questions will be resolved also by the meticulous accumulation of a large set of new
results. In a general sense, the work described in this dissertation is a contribution to
10
the body of experimental data necessary for formulating a fundamental description of
solid water. More specifically, it is an examination of the guest-host interactions
within ice (ASW and CI to be exact) designed to yield basic knowledge about
structure and molecular transport in ice. To set the results of this work in context, it
is constructive to review briefly some of the previous studies detailing the interaction
between various molecules and ice.

1.3 Icy Mixtures
The localization of an atom or molecule on the surface of a solid is termed
adsorption. This terminology is relatively unambiguous when it applies to a
nonporous solid that permits negligible diffusion into the bulk lattice by any atomic
or molecular species residing on the solid surface. Some solids can be prepared in a
manner that imbues them with an intrinsic porosity, e.g., zeolites. In this case, an
atom or molecule located on the surface of a pore can also be described as being
adsorbed. Several reports detailing ice-related experiments use the term adsorption to
describe accumulation in, or on, ice pores and on the outer molecular layer of an ice
film or cluster [27, 44, 53, 54]. It is this relatively loose definition that will be used in
this discussion. However, it is relevant to note that there is often little evidence in
these studies that allows a distinction between a molecule located on the surface of a
pore and within the bulk lattice.
There is a tendency to categorize interactions between ice and various species in
terms of interaction strength; these energies are derived from vibrational shifts of ice
11
bands that are observed as ice interacts with different molecules [53, 55]. The spectra
of ice in the OH stretching region is dominated by a broad (~300 cm
-1
) vibrational
feature at ~3250 cm
-1
, and, in some ice structures, a relatively weak feature at ~3700
cm
-1
is present as well [36, 55]. The latter feature is believed to originate from
dangling OH (hereafter denoted by d–OH) groups that do not participate in the
hydrogen bonding network present in ice, and thus this feature is observed only
when these groups are present in appreciable concentrations. Several groups have
reported no discernible change in the broad OH signal (at ~3250 cm
-1
) during the
interaction of various molecules with ice. However, the d–OH signal has displayed
vibrational shifts dependent upon the molecular species associating with the ice film,
and these shifts range from a few to a hundred wavenumbers [55-57].
Compared to the typical interaction energies and their corresponding vibrational
band shifts observed in rare gas matrices, the interaction energies derived from such
shifts are small (less than 20 kJ/mol ). Thus, there are some questions concerning the
validity of the process used to obtain the interaction strength of molecules interacting
with ice [33]. A less controversial method of characterizing these interactions is
simply in terms of the magnitude of the d–OH band shift. This approach sidesteps
the debate concerning the correlation of d–OH band shifts and interaction energies
and still affords a convenient categorization scheme.
Generally, the shift experienced by the d–OH band during the aforementioned
interactions is always to lower frequencies, i.e., a red shift. For example, the
exposure of Ar, Kr, or CF
4
to a freshly prepared ASW film results in d–OH band
12
shifts of roughly 12, 17, and 7 cm
-1
, respectively [58]. On the other hand, acetylene,
ethylene, and benzene each induce a shift of ~100 cm
-1
in ice clusters [53]. In fact,
the d–OH shift can become indiscernible when certain molecules are deposited onto
ice; Schaff and Roberts [52] observed this phenomenon during the infrared studies of
acetone on ice. In this case, the red shift occurs to such an extent that the weak d–OH
band overlaps with the dominant OH band, and therefore this weak band becomes
indistinguishable. These changes in the d–OH band position are attributed to
"hydrogen-bond-like" associations between these various molecules and ice. This
conclusion is based partly on the fact that the formation of hydrogen bonds in other
systems also tends to induce red shifts.
Certain molecules not only adsorb onto ice but also appear to induce molecular
rearrangements within the ice lattice. HCl, an atmospherically important species, is
one such molecule and elucidating the fundamental mechanism of HCl adsorption
has been especially challenging. Firstly, the nature of the adsorbed HCl was
vigorously debated, with early arguments centered upon whether the adsorbate was
molecular or ionic [50, 59-61]. Contrary to general intuition, there is experimental
evidence suggesting HCl can ionize once adsorbed molecularly on ice [50, 62], and
this implies a reorientation of neighboring water molecules to accommodate the
process. In addition, some studies observe an eventual disruption of the ice lattice
and the formation of stable ionic hydrates at appropriate HCl-water stoichiometries
[10]. There now appears to be little doubt that the initial adsorption state for HCl
serves as a precursor state from which subsequent interactions can occur.
13
Investigations of N
2
O
5
and ClONO
2
on ice films have arrived at similar conclusions
concerning the role of the initial adsorption state [19].
IR spectroscopy has been an invaluable tool in characterizing the nature of ice
and its interactions with other species. However, thermal desorption studies [28, 63-
67] have also been beneficial in providing novel insights into the sublimation
characteristics of various ice systems. Two disparate fields have driven such studies:
atmospheric chemistry and astrophysics. For astrophysics in particular, the
realization that ice could trap and retain molecules to temperatures much higher than
the sublimation temperatures exhibited by these molecules was an important one
[28].
Early studies focused on establishing the sublimation traits of icy mixtures via
thermal desorption experiments. One of the earliest reports on the trapping
capabilities of ASW showed that the release of species trapped in ice occurred at
distinct temperatures: O
2
trapped in ASW (prepared at 77 K) desorbed at roughly 95
K, 160 K and 214 K [28]. This behavior, in conjunction with similar observations for
other trapped species in ASW [64, 68], implied that species could be present in
interstellar ices at temperatures far beyond their sublimation temperatures. Sandford
and Allamandola recorded the IR spectra of ice mixtures for different temperatures
and "gas/H
2
O" ratios, and they found the release of trapped species did not transpire
continuously as the ice temperature was increased [66, 67, 69]. An extension of these
trapping studies was done by Hudson and Donn [63]. Monitoring gas release with a
14
mass spectrometer, they were able to correlate changes in the IR spectra of ice (e.g.,
the ASW-to-crystalline ice transition) with the desorption of trapped species.
Recent studies have extended this work through the development of simplified
models capable of quantitatively simulating experimental desorption data [70]. These
desorption models have taken into account the fact that gas release in certain
temperature regions is not indicative of the interaction strength between a molecular
(or atomic) species and ice. Instead, these expulsions are related to changes induced
in ice by thermal energy supplied during ice annealing, and consequently the
desorption attributes of a species leaving an ice film is indeed mediated by the film.
Several recent reports in this field have focused on the formation complex products
produced by UV photolysis of interstellar ices [11]. These investigations utilize
laboratory analogs of interstellar ice to explore the transformation in composition of
interstellar ice induced by UV radiation, and they represent a new front in ice-related
studies.
In general, the desorption of a molecular adlayer, consisting of multiple layers,
from a surface results in two distinct desorption trace features (a detailed explanation
of how desorption traces are obtained is provided in the ensuing chapter of this
dissertation). One feature originates from molecules directly in contact with the
underlying surface, i.e., the monolayer, while the other is from molecules separated
from the solid surface by one or more layers of molecules, i.e., the multilayer.
Because of a screening effect by the monolayer, the binding energy of molecules in
the multilayer to the surface is less than the energy binding the monolayer to the
15
surface. Consequently, less energy is required to remove the multilayer, and it
desorbs at lower temperatures than the monolayer.
The desorption characteristics of a species interacting with ice can be explained
relatively simply if the species does not display the complex adsorption dynamics
associated with molecules like HCl. In a thermal desorption study of several
astrophysically relevant molecules, Collings and coworkers [71] found that less than
four prominent desorption features were present in desorption traces of several
species deposited onto thin (< ~100 layers thick) ASW films. In addition to the
monolayer and multilayer desorption from the ice surface, they observed two
features from molecules that get trapped within ASW. For some molecules, however,
the monolayer and multilayer features appeared to overlap and were
indistinguishable from one another. One of the features from the trapped molecules
is associated with molecules that desorb during the ASW phase transition and the
other is from molecules that desorb during the cubic ice film desorption. The release
of trapped species during the aforementioned phase transition has been observed by
several groups [24, 63, 69, 72, 73] and this release is sometimes referred to as a
molecular volcano [24].
It is widely accepted that molecules trapped in ice can be released during the
ASW-to-CI transition. What is unknown is the process that mediates the ratio of
trapped species released during the phase transition to trapped species that are
retained within CI. Several investigators have observed both features [24, 63, 71-73]
during desorption studies, and yet the nature of site from which these molecules
16
desorb remains a source of speculation. Ayotte and coworkers [24] have suggested
that it may originate from guest molecules trapped in a clathrate-hydrate cage.
However, there are only a few species that appear to form clathrate-hydrates under
low temperature and pressure conditions [74]. Collings and coworkers [71] observe
that this co-desorption feature scales with ASW film thickness for "CO-ice"
mixtures, suggesting that these molecules are distributed throughout the ice bulk (be
it in pores or within the lattice). However, it is not clear if the amount of co-
desorbing species, which is proportional to the intensity of the co-desorption feature,
can be manipulated independent of ASW thickness, e.g., by changing deposition
conditions.

1.4 Thesis Overview
The nature of the interaction between CO
2
and N
2
O molecules and ice is the subject
of this dissertation, and it is work that was motivated primarily to understand the
trapping, adsorption, and desorption processes in ice. The rationalizations supporting
the choice of these molecules (CO
2
, N
2
O) is detailed in the ensuing chapters. For the
first time, the d–OH red shift observed for the "CO
2
-ice" and "N
2
O-ice" systems is
reported. Basically, this work
1. confirms the effectiveness of CO
2
and N
2
O as probes of ice morphology
2. expounds upon the nature of the co-desorbing species in ice
17
3. sets the stage for future experiments, utilizing these probes, that will explore
molecular transport at complex interfaces with ASW as a model for
amorphous systems
The theoretical constructs behind the two diagnostic techniques used in this study are
detailed in the following chapter, Chapter Two. In Chapter Three, a description of
the experimental apparatus is provided. Chapter Four discusses the intermolecular
interactions of two specific species (CO
2
or N
2
O) with ice. Preliminary work focused
on understanding the dynamics of ice crystallization is presented in Chapter Five.
Finally, Chapter Six provides a summary of future work.
18
1.5 Chapter 1 References
[1] P. V. Hobbs, Ice Physics, Claderon Press: Oxford, 1974.
[2] F. Franks, In Water: A comprehensive treatise, F. Franks, Ed.; Plenum Press,
New York, 1982, Vol. 7.

[3] F. Franks, Polywater, MIT press: Cambridge, 1981.

[4] M. Matsumoto, S. Saito and I. Ohmine, Nature, 416, 409–413, (2002).

[5] V. Buch and J. P. Devlin, In Water in Confining Geometries, V. Buch and J.
P. Devlin, Eds.; Springer, Berlin, 2003.

[6] Z. Y. Wang and S. K. Zhou, Progress in Chemistry, 16, 49, (2004).

[7] A. B. Horn, J. R. Sodeau, T. B. Roddis and N. A. Williams, Journal of
Physical Chemistry A, 102, 6107, (1998).

[8] G. Faraudo and D. E. Weibel, Progress in Reaction Kinetics and Mechanism,
26, 179, (2001).

[9] L. Chu and L. T. Chu, Journal of Physical Chemistry A, 103, 691, (1999).

[10] S. Haq, J. Harnett and A. Hodgson, Journal of Physical Chemistry B, 106,
3950, (2002).

[11] M. P. Bernstein, J. P. Dworkin, S. A. Sandford, G. W. Cooper and L. J.
Allamandola, Nature, 416, 401, (2002).

[12] H. Schober, M. M. Koza, A. Tolle, C. Masciovecchio, F. Sette and F. Fujara,
Physical Review Letters, 85, 4100, (2000).

[13] C. A. Angell, Annual Review of Physical Chemistry, 34, 593, (1983).

[14] C. A. Angell, Annual Review of Physical Chemistry, 55, 559, (2004).

[15] C. A. Angell, Science, 267, 1924, (1995).

[16] E. Mayer and R. Pletzer, Nature, 319, 298, (1986).

[17] R. S. Smith, Z. Dohnalek, G. A. Kimmel, G. Teeter, P. Ayotte, J. L.
Daschbach and B. D. Kay, In Water in Confining Geometries, V. Buch and J. P.
Devlin, Eds.; Springer, Berlin, 2003.
19

[18] T. L. Roush, Journal of Geophysical Research-Planets, 106, 33315, (2001).

[19] R. A. Baragiola, In Water in Confining Geometries, V. Buch and J. P. Devlin,
Eds.; Springer, Berlin, 2003.

[20] M. G. Sceats and S. A. Rice, In Water: A comprehensive treatise, F. Franks,
Ed.; Plenum Press, New York, 1982, Vol. 7.

[21] E. F. Burton and W. F. Oliver, Proceedings Royal Society, A153, 166, (1935).

[22] A. H. Narten, C. G. Venkatesh and S. A. Rice, The Journal of Chemical
Physics, 64, 1106, (1976).

[23] C. A. Tulk, C. J. Benmore, J. Urquidi, D. D. Klug, J. Neuefeind, B. Tomberli
and P. A. Egelstaff, Science, 297, 1320, (2002).

[24] P. Ayotte, R. S. Smith, K. P. Stevenson, Z. Dohnalek, G. A. Kimmel and B.
D. Kay, Journal of Geophysical Research-Planets, 106, 33387, (2001).

[25] K. P. Stevenson, G. A. Kimmel, Z. Dohnalek, R. S. Smith and B. D. Kay,
Science, 283, 1505, (1999).

[26] B. Schmitt, J. Ocampo and J. Klinger, Journal De Physique, 48, 519, (1987).

[27] R. Pletzer and E. Meyer, Journal of Chemical Physics, 90, 5207, (1989).

[28] J. A. Ghormley, Journal of Chemical Physics, 46, 1321, (1967).

[29] B. S. Berland, D. E. Brown, M. A. Tolbert and S. M. George, Geophysical
Research Letters, 22, 3493, (1995).

[30] D. Nordlund, H. Ogasawara, P. Wernet, M. Nyberg, M. Odelius, L. G. M.
Pettersson and A. Nilsson, Chemical Physics Letters, 395, 161, (2004).

[31] Y. Zubavichus, M. Zharnikov, Y. J. Yang, O. Fuchs, E. Umbach, C. Heske
and M. Grunze, Langmuir, 22, 7241, (2006).

[32] G. A. Kimmel, K. P. Stevenson, Z. Dohnalek, R. S. Smith and B. D. Kay,
Journal of Chemical Physics, 114, 5284, (2001).

[33] C. Manca, C. Martin and P. Roubin, Chemical Physics, 300, 53, (2004).

20
[34] P. Jenniskens, S. F. Banham, D. F. Blake and M. R. S. McCoustra, Journal of
Chemical Physics, 107, 1232, (1997).

[35] A. Givan, A. Loewenshuss and C. J. Nielson, Journal of Physical Chemistry
B, 101, 8696, (1997).

[36] C. Manca, C. Martin and P. Roubin, Chemical Physics Letters, 364, 220,
(2002).

[37] A. Hallbrucker, E. Mayer and G. P. Johari, Journal of Physical Chemistry,
93, 4986, (1989).

[38] G. P. Johari, A. Hallbrucker and E. Mayer, Nature, 330, 552, (1987).

[39] J. Dubochet and A. W. McDowall, Journal of Microscopy-Oxford, 124, RP3,
(1981).

[40] Y. P. Handa, O. Mishima and E. Whalley, Journal of Chemical Physics, 84,
2766, (1986).

[41] O. Mishima, L. D. Calvert and E. Whalley, Nature, 310, 393, (1984).

[42] O. Mishima, L. D. Calvert and E. Whalley, Nature, 314, 76, (1985).

[43] D. Laufer, E. Kochavi and A. Barnun, Physical Review B, 36, 9219, (1987).

[44] J. E. Schaff and J. T. Roberts, Journal of Physical Chemistry, 98, 6900,
(1994).

[45] P. U. Andersson, M. B. Nagard, G. Witt and J. B. C. Pettersson, Journal of
Physical Chemistry A, 108, 4627, (2004).

[46] M. E. Palumbo, Journal of Physical Chemistry A, 101, 4298, (1997).

[47] J. E. Bertie, L. D. Calvert and E. Whalley, Journal of Chemical Physics, 38,
840, (1963).

[48] R. S. Smith and B. D. Kay, Surface Review and Letters, 4, 781, (1997).

[49] K. L. Foster, M. A. Tolbert and S. M. George, Journal of Physical Chemistry
A, 101, 4979, (1997).

[50] J. D. Graham and J. T. Roberts, Journal of Physical Chemistry, 98, 5974,
(1994).
21

[51] T. Huthwelker, M. Ammann and T. Peter, Chemical Reviews, 106, 1375,
(2006).

[52] J. E. Schaff and J. T. Roberts, Journal of Physical Chemistry, 100, 14151,
(1996).

[53] S. C. Silva and J. P. Devlin, Journal of Physical Chemistry, 98, 10847,
(1994).

[54] T. Takoaka, M. Inamura, S. Yanagimachi, I. Kusunoki and T. Komeda,
Journal of Chemical Physics, 121, 4331, (2004).

[55] B. Rowland, M. Fisher and J. P. Devlin, Journal of Chemical Physics, 95,
1378, (1991).

[56] J. Sadlej, B. Rowland, J. P. Devlin and V. Buch, Journal of Chemical
Physics, 102, 4804, (1995).

[57] F. Borget, T. Chiavassa, A. Allouche and J. P. Aycard, Journal of Physical
Chemistry B, 105, 449, (2001).

[58] C. Martin, C. Manca and P. Roubin, Surface Science, 502, 275, (2002).

[59] G. J. Kroes and D. C. Clary, Journal of Physical Chemistry, 96, 7079, (1992).

[60] Y. A. Mantz, F. M. Geiger, L. T. Molina, M. J. Molina and B. L. Trout,
Chemical Physics Letters, 348, 285, (2001).

[61] M. Svanberg, J. B. C. Pettersson and K. Bolton, Journal of Physical
Chemistry A, 104, 5787, (2000).

[62] N. Uras, M. Rahman and J. P. Devlin, Journal of Physical Chemistry B, 102,
9375, (1998).

[63] R. L. Hudson and B. Donn, Icarus, 94, 326, (1991).

[64] A. Barnun, G. Herman, D. Laufer and M. L. Rappaport, Icarus, 63, 317,
(1985).

[65] A. Barnun, J. Dror, E. Kochavi, D. Laufer, D. Kovetz and T. Owen, Origins
of Life and Evolution of the Biosphere, 16, 220, (1986).

[66] S. A. Sandford and L. J. Allamandola, Icarus, 76, 201, (1988).
22

[67] S. A. Sandford and L. J. Allamandola, Astrophysical Journal, 355, 357,
(1990).

[68] A. Barnun, J. Dror, E. Kochavi and D. Laufer, Physical Review B, 35, 2427,
(1987).

[69] L. J. Allamandola, S. A. Sandford and G. J. Valero, Icarus, 76, 225, (1988).

[70] M. P. Collings, J. W. Dever, H. J. Fraser and M. R. S. McCoustra,
Astrophysics and Space Science, 285, 633, (2003).

[71] M. P. Collings, M. A. Anderson, R. Chen, J. W. Dever, S. Viti, D. A.
Williams and M. R. S. McCoustra, Monthly Notices of the Royal Astronomical
Society, 354, 1133, (2004).

[72] J. D. Graham, J. T. Roberts, L. A. Brown and V. Vaida, Journal of Physical
Chemistry, 100, 3115, (1996).

[73] J. E. Schaff and J. T. Roberts, Langmuir, 14, 1478, (1998).

[74] G. Notesco and A. Bar-Nun, Icarus, 148, 456, (2000).
23
Chapter 2: Fourier Transform Infrared Spectroscopy and Temperature
Programmed Desorption
2.1 Absorption Spectroscopy
Absorption spectroscopy, as its name suggests, is associated with the absorption
of radiation by a sample. The absorption of radiation promotes the constituents of a
sample (i.e., atoms, ions or molecules) from one internal state to another, and is a
process whereby radiant energy is transferred into the sample. The internal states of
atoms, ions, or molecules are quantized, or consist of discrete energy levels. One of
the stipulations for the efficient absorption of radiation is that the incident light be
resonant with an energy level separation within the absorbing entity. Because these
energy separations are unique for each absorbing entity, the constituents of a sample
can be characterized by determining the frequencies of radiation absorbed by a
sample.
Infrared (IR) absorption spectroscopy focuses on the absorption of IR radiation,
and it is restricted to the study of neutral or ionic molecules. This restriction arises
because only these species possess energy level separations that are resonant with IR
radiation; these separations can be traced to inter-atomic vibrations present in a
molecular species. Thus, IR absorption spectroscopy, which yields information about
the vibrational energy levels of a system, is commonly referred to as vibrational
spectroscopy.
The quantum mechanics of vibrational spectroscopy is covered in many
elementary spectroscopy texts [1-3] and it will not be discussed in any detail in this
24
dissertation. The usual approach involves first discussing the vibrations of diatomic
molecules and then detailing the vibrations of polyatomic molecules. The selection
rules, transition moments and spectral lineshapes associated with vibrational
spectroscopy are often presented as well.
Classically, the attenuation in irradiance as a wave traverses a medium is
explained by examining the time-averaged Poynting vector, S
v
, for time harmonic
fields [4]:

*
Re 5 . 0 H E S
v v v
⊗ = (2.1)
where E
v
and H
v
are the electric and magnetic vector fields, respectively.
For a plane wave [4],

) , (
0
) , (
t x i
e E t x E
v v
v
v
φ
= (2.2)

) , (
0
) , (
t x i
e H t x H
v v
v
v
φ
= (2.3)
Here, x
v
is the directional distance traveled from some specified origin, t is the time
interval relative to some reference time, ) , ( t x
v
φ is the phase of the wave, and
0
E
v
and
0
H
v
are vectors constant in time and space (commonly referred to as the amplitude
vectors). For such a wave

⎭
⎬
⎫
⎩
⎨
⎧
⎟
⎠
⎞
⎜
⎝
⎛
′ ′
− =
λ
π
μ
ε z n
E S
4
exp Re 5 . 0
2
0
v v
(2.4)
where ε is the complex permittivity, μ is the permeability, n ′ ′ is the imaginary
component of the complex refractive index, λ is the wavelength of the radiation,
25
and z is the optical path length traveled in a specified direction. The irradiance, I ,
which is equivalent to S
v
, is thus given by
) exp(
0
z I I α − = (2.5)
where the absorption coefficient is λ π α / 4 n ′ ′ ≡ . In principle, all processes that
remove radiant energy from a beam of radiation, whether by converting the energy
into some other energy form (e.g., absorption followed by non-radiative processes)
or changing the direction of energy flow (e.g., scattering), contribute to the
absorption coefficient.
The attenuation in irradiance when light passes through a sample is usually
quantified in terms of the transmittance or the absorbance of the sample. The
transmittance, ) ( ν T , of a sample at a specified frequency ν is defined as

) (
) (
) (
0
ν
ν
ν
I
I
T = (2.6)
where ) (
0
ν I and ) ( ν I are the irradiance of the incident and transmitted beam,
respectively. Although the terms intensity and irradiance are often used
interchangeably, this synonymous usage is confusing, and it is strictly incorrect;
spectral intensity (power per unit area per solid angle) possesses an angular
dependence while irradiance (power per unit area) does not [4]. The absorbance,
) ( ν A , of a sample is directly related to its transmittance, and it is defined as
)) ( log( ) ( ν ν T A − = (2.7)
An absorption spectrum is displayed typically by plotting absorbance versus
frequency or wavelength.
26
There are essentially two methods of obtaining an IR spectrum. One technique
involves measuring the absorbance of a sample at each individual wavelength
separately. To perform these measurements, radiation of suitable monochromaticity,
i.e., narrow bandwidth, is required, and thus this technique utilizes a dispersing
element to separate polychromatic IR radiation into its component wavelengths. An
IR spectrum can also be obtained without physically separating the electromagnetic
constituents of a polychromatic, or broadband, source. This technique utilizes a
mathematical technique, known as the Fourier transform, and a mechanical device,
known as an interferometer, to facilitate the measurement of a spectrum. As a result,
the latter method is referred to as Fourier transform infrared (FTIR) spectroscopy.

2.2 Fourier Transform Infrared Spectroscopy
The history, theory and practice of FTIR spectroscopy has been the subject of
many texts [5-7]. In fact, FTIR spectrometers have been commercially available for
almost half a century [8]. A thorough review of this technique necessarily includes
its numerous practical and theoretical issues; but such a review is intrinsically long,
and can be tedious to peruse. The goal in this chapter is by no means to present an
exhaustive analysis of FTIR spectroscopy. Instead, a basic introduction shall be
provided by summarizing some of the salient theoretical and practical aspects of this
method. The desired outcome is a coherent picture detailing the capabilities and
limitations of FTIR.

27
2.3 The Michelson Interferometer
While there are several different interferometer designs, they are all in effect
variations of the first interferometer constructed by Albert Michelson in 1891 [5, 7].
Consequently, a convenient approach to understanding how these devices work is to
first understand the operations of their progenitor, the Michelson interferometer. A
Michelson interferometer (Figure 2.1) consists of a beamsplitter and two mirrors.
Essentially, it is a device that splits a beam of radiation into two and then recombines
the two resultant beams after introducing a phase difference between them.

Fixed mirror
Moving
mirror
to sample
from the IR source

Figure 2.1. Radiation (shown in blue) from a broadband IR source is split by a
beamsplitter (red) and directed toward two mirrors. The radiation reflected from
these mirrors recombines at the beam splitter. The radiation reflected back to the
source during this recombination has been omitted for clarity. The moving mirror is
used to change the path difference between the two beams.

28
The effect of this beam splitting and re-combination is performed constructively
by considering the fate of a homogenous monochromatic plane wave incident on a
beamsplitter. As previously stated, the electric field, ) , ( t x E
v
v
, of a plane wave can be
represented classically by

) , (
0
) , (
t x i
e E t x E
v v
v
v
φ
= (2.8)
and the irradiance of this wave is proportional to
*
E E
v v
⋅ , i.e.,

*
E E K I
v v
⋅ = (2.9)
where K is the proportionality constant relating I and
*
E E
v v
⋅ . The phase of the
wave can be expressed as
ν π φ t x k t x 2 ) , ( − ⋅ =
v
v
v
(2.10)
where k
v
is the wave vector. For the sake of convenience, it is advantageous to
assume the beamsplitter divides the beam into two equal parts and the center of the
beamsplitter is the spatial origin. It is also advantageous to assume that the beam is
fully collimated. Under these assumptions, the two plane waves of equal irradiance
generated by the beamsplitter can be represented as

) , (
0
1
1
2
) , (
t x i
e
E
t x E
v
v
v
φ
= (2.11)

) , (
0
2
2
2
) , (
t x i
e
E
t x E
v
v
v
φ
= (2.12)
These are two waves traveling in different directions, each propagating towards one
of the mirrors (Figure 2.1) where they are reflected back to the beamsplitter. If the
29
path lengths traversed by wave ) , (
1
t x E
v
v
and ) , (
2
t x E
v
v
when they recombine at the
beamsplitter are
1
x and
2
x respectively, the resultant wave is
( )
) 2 ( ) 2 (
0
2 1
2 1
2
) , (
t kx i t kx i
e e
E
t x E
πν πν + +
+
+ =
v
v
(2.13)
The irradiance of this resultant wave is
)) cos( 1 (
2
) (
2
0
2 1
x k
E
K x I Δ + = Δ
+
(2.14)
where x Δ is the optical path difference,
2 1
x x − ,also known as the retardation. Using
the well–known relationship v k π 2 = , equation (2.14) transforms into
)) 2 cos( 1 (
2
) (
2
0
2 1
x v
E
K x I Δ + = Δ
+
π (2.15)
where v is the inverse wavelength. Since 1 ) 2 cos( 1 ≤ Δ ≤ − x v π , the irradiance varies
sinusoidally between zero and
2
0
5 . 0 E K as the path difference is changed. Half of
this irradiance is directed back to the source, and so the irradiance at the detector
oscillates between zero and
2
0
25 . 0 E K (Figure 2.2a). An extension of the above
analysis to encompass two plane waves of different wavelengths incident on the
beamsplitter is trivial, and the irradiance for this case is depicted in Figure 2.2 (b).
The overall effect of the interferometer is to modulate the source radiation.
Namely, the radiation at the detector is made to have a frequency much lower than
the frequency of the source. Prior to passing through the interferometer, 2000 cm
-1

radiation has a frequency of 6 × 10
13
Hz. As Figure 2.2 (a) shows, one cycle for this
radiation occurs at the detector when the phase difference is 5 × 10
-4
cm, i.e.,
30

v
x x
cycle
1
= Δ = Δ (2.16)

intensity
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
-2 0 2 4 -4
phase difference × 10
3
/ cm
(a)
(c)
(b)

Figure 2.2.The intensity at the detector oscillates as a result of interference. This is
shown in plots of ) 2 cos( 1 x vΔ + π versus phase difference for plane waves of light
with (a) one wavelength, 2000 cm
-1
, (b) two wavelengths, 2000 and 2200 cm
-1
, with
the same irradiance, and (c) two wavelengths, 2000 and 2200 cm
-1
, of different
irradiance (the 2000 cm
-1
wave has 0.6 the irradiance of the 2200 cm
-1
wave).

To accomplish this phase difference, the non-stationary mirror moves by

v
x
cycle
2
1
2
=
Δ
(2.17)
31
and the time period,
cycle
t Δ , for the moving mirror to travel this distance is

m
cycle
V v
t
2
1
mirror the of velocity
mirror by the traveled distance
= = Δ (2.18)
where
m
V is the mirror velocity. Thus, the frequency at the detector is
m
V v 2. For
2000 cm
-1
radiation and moving mirror velocities of 0.1-1 cm/s, detector signal
frequencies range from 400-4000 Hz. This modulation permits infrared transducers,
capable of following fluctuations in the audio frequency range (20 Hz-20 kHz), to be
used as detectors in FTIR spectrometers.
The magnitude of the electric field vector,
0
E , is typically different for each
frequency, and the irradiance at the detector for two plane waves of different
frequency and intensity is shown in Figure 2.2 (c). In general, the irradiance for each
wavelength (after passing through the interferometer) is
)) 2 cos( 1 )( ( ) , ( x v v B x v I Δ + = Δ π (2.19)
where save for ) 2 cos( 1 x vΔ + π , all terms in equation (2.15) have been grouped into a
frequency dependent factor. The last equation conveys simply the fact that the
irradiance between wavenumbers v and v d v + is v d x v I ) , ( Δ , and thus the total
irradiance at the detector is
v d x v v d v B v d v B x I
v v
) 2 cos( ) ( ) ( ) (
0 0
Δ + = Δ
∫ ∫
∞
=
∞
=
π (2.20)
This implies
v d x v v d v B x K v d v B x I
v v
) 2 cos( ) ( ) ( ) ( ) (
0 0
Δ = Δ = − Δ
∫ ∫
∞
=
∞
=
π (2.21)
32
The phase difference independent term, v d v B
v
∫
∞
=0
) ( , is the irradiance at zero path
difference, and its experimental determination is relatively straight forward. At
various phase differences, ) ( x K Δ can thus be evaluated by measuring ) ( x I Δ and
then subtracting v d v B
v
∫
∞
=0
) ( .

2.4 The Fourier Transform
The Michelson interferometer produces a detector signal dependent upon the
phase difference, and the critical step is converting this signal into one displaying the
irradiance, specifically ) (v B , at each frequency. In a holistic sense, the Fourier
transform can be described as a mathematical technique that converts a function
) (x F into some other function ) (y G ; for this transformation to be possible, the
variable x must have a dimension that is the inverse of the dimension of y . Two
symmetric functions, ) (x F and ) (y G are said to be a Fourier pair if [6]

∫
∞
−∞ =
=
v
dy xy y G x F ) 2 cos( ) ( ) ( π (2.22)
and

∫
∞
−∞ =
=
v
dx xy x F y G ) 2 cos( ) ( ) ( π (2.23)
As equation (2.21) shows, ) ( x K Δ is a symmetric function and thus

∫
∞
−∞ =
Δ = Δ = Δ
v
v d x v v B x K x L ) 2 cos( ) ( ) ( 2 ) ( π (2.24)
33
Furthermore, because x Δ and v do indeed have inverse dimensionality, ) (v B can be
written as

∫ ∫
∞
=
∞
−∞ =
Δ Δ Δ = Δ Δ Δ =
0
) 2 cos( ) ( 2 ) 2 cos( ) ( ) (
v v
x d x v x L x d x v x L v B π π (2.25)
Equation (2.20) shows that, in theory, ) (v B can be calculated once ) ( x K Δ (see
equation (2.21)) is measured over ∞ ≤ Δ ≤ x 0 .

2.5 Practical Aspects of FTIR Spectroscopy
The ideas presented up until this point have been primarily conceptual, and they
convey the underlying principles of FTIR spectroscopy. However, the practical
implementation of this theory is riddled with many challenges. Firstly, measuring the
irradiance to an optical path length of infinity is practically impossible. Placing
restrictions on the maximum retardation,
max
x Δ is equivalent to multiplying the
complete interferogram (extending from ∞ ≤ Δ ≤ ∞ − x ) by a truncation function
) ( x D Δ which is unity for
max max
x x x Δ ≤ Δ ≤ Δ − and zero elsewhere; this is the
boxcar function. The spectrum obtained in this situation is

∫∫
∞
∞ =
Δ
Δ −
Δ Δ Δ = Δ Δ Δ Δ = ′
v
x
x
x d x v x C x d x v x D x L v B
max
max
) 2 cos( ) ( ) 2 cos( ) ( ) ( ) ( π π (2.26)
The Fourier transform of ) ( x C Δ , which is ) (v B′ , is equal to the convolution of the
Fourier transforms of ) ( x L Δ and ) ( x D Δ ; this is a result of convolution theory [9,
10]. Basically, the measurement of a spectrum with a spectrometer of finite
resolution distorts the true spectrum, ) (v B .
34
The extent of this distortion depends upon the characteristics of the true spectrum
and the convoluting function. The Fourier transform of ) ( x L Δ is ) (v B , and because
) ( x D Δ is a boxcar function [10, 11] its Fourier transform, ) (v f , is the sinc(x)
function

max
max
max
2
) 2 sin(
2 ) (
x v
x v
v f
Δ
Δ
Δ =
π
π
(2.27)

From its maximum at 0 = v , ) (v f first becomes zero at
1
max
5 . 0
−
Δx and the width of
this function can be arbitrarily defined as
1
max
−
Δx (Figure 2.3). The interval between
the maxima of two sinc functions when they are placed as close as possible without
any overlap of their main feature (centered at 0 = v ) is
1
max
−
Δx . Specifically, two
such features are sufficiently resolved if they are separated by
1
max
−
Δx , and this value
is often defined as the nominal resolution of a FTIR spectrometer. The convolution
of ) (v B and ) (v f is
v d v v f v B v B ′ ′ − ′ = ′
∫
∞
∞ −
) ( ) ( )( (2.28)
For a ) (v B function that is has line width much narrower (i.e., over 3 times
narrower) than ) (v f , the effect of this convolution is to impart the ) (v f lineshape
to the measured spectrum. For example, the spectrum of a monochromatic line will
take on a sinc(x) lineshape. Essentially, the resolution of the spectrum is restricted to

35
a value of
1
max
−
Δx because of the convoluting boxcar function, and

1
max
−
Δ = Δ x v
res
(2.29)

2 Δx
max
1
+ Δx
max
- Δx
max
D(x)
v
(b) (a)
(v)
1
v
1
Δx
max
(c)
Δx
max
2
v
1
-
Δx
max
2
f
v
v
1
+

Figure 2.3. The Fourier transform of a boxcar function (shown in (a)) is the sinc
function (shown in (b)), and this function has several side lobes adjacent to its main
feature. (c) The spectrum of a monochromatic line takes on a sinc(x) lineshape as a
result of the inability to measure the interference signal over ∞ ≤ Δ ≤ ∞ − x . The
width of the feature obtained is shown.

Indeed, FTIR spectrometer software is configured such that a user specified nominal
resolution controls the maximum distance traveled by the moving mirror during each
scan.
36
The sinc(x) function is not particularly useful because of fairly large amplitudes at
wavenumbers well away from the central feature (Figure 2.3). These side lobes can
be mistaken for the main feature of a weak transition. The side lobes of a sinc(x)
function were once referred to as feet or podes, and as a result the process of
removing these features is known as apodization. Basically, apodization involves
substituting the sinc(x) convolution function with a simple weighting function to
reduce the side lobes. For example, a triangular apodization function [5], which is a
commonly used apodization function, distorts narrow absorption lines and imparts a
sinc
2
(x) lineshape to these lines. At a peak separation value of
1
max
−
Δx there is
significant overlap between two sinc
2
(x) functions (as opposed to no overlap in the
main feature of two sinc(x) functions separated by the same value). However, the
side lobes are minimized, and the two peaks are still easily discernible (Figure 2.4).
Several other apodization functions exist [12], and they are utilized routinely in
FTIR spectroscopy. An ideal apodization function would yield an infinitely narrow
function upon Fourier transformation. In general, as the main lope of the convoluting
function gets narrower, the ratio of the magnitude of the strongest side lobe to the
central lobe increases. Consequently, the choice of which apodization function to use
depends upon the experiment being performed, and some of the necessary
considerations are detailed by Griffiths and deHaseth [5].
Another issue of concern in the implementation of FTIR spectroscopy is how to
sample the signal reaching the detector. Specifically, the question involves how
regularly a signal must be sampled to detect all the frequencies present in the signal
37
.
1
- Δx
max
+ Δx
max
Δx
max
D(x)
(v)
2
Δx
max
v
1
Δx
max
(c)
(b) (a)
f
v

Figure 2.4. A triangular apodization function (shown in (a)) Fourier transforms to
yield a sinc
2
function (shown in (b)), which has side lobes that are small compared to
the main feature. Although there is significant overlap when two sinc
2
functions are
separated by the distance shown in (c), these two functions (blue) are still
distinguishable in the combined trace (red). At this separation, two sinc functions
would display no overlap in their main feature.

According to the sampling theorem [11], discrete sampling at a frequency of twice
the highest frequency present in a signal is necessary and sufficient to meet the
aforementioned objective. Consider a source with a maximum wavenumber
max
v .
The retardation distance for one irradiance cycle for a plane of wavenumber
max
v at
the detector is
1
max
−
v (see equation (2.16)), and this is the shortest retardation distance
per cycle amongst the various frequencies from the source. Thus, the minimum
38
sampling rate is a point every
1
max
) 2 (
−
v ; this is akin to specifying at least two data
points per the cycle for radiation of
max
v .
The number of points,
s
N , collected during a scan depends upon the nominal
resolution specified and the maximum wavelength of the source. This number of
points is equal to the maximum retardation divided by the sampling frequency. For a
resolution of
res
v Δ and a source with a maximum wavenumber,
max
v ,
s
N is given by

res
res
s
v
v
v
v
N
Δ
=
Δ
=
−
−
max
1
max
1
2
) 2 (
) (
(2.30)
It is important to make the distinction between the number of data points per
retardation and the number of data points per unit time. The former is a point every
1
max
) 2 (
−
v while the latter is equal to how fast the moving mirror travels a distance of
1
max
) 2 ( 5 . 0
−
v . For a moving mirror velocity of
m
V , this distance is traveled in
1
max
) 4 (
−
v V
m
.
The time taken to collect a scan at a specific mirror velocity is dictated by the
desired resolution. As previously mentioned, the moving mirror must travel
1
) (
−
Δ
res
v
for a nominal resolution of
res
v Δ . Consequently, regardless of the number of data
points obtained, the collection time at a specified nominal resolution is the same for
the same mirror velocity. The slight advantage in minimizing the number of data
points comes in the reduction of computer time for data processing. At typical mirror
velocities of 0.1-5 cm/s and a nominal resolution of 4 cm
-1
, a scan is obtained in
0.05-2.5 seconds.
39
A single scan often possesses too much noise to be useful, and therefore several
scans are collected and averaged to produce an interferogram with a suitable signal-
to-noise ratio (SNR). The SNR increase, over a single scan, is equal to the square
root of the number of scans averaged [13]. However, even when single scans possess
an adequate SNR, FTIR spectroscopy is limited in the type of dynamic events it can
follow in real time. In principle, a dynamic event with a half-life longer than 500 ms
can be characterized by multiple time-resolved spectra [7].
A major limitation in decreasing the scan time comes from the manner in which
the moving mirror is translated during data collection. This drive mechanism is
restricted to maximum mirror velocities of ~5 cm/s, and the technical difficulties
associated with modifying this system are, to say the least, challenging [7].
Nevertheless, several new FTIR spectroscopy methods increasing the rate of data
acquisition already exist [7].
Up until this point, it has been assumed that the interferogram is symmetric with
respect to retardation. In practice, instrumental artifacts often lead to an asymmetric
interferogram [5, 7]. As a result, the moving mirror must travel from
1
) (
−
Δ −
res
v to
1
) (
−
Δ
res
v to complete a scan at a specified nominal resolution of ) (
res
v Δ . The process
of removing these artifacts to obtain the true symmetric spectrum is known as phase
correction [5].
As discussed above, the intensity of the radiation incident upon the detector must
be sampled in a certain manner to obtain an accurate spectrum. During each
measurement, the detector converts this intensity into an electrical signal, which is
40
then processed, digitized and stored (along with the corresponding phase difference
value).This digitization process is performed with an analog-to-digital converter
(ADC). At the end of the scan, this digitized data, which is in fact the digitized
interferogram, is then transformed into a spectrum.
The dynamic range of an ADC is a measure of the difference between the largest
and smallest signals that can be digitized [5]. For example, a 16-bit ADC can be
thought of as possessing 2
16
levels, and therefore it can be described as having a
dynamic range of 0-2
16
. To digitize a signal ranging from 0-10 V, the difference in
magnitude between each level can be assigned a value of 10/2
16
V or ~0.1 mV. In
this case, any signal less than 0.1 mV cannot be measured, and it is assumed to be
zero.
Interferograms from a broadband source are intrinsically signals with a
centerburst irradiance that is high compared to regions of the interferogram well
displaced from zero retardation. For suitable digitization accuracy, the irradiance of
any signal should fall within the dynamic range of the digitizer. Thus, the software
for most spectrometers will only permit data collection when irradiance at the
detector for zero retardation is less than some maximum value, i.e., the upper limit of
the instrument's dynamic range. To limit the irradiance at the detector, screens or
grids are placed in the beam path.
Real information can sometimes be excluded during the digitization if the noise
level in the analog signal being processed falls outside the dynamic range of the
ADC [5]. Hence, the dynamic range of many spectrometers is configured in a
41
manner that affords digitization of the noise. In fact, the co-addition of spectra only
improves the SNR if the noise is large enough to be measured, i.e., the magnitude of
the noise falls within the dynamic range of the spectrometer.
The transformation of the discrete interferogram into a spectrum was, to a certain
degree, time-consuming prior to 1970. In fact, with a large amount of data points,
this transformation became prohibitively time-consuming. However, since the
publication of the Cooley-Tukey Fast Fourier Transform algorithm in 1965, this
issue has been suitably addressed with the implementation of software programs that
perform the discrete Fourier transform relatively quickly [5].
The improved performance of FTIR spectrometers over dispersive instruments is
attributed to the multiplex (or Fellgett) advantage and the throughput (or Jacquinot
advantage) [5]. For each data point, the sum of the irradiance of light reaching the
detector for all wavelengths is recorded (equation 2.20). Thus, at each data point
some information is gained about all the wavelengths. This is in contrast to data
collection with dispersive spectrometer, where each wavelength is determined
separately. The amount of noise at a specific wavelength is related to the time spent
observing that wavelength, i.e., collecting information about that wavelength. As a
result, with the same experimental parameters (e.g., collection time, source,
detector), a spectrum from a FTIR spectrometer will have a better SNR than a
dispersive spectrometer; this is called the multiplex advantage. The throughput
advantage arises from the fact that the optical throughput of an interferometer and a
dispersive spectrometer are different for the same beam of light. The former has a
42
better throughput, which is a measure of how much the source irradiance is
decreased by passing through a spectrometer.

2.6 Temperature Programmed Desorption
As an atom or molecule approaches a surface, it experiences an energetic
interaction with the surface constituents much like the interaction two gas phase
atoms experience as they approach each other. However, in this case, the incoming
particle interacts simultaneously with several surface or near-surface constituents.
There are a number of ways to model this event. A simplistic one-dimensional
potential scheme, in which the interaction energy depends solely upon the particle's
distance from the surface, is the model most widely invoked [14]. In this scheme, a
molecule or atom attaches to the surface if it approaches the surface, loses some of
its incident energy, and finds itself without enough energy to desorb from, or leave,
the surface. In this attached state, the particle is said to be adsorbed on the surface.
Temperature programmed desorption (TPD) is an experimental technique in
which the energy required by an adsorbed species to desorb is supplied by heating
the surface, and this desorption is monitored by a mass spectrometer. TPD data is
presented typically by plotting the mass detector signal, which is proportional to the
desorption flux, versus the surface temperature. These traces are referred to as TPD
spectra, and one of the primary uses of this technique is to estimate of the energetic
barrier for desorption from these traces.
43
The desorption of a species that remains intact upon adsorption can be
represented by the equation [14]

) ( ) ( g ads
A A ↔ (2.31)
where
) (ads
A and
) (g
A denote species A localized on a surface and in the gas phase,
respectively. The desorption flux,
lux f
R , in this case is given by

n
ads d
ads
flux
A k
dt
A d
R ] [
] [
)
)
(
(
= = (2.32)
where n is the desorption order, ] [
) (ads
A is the surface coverage of
) (ads
A with units
of number of species A per unit area, and
d
k is rate constant for the forward
reaction of equation 2.31. The desorption order, much like the order of a reaction,
cannot be deduced from a stoichiometrically balanced equation, and it is a value that
must be determined empirically. Generally, desorption is assumed to be a simple
activated process, and thus the rate constant for desorption is [14]

kT
d
E
d
e k
−
=ν (2.33)
where v is a pre-exponential factor for
d
k ,
d
E is the desorption energy, k is the
Boltzmann constant and T is the surface temperature. Combining equations 2.32 and
2.33 yields

kT
d
E
n
ads lux
e A R
f
−
= ] [
) (
ν (2.34)
As equation 2.34 shows, the desorption flux (and hence the form of the desorption
trace) depends on several factors;
d
E ,ν , and surface coverage.
44
There are several approaches to analyzing TPD spectra [15-18], and all of these
are attempts to extract information about relative surface coverage, desorption
energy, and the desorption order. By far the most widely used is the Redhead
analysis. The inherent simplicity of this technique is a major reason for its
widespread use. However, this simplicity is ultimately its undoing, and several more
sophisticated techniques exist to address the shortcomings of the Redhead method.
While the rate constant increases with increasing temperature (equation 2.34), the
number of species on the surface ultimately goes to zero when all species have
desorbed. This leads to desorption peaks in TPD spectra, with the peak maximum
depending upon the concentration of species originally on the surface. For example,
the TPD feature for a monolayer (possessing a desorption energy independent of
coverage) saturates when a full monolayer is present during the thermally induced
desorption.
TPD spectra with well-defined features that can be correlated with molecular, or
atomic, monolayer desorption provide an in-situ method of calibrating surface
coverage. The integrated area of a TPD spectral feature is proportional to the number
of desorbing species (associated with this feature). Thus, the integrated area of a
saturated monolayer feature is a convenient measure of the number of species within
a monolayer. The integrated areas of other TPD features can then be compared to
that of the monolayer, and the total coverage is often defined in terms of monolayers.
In practice, the flux within the mass sensor, which determines the intensity of a
mass signal, is actually not the desorption flux at the surface. The local flux at a
45
point away from the surface depends upon the trajectories of desorbing species.
Consequently, for the same surface coverage the measured flux often varies with the
position of the surface relative to the mass sensor. It is difficult, if not impossible, to
prove that the flux within the mass sensor is effectively the desorption flux at the
surface. Thus, most experiments are conducted with the surface at some constant
distance from mass detector, i.e., a constant distance for all traces. In this
experimental configuration, the measured signal is simply proportional to the actual
desorption flux and the proportionality constant relating this two quantities is the
same for all spectra.
46
2.7 Chapter 2 References
[1] W. S. Struve, Fundamentals of Molecular Spectroscopy, John Wiley and
Sons, Inc.: New York, 1989.

[2] M. J. Hollas, High Resolution Spectroscopy, John Wiley & Sons, Inc.:
Chichester, 1998.

[3] P. Atkins and R. Freidman, Molecular Quantum Mechanics, Oxford
University Press: Oxford, 1998.

[4] C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by
Small Particles, John Wiley & Sons, Inc.: New York, 1998.

[5] P. Griffiths and J. A. De Haseth, Fourier Transform Infrared Spectrometry,
John Wiley and Sons, Inc.: New York, 1986.

[6] B. C. Smith, Fourier Transform Infrared Spectroscopy, CRC Press: Boca
Raton, 1996.

[7] G. D. Smith and R. A. Palmer, In Handbook of Vibrational Spectroscopy, J.
M. Chalmers and P. R. Griffiths, Eds.; John Wiley and Sons, Inc., New York, 2002,
Vol. 1.

[8] L. Mertz, Astronomical Journal, 70, 685, (1965).

[9] G. B. Arfken and H. J. Weber, Mathematical Methods for Physicists,
Academic Press: San Deigo, 1995.

[10] B. Kusse and E. Westwig, Mathematical Physics, John Wiley and Sons, Inc.:
New York, 1998.

[11] J. F. James, A student's guide to Fourier transforms, Cambridge University
Press: Cambridge, 1998.

[12] R. H. Norton and R. Beer, Journal of the Optical Society of America, 66, 259,
(1976).

[13] D. A. Skoog and J. L. Leary, Principles of Instrumental Analysis, Harcourt
Brace College Publishers: Fort Worth, 1992.

[14] J. C. Tully, Surface Science, 299/300, 667, (1994).

[15] P. A. Redhead, Vacuum, 12, 203, (1962).
47
[16] D. A. King, Surface Science, 47, 384, (1975).

[17] A. M. Jong and J. W. Niemantsverdriet, Vacuum, 41, 232, (1990).

[18] C. M. Chan, R. Aris and W. H. Weinberg, Surface Science, 1, 360, (1978).

48
Chapter 3: Experimental Apparatus
3.1 Introduction
Investigations into the nature of molecular interactions with ice are often
undertaken using infrared (IR) spectroscopy. For such work, careful consideration
must be given to the optical properties of the materials in the IR beam path as it
travels from the source to the detector. Depending upon the substrate, IR
spectroscopy can be performed in several different modes, e.g., transmission,
reflection, and attenuated total reflectance. Naturally, each mode has its advantages
and limitations. With respect to ice studies, transmission and reflection absorption IR
spectroscopy are both viable experimental techniques, and they have been used with
varying degrees of success.
The work detailed in this dissertation was performed in a chamber designed and
built to house a host of diagnostic equipment. A thorough description of this
chamber can be found elsewhere [1], and only a brief description will be presented
here. An emphasis will be placed on the physical aspects of the chamber that relate
specifically to the two diagnostic techniques employed for the experiments detailed
in this dissertation: transmission Fourier transform IR (FTIR) spectroscopy and
temperature programmed desorption (TPD)

3.2 The UHV system
The ultra high vacuum (UHV) system is a three-tiered chamber, with a surface
manipulator affixed to the topmost tier (Figure 3.1). The two tiers at the bottom level
49
of the chamber are designed to permit molecular beam scattering experiments and to
accommodate instrumentation commonly required for surface analysis (e.g., Low
Energy Electron Diffraction, High Resolution Electron Energy Loss Spectroscopy,
and Auger Electron Spectroscopy). The numerous ports at these levels afford a
degree of versatility with respect to different experimental arrangements, and these
ports also allow the chamber to house several different diagnostics simultaneously.
Using a turbomolecular pump (TurboVac 361), pressures of ~2 × 10
-10
Torr are
attained routinely after heating the chamber at temperatures of ~400 K for 2-3 days.
During this heating process, which is commonly referred to as a chamber bakeout,
regions of the chamber at significantly lower temperatures than other areas of the
chamber can develop. At these regions, known as cold spots, the desorption rate is
relatively small compared to other areas of the chamber. These cold spots often
result in a virtual leak, which is the slow release of species from an area within the
UHV system interior. To prevent cold spots from occurring during bakeout, the
temperatures at different parts of the chamber are monitored periodically using
various thermocouples attached to the chamber, and this ensures a relatively even
heating process. Resistive heating tapes, permanently attached to the chamber, are
used to heat the chamber, and the heat dissipation from each of these tapes is
controlled via variable transformers (Variac).

50
gate valve
Tier 2
Tier 1
Tier 3
surface manipulator
z
y
x

Figure 3.1. A schematic of the UHV Chamber. The numerous ports on tiers 1 and 2
have been omitted for clarity. The entire UHV system is evacuated by a
turbomolecular pump located on the second tier.

51
3.3 The Top Tier: the FTIR Chamber
The top tier of the UHV system was configured specifically for IR experiments,
and it is hereafter referred to as the FTIR chamber. It is connected to the levels below
it via a gate valve (MDC, GV-4000M), and this permits the FTIR chamber to be
vented (i.e., brought up to atmospheric pressures) without necessarily having to vent
the entire chamber. After such a venting procedure, it would be desirable to reduce
the pressure differential across the aforementioned gate valve, by pumping out the
FTIR chamber, prior to opening this gate valve. However, under the existing
experimental configuration, the FTIR chamber is evacuated by a turbomolecular
pump that evacuates the entire chamber, and this pump is located on the second tier
of the chamber. Consequently, the entire system was vented even when changes
were made only in the top tier of the chamber.
As Figure 3.2 shows, two calcium flouride (CaF
2
) windows allow IR radiation to
enter and exit the FTIR chamber. The experimental set-up used for the IR and
thermal desorption experiments discussed in the ensuing chapters was one in which
leak valves and a mass spectrometer, a SRS residual gas analyzer (RGA 300), were
attached to the chamber at the FTIR level (Figure 3.2). This obviated the need to
move between different tiers, as all the equipment necessary for each experiment was
on the same level.

52
nose cone
CaF
2
windows
Leak valves
mass
spectrometer
nose cone
CaF
2
windows
Leak valves
mass
spectrometer
x
y

Figure 3.2. A schematic of the FTIR chamber. The axis system shown relates the
orientation of this view to the view depicted in Figure 1.

3.4 FTIR Optics and Set-up
To record the IR spectrum of a sample located within the chamber, the collimated
beam of a commercial FTIR spectrometer (Nicolet Protégé 740) was directed into
the chamber (Figure 3.3). The IR source was a resistively heated silicon nitride
filament enclosed in a tubular cavity with a square opening of ~5 mm. This device is
known as a glowbar, and its spectral output approximates that of a blackbody [2].
The source was situated at the focal point of a mirror with a 87.07 mm focal length.
The collimated beam leaving this mirror entered an interferometer, as shown in
Figure 3.3. Two mirrors, a mirror with a focal length of 152.40 mm and a flat mirror,
directed the beam exiting the interferometer into the chamber. As a result of this
53
experimental configuration, the beam was focused to ~9 mm (i.e., 5.1 × 152.40/87.07
mm) within the chamber.

source
interferometer
beamsplitter
sample focusing mirror
plane mirror
InSb
detector
substrate
Nicolet spectrometer

Figure 3.3. A schematic showing the IR beam path from the source to the detector.
All pertinent components outside the UHV chamber and the spectrometer are placed
in purge boxes (shown in green).

54
The size of the beam at the sample could not be reduced significantly by simply
replacing the mirror focusing the beam into the chamber (i.e, the sample focusing
mirror) with a mirror having a shorter focal length. This is because the manipulator
allowed the sample to be moved a maximum distance of 12.5 mm from the center of
the FTIR chamber when translating the sample along a direction perpendicular to the
face of the two CaF
2
windows (Figure 3.3). Hence, the beam could not be focused at
the sample if a focusing mirror with a focus length less than 139.9 mm (i.e, (157.4-
12.5) mm) was used, and this placed a limit on the beam size at the sample (~5.1 ×
139.90/87.07 mm). Instead, the beam size at the sample could be reduced by placing
an aperture between the sample focusing mirror and the plane mirror (Figure 3.3).
However, an aperture also reduces the intensity of the incident radiation at the
sample, and so it was not used in any of the work reported herein.
An IR detector, with a indium antimonide (InSb) element (~2 mm × 2 mm),
converted the intensity of the IR radiation exiting the chamber into an electrical
signal. To reduce noise from thermally induced transitions among close lying energy
levels of this element, it was cooled with liquid nitrogen. This liquid nitrogen was
located in a dewar that was in thermal contact with the InSb element; a dewar is a
container consisting of two walls with an evacuated space in between them to reduce
thermal energy loss. A filled dewar remained at liquid nitrogen temperatures for ~20
hours. As with all photoconductive detectors, the absorption of radiation by the InSb
element changes its conductivity, and this change is measured by an electrical
circuit.
55
CO
2
and H
2
O have relatively strong absorptions in the mid-IR region (4000-2000
cm
-1
), and FTIR spectrometers are often evacuated or purged to reduce the
concentrations of these species in the beam path. To this end, all optics outside the
spectrometer were placed in two Plexiglas boxes (Figure 3.3), and both the
spectrometer and these boxes were purged with air filtered by a Whatman (FTIR 75-
62) purge gas generator. An in-house air compressor supplied this generator with
unfiltered air.
A wire grid polarizer (Molectron, 93-98 % purity) placed in the IR beam path
produced polarized IR radiation. This polarizer, which consists of a regular array of
fine parallel aluminum wires on a barium fluoride (BaF
2
) support, only transmits
radiation of a certain polarization. Hence, there is a marked decrease in beam
intensity after it passes through the polarizer. The exact mechanism that allows such
an arrangement of wires to only transmit radiation with an electric field parallel to
them is currently unclear [3].
The intensity of the IR source tends to decrease with time, and this can result in
insufficient radiation at the detector. More importantly, an aging source may begin to
exhibit significant intensity fluctuations from scan to scan. An average of such scans
results in noise levels that are higher than the noise levels in an average, of the same
number of scans, from a more stable source. The peak-to-peak absorbance (Figure
3.4) of a spectrum in a region that does not coincide with any known spectral feature
is a reasonable measure of noise [4]. Using the experimental set-up described above,
a peak-to-peak absorbance of ~2 ×10
-4
was obtained routinely at ~2700 cm
-1
, and this
56
corresponds to a transmittance of 0.9998. The spectral feature of any species was
easily discernible as long as it was ~5 times larger than this typical noise level.

2750 2740 2730 2720 2710 2700
0.0001
wavenumber / cm
-1
absorbance

Figure 3.4. The peak-to-peak absorbance routinely obtained in the ~2700 cm
-1

region with the experimental configuration described in this chapter.

3.5 The Surface Manipulator
A sample manipulator was used to control the sample position in the chamber,
and this manipulator was a modified version of a manipulator designed and built by
K. J. Lesker from Vacuum Generator parts [5]. In brief, the manipulator has a
stainless steel tube that extends into the three-tiered vacuum chamber (Figure 3.5),
57
and this tube can be translated along three mutually perpendicular axes. Movement
along one of these axes simply permits the end of the tube to move between the
different tiers of the chamber, and the maximum distance the tube can be moved in
this direction is ~600 mm. A copper piece is brazed onto one end this tube while the
other end is open (Figure 3.5). Filling the tube with liquid nitrogen through its open
end, cools the tube and the copper piece attached to it. The manipulator also permits
the sample to be rotated by 360°.

3.6 The Substrate
The ice films studied were grown on a MgO(100) single crystal (~1 mm × 10 mm
× 10 mm). In principle, any substrate that transmits light in the spectral region of
interest (4000-2000 cm
-1
) could have been used as an ice film support. The
MgO(100) substrate was chosen because it could be prepared easily, and cleaning it
in-situ was relatively simple. This substrate was prepared by cleaving a MgO crystal
(with a thickness greater than 5 mm) twice to expose two fresh (100) crystal faces.
This procedure was performed in a Plexiglas box purged by nitrogen gas. The
substrate was inserted into the chamber, after attaching thermocouple wires to the
substrate and placing it in a copper sample holder.
Thermocouple wires, for a k-type thermocouple, were adhered to the edge of the
crystal face with a ceramic adhesive (Aremco Products 835M). Sometimes these
wires detached unexpectedly from the substrate when the substrate was under
58
vacuum. To limit this occurrence, several ceramic adhesives were evaluated to
determine which was the most suitable for binding thermocouple wires to a MgO
surface.

manipulator
tube
manipulator
tube

Figure 3.5. The surface manipulator that is attached to the UHV system. The
stainless steel tube of the manipulator and the copper piece attached to this tube are
shown explicitly in the inset.
59
3.7 The Sample Holder
The sample holder, which affixed the substrate to the manipulator, had to meet
certain requirements. In order to form amorphous ice by vapor deposition, the
substrate temperature had to be less than 120 K. Moreover, preliminary experiments
and available literature suggested substrate temperatures of 100 K or less were
needed to physisorb CO
2
and N
2
O on a cold substrate. In addition to this cooling
requirement, the sample holder had to be designed in a manner that allowed radiation
to be transmitted through the sample. This latter specification implied the substrate
could not be cooled by simply placing one of its surfaces on a cold metal surface.
Instead, the substrate temperature had to be controlled by heating or cooling the
edges of the crystal. The design and construction of a sample holder that confirmed
to these requirements was an evolutionary process that involved a systematic
improvement of several designs. Those various designs have been detailed already
[1], and the specific substrate holder used for the experiments reported in this
dissertation is shown in Figure 3.6.
Two copper parts (labeled A and B in Figure 3.6), separated by a ceramic spacer,
were attached to the copper piece at one end of the surface manipulator's stainless
steel tube (Figure 3.5). The three screws that affixed these two parts to the
aforementioned copper piece on the surface manipulator tube were electrically
insulated from all of these three parts. This was accomplished using ceramic pieces
known as ceramic hat washers (McAllister Technical Services). The two pieces were
60
also separated physically from the manipulator by a flat (~1 mm) spherical (~ 25 mm
in diameter) sapphire disc (Esco Products).

B
A
Pair of copper
plates
Tapped holes for
attaching heating leads

Figure 3.6. A drawing of the surface holder used to attach the surface to the
manipulator. The two main copper parts of the holder are labeled A and B. One of
the two copper plates that are used to sandwich the surface is shown.

The thermal conductivity of sapphire decreases with temperature. For example, it
conducts ~10 W cm
-1
K
-1
at 80 K, and this value decreases to ~0.3 W cm
-1
K
-1
at
61
400 K [6]. Thus, the sapphire disc separating the manipulator from the sample holder
ensured that very little heat was transferred from the holder to the manipulator during
substrate annealing. In turn, this meant that any species accumulated on the surface
of the cold manipulator rod, during vapor deposition onto the substrate, did not
desorb during annealing. However, the thermal conductivity of the sapphire disc at
~80 K was enough to permit substrate temperatures of ~90 K when helium gas was
bubbled through the liquid nitrogen cooling tube. Bubbling helium through a dewar
of liquid nitrogen is a well-known technique used to decrease the temperature in a
dewar [7]. In addition to acting as a thermal switch, the sapphire disc electrically
insulated the copper sample holder from the manipulator.
The MgO substrate was attached to the copper sample holder by sandwiching it in
between two pairs of copper plates (Figure 3.6). The two previously mentioned
copper parts of the sample holder that were attached to the manipulator each extend
into a flat plate (Figure 3.6), and each of these plates is paired with another copper
plate. Stainless steel screws aid each copper plate pair clamp, or sandwich, the
substrate at different edges, and together these plates hold the substrate rigidly in
place.
Two narrow (diameter ~0.25 mm) tantalum wires placed on one surface of the
substrate were used to increase the substrate temperature. These wires were the only
electrical connection between the two copper pieces of the sample holder, (i.e., the
parts labeled A and B in Figure 3.6). Two leads connected to an external power
supply were attached to each of these two sample holder parts, and because of the
62
high resistivity exhibited by tantalum (compared to copper), essentially all of the
resistance in this electrical circuit came from the wires. Hence, these wires generated
enough resistive heat to increase the sample temperature from ~90 K to ~400 K, at a
rate of ~2 Ks
-1
, when a current of ~18 A was passed through the circuit.

3.8 The Sample Holder: the Second Design
The sample holder design described above could not be used for thermal
experiments. To limit the species entering the mass spectrometer detection region to
those desorbing from the surface, the mass spectrometer was fitted with a stainless
steel "nose cone" with a small (~8 mm in diameter) aperture (Figure 3.2). During
thermal desorption experiments, the substrate was heated while it was directly in
front (~1 mm) of the mass spectrometer; the substrate face was roughly parallel to
the nose cone aperture. In addition, a fine (wire diameter ~0.025 mm) stainless steel
mesh (openings ~1.3 mm) was placed over the aperture of the cone to prevent
electrons, in the mass spectrometer ionization region, from reaching the surface. This
mesh was attached by soldering it onto the cone, which was grounded. Even with the
nose cone present, the close proximity of the copper plates to the sample made it
impossible to prevent species desorbing from these plates from entering the mass
spectrometer during thermal desorption. Hence, a new surface holder was
constructed, and this holder, hereafter referred to as a sample holder 2, permitted the
substrate to be cooled to ~90 K and allowed transmission IR and TPD experiments to
be performed with the same substrate (Figure 3.7).
63

A
B
C
Copper foil with two
edges that fold over
the crystal

Figure 3.7. A drawing of sample holder 2; the copper foil is attached to the two main
parts of the holder (A and B) using a screw (inserted into the tapped hole labeled C).

In this design, the MgO crystal was centered on a flat piece of copper foil (~0.2
mm × 10 mm × 14 mm), and this piece of foil had a protruding arm extending from
one of its edges and a hole (~6 mm in diameter) at its center (Figure 3.7). Two
64
opposite edges of the foil were folded over the crystal to hold it in place. The arm
extending from the copper foil was attached to one of two copper parts of the sample
holder (parts A and B of Figure 3.7). As with the previous design, these two copper
pieces, separated from the manipulator by a sapphire disc, were attached to the
manipulator using three screws. In addition, the sample holder, the manipulator and
each screw were all electrically insulated from each other.
As mentioned in the previous paragraph, the MgO substrate was placed on one
side of a flat piece of foil. A homemade heater was adhered (Aremco Products
835M) onto the other side of the foil, and basically this heater was a wire coil made
from narrow (diameter ~0.25 mm) tantalum wire (Figure 3.8). Several alumina
sleeves (Omega Engineering, ORX-132116) were placed around the wire used for
this coil to ensure that any current passed into this wire went through its entire
length, i.e., several portions of this wire were carefully insulated from each other.
This ensured that the heater was a relatively high resistive element. Each end of this
wire was connected to one of the two copper parts (Figure 3.7) attached to the
sample holder. Two power leads, from an external power supply, were each attached
to one the aforementioned copper parts, and consequently, electrical current could be
passed through the homemade resistive heater. Using an electrical current of ~14 A,
the substrate could be heated from ~90 K to ~500 K at a rate of ~2 K/s. Moreover,
the hole at the center of the copper foil permitted IR radiation to be transmitted
through the sample. During thermal desorption experiments, the face of the MgO
65
crystal with only two copper strips at two of its edges was placed directly in front (~1
mm) of the mass spectrometer nose cone.

Figure 3.8. The homemade heater is a wire coil with alumina sleeves. The sleeves
insulate the wire and ensure current travels throughout the entire length of the wire.

3.9 Sample Preparation
After attaching thermocouple wires to the crystal surface, the surface was
mounted onto the manipulator by placing it in a sample holder and connecting this
sample holder to the manipulator. The UHV chamber was then closed and evacuated.
Prior to pumping out the chamber with the turbomolecular pump attached to the
chamber, two cryogenic sorption pumps (Thermionics) were used to reduce the
chamber pressure to 10
-3
Torr. Typically, the turbomolecular pump evacuated the
chamber to 10
-6
Torr within an hour, and at this pressure the bakeout process was
initiated. Generally, the chamber pressure increases as the temperature of the walls
66
increases but eventually this pressure begins to reduce as the temperature of the
chamber walls stabilizes at ~400 K. When the chamber pressure stabilized, the
heaters were turned off and the chamber was allowed to cool to room temperature.
This cooling process took approximately 12 hours.
One of the advantages of using an MgO substrate is that it is a relatively inert
surface, and annealing it to 500 K essentially removes any contaminants from its
surface [8]. However, a contaminant free surface may still contain a lot of defect
sites, particularly oxygen vacancies. A relatively clean and defect free surface can be
produced by annealing an MgO surface to ~700 K in 10
-7
Torr for 1 hour [8]. All
surfaces used for the experiments reported in the ensuing chapters were subjected to
this procedure after heating the chamber.
H
2
O (distilled and purified by osmosis) was degassed by several freeze-pump-
thaw cycles and used to produce vapor deposited ice films. This deposition process
was performed typically at ~90 K, and this resulted in the formation of amorphous
ice films. N
2
O (Praxair, 99.999% purity) and CO
2
(Gilmore, 99.99% purity) were
used without further purification.
67
3.10 Chapter 3 References
[1] S. A. Hawkins: Fourier transform infrared spectroscopy and temperature
programmed desorption of water thin films on the MgO (100) surface, Ph. D. Thesis,
Department of Chemistry, University of Southern California, Los Angeles, 2004.

[2] D. A. Skoog and J. L. Leary, Principles of Instrumental Analysis, Saunders
College Publishing: Fort Worth, 1992.

[3] X. J. Yu and H. S. Kwok, Journal of Applied Physics, 93, 4407, (2003).

[4] B. C. Smith, Fourier Transform Infrared Spectroscopy, CRC Press: Boca
Raton, 1996.

[5] M. M. Suchan: Molecules-surface interactions in HCl/MgO and Water/MgO
Systems, Ph. D. Thesis, Department of Chemistry, University of Southern California,
Los Angeles, 2001.

[6] Sheikh, II and P. D. Townsend, Journal of Physics E-Scientific Instruments,
6, 1170, (1973).

[7] J. Yates, J. T., Experimental Innovations in Surface Science, AIP Press
Springer-Verlag: New York, 1998.

[8] L. K. Hodgson: Photodissociation, molecule-surface collision-induced
dissociation and direct adsorbate photolysis of nitroso molecules, Ph. D. Thesis,
Department of Chemistry, University of Southern California, Los Angeles, 1993.

68
Chapter 4: Transport and Guest-Host interactions in Ice
4.1 Introduction
Understanding interactions between molecules and ice continues to be the focus
of numerous experimental and theoretical studies, and the dependence of physical
properties on structure makes the determination of ice structures a prerequisite for
such research. At low temperatures, water can exist in a number of phases. Of these,
amorphous and cubic ices lend themselves readily to ultra high vacuum studies, and
consequently a considerable amount of research involving these forms has been
carried out.
Amorphous ice, commonly referred to as amorphous solid water (ASW), is a
metastable form of ice. It is the most abundant phase of water in interstellar clouds
[1, 2]. There is, however, little quantitative agreement among the many published
results detailing its physical properties. For example, specific surface areas have
been reported that range from ~12 to 640 m
2
g
-1
[3-5].
Recent results by Kimmel and coworkers. show that such disagreements may
arise from differences in ice preparation [5]. Their investigations indicated that ASW
film morphology depends on both the growth conditions and the thermal histories of
the films. Accordingly, meaningful comparisons between ASW studies performed by
different groups require clear specification of these parameters.
The main IR signature of ASW is not a good indicator of film structure. Despite
the fact that different preparation conditions lead to ASW samples having a number
of different physical properties, spectra obtained at different temperatures and by
69
various investigators display remarkably similar shapes [6, 7]. Thus, these spectra
are poor diagnostics. In the 2000–4000 cm
–1
region, ASW spectra are dominated by
a broad band that is ~300 cm
–1
wide and centered at ~3250 cm
–1
. A relatively weak
feature at 3696 cm
–1
is believed to originate from dangling OH (d–OH) groups [4, 8].
The hydrogen atoms of these groups do not participate in the H-bonding network
present in ice, and hence the term dangling bonds. Compared to the dominant ASW
band, the d–OH line width is relatively narrow. The broad nature of the main spectral
signature makes it difficult to discern subtle changes that occur in this band when
ASW interacts with added molecular species. Consequently, the properties of the
weak d–OH band are often monitored to glean information about such interactions
and to explore structural changes induced by annealing.
To circumvent problems that derive from the use of a relatively weak signal as
the observable, some studies have exploited IR spectral signatures of probe
molecules [9-13]. If these molecules interact weakly with water molecules in the ice
film, they may serve as good probes of certain properties. From a spectroscopic
standpoint, a good probe should possess certain characteristics: an IR band having a
narrow line width (to allow small shifts to be easily discerned), and reasonable
oscillator strength to facilitate detection. For example, CO and CF
4
have been used
as spectral probes of ice morphology because they fit the above criteria [9, 10, 13].
As probe molecules, both CO
2
and N
2
O possess desirable characteristics. For
example, in the gas phase, they both have strong IR absorptions in the mid-IR region.
It is unclear, however, if the presence of CO
2
or N
2
O leads to alterations in ice
70
structure. Sandford and Allamandola report the binding energy of CO to water to be
about one fifth of the energy that binds a water molecule to a water film (~ 0.5 eV)
[14]. Thus, it is reasonable to expect that the presence of CO molecules will not lead
to significant changes in ice morphology. In contrast, binding energies of around 0.2
eV have been determined for CO
2
desorbing from ice films [14, 15], and therefore
interactions between CO
2
and ice might (conceivably) alter ice structure. Though a
literature search did not find binding energies of N
2
O to ice, they are expected to be
similar to those of CO
2
.
This study explored the uptake and desorption of CO
2
and N
2
O from ASW films
by using IR spectroscopy as the main diagnostic. The similar physical properties of
CO
2
and N
2
O suggested that results obtained with each would be complementary.
The observations reported below show that, regardless of the CO
2
-ice and N
2
O-ice
interaction strengths, some spectroscopic characteristics exhibited by CO
2
and N
2
O
are indeed sensitive to ice morphology. Thus, these molecules can be used to monitor
or probe structural changes within the ice film. Moreover, these molecules do not
require exceptionally low (< 50 K) temperatures for their accumulation on ice. This
obviates the need for helium cooling systems, which are generally more expensive
than liquid nitrogen cooling systems.
The d–OH band shifts observed in each probe-ASW system are also reported.
Silva and Devlin used these shifts to estimate binding energies of certain molecules
to amorphous ice [12]. Recent theoretical results, however, suggest that there is no
direct correlation between interaction energy and the magnitude of the d–OH band
71
shift.

Manca and coworkers state that these shifts appear to be related to the local
electric field along the OH bond of a dangling OH group [16]. The d–OH band shifts
induced by CO
2
and N
2
O provide yet another experimental test of theoretical models
that attempt to elucidate the physical processes responsible for the d–OH vibrational
frequency. Moreover, the small sizes of CO
2
and N
2
O should facilitate their
inclusion into quantum mechanical models.

4.2 Experimental Details
Experiments were carried out using the ultrahigh vacuum chamber and
experimental configuration described in the previous chapter. The MgO (100) single
crystal sample (~1 mm × 10 mm × 10 mm) was prepared under dry nitrogen
conditions and quickly inserted into the chamber. A k-type thermocouple attached to
one edge of the crystal with a ceramic adhesive (Aremco 569) recorded temperature.
Sample holder 1, described in the previous chapter, was used to attach the sample to
the precision manipulator.
The importance of the minimum temperature (~90 K) obtained routinely with this
surface holder cannot be overstated, as maintaining temperatures less than 100 K was
crucial to the formation of stable N
2
O and CO
2
films. Namely, maintaining suitable
concentrations of these molecules on an ice or MgO(100) surface for the 5-10
minutes needed to record a FTIR spectrum required temperatures below 100 K.
Unfortunately, the efficient thermal contact between the sample and the sample
holder limited the maximum achievable surface temperatures to ∼400 K and 200 K,
72
with samples at room temperature and 90 K, respectively. In addition, the close
proximity of the surface holder components to the sample made it difficult to
determine the origin of desorbing molecules during temperature programmed
desorption (TPD). Thus, the ability to perform accurate and meaningful TPD was
sacrificed when using this sample holder.
Purified and de-ionized H
2
O was introduced into the chamber via a leak valve
following several freeze-pump-thaw cycles. N
2
O (Praxair, 99.999%) and CO
2

(Gilmore Liquid Air Company, 99.99%) were dosed through a separate leak valve.
In all of the experiments, the admitted gases were background dosed, thereby
subjecting both faces of the crystal to the gas, with continuous pumping during
dosing.
Each spectrum consisted of an average of 200-500 scans obtained within 5-10
minutes and recorded with 1 cm
-1
resolution using a Happ-Genzel apodization
function. A cooled surface, flashed to 400 K just before cooling, served as a
reference, i.e., the background traces. The InSb detector cut-off frequency of ~1850
cm
-1
set the lower limit on the observable frequency range.
Ice films were obtained by exposing the MgO(100) surface to a constant flux of
water vapor for a fixed period of time prior to recording sample spectra. Ice
thickness was estimated by comparing a film's integrated absorbance with the
integrated absorbance of a water monolayer on MgO(100). The IR spectrum of a
water monolayer on MgO (100) was determined in experiments preceding this
73
dissertation work; TPD was used to calibrate water coverage [17]. The adsorbate
surface residence time τ can be estimated by using:
τ = τ
0
e
ΔH
ad
/RT
s
(4.1)
where τ
0
typically is assumed to be 10
–13

s, T
s
is the surface temperature, and ΔH
ad

is the heat of adsorption [18]. The isosteric heat of adsorption of the water monolayer
on MgO(100) and the desorption barrier for water molecules on ice are roughly 0.8
and 0.5 eV, respectively [19, 20]. At a surface temperature of 100 K, both the
monolayer and multilayer films are stable for several hours because of long
residence times. At surface temperatures of 90 K, water vapor deposition resulted in
ASW formation. Crystalline (cubic) ice was obtained by annealing ASW films past
the phase transition temperature; this phase transition was signaled by changes in the
spectral characteristics of the annealed ASW film at ~3200 cm
-1
.
The introduction of N
2
O and CO
2
was achieved by exposing water films (and in
some cases the MgO(100) substrate) to a background pressure of N
2
O or CO
2
for
fixed times prior to recording spectra. All probe molecule depositions were
performed at 90 K. Ideally, the amount of probe molecules on or within the water
film can be estimated by dividing the integrated absorbance of a specific probe IR
feature with the integrated absorbance of that same IR feature at monolayer
coverage. However, TPD coverage calibrations for these molecules could not be
performed because the surface holder suitable for TPD experiments failed to reach
temperatures low enough (i.e., < 100 K) for a detectable accumulation of these
molecules

on the MgO(100) surface. Consequently, the background pressure and
74
exposure time for each N
2
O and CO
2
dose will be stated. The integrated absorbance
of a particular probe molecule IR feature (e.g., the asymmetric stretch band) obtained
immediately after dosing N
2
O or CO
2
onto the MgO(100) substrate at 90 K was
approximately proportional to the exposure time at a constant background pressure.
CO
2
has three active IR bands in the 2000-4000 cm
-1
region [21], the most intense
being the asymmetric stretch  ν
3
fundamental near 2349 cm
-1
. Studies of CO
2
thin
films show that this mode splits into the longitudinal optical (LO) and the transverse
optical (TO) modes in crystalline CO
2
[22]. The same type of splitting occurs with
the ν
3
fundamental mode of crystalline N
2
O [22]. The transition dipole moment
associated with the LO mode lies perpendicular to the face of the film, while the TO
mode transition dipole moment is parallel to the film surface. Consequently, to
observe both modes requires incident radiation capable of exciting transition dipole
moments perpendicular and parallel to the film surface. As p-polarized light meets
this criterion, it was used for most the experiments. For these polarization studies, a
wire grid polarizer (Molectron) was placed in the infrared beam path; the angle
between the surface normal and incident beam ranged from 48° to 58°.

4.3 Results: Interactions of CO
2
with ASW
The overall shapes of the absorption spectra obtained after depositing CO
2
onto
ASW films at 90 K depend on the amount of CO
2
deposited and the ASW film
thickness. Figure 4.1 shows the trend observed in the ν
3
region upon depositing
varying amounts of CO
2
on ASW films of constant thickness. A single band
75
4000 3500 3000 2500 2000
wavenumber / cm
-1
absorbance
0.02
(a)
(i)
(ii)
(iii)
4000 3500 3000 2500 2000
wavenumber / cm
-1
absorbance
0.02
(a)
(i)
(ii)
(iii)
2450 2400 2350 2300 2250
0.02
wavenumber / cm
-1
absorbance
(i)
(ii)
(iii)
(b)
2450 2400 2350 2300 2250
0.02
wavenumber / cm
-1
absorbance
(i)
(ii)
(iii)
(b)
2300 2290 2280 2270 2260 2250
(c)
0.002
(i)
(ii)
(iii)
wavenumber / cm
-1
absorbance
2300 2290 2280 2270 2260 2250
(c)
0.002
(i)
(ii)
(iii)
wavenumber / cm
-1
absorbance

Figure 4.1. (a) Varying amounts of CO
2
deposited at 90 K onto ASW films of
approximately constant thickness (∼65 layers). CO
2
exposure times (in minutes) at
constant pressure (4 × 10
-8
Torr) were: (i) 1.5 (ii) 3 (iii) 6. The broad water feature at
~3250 cm
-1
provides estimates of relative film thickness. The horizontal scale is
expanded to emphasize (b) the CO
2
and (c)
13
CO
2
ν
3
region.
76
(2325-2375 cm
-1
) was observed with small doses, with peaks at 2379 and 2345 cm
-1

evident with higher doses. At 90 K, equal exposures of CO
2
to ASW films that
differed only in thickness revealed that the thinner the ASW film the greater the
intensities of the features at 2345 and 2379 cm
-1
(Figure 4.2).

2450 2400 2350 2300 2250
0.02
(iv)
(iii)
(ii)
(i)
wavenumber / cm
-1
absorbance
(a)
2450 2400 2350 2300 2250
0.02
(iv)
(iii)
(ii)
(i)
wavenumber / cm
-1
absorbance
(a)
4000 3500 3000 2500
absorbance
wavenumber / cm
-1
0.02
(i)
(ii)
(iii)
(iv)
(b)
4000 3500 3000 2500
absorbance
wavenumber / cm
-1
0.02
(i)
(ii)
(iii)
(iv)
(b)

Figure 4.2. The same amount of CO
2
was deposited at 90 K onto ASW films of
varying thickness. The films were exposed to 4 × 10
-8
Torr CO
2
for 3 min. H
2
O
exposure times at 5 × 10
-8
Torr were: (i) 2 (ii) 4 (iii) 8 and (iv) 16 min, corresponding
to 10, 20, 40 and 80 layers, respectively. Entries (a) and (b) show spectral regions for
CO
2
and H
2
O, respectively.

77
Referring to Figure 4.1, a band was also observed in the
13
CO
2
ν
3
region. It
consisted of a peak at ~2278 cm
–1
at low dosage, and a relatively narrow feature at
~2282 cm
–1
that emerged as the amount of accumulated CO
2
increased [Figure
4.1(c)]. The overall shape of this isotopologue feature depended on both the amount
of CO
2
deposited and the ASW thickness. For equal amounts of CO
2
dosed at 90 K,
the thinner the ASW film the greater the intensity of the feature at ~2282 cm
-1
. The
two small peaks on the high frequency side of the water IR band are CO
2

combination bands [Figure 4.1(a), trace (iii)].
Raising the substrate temperature to 105 K for 15-20 minutes after CO
2
deposition
onto an ASW film resulted in a residual CO
2
band around 2340 cm
-1
. The
temperature of 105 K was chosen because experiments involving CO
2
films adsorbed
at 90 K on the MgO(100) substrate showed complete CO
2
desorption by 105 K. Both
the maximum peak intensity and the integrated absorbance of the residual band were,
to within experimental uncertainty, proportional to the ASW film thickness (Figure
4.3). For ASW films having the same thickness, however, increasing the CO
2
dosage
time at 90 K did not necessarily result in a concomitant increase in the integrated
absorbance of the 2340 cm
-1
band (Figure 4.4). In fact, the 2340 cm
-1
feature
appeared to saturate. Namely, for sufficiently long exposure times, the shape and
intensity of this band did not continue to increase with exposure. This suggests the
amorphous host has taken up the maximum amount of CO
2
that it is capable of
absorbing.

78

4000 3500 3000 2500
0.02
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance
(a)
4000 3500 3000 2500
0.02
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance
(a)
0.002
2380 2360 2340 2320 2300 2280
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance
(b)
0.002
2380 2360 2340 2320 2300 2280
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance
(b)

Figure 4.3. Spectra of the CO
2
that remained after depositing at 90 K equal amounts
of CO
2
(3 min at 4 × 10
-8
Torr) onto ASW films of varying thickness followed by
raising the temperature. The temperature was raised to 105 K where it was held for
15-20 minutes before each trace was obtained. Entries (i) – (iii) correspond to ASW
films of thickness 20, 40, and 80 layers, respectively. Panels (a) and (b) show the
same traces on different horizontal scales.
79
2400 2380 2360 2340 2320 2300 2280
0.002
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance
2400 2380 2360 2340 2320 2300 2280
0.002
(ii)
(i)
(iii)
wavenumber / cm
-1
absorbance

Figure 4.4. Spectra of the CO
2
that remained after depositing different amounts of
CO
2
onto ASW films of constant thickness (~65 layers) and then increasing the
temperature to 105 K. The spectra were recorded at 105 K, 15-20 min after the
reaching this temperature. The CO
2
exposure times at 4 × 10
-8
Torr were (i) 1.5 (ii) 3
and (iii) 6 min. See Figure 4.1 for the spectra recorded following CO
2
deposition at
90 K.

The residual CO
2
band was relatively stable up to the ASW phase transition
temperature. During the conversion of ASW to cubic ice, the intensity of this band
went almost to zero (Figure 4.5). A similar phenomenon involving the release of
molecules trapped in porous ice films has been reported previously [23].
It was found that the residual band intensity is sensitive to the amount of water
deposited on the CO
2
at 90 K. In a series of "sandwiching" experiments, a fixed
amount of CO
2
was deposited between two ASW films. This involved forming the
bottom ASW layer, dosing CO
2
, and then depositing the top ASW layer. The dosing
times for the two ASW layers were chosen in a manner that ensured that the total
amount of water in each sandwich remained constant. The results revealed that the
thicker was the ASW film on top of the accumulated CO
2
, the greater was the
80
intensity of the residual CO
2
peak. Similar sandwiching experiments could not be
performed with crystalline ice films because formation of the top cubic ice film

2450 2400 2350 2300 2250
0.002
170 K
155 K
145 K
105 K
125 K
wavenumber / cm
-1
absorbance
(a)
2450 2400 2350 2300 2250
0.002
170 K
155 K
145 K
105 K
125 K
wavenumber / cm
-1
absorbance
(a)
4000 3500 3000 2500
0.02
170 K
155 K
145 K
105 K
125 K
wavenumber /cm
-1
absorbance
(b)
4000 3500 3000 2500
0.02
170 K
155 K
145 K
105 K
125 K
wavenumber /cm
-1
absorbance
(b)

Figure 4.5. CO
2
(4 × 10
-8
Torr, 4 min) was deposited onto a 90 K ASW film of 40
layers. The temperature was then raised to 105 K for 15 min, at which time a
spectrum was recorded. The figure depicts the evolution of the 105 K spectra as the
temperature was increased in steps to the temperatures shown and held at these
temperatures for the duration of a scan (∼8 min). The spectra were recorded
immediately after each temperature increase and at the temperatures indicated.
Entries (a) and (b) show the spectral region for CO
2
and H
2
O, respectively. The
broad H
2
O feature (~3250 cm
-1
) changes upon annealing past 165 K because of the
ASW-to-cubic phase transition.

81
would require deposition or annealing temperatures greater than 150 K, and CO
2

desorption begins well below 150 K.
The
13
CO
2
ν
3
region showed no discernable IR signature after 105 K isothermal
desorption. This was most likely due to the relatively small percentage of these
isotopologues in the CO
2
sample and the small total amount of CO
2
remaining after
desorption. Naturally, increasing the total amount of CO
2
remaining at 105 K by
increasing film thickness should eventually lead to a distinguishable signature in the
isotopomer ν
3
region.
Converting the ASW film to cubic ice via flash-annealing to 170 K prior to CO
2

deposition drastically altered the outcome of isothermal desorption at 105 K. Though
cubic ice films were cooled to 90 K before CO
2
adsorption, no CO
2
spectral features
remained after 105 K desorption. Changing the thickness of the cubic film and the
CO
2
dosing time did not alter this result.
Experiments monitoring the d–OH feature utilized ASW films that were more
than 800 layers thick. This stemmed from the difficulty associated with observing the
d–OH IR band in thinner films (c.f., Figures. 4.5 and 4.6). The spectra presented in
Figure 4.6 were taken between successive doses of CO
2
. Each CO
2
deposition added
to previously accumulated CO
2
and a new feature (~3655 cm
-1
) emerged as the
amount of CO
2
increased. Though CO
2
combination bands (3708 and 3599 cm
-1
)
complicate the overall picture, the 3655 cm
-1
band is in fact the shifted d–OH band.
82
This conclusion is supported by "N
2
O-ASW" spectra that are presented below, in
which the absence of N
2
O combination bands in this spectral region simplifies the
spectrum.

4000 3600 3200 2800 2400 2000
wavenumber /cm
-1
absorbance
0.3
(iii)
(ii)
(i)
3800 3700 3600 3500
4000 3600 3200 2800 2400 2000
wavenumber /cm
-1
absorbance
0.3
(iii)
(ii)
(i)
wavenumber /cm
-1
absorbance
0.3
(iii)
(ii)
(i)
3800 3700 3600 3500

Figure 4.6. Different amounts of CO
2
on an ASW film (~1000 layers). The amount
of CO
2
was varied via the sequence: (i) no deposition (ii) 4 × 10
-8
Torr for 2 min and
(iii) 1 × 10
-7
Torr for 16 min. Spectra shown were recorded immediately after each
deposition. The inset shows the d–OH feature is red shifted as a consequence of CO
2

deposition; namely the feature at ~3700 cm
-1
moves to ~3650 cm
-1
.

4.4 Results: N
2
O Interactions with ASW
The experimental results obtained for N
2
O and CO
2
are similar. The nature of the
observed N
2
O ν
3
features after dosing N
2
O onto ASW films at 90 K depended on
both the amount of N
2
O dosed and the ASW film thickness. At low N
2
O coverages,
only a single band at 2215-2250 cm
-1
was present (Figure 4.7). Two additional peaks
at 2238 and 2256 cm
-1
were evident at higher coverage. At 90 K, depositing equal
83
amounts of N
2
O onto ASW films that differ only in thickness demonstrated that the
thinner the ASW film the greater the intensity of the 2238 and 2256 cm
-1
features
and the smaller the intensity of the 2215-2250 cm
-1
band (Figure 4.8).

4000 3500 3000 2500 2000
(a)
0.02
(iii)
(ii)
(i)
absorbance
wavenumber / cm
-1
4000 3500 3000 2500 2000
(a)
0.02 0.02
(iii)
(ii)
(i)
absorbance
wavenumber / cm
-1
2300 2270 2240 2210 2180
absorbance
wavenumber / cm
-1
0.02
(iii)
(ii)
(i)
(b)
2300 2270 2240 2210 2180
absorbance
wavenumber / cm
-1
0.02 0.02
(iii)
(ii)
(i)
(b)

Figure 4.7. (a) Varying amounts of N
2
O were deposited onto 90 K ASW films. The
broad H
2
O feature (~3250 cm
-1
) provides an estimate of relative film thickness,
which were approximately equal to 70 layers. The N
2
O exposure times at 4 × 10
-8

Torr were: (i) 2 (ii) 4 and (iii) 8 min. (b) The horizontal scale is expanded to
emphasize the N
2
O ν
3
region.

84
Isothermal desorption of the accumulated N
2
O (deposited onto an ASW film at 90
K) for 15-20 minutes resulted in a "residual" N
2
O band at 2222 cm
-1
(Figure.4.9). A
desorption temperature of 105 K was chosen because experiments involving N
2
O
films on the MgO substrate revealed that any N
2
O deposited at 90 K desorbed at

absorbance
wavenumber / cm
-1
3000 2500 3500 4000
0.04
(iii)
(ii)
(i)
(iv)
(a)
absorbance
wavenumber / cm
-1
3000 2500 3500 4000
0.04 0.04
(iii)
(ii)
(i)
(iv)
(a)
0.02
absorbance
wavenumber / cm
-1
2300 2270 2240 2210 2180
(iii)
(ii)
(i)
(iv)
(b)
0.02 0.02
absorbance
wavenumber / cm
-1
2300 2270 2240 2210 2180
(iii)
(ii)
(i)
(iv)
(b)

Figure 4.8. (a) ASW films of different thicknesses were prepared at 90 K by vapor
depositing H
2
O at 4 × 10
-8
Torr. (a) H
2
O spectra (i) – (iv) were recorded with
deposition times of (i) 4, (ii) 8, and (iii) 16 min, corresponding roughly to 15, 35, and
70 layers. (b) Each film was exposed to 4 × 10
-8
Torr N
2
O for 4 min, and the
corresponding N
2
O spectra are labelled (i)-(iv).
85

4000 3500 3000 2500 2000
0.04
wavenumber / cm
-1
absorbance
(iii)
(ii)
(i)
(iv)
(a)
4000 3500 3000 2500 2000
0.04 0.04
wavenumber / cm
-1
absorbance
(iii)
(ii)
(i)
(iv)
(a)
2300 2270 2240 2210 2180
wavenumber / cm
-1
absorbance
(iii)
(ii)
(i)
(iv)
0.002
(b)
2300 2270 2240 2210 2180
wavenumber / cm
-1
absorbance
(iii)
(ii)
(i)
(iv)
0.002 0.002
(b)

Figure 4.9. ASW films at 90 K were exposed to 4 × 10
-8
Torr N
2
O for 4 min. The
temperature was then raised to 105 K and held there for 15-20 min before a spectrum
was recorded. Approximate ASW thicknesses for (i)–(iv) were 15, 35, 70, and 150
layers, respectively. Entries (a) and (b) show H
2
O and N
2
O features, respectively.

~100 K. Thus, the spectra recorded in these experiments are due to N
2
O that is
present in the film, rather than on its surface. Figure 4.9 shows the sensitivity of this
residual N
2
O feature to ASW thickness. For an ASW film of a specific thickness,
86
doubling the amount of N
2
O dosed onto the film at 90 K did not necessarily double
the intensity of the spectral feature associated with the remaining N
2
O (Figure 4.10).
A similar effect was seen with CO
2
(Figure 4.4).

2280 2260 2240 2220 2200
0.002 0.002
absorbance
wavenumber / cm
-1
(iii)
(ii)
(i)

Figure 4.10. Spectra of the N
2
O that remained after depositing different amounts of
N
2
O onto ASW films of same thickness (70 layers) and then increasing the
temperature to 105 K. The temperature was then raised to 105 K and held there for
15–20 min before a spectrum was recorded. N
2
O exposure times at 4 × 10
-8
Torr for
(i)–(iii) were 2, 4, and 8 min, respectively.

The N
2
O residual feature was stable at temperatures below the ice transition
temperature. As shown in Figure 4.11, this IR band persisted even at a temperature
of 155 K which is much higher than the desorption temperature of pure N
2
O films
and only a few degrees below the ASW-to-crystalline transition temperature.
Converting the ASW film to crystalline ice prior to 90 K N
2
O adsorption resulted in
no remaining N
2
O features after 105 K isothermal desorption.

87

2300 2270 2240 2210 2180
0.003 0.003
absorbance
wavenumber / cm
-1
105 K
120 K
135 K
155 K
165 K 0 min
165 K 5 min
165 K 15 min

Figure 4.11. N
2
O (4 × 10
-8
Torr for 4 min) was deposited onto a 90 K ASW film of
40 layers. The temperature was then raised to 105 K for 15 min, at which time a
spectrum was recorded. The figure depicts the evolution of the 105 K spectra as the
temperature was increased in steps to the temperatures shown and held at these
temperatures for the duration of a scan (∼8 min). The spectra were recorded
immediately after each temperature increase and at the temperatures indicated. Three
scans (~5 min in duration) were recorded at 165 K; the time interval between
reaching 165 K and commencing each of these scans is specified for three relevant
spectra.

ASW films formed on top of deposited N
2
O led to larger residual N
2
O intensities,
after isothermal desorption, than ASW films situated beneath accumulated N
2
O. In a
series of experiments, N
2
O was sandwiched between two ASW films at 90 K, and,
aside from the amounts of water dosed prior to and after N
2
O deposition, all
variables (i.e., total amount of water and N
2
O dosed, deposition and desorption
temperatures) were the same in each experiment. All that changed was the respective
thicknesses of the lower and upper ASW films. The constancy in the amount of total
88
water is confirmed by the similarity in shape and intensity of the water IR band
(Figure 4.12). As shown in Figure 4.12, the amount of N
2
O remaining after 105 K
desorption increased with increasing thickness of the top ASW film in these
sandwiching experiments. Figure 4.12 can be reconciled by noting: (i) ASW films
grown on N
2
O overlayers are likely to be more porous than ASW films grown on
MgO(100), where there is good registry. (ii) At 105 K, ASW films undergo some
degree of annealing, which can trap dopants.

4000 3500 3000 2500 2000
wavenumber / cm
-1
absorbance
0.4
(iii)
(ii)
(i)
(iv)

Figure 4.12. Spectra were recorded after sandwiching N
2
O between two ASW films
at 90 K, then raising the temperature to 105 K, and keeping it there for 15 min. The
amount of deposited N
2
O and the total number of water layers (~80) is the same for
all spectra. The ratios of bottom layer to top layer thickness are: (i) 80:0, (ii) 60:20,
(iii) 40:40, and (iv) 20:60.

89

4000 3500 3000 2500 2000
wavenumber / cm
-1
absorbance
0.4
(i)
(ii)
(iii)
(iv)
3750 3700 3650 3600 3750 3700 3650 3600

Figure 4.13. Spectra were recorded for different amounts of N
2
O deposited onto an
ASW film of ~1000 layers at 90 K. The amount of N
2
O was increased via the
sequence: (i) no deposition, (ii) 4 × 10
-8
Torr for 4 min, (iii) 1 × 10
-7
Torr for 3 min,
and (iv) 2 × 10
-7
Torr for 7 min. Spectra were recorded immediately following each
deposition. The inset shows the d–OH feature is red shifted as a consequence of N
2
O
deposition.

The influence of N
2
O on the position of the d–OH peak is shown in Figure 4.13.
As mentioned earlier, experiments monitoring the d–OH band focused on thick ASW
films (> 800 layers) because this band was hard to detect in thin ASW films. The
spectra presented in Figure 4.13 were taken between successive doses of N
2
O and
they reveal a d–OH red shift of ~50 cm
-1
as the amount of N
2
O accumulates. Also,
though some N
2
O remains after isothermal desorption at 105 K, the d–OH peak
returns to its original position at ~3700 cm
-1
.

90
4.5 Discussion
When molecules are deposited on ice films, an issue to be addressed is where
they reside after deposition. Three simple scenarios can be considered. One is that
the molecules diffuse through the film and accumulate on the substrate via the
displacement of water molecules. A second scenario involves molecules penetrating
into the bulk of the ice and residing there. Localization on top of the water film
surface is yet another possibility. Nothing precludes a priori the simultaneous
participation of these processes.
The displacement of water molecules in the first adlayer (i.e., the layer in contact
with the MgO substrate) by CO
2

is unlikely. Theoretical studies predict that the heat
of adsorption of CO
2
on MgO(100) is ΔH
ad
~0.3 eV [23]. In sharp contrast, the
experimentally determined value of ΔH
ad
for monolayer water on MgO(100) is ~0.8
eV, which is very large for a physisorbed molecule [20]. It is this considerable
difference in the ΔH
ad
values that makes the

displacement of monolayer H
2
O by
CO
2
highly improbable on this surface. It is assumed that this argument also applies
to N
2
O, despite the lack of theoretical or experimental binding energies for N
2
O on
MgO(100).
The two IR peaks at 2379 and 2345 cm
-1
that emerge as CO
2
coverage increases
(Figure 4.1) lie at energies that are similar to the LO and TO resonances of thin,
polycrystalline CO
2
films [22]. These peaks also exhibit the same polarization
dependence as the aforementioned resonances, i.e., the peak at ~2380 cm
-1
is
91
drastically reduced in spectra recorded with s-polarized light. Moreover, the
13
CO
2

band at 2284 cm
-1
that emerges as the exposure time of CO
2
to an ASW film
increases, also appears in the spectra of CO
2
thin films. There is little doubt that
these observations imply the formation of a CO
2
thin film. Similarly, the two
additional IR features that grow in as N
2
O accumulates (Figure 4.7) are correlated
strongly with the LO and TO vibrational resonances of N
2
O thin films [22]. The
propensity of CO
2
and N
2
O dosed onto ASW films at 90 K to produce features
characteristic of polycrystalline CO
2
and N
2
O indicates that enough accumulation of
these molecules has occurred on top of the ASW film to yield thin films.
Some CO
2
and N
2
O molecules enter the film interior even at 90 K. It is intuitive
to assume that a property that varies with film thickness must be related to a
phenomenon occurring in the interior. Consequently, it is easy to understand why the
dependence of the CO
2
and N
2
O band shapes on ASW thickness (Figures 4.2 and
4.8) leads to the conclusion that some of these molecules diffuse into the film at 90
K. Both CO
2
and N
2
O are expected to possess significant surface mobility because
90 K is only ~10 K lower than the temperature required for appreciable desorption.
This mobility facilitates diffusion of small molecules into films [24].
Prior to the formation of a CO
2
or N
2
O thin film, the deposited molecules diffuse
into and reside within the ASW film at 90 K. Also, the amount of favorable
residence sites within ASW films rises as the film thickness grows. These
conclusions originate primarily from the fact that film formation is only evident after
92
a certain amount of CO
2
or N
2
O has already accumulated. The relatively broad
feature initially observed upon dosing these molecules onto ASW films (Figures 4.1
and 4.7) serves as evidence of this prior accumulation. This initial feature is assigned
to probe molecules within the ASW film. Moreover, the integrated area of this band
just prior to the emergence of the narrow LO and TO resonances scales linearly with
film thickness over two orders of magnitude (10 – 1000 ASW layers). The
possibility that CO
2
or N
2
O molecules are all situated entirely on top of the ASW
film and that the apparent influence of film thickness on probe molecule spectra is
simply a manifestation of a change in ASW film surface roughness with film
thickness is thus eliminated. This is because a surface roughening effect is not
expected to maintain a linear relationship with film thickness over such a large
range.
The conclusions stated above are consistent with the results depicted in Figures
4.1 and 4.7 where the increase in the LO and TO peak intensity can be correlated
with an increase in CO
2
or N
2
O film thickness. Specifically, the thinner the ASW
film, the less the amount of deposited molecules needed to fill the favorable sites
within the ice film before CO
2
or N
2
O thin film formation commences on the
"populated" ice film (Figures 4.2 and 4.8).
The last statement addresses the issue introduced at the beginning of this section.
Namely, where do the deposited molecules reside? Of the three scenarios under
consideration, only the latter two are active in the systems under study here.
93
Moreover, the experimental results obtained were able to distinguish between probe
molecule uptake into the bulk versus film formation.
Studies of the adsorption of small molecules often differentiate between
monolayer (2D) and multilayer (3D) spectral features [25-28]. In these studies,
interaction with the surface affects the energy levels of first-layer adsorbates in a
manner that is distinct from the ensuing layers. This leads to monolayer IR
signatures (at vibrational energies that differ from those of the multilayer) that
saturate because of the finite number of adsorption sites. In addition, many of these
investigations observe that the vapor pressure of the overlayers is at least an order of
magnitude greater than that of the monolayer. Thus, there is a temperature range in
which isothermal desorption removes the multilayer but little of the monolayer.
Neither CO
2
nor N
2
O displayed an isothermal desorption temperature that
facilitated a distinction between monolayer and multilayer IR features of these
molecules. Here, the term monolayer refers to the first layer of CO
2
or N
2
O atop the
ASW film surface. For example, isothermal desorption of CO
2
deposited onto ASW
resulted in a decrease in the CO
2
band intensity until either no CO
2

signature or only
the residual CO
2

feature peaked ~2340 cm
-1
remained; the outcome depended upon
the ASW film thickness. The fact that the residual 2340 cm
-1
peak scales with film
thickness indicates that it is not the monolayer band. IR features associated with the
monolayer probably coincide with features of the multilayer. Furthermore, such a
similarity in the peak positions and desorption temperatures of the monolayer and
multilayer support the assumption that the binding energies of molecules in these
94
two regimes are approximately the same. Martin and coworkers [29] found the
difference in adsorption energies between monolayer and multilayer CO adsorbed on
ASW to be < 3 kJ/mol.
The fact that residual band integrated absorbance varies linearly with ASW film
thickness suggests that this feature originates from CO
2
and N
2
O within the film.
Devlin noted that spectra of small molecules and water vapor co-deposited at
temperatures ~10 K are similar to spectra of molecules absorbed in ASW films [24].
Indeed, FTIR studies of CO
2
-H
2
O ices [14], formed at 10 K and warmed to
temperatures above 100 K demonstrate that these mixtures have a band at 2340 cm
-1
.
The substantial release of trapped molecules during the phase transition of ASW has
been observed in previous studies [23, 30]. Because this occurs as the film
crystallizes, it depends on the experimental conditions that influence crystallization,
e.g., heating rate and annealing temperature. In monitoring the main ASW band and
the remaining CO
2
and N
2
O IR signatures, it is clear that the release of any trapped
molecules occurs during the conversion from ASW to cubic ice (Figure 4.5).
A plausible process by which gases are trapped has been discussed by Ayotte and
coworkers [23]. They suggest that some of the molecules deposited onto the ASW
film enter pores that open to the vacuum, and two things happen simultaneously
when the film temperature is raised. The desorption rate of molecules accumulated
within these pores rises, due to the temperature increase, while the ASW film
undergoes a slight molecular rearrangement. This rearrangement closes avenues of
connectivity to the vacuum possessed by some pores and, thus, traps molecules that
95
have yet to desorb from the film interior. These trapped molecules remain enclosed
in the film until ASW crystallization when the rearrangement required for the phase
transition allows these molecules to escape.
The apparent saturation in the integrated absorbance of the residual band
intensity for a specific ASW film (Figures 4.5 and 4.11) can be accounted for with
such a model. Once the favorable sites within the film become populated, additional
molecules localize on the film surface. If CO
2
and N
2
O film formation does not
drastically influence the desorption rate of molecules residing within the ASW, then
isothermal desorption at a specific temperature will result in the trapping of
approximately the same amount of molecules when all the available interior sites are
populated. The assumption that both the desorption rate from ASW cavities and the
rate of ASW film pore closure depend on temperature forms the basis of this
inference. As stated earlier, the inability to calibrate CO
2
and N
2
O coverage via TPD
precluded a determination of CO
2
and N
2
O concentrations from integrated
absorbances. However, an estimate of 100 H
2
O molecules for each trapped CO
2

molecule was obtained by using the integrated absorption cross section per molecule
of the CO
2
ν
3
stretch in CO
2
films [31], and assuming an ASW density and layer
thickness of 0.9 g/cm
3
and 4 Å, respectively.
There is evidence of structural rearrangement in ASW films induced by
annealing at temperatures well below that of the amorphous to cubic ice transition.
Manca and coworkers [32] reported ASW reorganization commencing at 105 K, and
Kimmel and coworkers [3] demonstrated that the thermal history of an ASW film
96
influences its porosity. It was anticipated that these structural changes would lead to
a decrease in the amount of trapped CO
2
and N
2
O. To verify this, ASW films were
annealed to 110-130 K for 15 minutes and then cooled to 90 K prior to CO
2
or N
2
O
deposition. The CO
2
and N
2
O that remained after isothermal desorption was far less
than the amount trapped by films that were not annealed in this manner (Figure
4.14). Thus, the 110-130 K annealing of ASW films deposited at 90 K results in
structural rearrangements that influence the trapping of small molecules. It is
noteworthy that this happens even though the ASW spectral signature undergoes no
discernible change (Figure 4.14). Presumably these rearrangements are associated
with collapsing pores that are responsible for trapping molecules. As mentioned
earlier, crystalline films failed to trap a detectable amount of deposited molecules,
consistent with the fact that these films lack significant porosity.
The presence of CO
2
and N
2
O within the ASW film does not necessarily bring
about a shift in the d–OH feature. As Figure 4.13 shows, this band remains
unaffected by relatively small N
2
O accumulations within the film. However,
increasing the amount of deposited N
2
O eventually leads to a 50 cm
-1
red shift in the
d–OH peak position.
Brunauer-Emmett-Teller (BET) isotherms have been used to determine
adsorption energies of molecules interacting with ice films [29]. These studies report
no correlation between d–OH band shifts and heats of adsorption. Though no
inferences are made here about interaction strength between ice and CO
2
or N
2
O,
recent theoretical efforts have been directed at determining parameters responsible
97
for shifts of the d–OH group [16]. Thus, the data presented here provide grist for the
mill of theory.

wavenumber / cm
-1
absorbance
0.02
4000 3500 3000 2500 2000
2400 2350 2300 2250
10 ×
(i)
(ii)

Figure 4.14. 90 K ASW films of approximately the same thickness were exposed to
4 × 10
-8
Torr of CO
2
for the same duration (3 min). After CO
2
deposition, the
temperature was increased to 105 K and held there while a spectrum was recorded.
Traces are the spectra recorded for (i) a film that was not annealed prior to CO
2

exposure and (ii) a film that was annealed for 15–20 min at 120 K and then re-cooled
to 90 K prior to CO
2
exposure. The inset shows these spectra on an expanded
horizontal scale with the corresponding absorbance magnification.

Molecular layers sandwiched between non-porous ASW films have been used as
barriers to H
2
O inter-film mixing during thermal annealing [33]. In such studies, the
sandwiched molecular layer is believed to remain as deposited until the ASW phase
transition. During this transition, the molecules in this layer escape via pathways
98
created in the transforming ASW. Thus, up until the aforementioned transition, these
sandwiched layers are expected to physically separate non-porous ASW films.
However, by all indications, the 90 K ASW films prepared in this dissertation
work are porous. The IR features of sandwiched CO
2
and N
2
O films change
markedly once the sandwich is annealed past 100 K, which is the sublimation
temperature for these species; most noticeably, the integrated absorbance of these IR
features decreases indicating that some these species escape from the sandwich
structure. Moreover, the spectral feature of the remaining CO
2
and N
2
O molecules
suggest these molecules are located within the ASW bulk, i.e., these molecules
display the same IR signatures as guests molecules that enter via transport from the
surface into the ASW bulk (c.f., Figures. 4.9 and 4.12). Together these observations
make it unlikely that the remaining CO
2
or N
2
O molecules hinder diffusion between
porous ASW films.

4.6 Summary
The experimental study described in this chapter has explored issues of porosity,
uptake, and transport in ASW films. The probe molecules CO
2
and N
2
O provide IR
spectral signatures that enable the distinction between their surface films and their
bulk inclusions in ASW hosts, including nuances such as binding to dangling OH
sites. Good sensitivity is achieved because CO
2
and N
2
O have strong ν
3
absorptions
and relatively narrow spectral features, e.g., compared to those of the ASW host. The
conclusions listed below constitute a qualitative picture of the associated phenomena.
99
(1) The ASW films grown at the deposition conditions used in the experiments
reported herein are porous, i.e., ASW films grown at 90 K can be porous. The CO
2

and N
2
O guest molecules move throughout the bulk until essentially all of the
adsorption sites are occupied.
(2) When the temperature of a porous ASW film is increased from its deposition
temperature of 90 K to temperatures in the range 110-130 K, the porosity changes
dramatically. If the sample thus annealed is then cooled to 90 K, it traps much less
CO
2
or N
2
O than freshly grown 90 K samples, i.e., samples that have not been
annealed. It is interesting that this annealing brings about no discernible change in
the ASW spectral feature centered at ~3250 cm
-1
.
(3) Guest molecules trapped within ASW films remain there as the temperature
is increased from 90 K to 155 K. Thereafter, the ASW to cubic ice transition occurs
and these guest molecules are expelled. Once the ice film is crystalline, it is no
longer possible for it to take up the guest molecules within the film.
(4) Large inclusion regions can be prepared by using sandwiched films at 90 K:
ASW-guest-ASW and so forth. Raising the temperature to 105 K drives off most of
the guest molecules, though guest molecules are left behind that are included in the
bulk. These display the same spectral signature as guest molecules that enter via
transport from the surface into the bulk.
(5) Dangling OH bonds can be detected. Their spectral feature is sharp relative to
that of the ASW film, as they are not hydrogen bonded. Guest molecules interact
100
with these sites, resulting in a 50 cm
-1
red shift. This is reminiscent of gas phase
studies of hydrogen-bonded dimers, where such effects are common.
101
4.7 Chapter 4 References
[1] K. P. Stevenson, G. A. Kimmel, Z. Dohnalek, R. S. Smith and B. D. Kay,
Science, 283, 1505, (1999).

[2] P. Jenniskens and D. F. Blake, Science, 265, 753, (1994).

[3] R. Pletzer and E. Meyer, Journal of Chemical Physics, 90, 5207, (1989).

[4] C. Manca, C. Martin and P. Roubin, Chemical Physics Letters, 364, 220,
(2002).

[5] G. A. Kimmel, K. P. Stevenson, Z. Dohnalek, R. S. Smith and B. D. Kay,
Journal of Chemical Physics, 114, 5284, (2001).

[6] A. Givan, A. Loewenshuss and C. J. Nielson, Journal of Physical Chemistry
B, 101, 8696, (1997).

[7] L. Schriver-Mazzuoli, A. Schriver and A. Hallou, Journal of Molecular
Structure, 554, 289, (2000).

[8] B. Roland and J. P. Devlin, Journal of Chemical Physics, 94, 812, (1991).

[9] A. Allouche, P. Verlaque and J. Pourcin, Journal of Physical Chemistry B,
102, 89, (1998).

[10] M. E. Palumbo, Journal of Physical Chemistry A, 101, 4298, (1997).

[11] T. Takoaka, M. Inamura, S. Yanagimachi, I. Kusunoki and T. Komeda,
Journal of Chemical Physics, 121, 1 September 2004, (2004).

[12] S. C. Silva and J. P. Devlin, Journal of Physical Chemistry, 98, 10847,
(1994).

[13] V. Buch, L. Delzeit, C. Blackledge and J. P. Devlin, Journal of Chemical
Physics, 100, 3732, (1996).

[14] S. A. Sandford and L. J. Allamandola, Journal of Astrophysics, 355, 357,
(1990).

[15] P. U. Andersson, M. B. Nagard, G. Witt and J. B. C. Pettersson, Journal of
Physical Chemistry A, 108, 4627, (2004).

[16] C. Manca and A. Allouche, Journal of Chemical Physics, 114, 4226, (2001).
102
[17] S. Hawkins, G. Kumi, S. Malyk, H. Reisler and C. Wittig, Chemical Physics
Letters, 404, 19, (2004).

[18] G. A. Somorjai, Introduction to Surface Chemistry and Catalysis, John Wiley
and Sons, Inc: New York, 1994.

[19] R. S. Smith and B. D. Kay, Surface Review and Letters, 4, 781, (1997).

[20] D. Ferry, A. Glebov, V. Senz, J. Suzzane, J. P. Toennies and H. Weiss,
Journal of Chemical Physics, 105, 1697, (1996).

[21] G. Herzberg, Infrared and Raman Spectra, Van Nostrand Reinhold
Company, Inc: New York, 1945.

[22] M. A. Ovchinnikov and C. A. Wight, Journal of Chemical Physics, 99, 3374,
(1993).

[23] P. Ayotte, R. S. Smith, K. P. Stevenson, Z. Dohnalek, G. A. Kimmel and B.
D. Kay, Journal of Geophysical Research., 106, 33837, (2001).

[24] J. P. Devlin, Journal of Physical Chemistry, 96, 6185, (1992).

[25] J. Heidberg and B. Redlich, Surface Science, 368, 140, (1996).

[26] J. Heidberg, B. Redlich and D. Wetter, Berichte Bunsenges Physical
Chemistry, 99, 1333, (1995).

[27] O. Berg and G. E. Ewing, Surface Science, 220, 207, (1989).

[28] H. Chang, H. H. Richardson and G. E. Ewing, Journal of Chemical Physics,
89, 7561, (1988).

[29] C. Martin, C. Manca and P. Roubin, Surface Science, 502-503, 275, (2002).

[30] A. Bar-Nun, J. Dror, E. Kochavi and D. Laufer, Physical Review B, 35, 2427,
(1987).

[31] H. Yamada and W. B. Person, Journal of Chemical Physics, 41, 2478,
(1964).

[32] C. Manca, C. Martin and P. Roubin, Chemical Physics, 300, 53, (2004).

[33] S. M. McClure, E. T. Barlow, M. C. Akin, D. J. Safarik, T. M. Truskett and
C. B. Mullins, Journal of Physical Chemistry B, 110, 17987, (2006).
103
Chapter 5: The Nature of Trapping Sites in Ice
5.1 Introduction
Amorphous solid water (ASW) films are capable of influencing the desorption
characteristics of certain molecules deposited onto their surfaces or co-condensed
during their formation [1-4]. Moreover, most of the experimental results are
consistent with the idea that thermally induced structural changes in these films play
a major role in the aforementioned desorption process. In particular, these structural
changes trap molecules residing within ASW, and they inhibit the release of these
molecules until ASW crystallization and the sublimation of the crystallized ice. But
what are these structural relaxations and where do they occur, i.e., throughout the
film or localized to certain layers? Several groups [5-9] have observed that some of
these trapped species are retained within crystalline ice after the phase transition.
However, there are only speculations as to how or why this happens, and this is
because the processes occurring during this phase transition are poorly understood.
Indeed, the very nature of ice at the phase transition is in itself a source of
controversy.
ASW is often described as a glassy state of water, and implicit in such a
description is the suggestion that liquid water can be supercooled suitably enough to
induce a glass transition [10, 11]. Determining this glass transition, a key step in
establishing this relationship, has been hampered by the experimental difficulty of
probing liquid water's properties in the temperature range of 160-230 K. On one
hand, it is difficult to inhibit the crystallization of supercooled water at ~230 K, and
104
on the other, amorphous ice readily crystallizes at ~160 K. Thus, what emerges is an
experimentally inaccessible region, the so-called "no man's land" [12] that makes the
nature of supercooled water below ~230 K controversial.
There are reports suggesting amorphous ice undergoes a glass transition to
supercooled water at ~140 K, prior to crystallizing at ~160 K [13-16]. In this case,
rather than transforming glassy water to crystalline ice, the phase transition
transforms supercooled water to crystalline ice. In contrast, there are reports also that
suggest the glass transition temperature (Tg) is greater than ~160 K [1, 9, 11, 17],
and this implies that observing the glass transition is prevented simply by the fact
takes place in an experimentally inaccessible region.
If the glass transition does indeed occur before the phase transition, the question
then turns to the nature of the supercooled liquid formed by the glass transition.
Fragility is a term used to characterize the temperature dependence of relaxation
processes in liquids; the viscosity of a fragile liquid displays a non-Arrhenius
dependence on temperature, and a fragile liquid becomes very fluidic, relative to its
glassy state, in a short temperature range above Tg [10]. Recent thermal desorption
studies have interpreted inter-mixing in isotopically labeled ASW films (grown one
atop the other) at ~150 K as bulk diffusion [13, 14], and this led to the conclusion
that ASW behaves as a fragile liquid above ~140 K. However, subsequent studies
argue that this inter-mixing is a result of H
2
O transport through the interconnected
network of fractures created during crystallization and that this process is distinctly
different from bulk diffusion [9, 17].
105
While the details about the molecular rearrangements accompanying ASW
crystallization are important, obtaining these details has been difficult. Instead,
investigations have settled for a more practical goal; a microscopic description of
these dynamics [18-20]. The work detailed in this chapter takes this same approach,
and it focuses on the ASW crystallization dynamics by exploring how these
dynamics influence the release of trapped species. Specifically, this study examines
what factors (ASW film thickness, deposition technique, annealing rate, etc.) affect
the ratio of trapped CO
2
molecules that escape during crystallization to those that are
retained within ice. This knowledge will support rationalizations, consistent with the
observed experimental results, about likely processes occurring during
crystallization. Moreover, the IR spectroscopy of these trapped species will also
provide valuable insight into the nature of the local environment surrounding these
species.
Using a combination of TPD and IR experiments, two independent but
complementary means are used to explore differences in the properties of the trapped
species (aside from their differing desorption temperature). These results show that
thermal desorption experiments without accompanying IR experiments may lead to
erroneous conclusions about obtained results. As detailed in the previous chapter,
CO
2
is well-suited for these types of IR studies.

106
5.2 Experimental
The experiments were performed in the UHV chamber described in chapter 3 of
this dissertation. The MgO (100) substrate was prepared by cleaving an MgO crystal
in dry nitrogen. This substrate was inserted into the chamber, and then cleaned using
an established procedure. TPD studies were performed using a Stanford Research
Systems residual gas analyzer (SRS RGA 300). A stainless steel cone with a small
(~9 mm) aperture was used to screen out molecules not originating from the surface
during TPD.
The substrate was attached to a manipulator, which provided XYZ translation
and 360º rotation of the substrate, using a surface holder. Unless explicitly stated, the
surface holder used for these experiments is one that allowed IR and TPD
experiments to be performed using the same holder; this surface holder design is
detailed in chapter 3 (see Figure 3.7). For this holder, the lowest temperature
attainable was typically ~90 K. This was achieved by bubbling helium gas through a
filled liquid nitrogen reservoir which was in contact (thermally) with the surface
holder. The surface could be heated to a temperature of ~500 K using a homemade
heater, and this temperature was monitored by a k-type thermocouple adhered to the
surface.
FTIR experiments were carried out with a commercial spectrometer (Nicolet
Protégé 460), and the IR beam was directed in and out of the chamber using mirrors.
H
2
O (distilled and purified by osmosis) was degassed by several freeze-pump-thaw
107
cycles, and CO
2
(Glimore 99.99%) was used without further purification. These
gases were introduced into the chamber using separate leak valves.

5.3 Results
The formation of a CO
2
film at 90 K on MgO (100) yields a p-polarized IR
spectrum with distinctive LO and TO features. The TPD trace of such a film exhibits
one desorption feature at ~105 K (Figure 5.1), and this suggests there is little
difference in the binding energies of the CO
2
monolayer (i.e., CO
2
molecules in
contact with the MgO surface) and multilayers. This is in contrast to a
CO
2
/MgO(100) study by Heidberg and others [21] in which a CO
2
monolayer was

Figure 5.1. A TPD trace of a CO
2
film prepared at 90 K on MgO(100). The TPD
experiment was performed by heating the surface at ~1 K/s and monitoring a mass-
to-charge ratio (m/e) of 44. MgO exposure to CO
2
was carried out at a pressure 4 ×
10
-8
Torr for 3 min.

108
produced by exploiting a significant difference in the vapor pressure, and hence
binding energy, of the CO
2
monolayer and the ensuing multilayers. However, it is
well-known that the concentration of defect sites on the MgO(100) surface depends
upon how the surface is prepared (i.e., polished or unpolished, cleaved in vacuum or
in nitrogen, etc.) [22]. Thus, it is conceivable that these apparent differences stem
from different surface preparatory methods. No other TPD spectra of CO
2
desorbing
from MgO(100) appear to have been reported.
In agreement with recent studies [23], the TPD traces of CO
2
deposited onto
ASW consists of three desorption features (Figure 5.2). The low temperature (~105
K) peak is attributed to CO
2
multilayer and monolayer desorption from ice, while the
two features at higher temperatures (>150 K) are from CO
2
molecules trapped in
ASW. Some of these trapped CO
2
molecules desorb during crystallization at ~165 K,
while the remaining species co-desorb with crystalline ice film. The tail at T > 180 K
is a result of co-desorption from the sample holder. This was deduced by varying the
sample holder position relative to the mass spectrometer aperture.
The spectral characteristics of the CO
2
molecules that are retained within
crystalline ice can be explored, without interference from CO
2
released during the
phase transition, by ice annealing to appropriate temperatures after CO
2
deposition.
As the TPD trace in Figure 5.3 shows, annealing temperatures of 165 K are suitable
for this purpose; the annealed film was re-cooled to ~90 K prior to obtaining the
TPD trace. Also shown in this figure is a TPD trace of a sample prepared in the same

109

Figure 5.2. CO
2
was deposited

(4 × 10
-8
Torr for 3 min) onto an ASW film (5 × 10
-8

Torr for 7 min) ~80 layers thick. The H
2
O desorption was monitored by measuring
m/e 18. Panels (a) and (b) show the TPD traces of CO
2
and H
2
O respectively. The
TPD trace of H
2
O was scaled for clarity.

way (i.e., same ASW thickness, deposited CO
2
, etc.) but annealed to ~105 K. The IR
spectra of these two samples (Figure 5.4), recorded after they had been re-cooled to
90 K and before obtaining the aforementioned TPD traces, suggests that CO
2

molecules that remain in crystalline ice possess significantly smaller oscillator
strengths than the CO
2
molecules escaping during the phase transition. Indeed, for
these samples, there is no distinguishable IR feature for the former species (Figure
5.4), in spite of the TPD data suggesting there are more of these species than that of
the latter (Figure 5.3).
110

Figure 5.3. TPD traces of CO
2
(4 × 10
-8
Torr for 3 min) deposited onto ASW films
(~80 layers). The temperature to which each sample was annealed prior to recording
FTIR spectrum is indicated. As the traces show, the CO
2
TPD feature at 185 K was
not affected by the annealing to 105 K.

Figure 5.4. Spectra (p-polarized) of ASW films (~80 layers) exposed to CO
2
(4 ×
10
-8
Torr for 3 min). Each sample was annealed to the temperature indicated and then
re-cooled to 90 K. The broad H
2
O feature (~3250 cm
-1
) changes upon annealing past
165 K because of the ASW-to-cubic phase transition. The inset shows the expanded
scale of the CO
2
ν
3
band absorbance region with the corresponding magnification
factor.

111
5.4 Discussion
Morphological changes in porous ASW films can be induced thermally, and
molecules that have to desorb by diffusing through ASW as these changes takes
place can become trapped within ASW. There are three ways to create a situation
whereby a species has to diffuse through ASW to desorb. One procedure involves
co-condensing H
2
O and another gas at low temperatures (less than 100 K) to form a
solid film with two different molecular species. Another method is to form an ASW
film on top of a solid film comprised of a different molecular species. Finally,
molecules can be deposited onto ASW at temperatures where they can adsorb into
the ASW surface but are able to diffuse relatively freely on this surface. These
molecules diffuse into ASW and have to make their way out of this ASW to desorb.
For each of these methods, the amount of guest molecules trapped within the film
depends on a set of factors unique to the method. For example, the composition of
the gas mixture and the deposition temperature influence the composition of the solid
film formed by co-condensing ASW and another gas species. In turn, this influences
the quantity of guest molecules that are subsequently trapped by annealing the film.
Some of these trapped molecules escape during the crystallization, and the
general consensus is that these molecules reside within ASW pores. However, the
location and nature of the molecules that are not released during this transition have
been a source of speculation. Initial considerations suggested these molecules may
be part of CO
2
clathrate hydrates domains within ice; whether these domains are
created during crystallization or they exist even before this process is not something
112
that has been addressed. While the IR signatures of CO
2
molecules in hydrates have
been reported, the relative absorption strengths of these CO
2
molecules compared to
CO
2
molecules in a CO
2
film appear to be unknown [24]. Thus, it is unclear if the
lack of a discernible IR feature stems from the fact that these are CO
2
molecules in a
clathrate hydrate, and if in these structures the absorption strength of CO
2
is severely
attenuated (relative to other CO
2
molecules in ice). Nevertheless, studies indicate that
CO
2
does not form clathrate hydrates readily at low pressures [24, 25]. As a result, it
is very unlikely that CO
2
molecules co-desorbing with H
2
O are from CO
2
hydrate
structures within ice.
There are two likely explanations as to why CO
2
molecules that co-desorb with
crystalline ice exhibit no discernible IR feature. The oscillator strength of the two
trapped species (i.e., those that escape during crystallization and those that do not)
could be truly different, and this suggests that the immediate surroundings for these
species are dissimilar. However, it is also possible that the CO
2
molecules
responsible for the TPD peak at ~185 K do not originate from the surface and that is
why they exhibit no IR feature. Investigations with a similar but different molecule,
specifically a CO
2
isotopologue (
13
CO
2
), are currently underway to determine which
of the two scenarios is more likely.
If a distinguishable IR feature for
13
CO
2
trapped in crystalline ice is observed,
this would indicate that the
12
CO
2
molecules detected in the
12
CO
2
TPD trace (see
Figure 5.3) do not originate from the surface. It is improbable that IR features of CO
2
are influenced strongly by the surrounding H
2
O molecules while the IR features of
113
13
CO
2
are not. An IR or TPD feature that can be ascribed to molecules retained
within crystalline ice will be used to explore whether these molecules are distributed
throughout the crystalline bulk, i.e., whether the intensity of this feature scales with
film thickness. Moreover, a comparison of the IR features from these two sets of
trapped species (i.e., those that escape during crystallization and those that do not)
will yield insight into the similarity, or dissimilarity, in the environment surrounding
these two groups. For example, it is possible that not all the ice crystallizes and that
it is ASW domains imbedded within the polycrystalline ice film that host the guest
molecules co-desorbing with the crystalline ice film. Finally, the factors that mediate
the ratio of the molecules released during the phase transition to molecules retained
within crystalline ice will be investigated by exploring how this ratio varies with
deposition technique (co-dosing or depositing CO
2
atop ASW films), film thickness,
and annealing rate. The last factor listed influences the crystallization kinetics and
which, in turn, may be a mediating factor with regards to the amount of molecules
released during the phase transition.

5.5 Epilogue
The aforementioned
13
CO
2
studies were completed during the process of writing
this dissertation.
13
CO
2
was deposited onto ASW films (~80 layers thick), and the
absence of a
13
CO
2
ν
3
IR feature after the phase transition agreed with the CO
2

results described in this chapter. However, the TPD for the
13
CO
2
isotopologue
revealed no
13
CO
2
molecules desorbing with the cubic ice film, and this was
114
consistent with the
13
CO
2
IR results. This suggests that the CO
2
TPD signal (i.e., m/e
44) observed during desorption of the H
2
O film does not originate from CO
2

molecules trapped in ice. The source of this CO
2
is currently unclear.
Experiments with thicker (> 80 layers) ASW films show that the thicker the
ASW film, the larger the amount of
13
CO
2
molecules that co-desorbs with the cubic
ice film; this is true when all factors except ASW film thickness are kept constant.
Moreover, the IR signatures for molecules released during the phase transition and
molecules retained within cubic ice are similar. Based on this similarity, it is
concluded that there is little difference in the local environments for these two sets of
molecules. Most likely, both groups consist of molecules trapped in pores within ice,
be it ASW or cubic ice.
115
5.6 Chapter 5 References
[1] J. A. Ghormley, Journal of Chemical Physics, 46, 1321, (1967).

[2] A. Barnun, J. Dror, E. Kochavi, D. Laufer, D. Kovetz and T. Owen, Origins
of Life and Evolution of the Biosphere, 16, 220, (1986).

[3] G. A. Kimmel, K. P. Stevenson, Z. Dohnalek, R. S. Smith and B. D. Kay, J.
Chem. Phys., 114, 5284, (2001).

[4] P. Ayotte, R. S. Smith, K. P. Stevenson, Z. Dohnalek, G. A. Kimmel and B.
D. Kay, Journal of Geophysical Research-Planets, 106, 33387, (2001).

[5] J. E. Schaff and J. T. Roberts, Langmuir, 14, 1478, (1998).

[6] M. P. Collings, M. A. Anderson, R. Chen, J. W. Dever, S. Viti, D. A.
Williams and M. R. S. McCoustra, Monthly Notices of the Royal Astronomical
Society, 354, 1133, (2004).

[7] J. D. Graham, J. T. Roberts, L. A. Brown and V. Vaida, Journal of Physical
Chemistry, 100, 3115, (1996).

[8] R. L. Hudson and B. Donn, Icarus, 94, 326, (1991).

[9] S. M. McClure, E. T. Barlow, M. C. Akin, D. J. Safarik, T. M. Truskett and
C. B. Mullins, Journal of Physical Chemistry B, 110, 17987, (2006).

[10] C. A. Angell, Science, 267, 1924, (1995).

[11] C. A. Angell, Annual Review of Physical Chemistry, 55, 559, (2004).

[12] O. Mishima and H. E. Stanley, Nature, 396, 329, (1998).

[13] R. S. Smith, Z. Dohnalek, G. A. Kimmel, K. P. Stevenson and B. D. Kay,
Chemical Physics, 258, 291, (2000).

[14] R. S. Smith and B. D. Kay, Nature, 398, 788, (1999).

[15] G. P. Johari, A. Hallbrucker and E. Mayer, Nature, 330, 552, (1987).

[16] S. Bahr, A. Borodin, O. Hofft, V. Kempter, A. Allouche, F. Borget and T.
Chiavassao, Journal of Physical Chemistry B, 110, 8649, (2006).

116
[17] S. M. McClure, D. J. Safarik, T. M. Truskett and C. B. Mullins, Journal of
Physical Chemistry B, 110, 11033, (2006).

[18] R. S. Smith, C. Huang, E. K. L. Wong and B. D. Kay, Surface Science, 367,
L13, (1996).

[19] R. S. Smith, C. Huang and B. D. Kay, Journal of Physical Chemistry B, 101,
6123, (1997).

[20] S. Mitlin and K. T. Leung, Canadian Journal of Chemistry-Revue
Canadienne De Chimie, 82, 978, (2004).

[21] J. Heidberg and B. Redlich, Surface Science, 368, 140, (1996).

[22] S. A. Hawkins: Fourier transform infrared spectroscopy and temperature
programmed desorption of water thin films on the MgO (100) surface, Ph. D. Thesis,
Department of Chemistry, University of Southern California, Los Angeles, 2004.

[23] M. P. Collings, J. W. Dever, H. J. Fraser and M. R. S. McCoustra,
Astrophysics and Space Science, 285, 633, (2003).

[24] F. Fleyfel and J. P. Devlin, Journal of Physical Chemistry, 95, 3811, (1991).

[25] G. Notesco and A. Bar-Nun, Icarus, 148, 456, (2000).

117
Chapter 6: Future Experiments
6.1 Amorphous Materials
An amorphous material, in the popular and essentially correct conception, is a
solid that possess no long range order with respect to the position of its constituents.
This relatively loose description allows a plethora of materials to be catalogued as
amorphous materials, and it sidesteps the current confusion in the literature
concerning the distinction between a glass and an amorphous material; there is a
tendency to use these two terms interchangeably, but some argue that this in
incorrect [1, 2]. Amorphous materials are not new. The manufacture of glassy
materials from silica commenced thousands of years ago. What is new is the
increased interest in these substances as novel materials, produced in amorphous
form, have become technologically important [1, 3-5]. In particular, there has been
increased focus on the fundamental processes that occur within such systems [3, 6-
8]. In attempts to predict the glass transition of various amorphous materials, these
studies are addressing why and how glasses form.
Amorphous ice has been a model system for studying amorphous materials [9-
11]. It is easily prepared, relatively free of contaminants, in simplified environments
(ultrahigh vacuum), and structural changes in this amorphous state can be monitored
using small probe molecules [12-15]. Experiments using this amorphous system have
supported the current consensus that the structure of all amorphous materials
depends upon formation conditions, and they have revealed that even at temperatures
well-below (~50 K) known phase transition temperatures, amorphous materials can
118
undergo molecular rearrangements [13]. In turn, these rearrangements influence the
molecular transport through this phase.
How can this system be modified to provide even more subtle details about
amorphous materials? One strategy is to introduce spatial heterogeneities into an
amorphous solid water (ASW) film in a controlled fashion, and then to examine how
these induced effects modify the characteristics of the system. The rational behind
such a strategy is easy to comprehend. The amorphous state is influenced by its
immediate surroundings, and to examine this influence, these surroundings should be
altered in a systematic way. For example, the creation of boundaries within an
amorphous system enhances the participation of these entities in processes occurring
within the system. This approach is akin to that in which ice nanocrystals were
created to enhance the ratio of surface to bulk H
2
O molecules (relative to that in an
H
2
O film) [16-18]. In those investigations, this enhanced ratio allowed surface
effects to be more easily distinguished.

6.2 Amorphous Ice: Transport and the Amorphous-Crystalline Interface
The creation of domains of crystalline ice distributed, essentially in a
homogenous fashion, throughout an ASW film would constitute a novel model
system from which additional information about amorphous materials may be
obtained (Figure 6.1). An ASW film annealed to ~130 K appears nonporous when
examined by probe molecules deposited onto the ASW surface. That said, it
important to bear in mind that this lack of porosity may not pervade throughout the
119
film. Morphological changes restricted to the first few film layers can in principle
limit the accessibility of probe molecules, deposited on ASW, to pores located within
the ASW interior. It has been suggested that during crystallization fractures are
created as the more compact crystalline ice phase "shrinks" away from the
surrounding ASW [19-21]. Thus, it is anticipated that creating crystalline domains
within a relatively nonporous ASW film will create fractures (or voids), and the
presence of these crevices will make the film porous. This porosity can be examined
using probe molecules, and it would be of interest to determine if the anticipated
results are realized.

Figure 6.1. An ASW film with domains of crystalline ice (shown in red) may
possess cracks or voids (shown in blue) at the amorphous-crystalline interface.

It would be constructive also to examine the crystallization dynamics of this
novel system, i.e., domains of crystalline ice embedded within ASW. Appropriate
annealing temperatures can be selected such that the process of ASW crystallization
occurs over minutes, and this process can be monitored using the IR spectra of ice or
probe molecules [14, 21-23]. In these isothermal experiments, heat is supplied
through the supporting substrate, i.e., supporting the film. Thin (~15 layers) ASW
120
films deposited onto crystalline ice exhibit dramatic increases in their crystallization
rate relative to films deposited onto a Pt(111) substrate. For example, it takes ~1000
seconds for a 15 layer ASW film deposited on Pt(111) to be half way to complete
crystallization at ~140 K. In contrast, a film of the same thickness formed on
crystalline ice is half crystallized in ~30 seconds at the same temperature. This result
is interpreted as a "seeding" effect in which the underlying crystalline ice film
obviates the need for a nucleation step to initiate crystallization [23]. Within the
semiconductor industry, it is known that substrate-induced crystallization can be
used to lower the crystallization temperature of a film [24]. Hence, it is expected that
an "ASW-crystalline ice array" (Figure 6.1) crystallizes over a shorter time scale
than a regular ASW film (containing the same amount of ASW that is in the array).
Indeed, it may be possible to find a temperature where these samples crystallize on
very different time scales, e.g., one in minutes and the other in hours.
The preceding discussion concentrated on a system comprising of crystalline ice
domains within ASW, and it addressed specifically the crystallization process. A
system in which isolated regions of ASW are formed on a supporting substrate
represents another novel system from which many subtleties about transport and
flow in amorphous materials may be gleaned (Figure 6.2). Based on the stability of
ASW films at temperatures less than 100 K (see chapter 4), it is anticipated that these
ASW regions of ~50 µm × 50 µm will be stable for several hours. But what occurs as
the substrate is heated slowly? If the glass transition of amorphous ice does indeed
occur prior to the phase transition and supercooled water in this temperature range is
121
a fragile liquid, then the voids surrounding these ASW "patches" may be filled as
H
2
O molecules diffuse into this region (Figure 6.2). Presumably, it is the fluidic
nature a fragile liquid exhibits at temperatures above its glass transition that would
permit such lateral mobility and changes in the system's boundary.
An amorphous ice film retains trapped species up until the phase transition [14,
25]. Consequently, if there is a glass transition before the phase transition, these
trapped species do not escape during the first process. However, it is unclear if this
would be the case still if the supercooled liquid formed during the glass transition
commences to flow. Naturally, this depends upon the dynamics of the flow process.
These molecules may be carried along, or they may be able to escape because of the
mechanism by which the molecular transport occurs. A priori, the outcome is not
apparent. And yet it is an outcome that has important implications for how molecular
transport in amorphous materials takes place just above the phase transition
temperature.

Figure 6.2. The slabs (shown in blue) represent isolated regions of ASW on a
supporting substrate. As these isolated regions approach the phase transition
temperature, there may be a significant increase in lateral mobility (indicated by the
black arrows) and material may be transported to the surrounding regions.

122
6.3 Experimental Strategy
An approach to obtaining new insights into the crystallization dynamics and
molecular transport in ice has been described. What has yet to be discussed is how
the samples needed for such an endeavor can be created. Specifically, this issue
concerns the introduction of morphological changes into an ASW film in a selective
and controlled manner.
The proposed experimental strategy is to fabricate such samples by pulsed laser
irradiation of ASW films through fine meshes with apertures of ~50 µm × 50 µm.
The wavelength of the laser radiation is chosen such that it excites H
2
O vibrations
specifically, and it does not excite the substrate. The spectral width of the ASW
absorption band at ~3400 cm
-1
is relatively large (~300 cm
-1
), and any radiation
within this range will be absorbed. The details about how energy pumped into a
vibrational mode of an H
2
O molecule, within an ASW film, leads to film annealing
are complex but qualitatively understood. Basically, energy implanted into an OH
excitation stretch degrades to heat which is then transferred to the surroundings, i.e.,
neighboring H
2
O molecules and the underlying substrate (which is kept cold). As a
result, once the irradiation ceases, the system cools efficiently. A similar laser
heating technique has been used to desorb water from the Ru(001) surface [26-28].
The proposed process involves first forming ASW films on a MgO(100)
substrate at low temperatures (less than 100 K). This substrate permits transmission
IR experiments in the region of 4000-2000 cm
-1
. Next, with a stainless steel mesh
positioned in front of the cold sample, the sample is irradiated with a pulsed laser
123
beam (Figure 6.3). Depending upon the laser fluence and duration of irradiation, all
H
2
O behind the exposed areas of the mesh will either be desorbed, remain
unchanged (e.g., if the laser fluence is insufficient), or undergo morphological
changes as a result of heating. Figure 6.3 depicts the situation in which all the H
2
O
exposed to the irradiation desorbs.

(a)
(b) (c)
y
x

Figure 6.3. (a) An ASW film is formed on a supporting substrate. (b) A mesh
(shown in black) is placed in front of the film, and the film is irradiated. (c) All the
ASW in the exposed areas desorb, leaving the structure depicted in blue. To form
isolated rows of ASW with the axes of the rows parallel to the y-axis shown, the
mesh can be translated in small increments along the x-axis. At each increment the
substrate is irradiated to desorb any exposed ASW. To form isolated areas of ASW
(the black squares), this process has to be done along both axes of the mesh (shown
as x and y).

Ideally, the control of the laser fluence and irradiation time would be suitable
enough to vary the amount of morphological change in the irradiated areas from
small (i.e., in the case of a relatively small increase in local temperature) to very
significant (i.e., induce a phase change). In addition, this control would permit also
124
laser-induced desorption of H
2
O. Attaining some semblance of this ideal situation
will be challenging, but it may be possible. A logical approach is to examine first the
effects of laser irradiation on ASW films without the mesh in place.
With regard to ASW films irradiated without a mesh in place, molecules still
remaining on the surface after irradiation are expected to be cooled efficiently by the
substrate once this process ceases. For the situation in which a significant amount of
H
2
O desorbs as a result of irradiation, a comparison of the H
2
O infrared (IR)
spectrum before and after irradiation should provide a good indication of this
change; there will be a marked reduction in H
2
O absorbance, and, most likely, the
ASW IR spectrum will have transformed to that displayed by crystalline ice. The
latter statement assumes the temperature within the film is relatively even. In
addition to IR spectroscopy, temperature programmed desorption (TPD) can be used
also to confirm desorption, and this involves the comparison of TPD spectra from
two different samples, prepared in the same manner, having the same thickness. Only
one sample is subjected to irradiation, and the TPD spectrum for this sample will
display a decrease in the H
2
O TPD intensity, relative to the other sample, if
desorption occurred during the irradiation. It is also possible to place the ASW
sample in front of the mass spectrometer while it is being irradiated to detect any
H
2
O desorption.
While there are several ways to confirm desorption, the detection of any
transformations from ASW to crystalline ice without significant desorption is limited
essentially to monitoring changes in the H
2
O IR spectrum. These changes may be
125
difficult to perceive. The main IR band of both its amorphous and crystalline phases
is broad, and there is a significant overlap between these two features. Nevertheless,
using these two bands, it is possible to de-convolute the IR band of a H
2
O film to
determine the ratio of crystalline to amorphous ice.
To detect the formation of crystalline ice within the sample in a more sensitive
manner, probe molecules with narrow IR features (e.g., CO
2
and N
2
O) can be
trapped in ASW films prior to the laser irradiation. This involves forming an ASW
film at low temperatures, exposing the film to a specific probe species, and then
annealing the film at ~110 K to trap these molecules within the ASW film. Any
regions of ASW converted to crystalline ice during laser irradiation will expel these
trapped species, and the IR feature of these molecules will display a decrease, when
compared to the spectrum taken prior to any laser irradiation.
Hence, a systematic variation in the irradiation parameters (fluence and duration
of laser exposure) can be used to determine a suitable set of conditions over which
crystalline ice is formed with little desorption. The conditions necessary to desorb an
ASW film can be ascertained in a similar fashion. Morphological changes that do not
result in desorption or crystallization will have to be detected using probe molecules.
Specifically, such changes will have to be followed by monitoring the trapping
capabilities of a film.
In this process, an ASW film of a certain thickness is created at low temperature
(< 100 K), and then this film is induced to trap probe molecules deposited onto its
surface by annealing to ~110 K (see chapter 4). Next, an IR and TPD spectrum of the
126
trapped species in this ASW sample is obtained. After desorbing this film, another
ASW film of the same thickness is formed under identical conditions, but this film is
then subjected to laser irradiation (for some specified laser fluence and duration).
Probe molecules are deposited onto this film, and subsequently it is induced to trap
some of these molecules. After this, the IR and TPD spectrum of the trapped species
is recorded. A comparison of the obtained IR or TPD spectra should be sufficient to
determine if laser irradiation produced any morphological changes that subsequently
influenced the trapping capabilities of the ASW film.
Once a set of conditions for inducing various types of changes in an ASW film
has been established, the effects of irradiation through a mesh can be explored.
Ideally, the goal is to limit the induced changes to regions of the film directly behind
the apertures of the mesh. However, there is the issue of thermal flow to the
surrounding areas not directly behind these apertures. For openings of 50 µm × 50
µm, the area of each irradiated region in contact with the surface is 2.5 × 10
-3
mm
2
,
and for a thin (~5 nm thick) film the combined areas of the edges for this region
(assuming a square slab shape) in contact with the surrounding H
2
O is 10
-6
mm
2
.
Consequently, it is anticipated that most of the heat flow will be to the substrate in
this situation. For thicker (~0.1 µm) films, it is unclear how much heat will flow
laterally, and this effect will have to be evaluated empirically.
The binding energy of the H
2
O monolayer (~0.7 eV) to a MgO(100) substrate is
significantly larger than the binding energy of the multilayers (~0.5 eV). In fact, this
difference in binding energies is exploited to produce H
2
O monolayers on this
127
substrate using thermal desorption [29]. Consequently, it is likely that there may be
experimental conditions over which laser induced multilayer desorption occurs but
the monolayer remains on the surface. Using a mesh, it may be possible to create
monolayer regions of H
2
O in between isolated regions of multilayer amorphous ice
(Figure 6.2). A comparison of experiments from this type of arrangement to those
from exactly the same arrangement but with the monolayer removed might yield
insights into the diffusion dynamics of H
2
O on a water monolayer as opposed to the
MgO surface.

6.4 Experimental Details: the Current Configuration
The UHV chamber that will be used for these experiments was described in
chapter 3 of this dissertation. Basically, it is a three-tiered chamber with a surface
manipulator attached to the top tier. This manipulator allows the surface to be moved
between tiers, and it provides XYZ translation and 360º rotation of the sample. The
sample is cooled typically by bubbling helium gas through a liquid nitrogen
reservoir.
The top tier of this chamber is used for FTIR experiments, and in these
experiments optical mirrors direct the IR beam of a commercial spectrometer
(Nicolet Protégé 460) in and out of the chamber using CaF
2
windows. The beam size
at the surface is ~8 mm. Also, this tier can accommodate leak valves and a mass
spectrometer. A mass spectrometer, fitted with a stainless steel cone having a small
(~9 mm in diameter) aperture is used for TPD experiments. This cone limits the
128
detected desorption during TPD to species originating from the surface. A grounded
stainless steel mesh (wire diameter ~0.025 mm, openings ~1.3 mm) covering the
aperture of this cone prevents stray electrons leaving the mass spectrometer
ionization region from reaching the adsorbate. The bottom tier of the chamber is
designed to house several diagnostics simultaneously, and hence it has numerous
ports.
The MgO(100) crystal substrate is prepared under dry nitrogen conditions,
quickly inserted into the chamber, and cleaned following established procedure. A
thermocouple, fixed to one edge of the crystal with a high temperature adhesive,
records temperature. A surface holder that permits TPD and IR experiments to be
performed mounts the sample onto the manipulator, and this holder allows the
sample to be cooled to ~90 K. The surface is heated using a homemade heater
attached to the surface holder, and with this heater the sample can be heated to ~500
K at rate of ~2 K/s.
Water coverage is obtained by exposing the substrate to a constant flux of water
vapor for fixed time periods, and the thickness of the water films is estimated by
comparing a film's integrated absorbance to that of a water monolayer on MgO(100).
Probe molecules are introduced into the chamber using a separate leak valve from
that used to vapor deposit water.

129
6.5 Experimental Details: the Modifications
None of the experiments proposed can take place without laser radiation of an
appropriate wavelength ( λ), i.e., within the region 3.225-2.857 µm (3100-3500 cm
-1
),
and hence generating this laser light is the first logical step in attempting the
proposed experiments. Laser light in this wavelength range has been produced using:
an Er:YAG laser at 2.94 µm [26], an infrared free electron laser (IRFEL) at 3 µm
[28], and a Nd:YAG laser (1.064 µm) coupled to a Raman cell to produce 2.93 µm
light (the second Stokes line) [30]. Essentially all the components necessary for the
last technique have already been acquired, and consequently, it is currently the most
convenient method of choice.
However, there are issues of how much laser fluence will be necessary, and how
much this technique can be expected to produce. Foster and others [30] generated
2.93 µm light by Raman shifting 1.064 µm light from a pulsed Nd:YAG (20 Hz, 8 ns
pulses) using a 1 m D
2
Raman cell at 900 psi, and the power produced was ~2.5 ×
10
3
W (20 µJ/pulse, 8 ns pulses). However, the Nd:YAG energy per pulse used to
generate this power was not specified, and so it is unclear if this was the maximum
power attainable. It is the vibrational energy spacings within D
2
(~3.34 µm (2990
cm
-1
)) that permits this Raman shift in wavelength.
Studies using an Er:YAG laser found that powers of ~2.5 × 10
3
W (0.25
mJ/pulse, 100 ns pulses) focused onto a spot of ~200 µm in diameter were required
to desorb H
2
O from an ice film at 140 K [26]. A power of ~1.5 × 10
6
W will be
required to match this energy density (~6 × 10
10
W/m
2
) over an area of ~25 mm
2
(5
130
mm × 5 mm), which is a suitable area for conducting the proposed experiments, and
this implies ~15 mJ/pulse for a 10 ns pulse. This is less than 2 % of the maximum
energy per pulse (~800 mJ/pulse, 10 ns pulses) for the 1.064 µm Nd:YAG radiation
that will be used to generate the second D
2
Stokes line in the proposed experiments.
Thus, it is likely that this energy requirement can be met. Because water adsorbed on
surfaces effectively absorbs radiation, precautions may have to be taken to protect
the optical components of the Raman shifter.
However, in the event that 15 mJ/pulse is not attained, a beam with a
significantly less energy per pulse than this value can be focused tightly (e.g., to a
diameter ~200 µm) onto the surface, and then this beam can be translated across the
film using mirrors on a piezoelectric translator. It is also plausible that this power
requirement is specific to the experimental configuration used. For example, the
desorption rate might have been too low to detect at power densities less than ~6 ×
10
10
W/m
2
, and so desorption, in these situations, is low as opposed to non-existent.
In addition, the proximity of a surface to the entrance of the mass spectrometer is a
factor in the detection of species desorbing from the surface, and the distance
separating the surface and mass spectrometer during the aforementioned Er:YAG
laser experiments [26] was not specified. Hence, these issues of how much power
can be generated and whether or not this power is enough must, most likely, be
determined empirically.
The laser light can be introduced into the chamber through an ultraviolet grade
sapphire view port positioned at one of the ports on the middle tier. The FTIR set-up
131
will be located on the top tier, and so the sample will have to be moved between tiers
for IR experiments. However, a mass spectrometer can be placed on the middle tier
to obviate the need to move between tiers for TPD experiments. To create domains
of crystalline ice within an ASW film, a wire mesh can be mounted on a separate
manipulator, and the ASW film can be moved behind it prior to laser radiation.
The experiments involving isolated regions of ASW on MgO(100) require a little
more care with respect to the placement of the mesh. It is anticipated that the specific
experimental set-up used to undertake these experiments will be an arrangement
primarily based on the knowledge accrued from the experiments before it.
Experiments, with a ~50 µm × 50 µm mesh, have already been performed to ensure
that IR spectroscopy can be carried out without moving the mesh from in front of the
ASW film. The open areas of the mesh allow enough light to reach the detector, and
diffraction is not an issue.

6.6 Summary
The increased interest in amorphous materials is focusing attention on the
fundamental physics of such systems. In particular, amorphous ice has garnered
special attention because of its applicability as a model for liquid water [9], its
existence as a major component of interstellar ices [31], and its debated physical
properties. In addition, it has been a model system on which to explore the physics of
amorphous materials.
132
This chapter has outlined a strategy to modify the currently explored amorphous
and crystalline ice systems to create a novel system that will provide new insights
into the properties of amorphous materials. Specifically, it is about introducing
spatial heterogeneities into an ASW film in a controlled fashion and probing the
morphological changes and molecular transport in this amorphous material. The
experiments proposed, although specific to amorphous ice, are designed to address
questions germane to all amorphous systems.
133
6.3 Chapter 6 References
[1] S. R. Elliot, Physics of Amorphous Materials, 2 ed., Longman Scientific and
Technical: Essex, 1990.

[2] R. Zallen, The Physics of Amorphous Solids, John Wiley and Sons, Inc.: New
York, 1983.

[3] C. A. Angell, Science, 267, 1924, (1995).

[4] V. M. Fokin, E. D. Zanotto, N. S. Yuritsyn and J. W. P. Schmelzer, Journal
of Non-Crystalline Solids, 352, 2681, (2006).

[5] G. Frenzer and W. F. Maier, Annual Review of Materials Research, 36, 281,
(2006).

[6] M. Wyart, Annales De Physique, 30, 1, (2005).

[7] F. H. Stillinger, Science, 267, 1935, (1995).

[8] P. G. Debenedetti and F. H. Stillinger, Nature, 410, 259, (2001).

[9] C. A. Angell, Annual Review of Physical Chemistry, 55, 559, (2004).

[10] C. A. Angell, Annual Review of Physical Chemistry, 34, 593, (1983).

[11] G. P. Johari, A. Hallbrucker and E. Mayer, Nature, 330, 552, (1987).

[12] C. Manca, C. Martin and P. Roubin, Chemical Physics, 300, 53, (2004).

[13] C. Martin, C. Manca and P. Roubin, Surface Science, 502, 275, (2002).

[14] G. A. Kimmel, K. P. Stevenson, Z. Dohnalek, R. S. Smith and B. D. Kay, J.
Chem. Phys., 114, 5284, (2001).

[15] D. Laufer, E. Kochavi and A. Barnun, Physical Review B, 36, 9219, (1987).

[16] B. Rowland and J. P. Devlin, Journal of Chemical Physics, 94, 812, (1991).

[17] B. Rowland, M. Fisher and J. P. Devlin, Journal of Chemical Physics, 95,
1378, (1991).

[18] V. Buch, J. Sadlej, N. Aytemiz-Uras and J. P. Devlin, J. Phys. Chem. A, 106,
9374, (2002).
134
[19] S. M. McClure, D. J. Safarik, T. M. Truskett and C. B. Mullins, Journal of
Physical Chemistry B, 110, 11033, (2006).

[20] S. M. McClure, E. T. Barlow, M. C. Akin, D. J. Safarik, T. M. Truskett and
C. B. Mullins, Journal of Physical Chemistry B, 110, 17987, (2006).

[21] R. S. Smith, C. Huang, E. K. L. Wong and B. D. Kay, Physical Review
Letters, 79, 909, (1997).

[22] K. P. Stevenson, G. A. Kimmel, Z. Dohnalek, R. S. Smith and B. D. Kay,
Science, 283, 1505, (1999).

[23] Z. Dohnalek, G. A. Kimmel, R. L. Ciolli, K. P. Stevenson, R. S. Smith and B.
D. Kay, Journal of Chemical Physics, 112, 5932, (2000).

[24] C. Spinella, S. Lombardo and F. Priolo, Journal of Applied Physics, 84, 5383,
(1998).

[25] G. Kumi, S. Malyk, S. Hawkins, H. Reisler and C. Wittig, Journal of
Physical Chemistry A, 110, 2097, (2006).

[26] F. E. Livingston, J. A. Smith and S. M. George, Analytical Chemistry, 72,
5590, (2000).

[27] F. E. Livingston and S. M. George, Journal of Physical Chemistry A, 105,
5155, (2001).

[28] A. Krasnopoler and S. M. George, Journal of Physical Chemistry B, 102,
788, (1998).

[29] S. Hawkins, G. Kumi, S. Malyk, H. Reisler and C. Wittig, Chemical Physics
Letters, 404, 19, (2005).

[30] N. S. Foster, J. E. Amonette, T. Autrey and J. T. Ho, Sensors and Actuators
B-Chemical, 77, 620, (2001).

[31] E. Mayer and R. Pletzer, Nature, 319, 298, (1986).

135
Bibliography
[1] L. J. Allamandola, S. A. Sandford and G. J. Valero, Icarus, 76, 225, (1988).

[2] A. Allouche, P. Verlaque and J. Pourcin, Journal of Physical Chemistry B,
102, 89, (1998).

[3] P. U. Andersson, M. B. Nagard, G. Witt and J. B. C. Pettersson, Journal of
Physical Chemistry A, 108, 4627, (2004).

[4] P. U. Andersson, M. B. Nagard, G. Witt and J. B. C. Pettersson, Journal of
Physical Chemistry A, 108, 4627, (2004).

[5] C. A. Angell, Annual Review of Physical Chemistry, 34, 593, (1983).

[6] C. A. Angell, Science, 267, 1924, (1995).

[7] C. A. Angell, Annual Review of Physical Chemistry, 55, 559, (2004).

[8] G. B. Arfken and H. J. Weber, Mathematical Methods for Physicists,
Academic Press: San Deigo, 1995.

[9] P. Atkins and R. Freidman, Molecular Quantum Mechanics, Oxford
University Press: Oxford, 1998.

[10] P. Ayotte, R. S. Smith, K. P. Stevenson, Z. Dohnalek, G. A. Kimmel and B.
D. Kay, Journal of Geophysical Research-Planets, 106, 33387, (2001).

[11] P. Ayotte, R. S. Smith, K. P. Stevenson, Z. Dohnalek, G. A. Kimmel and B.
D. Kay, Journal of Geophysical Research, 106, 33837, (2001).

[12] S. Bahr, A. Borodin, O. Hofft, V. Kempter, A. Allouche, F. Borget and T.
Chiavassao, Journal of Physical Chemistry B, 110, 8649, (2006).

[13] A. Bar-Nun, J. Dror, E. Kochavi and D. Laufer, Physical Review B, 35, 2427,
(1987).

[14] R. A. Baragiola, In Water in Confining Geometries, V. Buch and J. P. Devlin,
Eds.; Springer, Berlin, 2003.

[15] A. Barnun, G. Herman, D. Laufer and M. L. Rappaport, Icarus, 63, 317,
(1985).

136
[16] A. Barnun, J. Dror, E. Kochavi, D. Laufer, D. Kovetz and T. Owen, Origins
of Life and Evolution of the Biosphere, 16, 220, (1986).

[17] A. Barnun, J. Dror, E. Kochavi and D. Laufer, Physical Review B, 35, 2427,
(1987).

[18] O. Berg and G. E. Ewing, Surface Science, 220, 207, (1989).

[19] B. S. Berland, D. E. Brown, M. A. Tolbert and S. M. George, Geophysical
Research Letters, 22, 3493, (1995).

[20] M. P. Bernstein, J. P. Dworkin, S. A. Sandford, G. W. Cooper and L. J.
Allamandola, Nature, 416, 401, (2002).

[21] J. E. Bertie, L. D. Calvert and E. Whalley, Journal of Chemical Physics, 38,
840, (1963).

[22] C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by
Small Particles, John Wiley & Sons, Inc.: New York, 1998.

[23] F. Borget, T. Chiavassa, A. Allouche and J. P. Aycard, Journal of Physical
Chemistry B, 105, 449, (2001).

[24] V. Buch, L. Delzeit, C. Blackledge and J. P. Devlin, Journal of Physical
Chemistry, 100, 3732, (1996).

[25] V. Buch, J. Sadlej, N. Aytemiz-Uras and J. P. Devlin, Journal of Physical
Chemistry A, 106, 9374, (2002).

[26] V. Buch and J. P. Devlin, In Water in Confining Geometries, V. Buch and J.
P. Devlin, Eds.; Springer, Berlin, 2003.

[27] E. F. Burton and W. F. Oliver, Proceedings Royal Society, A153, 166, (1935).

[28] C. M. Chan, R. Aris and W. H. Weinberg, Surface Science, 1, 360, (1978).

[29] H. Chang, H. H. Richardson and G. E. Ewing, Journal of Chemical Physics,
89, 7561, (1988).

[30] L. Chu and L. T. Chu, Journal of Physical Chemistry A, 103, 691, (1999).

[31] M. P. Collings, J. W. Dever, H. J. Fraser and M. R. S. McCoustra,
Astrophysics and Space Science, 285, 633, (2003).
137
[32] M. P. Collings, M. A. Anderson, R. Chen, J. W. Dever, S. Viti, D. A.
Williams and M. R. S. McCoustra, Monthly Notices of the Royal Astronomical
Society, 354, 1133, (2004).

[33] P. G. Debenedetti and F. H. Stillinger, Nature, 410, 259, (2001).

[34] J. P. Devlin, Journal of Physical Chemistry, 96, 6185, (1992).

[35] Z. Dohnalek, G. A. Kimmel, R. L. Ciolli, K. P. Stevenson, R. S. Smith and B.
D. Kay, Journal of Chemical Physics, 112, 5932, (2000).

[36] J. Dubochet and A. W. McDowall, Journal of Microscopy-Oxford, 124, RP3,
(1981).

[37] S. R. Elliot, Physics of Amorphous Materials, 2 ed., Longman Scientific and
Technical: Essex, 1990.

[38] G. Faraudo and D. E. Weibel, Progress in Reaction Kinetics and Mechanism,
26, 179, (2001).

[39] D. Ferry, A. Glebov, V. Senz, J. Suzzane, J. P. Toennies and H. Weiss,
Journal of Chemical Physics, 105, 1697, (1996).

[40] F. Fleyfel and J. P. Devlin, Journal of Physical Chemistry, 95, 3811, (1991).

[41] V. M. Fokin, E. D. Zanotto, N. S. Yuritsyn and J. W. P. Schmelzer, Journal
of Non-Crystalline Solids, 352, 2681, (2006).

[42] K. L. Foster, M. A. Tolbert and S. M. George, Journal of Physical Chemistry
A, 101, 4979, (1997).

[43] N. S. Foster, J. E. Amonette, T. Autrey and J. T. Ho, Sensors and Actuators
B-Chemical, 77, 620, (2001).

[44] F. Franks, Polywater, MIT press: Cambridge, 1981.

[45] F. Franks, In Water: A comprehensive treatise, F. Franks, Ed.; Plenum Press,
New York, 1982, Vol. 7.

[46] G. Frenzer and W. F. Maier, Annual Review of Materials Research, 36, 281,
(2006).

[47] J. A. Ghormley, Journal of Chemical Physics, 46, 1321, (1967).

138
[48] A. Givan, A. Loewenshuss and C. J. Nielson, Journal of Physical Chemistry
B, 101, 8696, (1997).

[49] J. D. Graham and J. T. Roberts, Journal of Physical Chemistry, 98, 5974,
(1994).

[50] J. D. Graham, J. T. Roberts, L. A. Brown and V. Vaida, Journal of Physical
Chemistry, 100, 3115, (1996).

[51] P. Griffiths and J. A. De Haseth, Fourier Transform Infrared Spectrometry,
John Wiley and Sons, Inc.: New York, 1986.

[52] A. Hallbrucker, E. Mayer and G. P. Johari, Journal of Physical Chemistry,
93, 4986, (1989).

[53] Y. P. Handa, O. Mishima and E. Whalley, Journal of Chemical Physics, 84,
2766, (1986).

[54] S. Haq, J. Harnett and A. Hodgson, Journal of Physical Chemistry B, 106,
3950, (2002).

[55] S. A. Hawkins: Fourier transform infrared spectroscopy and temperature
programmed desorption of water thin films on the MgO (100) surface, Ph. D. Thesis,
Department of Chemistry, University of Southern California, Los Angeles, 2004.

[56] S. Hawkins, G. Kumi, S. Malyk, H. Reisler and C. Wittig, Chemical Physics
Letters, 404, 19, (2004).

[57] S. Hawkins, G. Kumi, S. Malyk, H. Reisler and C. Wittig, Chemical Physics
Letters, 404, 19, (2005).

[58] J. Heidberg, B. Redlich and D. Wetter, Berichte Bunsenges Physical
Chemistry, 99, 1333, (1995).

[59] J. Heidberg and B. Redlich, Surface Science, 368, 140, (1996).

[60] G. Herzberg, Infrared and Raman Spectra, Van Nostrand Reinhold
Company, Inc: New York, 1945.

[61] P. V. Hobbs, Ice Physics, Claderon Press: Oxford, 1974.

[62] L. K. Hodgson: Photodissociation, molecule-surface collision-induced
dissociation and direct adsorbate photolysis of nitroso molecules, Ph. D. Thesis,
Department of Chemistry, University of Southern California, Los Angeles, 1993.
139
[63] M. J. Hollas, High Resolution Spectroscopy, John Wiley & Sons, Inc.:
Chichester, 1998.

[64] A. B. Horn, J. R. Sodeau, T. B. Roddis and N. A. Williams, Journal of
Physical Chemistry A, 102, 6107, (1998).

[65] R. L. Hudson and B. Donn, Icarus, 94, 326, (1991).

[66] T. Huthwelker, M. Ammann and T. Peter, Chemical Reviews, 106, 1375,
(2006).

[67] J. F. James, A student's guide to Fourier transforms, Cambridge University
Press: Cambridge, 1998.

[68] P. Jenniskens and D. F. Blake, Science, 265, 753, (1994).

[69] P. Jenniskens, S. F. Banham, D. F. Blake and M. R. S. McCoustra, Journal of
Chemical Physics, 107, 1232, (1997).

[70] G. P. Johari, A. Hallbrucker and E. Mayer, Nature, 330, 552, (1987).

[71] A. M. Jong and J. W. Niemantsverdriet, Vacuum, 41, 232, (1990).

[72] G. A. Kimmel, K. P. Stevenson, Z. Dohnalek, R. S. Smith and B. D. Kay,
Journal of Chemical Physics, 114, 5284, (2001).

[73] D. A. King, Surface Science, 47, 384, (1975).

[74] A. Krasnopoler and S. M. George, Journal of Physical Chemistry B, 102,
788, (1998).

[75] G. J. Kroes and D. C. Clary, Journal of Physical Chemistry, 96, 7079, (1992).

[76] G. Kumi, S. Malyk, S. Hawkins, H. Reisler and C. Wittig, Journal of
Physical Chemistry A, 110, 2097, (2006).

[77] B. Kusse and E. Westwig, Mathematical Physics, John Wiley and Sons, Inc.:
New York, 1998.

[78] D. Laufer, E. Kochavi and A. Barnun, Physical Review B, 36, 9219, (1987).

[79] F. E. Livingston, J. A. Smith and S. M. George, Analytical Chemistry, 72,
5590, (2000).
140
[80] F. E. Livingston and S. M. George, Journal of Physical Chemistry A, 105,
5155, (2001).

[81] C. Manca and A. Allouche, Journal of Chemical Physics, 114, 4226, (2001).

[82] C. Manca, C. Martin and P. Roubin, Chemical Physics Letters, 364, 220,
(2002).

[83] C. Manca, C. Martin and P. Roubin, Chemical Physics, 300, 53, (2004).

[84] C. Manca, C. Martin and P. Roubin, Chemical Physics, 300, 53, (2004).

[85] Y. A. Mantz, F. M. Geiger, L. T. Molina, M. J. Molina and B. L. Trout,
Chemical Physics Letters, 348, 285, (2001).

[86] C. Martin, C. Manca and P. Roubin, Surface Science, 502, 275, (2002).

[87] C. Martin, C. Manca and P. Roubin, Surface Science, 502-503, 275, (2002).

[88] M. Matsumoto, S. Saito and I. Ohmine, Nature, 416, 409–413, (2002).

[89] E. Mayer and R. Pletzer, Nature, 319, 298, (1986).

[90] S. M. McClure, E. T. Barlow, M. C. Akin, D. J. Safarik, T. M. Truskett and
C. B. Mullins, Journal of Physical Chemistry B, 110, 17987, (2006).

[91] S. M. McClure, D. J. Safarik, T. M. Truskett and C. B. Mullins, Journal of
Physical Chemistry B, 110, 11033, (2006).

[92] L. Mertz, Astronomical Journal, 70, 685, (1965).

[93] O. Mishima, L. D. Calvert and E. Whalley, Nature, 310, 393, (1984).

[94] O. Mishima, L. D. Calvert and E. Whalley, Nature, 314, 76, (1985).

[95] O. Mishima and H. E. Stanley, Nature, 396, 329, (1998).

[96] S. Mitlin and K. T. Leung, Canadian Journal of Chemistry-Revue
Canadienne De Chimie, 82, 978, (2004).

[97] A. H. Narten, C. G. Venkatesh and S. A. Rice, Journal of Chemical Physics,
64, 1106, (1976).

141
[98] D. Nordlund, H. Ogasawara, P. Wernet, M. Nyberg, M. Odelius, L. G. M.
Pettersson and A. Nilsson, Chemical Physics Letters, 395, 161, (2004).

[99] R. H. Norton and R. Beer, Journal of the Optical Society of America, 66, 259,
(1976).

[100] G. Notesco and A. Bar-Nun, Icarus, 148, 456, (2000).

[101] M. A. Ovchinnikov and C. A. Wight, Journal of Chemical Physics, 99, 3374,
(1993).

[102] M. E. Palumbo, Journal of Physical Chemistry A, 101, 4298, (1997).

[103] M. E. Palumbo, Journal of Physical Chemistry A, 101, 4298, (1997).

[104] R. Pletzer and E. Meyer, Journal of Chemical Physics, 90, 5207, (1989).

[105] P. A. Redhead, Vacuum, 12, 203, (1962).

[106] B. Roland and J. P. Devlin, Journal of Chemical Physics, 94, 812, (1991).

[107] T. L. Roush, Journal of Geophysical Research-Planets, 106, 33315, (2001).

[108] B. Rowland, M. Fisher and J. P. Devlin, Journal of Chemical Physics, 95,
1378, (1991).

[109] B. Rowland and J. P. Devlin, Journal of Chemical Physics, 94, 812, (1991).

[110] J. Sadlej, B. Rowland, J. P. Devlin and V. Buch, Journal of Chemical
Physics, 102, 4804, (1995).

[111] S. A. Sandford and L. J. Allamandola, Icarus, 76, 201, (1988).

[112] S. A. Sandford and L. J. Allamandola, Astrophysical Journal, 355, 357,
(1990).

[113] S. A. Sandford and L. J. Allamandola, Journal of Astrophysics, 355, 357,
(1990).

[114] M. G. Sceats and S. A. Rice, In Water: A comprehensive treatise, F. Franks,
Ed.; Plenum Press, New York, 1982, Vol. 7.

[115] J. E. Schaff and J. T. Roberts, Journal of Physical Chemistry, 98, 6900,
(1994).
142
[116] J. E. Schaff and J. T. Roberts, Journal of Physical Chemistry, 100, 14151,
(1996).

[117] J. E. Schaff and J. T. Roberts, Langmuir, 14, 1478, (1998).

[118] B. Schmitt, J. Ocampo and J. Klinger, Journal De Physique, 48, 519, (1987).

[119] H. Schober, M. M. Koza, A. Tolle, C. Masciovecchio, F. Sette and F. Fujara,
Physical Review Letters, 85, 4100, (2000).

[120] L. Schriver-Mazzuoli, A. Schriver and A. Hallou, Journal of Molecular
Structure, 554, 289, (2000).

[121] Sheikh, II and P. D. Townsend, Journal of Physics E-Scientific Instruments,
6, 1170, (1973).

[122] S. C. Silva and J. P. Devlin, Journal of Chemical Physics, 98, 10847, (1994).

[123] D. A. Skoog and J. L. Leary, Principles of Instrumental Analysis, Harcourt
Brace College Publishers: Fort Worth, 1992.

[124] B. C. Smith, Fourier Transform Infrared Spectroscopy, CRC Press: Boca
Raton, 1996.

[125] R. S. Smith, C. Huang, E. K. L. Wong and B. D. Kay, Surface Science, 367,
L13, (1996).

[126] R. S. Smith, C. Huang and B. D. Kay, Journal of Physical Chemistry B, 101,
6123, (1997).

[127] R. S. Smith, C. Huang, E. K. L. Wong and B. D. Kay, Physical Review
Letters, 79, 909, (1997).

[128] R. S. Smith and B. D. Kay, Surface Review Letters, 4, 781, (1997).

[129] R. S. Smith and B. D. Kay, Nature, 398, 788, (1999).

[130] R. S. Smith, Z. Dohnalek, G. A. Kimmel, K. P. Stevenson and B. D. Kay,
Chemical Physics, 258, 291, (2000).

[131] G. D. Smith and R. A. Palmer, In Handbook of Vibrational Spectroscopy, J.
M. Chalmers and P. R. Griffiths, Eds.; John Wiley and Sons, Inc., New York, 2002,
Vol. 1.
143
[132] R. S. Smith, Z. Dohnalek, G. A. Kimmel, G. Teeter, P. Ayotte, J. L.
Daschbach and B. D. Kay, In Water in Confining Geometries, V. Buch and J. P.
Devlin, Eds.; Springer, Berlin, 2003.

[133] G. A. Somorjai, Introduction to Surface Chemistry and Catalysis, John Wiley
and Sons, Inc: New York, 1994.

[134] C. Spinella, S. Lombardo and F. Priolo, Journal of Applied Physics, 84, 5383,
(1998).

[135] K. P. Stevenson, G. A. Kimmel, Z. Dohnalek, R. S. Smith and B. D. Kay,
Science, 283, 1505, (1999).

[136] F. H. Stillinger, Science, 267, 1935, (1995).

[137] W. S. Struve, Fundamentals of Molecular Spectroscopy, John Wiley and
Sons, Inc.: New York, 1989.

[138] M. M. Suchan: Molecules-surface interactions in HCl/MgO and Water/MgO
Systems, Ph. D. Thesis, Department of Chemistry, University of Southern California,
Los Angeles, 2001.

[139] M. Svanberg, J. B. C. Pettersson and K. Bolton, Journal of Physical
Chemistry A, 104, 5787, (2000).

[140] T. Takoaka, M. Inamura, S. Yanagimachi, I. Kusunoki and T. Komeda,
Journal of Chemical Physics, 121, 1 September 2004, (2004).

[141] C. A. Tulk, C. J. Benmore, J. Urquidi, D. D. Klug, J. Neuefeind, B. Tomberli
and P. A. Egelstaff, Science, 297, 1320, (2002).

[142] J. C. Tully, Surface Science, 299/300, 667, (1994).

[143] N. Uras, M. Rahman and J. P. Devlin, Journal of Physical Chemistry B, 102,
9375, (1998).

[144] Z. Y. Wang and S. K. Zhou, Progress in Chemistry, 16, 49, (2004).

[145] M. Wyart, Annales De Physique, 30, 1, (2005).

[146] H. Yamada and W. B. Person, Journal of Chemical Physics, 41, 2478,
(1964).

144
[147] J. Yates, J. T., Experimental Innovations in Surface Science, AIP Press
Springer-Verlag: New York, 1998.

[148] X. J. Yu and H. S. Kwok, Journal of Applied Physics, 93, 4407, (2003).

[149] R. Zallen, The Physics of Amorphous Solids, John Wiley and Sons, Inc.: New
York, 1983.

[150] Y. Zubavichus, M. Zharnikov, Y. J. Yang, O. Fuchs, E. Umbach, C. Heske
and M. Grunze, Langmuir, 22, 7241, (2006).

Abstract (if available)

Abstract Guest-host interactions in amorphous solid water (ASW) films doped with CO2 or N2O were examined experimentally. Investigations focused on exploring molecular transport and morphology in ASW. The main diagnostics were Fourier transform infrared (FTIR) spectroscopy and temperature programmed desorption (TPD).

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Transport and guest-host interactions in amorphous and crystalline ice

PDF

An investigation of morphology and transport in amorphous solid water via guest-host interactions

PDF

Energy transfer in amorphous solid water: light-mediated expulsion of N₂O₄ guest molecules

Asset Metadata

Creator Kumi, George A. (author)

Core Title Fourier transform infrared studies of guest-host interactions in ice

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry (Chemical Physics)

Publication Date 06/05/2007

Defense Date 05/03/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag amorphous ice,ASW,guest-host interactions,infrared spectroscopy,OAI-PMH Harvest

Language English

Advisor Reisler, Hannah (committee member), Wittig, Curt (committee member), Zhou, Chongwu (committee member)

Creator Email gkumi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m508

Unique identifier UC1421362

Identifier etd-Kumi-20070605 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-496069 (legacy record id),usctheses-m508 (legacy record id)

Legacy Identifier etd-Kumi-20070605.pdf

Dmrecord 496069

Document Type Dissertation

Rights Kumi, George A.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu