STUDYING THE DISSOCIATION DYNAMICS

OF CARBON DIOXIDE
Supervisor: Rami Sankari

Claudiu Bulbucan

Master Thesis
2016
STUDYING THE DISSOCIATION DYNAMICS OF CARBON DIOXIDE

© 2016/17... Claudiu Bulbucan

All rights reserved
Printed in Sweden by PRINT NAME, Lund, 201/17...

Cover: Picture of MAX IV Laboratory, 2016

MAX IV Laboratory, Lund University

P.O. Box 118
SE–221 00 Lund
Sweden
[Link]
If I were not a physicist, I would probably be
a musician. I often think in music. I live my
daydreams in music. I see my life in terms of
music.
Albert Einstein
ABSTRACT

Small molecules, and their fragmentation to smaller constituents

upon excitation, are often used to understand how the bonds in a
molecule are formed and how they break. One of such widely stud-
ied model molecule is CO2 .
This study aims at extending the knowledge of dissociation of car-
bon dioxide by measuring also the negatively charged fragments. In
addition, these anions are linked to positive ions (cations) which are
more abundant, and their production pathways are widely studied.
The measurements were carried at Elettra Synchrotron Radiation
Laboratory in Trieste, Italy. Provided that anion production at core-
level excitation is far less abundant than that of electrons and cations,
a new spectrometer was developed precisely for the purpose of de-
tecting the negative ions. It works in tandem with a positive ion spec-
trometer, mounted together for coincidence measurements. The
properties and advantages of the present instrument, such as high
acceptance angle and magnetic deflection of unwanted electrons are
described.
CO2 molecule is excited by photons which are energetic enough
to promote core electrons of the oxygen atom to unoccupied molec-
ular orbitals in CO2 - the following dissociation pathways are then
studied by time-of-flight based mass spectroscopy. Both singly-
charged and doubly-charged cations are found, in varying propor-
tions, at different photon energies. Also, negative ions are seen, with
C – being present only at the O1s edge but not at lower excitation
energies. Furthermore, metastable states and triply-charged cations
are a possibility, however, their complete analysis requires more at-
tention. They are nonetheless discussed and analyzed to some ex-
tent.

v
POPULAR SCIENCE WRITING (IN
ROMANIAN)

Dioxidul de carbon este unul dintre cei mai important, i compus, i

chimici pentru sust, inerea viet, ii pe Terra. Acesta joacă un rol crucial
în procesul de fotosinteză prin care plantele reglează natural canti-
tatea necesară de oxigen din atmosferă pentru fiint, ele aerobe, inclu-
siv oamenii.
Acesta a fost descoperit in anul 1640 de către Jan Baptist
van Helmont care a observat că în urma procesului de ardere
al cărbunelui, este o diferent, ă substant, ială de masă între pro-
dusul initial (cărbunele) s, i rămăs, it, ele rezultate (cenus, a). Fizicianul
antement, ionat a atribuit aparit, ia decalajului unei substant, e inviz-
ibile pe care a numit-o gaz sau „spiritus silvestre” (spirit sălbatic),
CO2 devenind astfel primul gaz considerat compus chimic de sine
stătător. Odată cu dezvoltarea s, tiint, elor naturii (fizică, chimie)
s-a ajuns la concluzia că este format din micro-particule numite
molecule care, la randul lor sunt formate din as, a numit, ii atomi, cei
din urmă formând legături pentru a constitui moleculele. Aceste
legături au putut fi explicate mai târziu (secolul XX) cu ajutorul
mecanicii cuantice.
Dioxidul de carbon este un gas incolor, inodor în concentrat, ii
scăzute, prezentând un miros înt, epător la concentrat, ii ridicate. Este
de aproximativ 1.6 ori mai dens decât aerul s, i nu poate fi lichefiat în
condit, ii normale de presiune (doar la presiuni înalte). Racit la tem-
peraturi sub -78.50 C acesta devine solid, formând astfel gheat, a car-
bonică sau gheat, a „uscată”, întâlnită deseori în industrie.
O posibilitate de a studia molecula de CO2 este prin iradiere cu
raze X. Razele X sunt lumină de energie mai înaltă comparativ cu lu-
mina vizibilă percepută de ochiul uman, iar energia este purtată de
as, a numit, ii fotoni. Atomii de oxigen s, i carbon ce constituie molec-
ula sunt format, i din protoni, neutroni s, i electroni. Protonii s, i neu-
tronii formează nucleele atomice iar electronii „orbitează” în jurul
acestora pe orbitali electronici. În interact, iunea lor cu radiat, ia, elec-
tronii pot „sări” pe un orbital superior sau pot părăsi molecula, pro-
movând astfel sistemul într-o stare energetică superioară sau într-o

vii
Popular science writing (in Romanian)

stare de ionizare. Mecanica cuantică spune că energia luminii inci-

dente trebuie să fie precis potrivită cu diferent, ele valorilor energetice
aferente orbitalilor în discut, ie, energii de obicei în domeniul razelor
X.
Datorită nevoilor ridicate de precizie a energiei radiat, iei, precum
s, i intensităt, ii luminoase, sursele convent, ionale de raze X nu sunt în-
totdeauna cele mai bune alegeri în astfel de experimente. Se poate
apela însă la radiat, ia sincrotronului, un accelerator de particule cir-
cular care cres, te energia electronilor la valori înalte (la viteze apropi-
ate de viteza luminii). Pentru a avea o idee de vitezele implicate, luăm
un calcul simplu pentru actualul sincrotron de la MAX IV din Lund.
Având în vedere că acesta are 532 m în circumferint, ă s, i electronii se
deplasează cu 99.99 % din viteza luminii (300.000 km/s), vor par-
curge acceleratorul de peste 500.000 de ori pe secundă. Electronii
sunt trimis, i în elemente magnetice construite pentru a imprima
oscilat, ii traiectoriei lor, fiecare electron producând astfel un foton
la fiecare „curbă” pe care o face. Energia radiat, iei produse depinde
de energia electronilor s, i de intensitatea câmpului magnetic la care
sunt supus, i, facând astfel posibilă reglarea ei cu mare acuratet, e. În
plus, numărul fotonilor produs, i este ridicat, asigurând astfel un flux
cu câteva ordine de mărime mai consistent decât în cazul surselor
obis, nuite (i.e. intensitate luminoasă ridicată).
As, adar, avem calea de a aduce molecula într-o stare energetică
excitată sau ionizată. În cazul ionizării, molecula nou formată (cu
un electron lipsă) devine instabilă s, i, ca toate entităt, ile din natură,
tinde la echilibru. Una dintre metodele de dezexcitare este ruperea
în fragmente ionice moleculare formate din atomii constituent, i.
Având în vedere că fragmentele nu sunt neutre d.p.d.v. electric
(de aceea numite ionice), în prezent, a unui câmp electric exterior
vor putea fi accelerate in direct, ii diferite (diferent, a venind în urma
semnului sarcinii + sau -). Spectroscopia bazată pe „timpul de
zbor” măsoară timpul pe care aceste fragmente îl fac de la produc-
ere până la detectarea lor. Astfel, se pot identifica fragmentele re-
spective, informat, ii utile pentru determinarea modelului după care
moleculele disociază, obt, inând astfel informat, ii esent, iale despre
structura moleculară. Această lucrare descrie în detaliu metodele
folosite la investigarea proprietăt, ilor ment, ionate mai sus precum s, i
rezultatele experimentale despre disocierea moleculei de dioxid de
carbon.

viii
POPULAR SCIENTIFIC INTRODUCTION

Carbon dioxide is one of the most important chemical compounds

to support life on Earth. It plays a crucial role in the process of photo-
synthesis through which plants naturally regulate the oxygen quan-
tity from the atmosphere for aerobic beings, including humans.
CO2 was discovered in 1640 by Jan Baptist van Helmont who ob-
served that after burning coal, there is a noticeable difference in mass
between the initial product (coal) and the remains (ash). He then
suggested the occurrence of an invisible substance he later named
gas or "spiritus silvestre" (wild spirit), making carbon dioxide the first
gas to be considered as an isolated chemical compound. With the
development of natural sciences, especially physics and chemistry, it
was concluded that there are micro-particles that form the gas, called
molecules, which have atoms as constituent parts, held together by
molecular bonds. This could only be explained later on, with the de-
velopment of quantum mechanics.
Carbon dioxide is a transparent gas, odorless in low concentra-
tions, presenting a sharp odor in high concentrations. It has a density
roughly 1.6 times higher than air and cannot be liquefied in normal
pressure conditions (only at high pressures). When cooled under a
temperature of -78.5o C it becomes solid, thus forming carbonic ice,
also known to industry as "dry ice".
One way of studying the CO2 molecule is by the means of X-ray
irradiation, thus allowing for processes such as photoabsorption (ab-
sorption of light), photoionization (ejecting electrons by photoab-
sorption) and photodissociation ("breaking" the molecule after pho-
toabsorption and ionization). X-ray light is outside the human visi-
ble spectrum, it has higher energy (shorter wavelengths) and has the
capacity to penetrate matter.
The oxygen and carbon atoms which make the molecule are
formed by electrons, protons and neutrons. The latter two form the
atomic nuclei, having electrons "orbit" around them on circular-like
paths called orbitals (this is not exactly true, but it is accurate enough
for present explanations). When interacting with X-rays, electrons
can "jump" on an energetically higher orbital or can even be com-
pletely ejected from the system, thus promoting the atom/molecule

ix
Popular scientific introduction

in an excited or ionized state. Quantum mechanics states that the en-

ergy of the light must be precisely matched with the transition energy
and for inner shell electrons these energies fall in the X-ray regime.
Due to the high precision demands in tuning the light, conven-
tional X-ray sources are not the desired choices for such experiments.
One way of satisfying the strict requirements is by using light emitted
by a synchrotron, a circular accelerator which increases the energy of
electrons to very high values (attaining velocities close to the speed
of light). In order to form an idea about the magnitudes involved,
we consider a simple situation at the new MAX IV synchrotron in
Lund. We know it had a circumference of 532 m and that electrons
moving at roughly 99.99% of the speed of light (approx. 300.000.000
m/s) will do more than 500.000 turns in the accelerator every second.
Energetic electrons are passing magnetic arrays, called undulators
that will force a curved trajectory on the beam. With every such os-
cillation, each electron will emit photons in the plane of oscillation
with energy defined by the energy of the electrons and the magnetic
field strength. The latter is adjustable, thus assuring a wide range of
energies for the produced light. Furthermore, the photon flux is con-
siderably higher than at any conventional sources, so synchrotrons
provide high intensity X-rays, should they be required.
So far, we have seen the ways of promoting the molecule into
energetically higher states. If ionization is considered, the newly
formed positively charged ion (with at least one electron missing) is
unstable and, like all entities from nature, it tends to reach a state of
equilibrium. One of the decay mechanisms is breaking into smaller
fragments formed by its constituent atoms. Considering the frag-
ments are not electrically neutral (this being the reason they are
called ionic), in the presence of an electric field they will be acceler-
ated in different directions (the difference arising from the sign of the
charge + or -). Time of flight mass spectroscopy measures the time
these fragments take from creation to detection, thus identifying
them by coincidence measurements. This information is then used
to determine the dissociation pathways of the studied molecule. This
work describes the methods used in investigating the aforemen-
tioned properties, followed by conclusions and discussion regarding
the interpretation of data.

x
ACKNOWLEDGEMENTS

The process of writing a master thesis offers a mixture of positive and

negative feelings, spanning from self content all the way to chaos.
However, the present work could have never existed without the help,
support and inspiration of several people.
Firstly, I would like to present my gratitude to my supervisor,
Rami Sankari, without whom none of this would have been possi-
ble. He manifested a great deal of interest and valuable contribu-
tions throughout all the stages of my thesis. I would specially like
to address my sincere acknowledgments for providing me with the
chance to be a part of a research team and working in a synchrotron
laboratory environment.
I wish to continue by offering my thanks and recognition to the
research team at Elettra laboratory: Antti Kivimäki, Robert Richter
and Marcello Coreno for putting up with me during the week I took
part at the measurements for this work. They took their time to ex-
plain crucial details from beamline operation to experimental sta-
tion and spectra interpretation.
Furthermore, I am grateful to Christian Stråhlman for offering
hours of his time to help with acceptance-angle simulations and pro-
viding additional information regarding the spectrometers. I would
also like to extend my gratitude to Sverker Werin for assistance with
choosing a master project at MAX IV.
Next, I would like to thank my colleague and friend, Mihai Pop,
for offering both his vast physics expertise and moral support in the
process of "creation".
On a more personal level, a special thanks and deep apprecia-
tion go to my girlfriend Delia Ivan, for her unconditional love and
support throughout this process, lessening the burden of writing the
thesis by bringing innumerous smiles on my face. I would also like
to thank my parents for all the moral support over the years. Last but
not least, I wish to express my gratitude to the person who first in-
spired and introduced me into the wonderful world of physics, my
school teacher, Daniel Lazăr.

xi
ABBREVIATIONS

TOF Time of flight

PEPICO Photo-electron positive ion coincidence
NIPICO Negative ion positive ion coincidence
PIPICO Positive ion positive ion coincidence
NIPIPICO Negative ion positive ion positive ion coincidence
PEPIPICO Photo-electron positive ion positive ion coincidence
LCAO Linear combinations of atomic orbitals
SALC Symmetry adapted linear combinations
MO Molecular orbital
TDC Time to digital converter
PIY Positive ion yield
NIY Negative ion yield
IP Ionization Potential

xiii
LIST OF FIGURES

2.1 Potential curves in the neutral and excited states of a sys-

tem with the vibrational wave functions and a probable
transition, showing the dependence of energy (E) as a
function of internuclear distance R. . . . . . . . . . . . . . . . . 5
2.2 Light-matter interaction. Case a), the incident photon is
absorbed and an electron is promoted to a higher energy
state; b) shows the process of ionization when, after ab-
sorption, the electron is completely relocated outside the
atomic or molecular system. . . . . . . . . . . . . . . . . . . . . . 6
2.3 Carbon dioxide absorption spectrum at core-level ener-
gies, with the resonance at 535.4 eV [1] and other possi-
ble transitions, including the Ionization Potential (IP) of
541.2 eV [2]. Data recorded at Elettra, Gas Phase beam-
line, in Trieste, Italy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Carbon dioxide molecule planes and axes of symmetry. . 8
2.5 Coordinate axes as chosen for carbon dioxide molecule. . 9
2.6 A sketch of carbon atom orbitals with their symmetry la-
bels corresponding to the D2h group - to be matched to
orbitals from the oxygen atoms. . . . . . . . . . . . . . . . . . . . 10
2.7 Schematic presentation of orbitals of the oxygen atoms. . 10
2.8 Carbon dioxide molecular orbitals with specifications of
bonding, non-bonding and anti-bonding orbitals. . . . . . 11
2.9 Fluorescence process with core-level photo-ionization. . 12
2.10 Illustration of Auger decay mechanism, from absorption
of an X-ray photon. The core-electron is removed and the
core-hole is refilled by a relaxation process from another
shell. The excess energy is released via an ejected Auger
electron. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.11 Example of a Coulomb explosion of carbon dioxide form-
ing the unstable CO2 ++ doubly-charged cation, with re-
sulting CO+ /O+ ion pair. . . . . . . . . . . . . . . . . . . . . . . . . 16

3.1 Schematic of light produced in a bending magnet. . . . . . 18

xiv
 List of Figures

3.2 Schematic of an undulator layout, with red and blue be-

ing the magnetic poles (with the field direction indicated
by arrows). The electron beam is drawn in green and the
yellow arrows at each bend represent the synchrotron ra-
diation being produced. . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 A schematic of the Gas Phase Photoemission beamline
optical layout. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 A schematic layout of a time-of-flight spectrometer. . . . . 21
3.5 Positive and negative acceptance angles for H – ions mea-
sured with respect to the perpendicular direction on the
spectrometer axis, for the negative ion spectrometer. . . . 22
3.6 Time-of-flight difference spectrum recorded at two dif-
ferent photon energies for low pressure carbon dioxide
gas. Feature assigned with (*) is unknown, as it does not
match any time-difference for CO2 ionic fragments. . . . . 23
3.8 A simplified schematic of the continuous acquisition
mode electronic setup for coincidence detection. . . . . . . 24
3.7 A schematic layout of the tandem time-of-flight ion spec-
trometer. Picture from Christian Stråhlman. . . . . . . . . . . 26

4.1 Coincidence spectrum for CO2 molecule core-excited

with 535 eV (blue) and 538.79 (red) radiation. Narrow
peak (assigned *) just below 4000 ns appeared in all
measurements, including valence excitations at 18 eV,
thus was attributed to noise/malfunction in the device.
Structure marked with (**) does not point to any real time
difference. Acquisition time was 120 minutes for each
photon energy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.2 Singly-charged positive fragments in coincidence with
negative ions as a function of photon energy at O1s edge.
Dotted lines are added to guide the eye; the vertical bars
show the excitation energies for O1s to 3p, 4s and 4p or-
bitals, and the O1s ionization potential in CO2 . . . . . . . . . 29
4.3 Singly-charged positive fragments in coincidence with
electrons as a function of photon energy at O1s edge. . . . 30
4.4 Doubly-charged positive fragments in coincidence with
negative ions as a function of photon energy at O1s edge. 32
4.5 Doubly-charged positive fragments in coincidence with
electrons as a function of photon energy at O1s edge. . . . 33
4.6 Diagonal map of coincidences for the CO2 molecule with
several events and their intensities encircled in red, at 50
eV excitation. Scale chosen to show region for e-/cation
coincidences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.7 Multiple coincidences for singly-charged cations with an-
ions and electrons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.8 Multiple coincidences for doubly-charged cations with
anions and electrons. . . . . . . . . . . . . . . . . . . . . . . . . . . 37

xv
CONTENTS

List of Figures xiv

1 Introduction 1

2 Molecular theory 3
2.1 Quantum approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Light-matter interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Franck-Condon approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.4 Molecular Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5 Molecular orbitals of CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.6 Decay of excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6.1 Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6.2 Auger decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.6.3 Fragmentation of CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 Methods and instrumentation 17

3.1 Synchrotron radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Gas Phase Beamline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Ion Time-of-Flight spectrometer . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Coincidence measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

4 Analysis 27
4.1 Anion-cation coincidence spectra . . . . . . . . . . . . . . . . . . . . . . . ... 27
4.1.1 Core-level photon energies - O1s edge . . . . . . . . . . . . . ... 29
4.2 Coincidence results including multiple positive fragments . . . . . ... 34
4.2.1 Singly-charged fragments in coincidence with anions or
electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 35
4.2.2 Doubly-charged fragments in coincidence . . . . . . . . . . ... 36

5 Discussion and conclusions 39

5.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

References 43
 CHAPTER 1

INTRODUCTION

Carbon dioxide is one of the most important compounds to sustain

life on Earth. It is a crucial gas in the process of photosynthesis, by
which plants naturally regulate the oxygen balance required by aer-
obic beings.
In 1640 [3] Jan Baptist van Helmont noticed that from the pro-
cess of burning charcoal, the mass of the remains was significantly
lower than the mass of the initial charcoal. He deduced that the dif-
ference rose due to an invisible substance which he called "gas" or
"spiritus silvestre" (wild spirit [4]), making CO2 the first gas consid-
ered as a discrete chemical compound. Scientific advances in the
fields of physics and chemistry made it clear that carbon dioxide gas
is made up of small constituents called molecules which, in turn, are
formed by even smaller parts called atoms that bond together to form
molecules. Later, by the development of quantum mechanics, the
bonding mechanism could be explained.
Carbon dioxide is a colorless, odorless (at lower concentrations)
gas, having a standard conditions density of 1.98 k g /m 3 [4]. Below
pressures of 5.1 atm, CO2 has no liquid state at low temperatures, it
becomes solid at roughly 194 K (∼ -80o C), state commonly known as
dry ice.
A suitable way of studying and analyzing the CO2 molecule is by
investigating its interaction with light. In order to successfully do so,
one needs to carefully tune the wavelength of the light so as to match
the discrete energy levels of the molecule. This can be achieved at a
synchrotron light facility, where energy adjustment is solved by the
use of tunable undulator light sources. With photo-absorption ex-
periments, the molecule can be excited to higher energy states or
ionized. More energetic radiation induces the creation of a so called
core-hole state, where an electron from the core of the atom (in-
ner shells) is promoted to a higher energy or continuum state, call-
ing it core-electron excitation or ionization. From these states, often

1
 via some intermediary states, the molecule will rupture into smaller
fragments via dissociation.
When dissociation leads to ionic fragments, time of light (TOF)
based mass spectroscopy can be used to analyze them. Charged par-
ticles with different masses are discerned by distinct flight times re-
flecting different mass-to-charge ratios (M/q). In order to study dis-
sociation dynamics, one makes use of coincidence measurements
with TOF spectroscopy, where charged particles are detected in co-
incidence with each other so as to ascertain which originated from
a certain event, thus receiving more information from the ionization
after interaction with light to the constituent fragments and, allow-
ing for the construction of the dissociation pathways more reliably.
This work aims to describe fragmentation and dissociation of car-
bon dioxide molecule at core-level energies (O1s) by means of time
of flight mass spectroscopy.

2
 CHAPTER 2

MOLECULAR THEORY

A system consisting of bound electrons and two or more atomic

nuclei is called a molecule. They are held together by chemical
bonds, such as covalent bonds and ionic bonds. In order to describe
molecules theoretically, one needs to adopt a quantum mechani-
cal approach. This chapter aims at briefly describing the molecule,
molecular symmetry and orbitals as theoretical grounds to the level
required to follow the motivation for the experiments presented in
this work.

2.1 Quantum approach

In quantum mechanics, any system can be described by the time-
independent Schrödinger equation:

H Ψ(r ) = E Ψ(r ) (2.1)

with Ψ the wave function, E the system energy and H the
Hamiltonian. For molecules, the Hamiltonian can be written as [5]:

ħ2 X 2 X ħ
h h2
H =− ∇i − ∇2N + V (R~ , r~) (2.2)
2m i N
2M N

which takes into account i electrons, N nuclei, each with mass M N

and the interactions between them. The summation over i repre-
sents the electronic kinetic energies and V the Coulomb potential
due to electrostatic attraction and repulsion forces. In other words,
it describes the position and motion of electrons as they are at-
tracted towards nuclei and repelled by other surrounding electrons.
This being said, with increasing number of particles, solving the
Schrödinger equation analytically proves to be very tedious work and
is furthermore impossible for more complex systems, such as poly-
atomic molecules, therefore approximations are required for simpli-
fication.

3
2.2 Light-matter interaction

A first assumption to be made is that the wave functions of elec-

trons and nuclei are separable and can be written as:

Ψm o l (r~i , R~N ) = ψe l (r~i , R~ )χn u c (R~ ) (2.3)

where Ψm o l denotes the molecular wave function, ψe l the electronic

wave function and χn u c the nuclear wave function. With this in
mind, the following approximation states that ∇N ψe l (the deriva-
tive of the electronic wave function with respect to nuclear coor-
dinates) is negligible or, in other words, due to the fact that nu-
clei are considerably heavier than electrons (at least by a factor of
a thousand - the nucleus of the hydrogen atom, i.e. a proton, is 1836
times heavier than an electron) they can be thought of as station-
ary. Furthermore, electrons are expected to relocate according to
the potentials created by the nuclei instantaneously. This is called
the Born-Oppenheimer approximation, named after physicists Max
Born and R. J. Oppenheimer in 1926 [6].
Now, if we rewrite the Schrödinger equation with the separable
wave functions, it yields a new set of wave functions which in turn
correspond to a set of discrete energy levels. This tool made it eas-
ier to perform calculations of wave functions and energies for more
complex molecular systems. Should one plot the aforementioned
energy values as a function of internuclear distance R, the potential
energy curves are revealed, as illustrated in Fig. 2.1.
One immediate observation states that, if R → 0 then E → ∞ or
the repulsion between the nuclei grows with decreasing internuclear
distance (of course, due to electrostatic repulsion). In Fig. 2.1 one
notices minima in the neutral and excited state curves. This is a state
where the molecule reaches an energetically stable geometry, there-
fore it is referred to as bonding. The other curve (with no minimum)
can be anti-bonding or non-bonding, states which do not allow for
an energetically stable configuration.

2.2 Light-matter interaction

The main tool in studying molecules and their dynamics is by their
interaction with light. An atom or a molecule that sees incident light
may absorb one or several photons, should they be of appropriate
energy. This section aims at briefly explaining the interaction of elec-
tromagnetic radiation with matter.
By absorption of a photon, the initial molecular energy state Ei is
altered to a different state (or value) E f by the following condition:

E f − Ei = h ν (2.4)

where h ν is the photon energy. One can also consider an ionized

state as a final state, then consisting of a nucleus, bound and emitted
electrons, as illustrated in Fig. 2.2.

4
 Molecular theory

Figure 2.1: Potential curves in the neutral and excited states of a sys-
tem with the vibrational wave functions and a probable transition,
showing the dependence of energy (E) as a function of internuclear
distance R.

Other excitation mechanisms can be by particle collisions (e.g.

electrons, protons, ions) but describing them is beyond the pur-
pose of this work, thus I shall restrain to photons. By studying the
aforementioned process and the absorption spectrum, it allows to
construct the related energy levels of the molecule. An example of
CO2 absorption at oxygen core-level energies is depicted in Fig. 2.3.
The spectrum shows structures corresponding to exciting the core-
electrons to unoccupied molecular orbital (π∗ ) and excitations to
more atomic-like orbitals (3s, 4s and 4p, for example).
During absorption, an electron is removed from a filled molec-
ular orbital and relocated to a different unoccupied orbital corre-
sponding to a higher energy. If the bound electron is promoted to
an empty bound orbital, this will take the molecule into a neutrally
excited state and the process will be referred to as resonant photo-
absorption.
If the incident radiation is energetic enough, an electron from the
outer shells will be completely expelled from the atomic/molecular
orbital and this process is known as ionization. Should the energy
increase even further (several hundred eV) a core-electron can be

5
2.3 Franck-Condon approximation

Figure 2.2: Light-matter interaction. Case a), the incident photon is

absorbed and an electron is promoted to a higher energy state; b)
shows the process of ionization when, after absorption, the electron
is completely relocated outside the atomic or molecular system.

ejected, i.e. core-level ionization can occur. The energy of the in-
coming X-ray photon must be tuned to the ionization energy of the
said core-electron and this tunability is attainable at a synchrotron
radiation facility (as described later in Ch. 3.1).
Another matter to consider when it comes to photo-absorption
is the cross-section, the probability a photon has to be absorbed
by the studied compound. For the C1s and O1s core excitations of
CO2 , X-ray absorption cross-section measurements were performed
by Truesdale et al and the results reveal a quick drop after resonance
energies i.e. ionization cross section is at maximum at the ionization
threshold of each orbital and it drops monotonically towards higher
photon energies. Therefore, valence ionization cross-section show
very low values at these energies and thus, only core-electron excita-
tions are generally expected to be dominant in these cases.

2.3 Franck-Condon approximation

Similar to the Born-Oppenheimer approximation, electrons being
over a thousand times lighter than nuclei, electronic transitions can
occur faster than nuclei have time to adjust (or relocate accordingly).
This is called the Franck-Condon approximation, and it assumes a
stationary nucleus during electronic transitions on different energy
levels. In actuality, the nucleus vibrates, however, due to the enor-
mous weight difference and short electronic transition times (can
be thought of as almost instantaneous) one can assume a nucleus
as "fixed" during this process. Another implication is that in the

6
 Molecular theory

Figure 2.3: Carbon dioxide absorption spectrum at core-level ener-

gies, with the resonance at 535.4 eV [1] and other possible transitions,
including the Ionization Potential (IP) of 541.2 eV [2]. Data recorded
at Elettra, Gas Phase beamline, in Trieste, Italy.

frame of potential curves, electronic transitions can be considered

as "jumps" in the vertical direction, as illustrated in Fig. 2.1.
Fig. 2.1 also shows the so-called Franck-Condon region (delim-
ited by the two most probable transitions) a zone where transitions
are more likely to occur. If one takes into account the wave functions
of the initial and final states, one notices a vibrational part |vi > and
|v f >. The maximum likelihood of an electronic transition to take
place is where the vibrational wave functions have the most overlap
or, | < vi |v f > |2 is largest. The latter is called a Franck-Condon Factor
(or FCF) and can be calculated numerically or deduced experimen-
tally.

2.4 Molecular Symmetry

When evaluating molecular orbitals (MOs), symmetry plays an im-
portant role due to the fact that it simplifies calculations by a great
amount, as shall be seen in the following. A symmetry property states
that when a certain symmetry operation is applied to an object, it will
leave said object unchanged. One can distinguish five types of sym-
metry operations:
- E - the identity operation
- Cn - n-fold rotation around an axis at angles 2 π/n
- σ - reflection through a mirror plane: vertical (σv ), horizontal
(σh ), diedral plane (vertical plane that bisects the angle between two
C2 axes, σd ) [7]
- i - inversion through a center of inversion

7
2.4 Molecular Symmetry

- Sn - n-fold improper rotation - rotation of the molecule by 2π/n

and reflection on the perpendicular plane of the rotation axis
Cn , the n-fold rotation with respect to an axis of symmetry con-
sists of a rotation through angles 2 π/n and it is of particular inter-
est for the carbon dioxide molecule. Provided that CO2 has a lin-
ear structure, as depicted in Fig. 2.4, it exhibits C∞ due to the fact
that any rotation about its symmetry axis will leave it unchanged.
Furthermore, it also has mirror plane reflection σ, in the center of in-
version i (the carbon atom), perpendicular to its principal axis (hor-
izontal), therefore denoted σh .
Molecules can be classified based on their symmetry properties
into groups such as C s (DNA molecule), C2 (H2 O2 ), C2v (H2 O) [8].
Given that carbon dioxide molecule has C∞ infinite fold rotation
axis, infinite other C2 two-fold axes perpendicular to C∞ and a hor-
izontal mirror plane σh , it belongs to D∞h symmetry group. The
molecule along with the aforementioned symmetries is illustrated in
Fig. 2.4.

Figure 2.4: Carbon dioxide molecule planes and axes of symmetry.

Molecular orbitals, like atomic orbitals, are defined by the elec-

tronic wave functions. However, deducing molecular orbitals from
performing calculations in the molecular Hamiltonian can be-
come very complicated, even in the previously discussed Bohr-
Oppenheimer approximation. What one does is rather construct
them by using linear combinations of the constituent atomic orbitals
(or LCAO) [9]:
X
Ψ= c r Φr (2.5)
r

where Ψ is the molecular orbital (also denoted MO), Φr atomic or-

bitals and c r numerical factors. Even in this situation, one can end

8
 Molecular theory

up with plenty possible combinations of atomic orbitals. The likeli-

hood of certain MOs to form is further determined by symmetry con-
ditions, reducing the complexity in defining the MOs.
If a molecule holds particular symmetry properties (other than
identity), group theory can state which atomic orbitals make con-
tributions in forming the molecular orbitals due to net overlap-
ping (having non-zero overlap integrals S, S6=0). Orthogonal atomic
orbitals cannot combine, as they will have zero overlap, S=0.
Furthermore, even with fulfilling the symmetry conditions, there are
several other aspects one needs to consider when constructing the
MOs. It is natural to think that the overlap will be small for atomic or-
bitals of very different energies thus, the atomic orbitals of the same
energy dominate the bonding. Also, if one orbital is very compact
(strongly localized inner-shell orbital), and the other is a diffuse va-
lence orbital, their overlap contribution will not be sufficient.

2.5 Molecular orbitals of CO2

I shall further summarize an approach to constructing the CO2

molecular orbitals by considering the overlap between oxygen or-
bitals with the orbitals of the central carbon atom. As a simplifica-
tion, it is convenient to reduce the D∞h to D2h , so as not to perform
calculations using infinite-fold axes. Also, assigning coordinates to
the atoms in the molecule makes the procedure more convenient.

Figure 2.5: Coordinate axes as chosen for carbon dioxide molecule.

The combinations of the two oxygen orbitals chosen that respect

the symmetry conditions are based on 2s and 2p (2pz , 2py , 2px ) [10],
as seen in Fig. 2.7.
Now considering symmetry properties, one finds that these rep-
resentations are reducible and can be matched to the orbitals of the
central atom, in this case carbon, illustrated in Fig. 2.6.
Combining all reduced representations of the group orbitals cor-
responding to the oxygen atoms and the carbon atom orbitals, one

9
2.6 Decay of excited states

Figure 2.6: A sketch of carbon atom orbitals with their symmetry la-
bels corresponding to the D2h group - to be matched to orbitals from
the oxygen atoms.

Figure 2.7: Schematic presentation of orbitals of the oxygen atoms.

ends up with molecular orbitals for the carbon dioxide molecule, as

seen in Fig. 2.8.

2.6 Decay of excited states

The previously described excitations can occur both with electrons
from the outer valence shells or inner shells, depending on the in-
coming radiation or energy of the incident particles. The electrons
from the inner orbitals are called core-electrons and they are of in-
terest in the field of photo-absorption spectroscopy, as such an ex-
citation will form a core-hole, bringing the system in a very unsta-
ble, high energy state, from which it will decay rather quickly (a few
femtoseconds for CO2 molecule [11]) via radiative or non-radiative
processes.

10
 Molecular theory

Figure 2.8: Carbon dioxide molecular orbitals with specifications of

bonding, non-bonding and anti-bonding orbitals.

2.6.1 Fluorescence

A radiative decay process is called fluorescence. It consists of exciting

a core-electron above the ionization threshold, thus ejecting it, leav-
ing the ionic fragment in an unstable energy state. The created core-
hole will be filled by an outer shell electron and the excess energy will
be released as a photon (typically in the X-ray regime). Fluorescence
process is illustrated in Fig. 2.9. This decay can also occur via excita-
tion onto a bound orbital (non-ionization), when the electron decays
to a lower energy orbital, process known as resonant inelastic X-ray
scattering.

11
2.6.2 Auger decay

Figure 2.9: Fluorescence process with core-level photo-ionization.

2.6.2 Auger decay

In the field of electron/ion spectroscopy the non-radiative decay
mechanism is of particular importance. This presumes the core-hole
to be filled by an outer-shell electron, creating enough excess en-
ergy (from the relaxation) to completely eject another electron from
the atomic or molecular system. This process is called Auger decay,
named after physicist Pierre Auger who firstly explained it in 1923
[12] (having been observed and published by the Austrian-Swedish
physicist Lise Meitner in 1922 [13]), shown in Fig. 2.10. There is also
the possibility of exciting a core-electron to a bound, empty orbital
and the decay process from such a state shall be referred to as reso-
nant Auger decay.

Figure 2.10: Illustration of Auger decay mechanism, from absorption

of an X-ray photon. The core-electron is removed and the core-hole
is refilled by a relaxation process from another shell. The excess en-
ergy is released via an ejected Auger electron.

However, emission of a photon might also arise in this situation

12
 Molecular theory

due to the so called radiative Auger decay [14]. This presumes the
relaxation to happen through emission of an X-ray photon along
with promoting an electron either to a bound or continuum state.
However, the probability of occurrence for such an event is consid-
erably lower than for fluorescence, as the relaxation takes place via
dipole-allowed transitions [15].

2.6.3 Fragmentation of CO2

Molecular dissociation is a process in which molecules break into
their constituents (smaller molecules, atoms, ions) by different
means of external excitation. Fragmentation is a particular type
of molecular relaxation and it implies the dissociation of unstable
molecular entities. This section introduces the reader in both known
and expected patterns of fragmentation for carbon dioxide, but also
less common ones to be investigated.
For fragmentation, a good initiating mechanism is electronic ex-
citation at the core ionization thresholds (e.g. C1s and O1s for CO2 ).
Going over this barrier one reaches the ionized state, in which the ex-
cess energy is most often taken care of by emission of an Auger elec-
tron, thus forming doubly charged parent ion. Triply charged ions
are also possible, due to double Auger decay or, in some cases by a
cascade of Auger decays, but indeed, the primary de-excitation path-
ways produce singly and doubly charged ions, below and above the
ionization threshold, respectively [16].
Fragmentation of carbon dioxide after core excitations has been
studied, among others, by Öhrwall et al [17] and their results are re-
viewed in the following, not least to provide some motivation for
present experiments. The most probable reaction producing ions af-
ter core-level electron excitation is the resonant Auger decay:

CO2 ∗ → CO2 + + eresA

Furthermore, there is the possibility that the excited molecule emits

a photon via fluorescence, before the dissociation. Such event is
shown below, CO2 * and CO2 ** representing initial excited state and
the (still excited) state after the photon emission, respectively.

CO2 ∗ → CO2 ∗∗ + h ν

In the case of core ionization, the Auger decay dominates:

CO2 + → CO2 ++ + eA

thus forming the doubly-charged parent ion, CO2 ++ . From these par-
ent ions, singly or doubly charged, the fragmentation usually starts,

13
2.6.3 Fragmentation of CO2

yielding C+ , O+ , CO+ and CO2 + [17] [18] [19], but also negatively
charged fragments, such as C – and O – , can appear [17].
I shall now construct the probable dissociation pathways, tak-
ing into account the previously presented steps. As known, inner-
shell photoionization is followed by a core-hole decay and emission
of an Auger electron. This process creates usually a highly unsta-
ble, doubly-charged parent ion. Due to electrostatic repulsion, the
molecule can undergo Coulomb explosion, in which they break into
energetic, singly charged fragments [20]. This process is illustrated
in Fig. 2.11. Two of the probable pathways forming singly-charged
cations following Auger decay are shown here:

CO2 ++ → O+ + C+ + O and

CO2 ++ → O+ + C + O+ .

Doubly charged ion fragments are also a possibility at, and especially
above the O1s edge (C++ and O++ ).
Furthermore, anion fragments can also form following core exci-
tation or ionization, thus they are included in the possible fragmen-
tation pathways. If one takes into account core excitations (below
the O1s ionization threshold), negative ions are created by ion-pair
formation from CO2 + parent ion when it fragments into a CO++ di-
cation [17], such as:

CO2 + → CO++ + O− and

→ C+ + O+ + O− .

where CO++ breaks into two singly charged cations. In addition, an-
ions are also created following ion-pair formation from a neutral ex-
cited and unstable CO fragment, yielding the following pathways:

CO2 + → CO + O+ ,

→ C+ + O− + O+ , and

→ C− + O+ + O+ .

Above the O1s ionization potential, where Auger decay is most

likely (thus forming the CO2 ++ parent ion), anions can also be formed
along with singly or doubly charged cations, and some possible path-
ways are given below:

14
 Molecular theory

CO2 ++ → C++ + O− + O+ or

CO2 ++ → O++ + C− + O+ .

Again, the information from both the negative and positive ions al-
lows to separate these fragmentation paths and, they will indeed be
taken into account when identifying the molecular fragments in the
recorded time-of-flight mass spectra, to be presented in the follow-
ing sections.
Furthermore, there is a possibility for electron recapture events,
at the onset of Auger decay just above the ionization threshold:

[CO2 + h ν] → CO2 + (1s −1 ) + eph

→ CO2 ++ + eph + eA

with 1s −1 illustrating that the molecular parent ion has a core hole on
the O1s orbital and the second line showing the Auger decay. From
here, CO2 ++ can recapture a slow photoelectron and then dissociate
as follows:

CO2 +∗ → CO+ + O∗

with O* representing the fragment that received the recaptured pho-

toelectron after dissociation. From here, through the fragmentation
of CO+ cation, the pathways can become:

[CO+ + O∗ ] → C+ + O + O∗ , or C + O+ + O∗

As stated in [21], these fragments can be metastable, in which case

they have a lifetime of microseconds, or fully stable but, as present
setup is not suitable for detecting photons typically emitted in their
decay, they shall not be explored further.
The fragmentation pathways considered so far assume there is no
external processes such as gas pressure dependency of electron re-
attachment to a neutral fragment or of electron capture by positive
fragments. This issue was briefly discussed in [17] [22] and suggested
that pressures in the range of 1∗10−5 mbar were low enough to ensure
single-collision conditions. Even though our pressures read values
around 4 ∗ 10−7 mbar, the pressure can be higher at the exit of the
gas needle, however less than two orders of magnitude [23], which
means that the single collision condition was met well in the present
measurements.
This work aims to experimentally investigate the dissociation
dynamics of CO2 by using time of flight spectroscopy, method de-
scribed in the following section, and especially emphasizing the ad-
ditional information reachable with negative ion measurements.

15
2.6.3 Fragmentation of CO2

Figure 2.11: Example of a Coulomb explosion of carbon dioxide

forming the unstable CO2 ++ doubly-charged cation, with resulting
CO+ /O+ ion pair.

16
 CHAPTER 3

METHODS AND INSTRUMENTATION

In order to produce core level excited and ionized states for the car-
bon dioxide molecule, one needs a high-flux soft X-ray radiation
source with narrow spectral width. This is attainable in practice at a
synchrotron radiation facility. The high-speed electrons in the stor-
age ring (synchrotron) are used to produce light which is then ex-
tracted and fed to the experiment through a beamline.
The measurements for this work were carried at Elettra syn-
chrotron radiation source (Trieste, Italy) at the Gas Phase Beamline
(described in Ch. 3.2), using two ion time-of-flight (TOF) mass
spectrometers. This chapter describes the instrumentation methods
used to acquire our data.

3.1 Synchrotron radiation

Core-excitation of the gas-phase carbon dioxide molecule is done in
the presence of X-ray light at more than ca. 300 eV photon energy.
Furthermore, good energy resolution and high flux are required for
good experimental data acquisition. These conditions are achiev-
able at modern synchrotron radiation facilities.
A synchrotron is a particle accelerator designed, in this case, to
store electrons for light production. It consists of straight sections,
bending magnets, acceleration cavities and insertion devices (un-
dulators and wigglers), electrons traveling at all times in a vacuum
chamber (ultra high vacuum conditions, otherwise they would have
a very short free path).
Synchrotron light is generated when charged particles are accel-
erated perpendicularly to the velocity vector by magnetic fields. It
was firstly created in bending magnets (1947 [24]) and it was consid-
ered a "parasitic" event as it drained the beam of energy with every
turn. Nowadays, synchrotrons are built solely for light generation as
it exhibits outstanding characteristics.

17
3.1 Synchrotron radiation

Figure 3.1: Schematic of light produced in a bending magnet.

Light is extracted from bending magnets or insertion devices

such as undulators and wigglers. A schematic of an undulator is
shown in Figure 3.2. It consists of an array of permanent magnets
which are laid over and under the vacuum chamber, with adjustable
gap (for energy tuning). The magnets are arranged such that the re-
sulting magnetic field vector has an alternating direction. Electrons
feel the magnetic field and oscillate accordingly, thus emitting a pho-
ton with each oscillation.

Figure 3.2: Schematic of an undulator layout, with red and blue being
the magnetic poles (with the field direction indicated by arrows). The
electron beam is drawn in green and the yellow arrows at each bend
represent the synchrotron radiation being produced.

Photons are emitted in the direction of electron propagation and

the small electron bunch is the actual source of emission. Therefore,
the resulting beam size and divergence are considerably smaller than
conventional X-ray sources. As polarization is linked to the motion
of electrons, it is also well defined (e.g. linearly polarized light due
to the fact that the electron oscillation occurs in one plane). As the

18
 Methods and instrumentation

gap between the magnetic arrays is adjustable there is a wide range

of tunability in photon energy (or its wavelength λ); the relation gov-
erning this is called the undulator equation [25]:

λu K2
λ= (1 + + γ2 θ 2 ) (3.1)
2n γ2 2
1
where the K factor is K = e B0 λu /2πm c , γ = p the Lorentz factor,
1−β 2
θ the angle relative to the center of the radiation cone (usually very
small), n the harmonic and λu the undulator period. Thus, by ad-
justing the gap between the magnets, one varies the magnetic field
strength and the K factor implicitly, allowing to tune to the desired
wavelength of light and its higher harmonics.
An important quality measure for the light source is called bril-
liance and it is defined as [24]:

photons
Brilliance = (3.2)
second ∗ mrad2 ∗ mm2 ∗ 0.1%BW

taking into account the photon flux (number of photons/sec), the

angular divergence, source cross-section and number of photons in
0.1% bandwidth of the selected central wavelength. For a higher flux
at high photon energies, one would close the undulator gap to the
minimum value: although the fundamental wavelength gets longer
(smaller photon energy) the high energy part of the undulator spec-
trum is fairly smooth and intense, resembling that of a wiggler.

3.2 Gas Phase Beamline

Our measurements were carried at the Gas Phase Photoemission
beamline at Elettra Synchrotron, Italy, in February 2016. It is a
beamline designed for gaseous samples, to perform photo-emission,
photo-absorption and coincidence measurements, the latter being
of interest for us.
The light source is a 36 period undulator divided into 3 sections
of 12 with a λU =12.5 cm, the maximum K factor value K=5.3 and
Bm a x =0.456 T [26]. It is capable of providing light with energies be-
tween 13-900 eV.
A switching mirror and a pre-focusing mirror ensure that the
photon beam from the undulator is focused at the entrance slit of the
monochromator. Then, after the wavelength dispersing grating (and
plane mirror coupled to the grating motion), the bandwidth of dis-
persed radiation is set with the exit slit. After that, the narrow band-
width beam is refocused by Kirkpatcrick-Baez pair, vertical and hor-
izontal refocusing mirrors, thus guaranteeing a focused, small spot-
size at the experimental stations. A schematic of the optical layout is
illustrated in Figure 3.3.

19
3.3 Ion Time-of-Flight spectrometer

Figure 3.3: A schematic of the Gas Phase Photoemission beamline

optical layout.

The Variable Angle Spherical Grating Monochromator (VASGM)

allows for a resolving power of over 10.000 over the whole energy
range by using 5 gratings to select from, depending on the desired
energy range.
All this provides high photon flux and good resolving power
(>10.000 over the whole energy range [27] [28]). The spot-size at the
end station is relatively small (200 µm x 200 µm). However, due to
refocusing solution, the end station is 1.762 m above the floor level,
and the beam has an angle of 4o relative to the floor.

3.3 Ion Time-of-Flight spectrometer

A commonly used method for analyzing ionic fragments is by the

use of ion time-of-flight mass spectroscopy, which differentiates be-
tween the fragments by considering their mass-to-charge ratio M/q.
A major improvement in TOF instruments, so-called space focusing,
was introduced by Wiley and McLaren [29]. Such mass spectrom-
eters have good mass resolution and, in other words, the ability to
differentiate small mass differences. A typical layout of such a device
is shown in Figure 3.4.
Figure 3.4 illustrates the constituent parts of a TOF spectrometer:
extraction region, acceleration region, field free region and detector.
The extraction region consists of a space of length L where an ex-
traction voltage V is applied between two grids, pusher and extractor
[30]. This is the location where negative or positive fragments, de-
pending on the polarity of the grids, are pushed towards the acceler-

20
 Methods and instrumentation

Figure 3.4: A schematic layout of a time-of-flight spectrometer.

ation region. In the process of dissociation, the resulting fragments

will have an initial kinetic energy therefore, the extraction voltage is
carefully chosen so as to ensure a high transmission/acceptance an-
gle, in practice the ability to "collect" ions and send them towards the
detector before they have time to escape the extraction region due
to initial kinetic energy - this matter will be given more attention in
the following, especially how it is related to present instrument. The
acceleration region provides a final increase in kinetic energy of the
particles and the field free region (or drift-tube) leads to the detector,
ensuring longer flight times and better separation of the fragments
with different velocities for detection resolution. The total flight du-
ration of an ion will be the sum of the times the fragment spends in
each part of the spectrometer:

T = Te + Ta + TD (3.3)

Measuring this time provides the information about charge to mass

ratio M/q. Knowing the system parameters (e.g. L, D, and V) one can
simulate flight times for certain fragments in the device, thus allow-
ing for identification of ions in spectra.
The device used for our measurements was a tandem time-of-
flight ion spectrometer. It consists of two TOF spectrometers used to
detect positive and negative fragments in coincidence by soft X-ray
excitation. Instrument is fully described by Stråhlman et al [23] and
a schematic layout is depicted in Figure 3.7.
The negative-ion spectrometer was designed to work with the
previously built positive-ion spectrometer [31] but also indepen-
dently as a stand-alone instrument. It consists of a 15 mm extraction
region, a 5 mm long acceleration region and a drift tube of 320 mm.
The voltages applied for these measurements were ±217 V at the ex-

21
3.3 Ion Time-of-Flight spectrometer

Figure 3.5: Positive and negative acceptance angles for H – ions mea-
sured with respect to the perpendicular direction on the spectrome-
ter axis, for the negative ion spectrometer.

traction plates, 1985 V over the MCP detector and +2000 V at the drift
tube.
Due to the fact that negative ions are considerably less abundant
than positive fragments, the corresponding spectrometer was built
to have as high transmission as possible, so as to compensate for
the low yield of negative ions. In the design phase, the goal was to
have a large detector (for high acceptance angle) and as short as pos-
sible drift tube while still preserving the required mass resolution.
As cation fragments are plenty, the conditions for the positive ion
spectrometer were not so stringent. I have performed simulations in
SIMION software to quantify the acceptance of the instrument and
the results are shown in Figure 3.5.
Given that lighter ions tend to have higher kinetic energies [23]
and, even at the same kinetic energy, a lighter ion would have greater
velocity (thus a higher probability of hitting the outer walls of the
spectrometer) the simulations were carried using negative hydrogen
ions H – with initial kinetic energies ranging between 0 and 20 eV. The
results were promising due to the fact that all the ions with initial ki-
netic energies up to 12 eV were collected at the detector.
Simulations were also performed for the positive-ion spectrom-

22
 Methods and instrumentation

Figure 3.6: Time-of-flight difference spectrum recorded at two differ-

ent photon energies for low pressure carbon dioxide gas. Feature as-
signed with (*) is unknown, as it does not match any time-difference
for CO2 ionic fragments.

eter, using H+ positive fragment, due to the fact that it is the lightest
possible positive ion, thus most prone to hitting the outer walls of the
instrument; it was found to be capable to record all ions with veloc-
ities perpendicular to its axis and initial kinetic energies up to 5 eV.
Although it has considerably lower transmission than the negative-
ion spectrometer, the higher abundance of positive fragments com-
pensates for this. Indeed, the instrument was found to perform well
in the initial experiments [31] and in several coincidence studies,
such as [23].

3.4 Coincidence measurements

A coincidence measurement is a method used to determine whether
two or more events occurred within a given time frame or, several
detections arose from the same event. The start signal (start of tim-
ing electronics) is provided, in this case, by the detection of a nega-
tive fragment. The stop signal is given by the positive ion detection
but that is actually delayed to make sure that even the signal from
the lightest positive ion comes to the measuring electronics after any
possible negative ion start. If there is no stop signal in that time
duration, the event is ruled out in later analysis. Detection of two
consecutive negative fragments (i.e. two start signals) also restarts
the measurement. Even though this scheme is highly effective for
two fragment coincidences (e.g. NIPICO - Negative Ion Positive Ion
Coincidence), three or four simultaneous coincidences are not possi-
ble due to the delay unit used, being able to pass only the first pulse
arriving to it. This limitation can be circumvented by recording all

23
3.4 Coincidence measurements

the events at both detectors and using computer software to find co-
incidences after the data collection is finished; this became actually
the main method for measuring coincidence events. In this contin-
uous acquisition mode, the start signal is given by a pulse genera-
tor operating at 10 Hz, and the signals from both detectors are fed
into the TDC unit (Time to Digital Converter - ACAM Messelectronic)
after pre-amplifier/constant fraction discriminator stage. The TDC
unit used in this case has a time resolution of 80 ps [31]. The simpli-
fied schematic of the aforementioned setup is depicted in Fig. 3.8.

Figure 3.8: A simplified schematic of the continuous acquisition

mode electronic setup for coincidence detection.

As negative ions are not as abundant as positive fragments, this

also means that they are not as abundant as electrons either (posi-
tive ion formation is often connected to direct or indirect photoion-
ization). Taking into account that the electrons travel in the same
direction as the negative ions, they would saturate the correspond-

24
 Methods and instrumentation

ing detector. In order to avoid that, a deflecting magnetic field was

introduced between the source region and the negative ion detector.
This was achieved by mounting two permanent neodymium mag-
nets on the drift tube of the negative ion spectrometer (outside the
vacuum chamber), thus ensuring the deflection of unwanted elec-
trons whilst not affecting the flight times of even the lightest negative
ion fragments [23] (simulations also performed by the author to con-
firm this matter). A typical spectrum obtained is illustrated in Figure
3.6. Several ion coincidences are marked next to their corresponding
peaks.

25
3.4 Coincidence measurements

Figure 3.7: A schematic layout of the tandem time-of-flight ion spec-

trometer. Picture from Christian Stråhlman.
26
 CHAPTER 4

ANALYSIS

The results presented here were recorded using an ACAM card in

continuous acquisition mode (as described in Chapter 3.4), thus
ensuring the possibility of detecting multiple coincidences. The
recorded spectra were imported and analyzed with Igor software.
This section describes the analysis and interpretation of our data.

4.1 Anion-cation coincidence spectra

As explained in Chapter 3, the data consists of the number of events

(normalized to experimental conditions) plotted against time differ-
ence between the detected signals at the two spectrometers. For this
data, the timespan allowed for finding coincidences was 12000 ns
and the zero was set at 5000 ns in the analysis. The data treatment
was such that a time-stamp of a negative fragment was regarded as a
start signal. From that, the presence of a positive fragment is checked
within the time window and the time difference between these two
(or more) time-stamps is calculated and saved, thus explaining the
occurrence of the coincidence events. A result of analysis of coin-
cidences recorded at 535 and 538.79 eV (O1s edge core-excitations)
showing the identified ion-pairs is illustrated in Fig. 4.1. Even though
there was a magnetic field for electron deflection, coincidences be-
tween electrons and positive ions are still observed due to the high
abundance of the latter (the magnetic field deflects most electrons to
avoid detector saturation, but a small fraction still reach the detec-
tor).
As the measurements were done in continuous acquisition mode,
which explains why the recorded ion coincidences can show below
and above the zero-point, 5000 ns: there could be a case of a pos-
itive fragment hitting its detector before the negative ion start, and
still resulting from the real event. For example, if one considers the
coincidence pair C++ /O – , the start signal was the detection of O – .

27
4.1 Anion-cation coincidence spectra

Figure 4.1: Coincidence spectrum for CO2 molecule core-excited

with 535 eV (blue) and 538.79 (red) radiation. Narrow peak (assigned
*) just below 4000 ns appeared in all measurements, including va-
lence excitations at 18 eV, thus was attributed to noise/malfunction
in the device. Structure marked with (**) does not point to any real
time difference. Acquisition time was 120 minutes for each photon
energy.

Provided that C++ is doubly charged and taking into account the
mass to charge ratio (M/q), C++ has 6 and O – 16, in absolute values,
C++ will fly and reach the detector much faster than O – , considering
the positive and negative voltages applied are somewhat similar.

On the other hand, in the case of positive ion/electron coinci-

dence, one can see that they tend to be further in the "positive zone"
(i.e. far above 5000 ns time difference; this is the reason why the zero-
point is asymmetrically in the time window, to allow for a wider range
to detect PEPICO events). This is due to the fact that electrons are
very light, so they will reach the detector considerably faster; e.g. the
CO+ /e – coincidence has a time difference of over 3000 ns due to the
fact that CO+ has M/q = 28 and the electron time of flight can be
though of as almost instantaneous. The lighter the positive fragment
(or lower mass to charge ratio) the shorter flight time difference they
have, relative to electrons.

As can be seen in Fig. 4.1, there are positive, negative and dou-
bly charged fragments present. This is in good agreement with the
non-coincidence study performed by Öhrwall et al [17] for the core-
excitation at the O1s edge, but provides complementary information
directly about the probability of various fragmentation pathways, as
ions are measured in coincidence with each other.

28
 Analysis

4.1.1 Core-level photon energies - O1s edge

Singly-charged fragments in coincidence

The experiment described above is repeated for many photon en-

ergies in order to find out how fragmentation develops at core-
excitations and after core-ionization. For further interpretation, the
intensities of the coinciding ion pairs as a function of photon en-
ergy are plotted and analyzed. Even though the intensity is displayed
in arbitrary units, there is normalization between all the measure-
ments, as they were analyzed in the same manner. The NIPICO
events for singly charged positive ions created around the O1s ion-
ization threshold of 541.2 eV [1] are shown in Fig. 4.2.

Figure 4.2: Singly-charged positive fragments in coincidence with

negative ions as a function of photon energy at O1s edge. Dotted
lines are added to guide the eye; the vertical bars show the excita-
tion energies for O1s to 3p, 4s and 4p orbitals, and the O1s ionization
potential in CO2 .

Such a plot provides valuable insights about the dissociation

channels and their relative probability across the photon energy
range. From Fig. 4.2, one can observe that the most favorable ionic
fragment pairs are: C+ /O – and O+ /O – , revealing considerably higher

29
4.1.1 Core-level photon energies - O1s edge

intensity (directly correlated to abundance) than CO+ or O2 + in coin-

cidence with any anion; the weakest, but still observable ion-pair sig-
nal includes C- ion. As presented in Ch. 2.6.3, above IP, normal Auger
decay is most likely, thus producing CO2 ++ doubly-charged parent
ion. However, there is the possibility of double excitations where the
incident photon produces a O1s→ π∗ excitation and simultaneously
a valence excitation to an unoccupied orbital; the decay in this case
can be resonant Auger thus resulting in CO2 + and explaining the in-
crease in counts at roughly 555 eV.
Although not shown here, present measurements at valence-level
excitations reveal NIPICO cases only with O – as the anion, rather
than C – . This is in agreement with the results from [17]; they con-
cluded that C – was not observed below the O1s edge; coincidences
such as O2 + / C – and O+ / C – were not found at the vicinity of C1s
ionization threshold.
Furthermore, PEPICO events were given the same treatment and
the results are plotted in Fig. 4.3. An immediate observation is the

Figure 4.3: Singly-charged positive fragments in coincidence with

electrons as a function of photon energy at O1s edge.

higher number of cation/electron coincidences (PEPICO) above IP.

After IP, the doubly-charged CO2 ++ parent ion can fragment to two

30
 Analysis

singly-charged cations, meaning that for each electron there are two
possible cations to be detected in coincidence. Another matter is the
low probability for CO2 + production at high photon energies, quite
different to valence ionization where it is the most likely fragment.
This is actually a good demonstration related to the cross-section re-
sult presented in Chapter 2.2: as the valence ionization cross section
decreases towards higher photon energies, it is clear that the corre-
sponding fragmentation channel almost disappears as the energy in-
creases, having close to zero intensity at O1s core-level excitation.
This is due to the valence ionization cross-section values decreas-
ing as the energy reaches higher values [32] and the fact that core
level ionization, and excitation, also bring the molecule into a very
unstable state with energy clearly higher than the double ionization
energy. Here the inner shell electrons have an increasing photoab-
sorption cross-section, thus explaining the low probability for ionic
fragments produced via valence ionization at the O1s edge; they are
still detectable but exhibit the behavior of a slowly decreasing back-
ground, as already seen in [17].

Although it is easy to notice O1s → 3p and O1s → 4p excitations,

the electron/positive ion yields are relatively strong at O 1s → π∗ res-
onance too. Below IP, the most likely decay mechanism, forming
CO2 + parent ion, is resonant Auger and the resulting energetic elec-
trons had the highest probability of bypassing the deflecting mag-
netic field (see Chapter 3.3), thus explaining their high intensity in
a PEPICO plot, in agreement to [17]. Furthermore, it is easy to no-
tice the O 1s → 4p shoulder at 539.96 eV [2], where there is a peak in
abundance for most of the positive fragments, also suggested in [17].

Doubly-charged fragments in coincidence with anions or

electrons

Doubly-charged cations were also found in coincidence with anions

and they are shown in Fig. 4.4. The O 1s → π∗ resonance shows non-
zero intensity, but one can clearly notice the O 1s → 4s , O 1s → 3p
(538.53 eV [2]) and O 1s → 4p (540.3 eV [2]) absorption edges for the
CO2 molecule as peaks of higher intensity.

31
4.1.1 Core-level photon energies - O1s edge

Figure 4.4: Doubly-charged positive fragments in coincidence with

negative ions as a function of photon energy at O1s edge.

As seen in Fig. 4.4, there is considerable intensity of this type of

pairs above IP. Even though the negative ions are not as abundant
as positive ions, the high transmission of the negative ion spectrom-
eter allowed for their detection. Another valuable observation to be
made in this case is the relative abundance between coincidences in-
cluding singly and doubly-charged cations. It is quite noticeable that
the signal coming from singly-charged positive ions is higher than for
doubly-charged cations. This is due to the fact that the fragmenta-
tion of such a dication can occur before the detection (almost imme-
diately in the extraction region, or later on in the accelerating region
or drift tube of the spectrometer), thus reducing its signal strength.
Last but not least, PEPICO events including doubly-charged
cations were also extracted and the events plotted as a function of
photon energy are presented in Fig. 4.5. One notices that the inten-
sity for these events is roughly double than that for NIPICO events
with doubly-charged positive ions, even though they are the same
cationic fragments. This is expected to be due to the fact that elec-
trons are much more abundant than negative ions and, even though
there were measurements taken with the deflecting magnetic field
(see Chapter 3.3) adjusted in such a way that only a small fraction

32
 Analysis

of electrons was able to go through, coincidences with the latter are

more probable than with anions.

Figure 4.5: Doubly-charged positive fragments in coincidence with

electrons as a function of photon energy at O1s edge.

The O 1s → π∗ , O 1s → 4s and O 1s → 4p features show in-

tensity in this case. After the ionization potential of 541.2 eV [1],
doubly-charged states are very likely to be produced by Auger de-
cay, and this can be seen in Fig. 4.5. However, as anions are not
expected in as high abundance as dications, this process is more
obvious in the PEPICO plots, also showing the usefulness of having
electrons at the negative ion detector. Furthermore, electron coinci-
dences can offer information about the parent ion: in the case of a
doubly-charged cation found in coincidence with electron, the par-
ent ion can be doubly-charged, whereas for dication/anion events,
the molecular parent ion is singly-charged. They are nevertheless
noticed in NIPICO events too, due to the high transmission of our
instrument.
Such an analysis is already found helpful when compared to just
measuring ion yields: aside from seeing which ions are present, one
also gets a more detailed view of the fragmentation process, as it is
possible to identify which ions (fragments) originate from the same

33
4.2 Coincidence results including multiple positive fragments

event. However, a more complete analysis, including multiple coin-

cidences, is presented in the following.

4.2 Coincidence results including multiple positive

fragments
So far, single coincidences were investigated only and the possible
fragments originating from the same event were found in different
abundances across the photon energy range studied here. However,
for a triatomic molecule that is not a very complete analysis. To im-
prove that, one needs to find which two or more coincidences origi-
nate from the same event. This section describes the analysis tool to
do so and presents the results found in these measurements.

Figure 4.6: Diagonal map of coincidences for the CO2 molecule with
several events and their intensities encircled in red, at 50 eV excita-
tion. Scale chosen to show region for e-/cation coincidences.

In order to find the coincidences that originate from the same

event, one plots two time difference spectra and related data as
shown in Fig. 4.6 and obtains a diagonal map of scattered points.
This kind of plot tells what the first observed cation was, in coinci-
dence with a negative particle and, the second axis reveals another
observed cation with the same negative particle as trigger. Here, the
data is taken at 50 eV photon energy; this low excitation energy was
used mainly for instrument tests but it also provides reference data
to compare with at core excitation energies, and the following inter-
pretation of results is used here as an example. Where coincidences
occur, there is a higher number of points, as can be seen in Fig. 4.6 in
the encircled areas. On a closer look, one notices how the multitude
of data points tends to form a tilted pattern of a certain slope. This

34
 Analysis

indicates the presence of momentum correlation between the pro-

duced fragments and, analyzing the position, length and width pro-
vides information regarding the masses of the two fragments and the
magnitude of the released kinetic energy both to the ions and neutral
fragments, see, e.g. [33]. However, this work shall restrain to identify-
ing the dissociation pathways, leaving the aforementioned analysis
for a future, more complex study.
Seeing that the 50 eV photon energy used for recording data pre-
sented in Fig. 4.6 is well above the double IP of CO2 (37.2 eV [34]), the
molecule can become doubly ionized:

[CO2 + h ν] → CO2 ++ + 2eph

From this state, the doubly-charged parent ion can fragment to

cations, and data shown in Fig. 4.2 can be explained by the following
fragmentation pathways:

CO2 ++ → CO+ + O+ ,

CO2 ++ → O+ + O+ + C and

CO2 ++ → O+ + C+ + O.

Although not shown in Fig. 4.2, the data also points to direct ion-
pair formation considering that O – anion was observed too, namely:

CO2 + → C+ + O+ + O−

4.2.1 Singly-charged fragments in coincidence with anions

or electrons
Now, moving to core-level excitation energies, PEPIPICO and
NIPIPICO events for singly-charged positive ions at O1s edge are il-
lustrated in Fig. 4.7. The data plotted here represents the summation
of all the counts for each dissociation channel over the whole pho-
ton energy range (O1s edge as illustrated in previous plots). Due to
poor statistics, the normalized counts as a function of energy for each
channel are not presented here.
These present the observed channels of molecular fragmentation
at this photon energy range and the results point to the same dissoci-
ation pathways here, as shown in PEPIPICO at valence-level energies:

CO2 ++ → CO+ + O+ ,

CO2 ++ → C+ + O+ + O and

35
4.2.2 Doubly-charged fragments in coincidence

CO2 ++ → O+ + O+ + C.

It is highly likely that the second and third pathways presented above
are produced via an intermediate CO+ fragment, yielding either
C+ /O or C/O+ pairs.

Figure 4.7: Multiple coincidences for singly-charged cations with an-

ions and electrons.

Less abundant NIPIPICO events were also searched for at this re-
gion and they are also plotted in Fig. 4.7; the probability for these
channels is about one fourth of that for the PEPIPICO events. For
this case the results suggest the following fragmentation pathways:

CO2 + → C+ + O+ + O− and

CO2 + → O+ + O+ + C− .

4.2.2 Doubly-charged fragments in coincidence

Multiple coincidences involving doubly-charged cations were also
explored at core-level excitation energies in both NIPIPICO and

36
 Analysis

PEPIPICO events and the results are illustrated in Fig. 4.8. The ob-
served dissociation pathways including doubly-charged cations in
coincidence with anions are:

CO2 ++ → C++ + O+ + O− ,

CO2 ++ → O++ + C+ + O− and

CO2 ++ → O++ + O+ + C− .

Figure 4.8: Multiple coincidences for doubly-charged cations with

anions and electrons.

It is however obvious that the probability for the carbon diox-

ide molecule to form dissociation pathways consisting of doubly-
charged cations is considerably lower than the case of singly-charged
fragments.
Furthermore, PEPIPICO events including doubly-charged
cations were also considered as can be seen in Fig. 4.8. The relative
strength of these fragmentation patterns is roughly similar to the
aforementioned NIPIPICO case. This is due to the fact that, even

37
4.2.2 Doubly-charged fragments in coincidence

though they were identified in different coincidences (NIPIPICO

and PEPIPICO), they can actually represent the same dataset. For
example, C++ /O+ /O – NIPIPICO event and C++ /e – /O+ PEPIPICO
event can follow form the same dissociation channel C++ /O+ /O – ,
solely the analysis approach is different.
It must be noted that in producing a singly-charged cationic frag-
ment, metastable neutral states can be achieved via the process of
electron recapture: close to the ionization edge, at the onset of Auger
decay [17], it results in a fast Auger electron and a much slower pho-
toelectron. The latter can be recaptured (as explained in Ch. 2.6.3)
and, following dissociation, form a metastable excited neutral state
(atom), or even a negative ion therefore, the pathways could be:

CO2 +∗ → C+ + O + O∗ ,

CO2 +∗ → O+ + C+ + O−∗ ,

CO2 +∗ → O+ + O+ + C−∗ and

CO2 +∗ → CO++ + O−∗ .

Even though the starting parent ion CO2 ++ was indeed doubly-
charged, due to recapture of the slow photoelectron, it becomes
highly likely for the dissociation pathways to yield only one singly-
charged cationic fragment; this is something that cannot be ob-
served with present setup.
To sum up, the CO2 molecule and its dissociation dynamics were
studied both at valence-level excitation energies and at O1s edge. As
expected, both singly-charged and doubly-charged cations were ex-
tracted and anions were recorded. The conclusions that arose from
the recorded and analyzed data are dealt with in the following chap-
ter.

38
 CHAPTER 5

DISCUSSION AND CONCLUSIONS

The dissociation pathways of CO2 were studied, both at low excita-

tion energies (close to valence ionization level) and especially core-
level excitations (around the O1s ionization threshold) and the re-
sults have been reported in the previous section. Both positive and
negative fragments were identified and associated with the proposed
fragmentation channels of varying strengths, as shall further be dis-
cussed in the following.

5.1 Discussion
At valence level excitation energies, the strongest fragment seen is
CO2 in coincidence with an electron, corresponding to photoioniza-
tion where the parent ion does not fragment; the pathway is:

CO2 + h ν → CO2 + + eph

This channel becomes weaker towards higher photon energies, as

the cross-section for valence ionization decreases. Besides the
strongest channel presented above, other fragments were observed.
For photon energies above 37.2 eV (double IP for CO2 [34]) the
molecule can become doubly photoionized:

CO2 + h ν → CO2 ++ + 2 eph

The doubly-charged parent ion usually dissociates yielding CO+ /O+ :

CO2 ++ → CO+ + O+

with CO+ fragmenting further, thus forming the following pathways:

CO+ + O+ → O+ + O+ + C and

39
5.1 Discussion

CO+ + O+ → O+ + C+ + O.

Furthermore, O – anion was observed [17] at valence level pho-

ton energies (below double photoionization) in C+ /O+ /O – channel.
The anion was produced following ion-pair formation, likely via an
intermediary neutral excited fragment CO:

CO2 + → CO + O+

[CO + O+ ] → C+ + O− + O+ .

Moving now to O1s → π∗ excitations, the presence of O2 + /C –

channel is observed. It was proposed [17] that O2 + fragment appears
at this photon energy due to overlapping of 3s Rydberg orbitals. The
conclusions to be drawn are, on one hand, the earlier predictions
and results [17] are confirmed and, on the other hand, the instru-
ment works very well and has a high sensitivity for the detection of
very weak channels (as seen in Fig. 4.2, O2 + /C – channel shows a rel-
atively weak intensity).
Here, the core excited molecule is likely to form a singly-charged
parent ion via resonant Auger process:

CO2 + h ν → CO2 + + eresA

and the most probable dissociation channels include singly-charged

positive ions in coincidence with anions, such as:

CO2 + → C+ + O− + O+ and

CO2 + → C− + O+ + O+ .

Weaker, but still possible channels are doubly-charged positive ions

in coincidence with negative ions such as C++ , O++ and also frag-
ments like CO++ . The latter tend to be very unstable but we have
seen them in coincidence with anions as follows:

CO2 + → CO++ + O−

Above O1s ionization threshold, the most probable event after

photoionization is Auger decay. As has been seen in Chapter 4.1.2,
normal Auger process produced CO2 ++ which is expected to decay
as illustrated by the following [11]:

CO2 ++ → CO+ + O+ → O+ + O+ + C or O+ + O + C+

which yields neutral atomic fragments (O and C).

40
 Discussion and conclusions

Anions were also identified above the O1s IP (C – and O – ) and

their presence was expected to be due to recapture processes, but
our data does not point to that. The likelihood of anion production
at these energies comes from valence and core double excitations.
However, above IP, a strong signal came from doubly-charged
cationic fragments in coincidence with anions and this is easiest to
notice in Fig. 4.4. The dissociation pathways (additionally to the
ones already presented) with anions are:

CO2 + → C− + O+ + O+ ,

CO2 + → C++ + O+ + O− ,

CO2 ++ → O++ + C+ + O− and

CO2 ++ → O++ + O+ + C− .

Triply-charged states are also a possibility, they can be created

in a double Auger decay. [11]. The most likely dissociation chan-
nels for a triply-charged CO2 +++ are C++ /O+ or C+ /O++ [17] which we
have found. However,statistically more reliable results are required
to safely argue the presence of such fragment.

5.2 Conclusions
This work discusses the photoionization of carbon dioxide molecule
at core-level photon energies, with a focus on studying the disso-
ciation dynamics. There has been considerable effort in exploring
the photoionization and ion yields of CO2 , see e.g. [17; 18; 22] all of
them being non-coincidence studies. The newly developed instru-
ment is based on coincidence detection, thus allowing to pinpoint
more precisely how the negative ions are produced (i.e. their path-
ways). This paper confirmed earlier suggestions and offered a more
complete description of the dissociation patterns.
First and foremost, the newly designed negative-ion spectrome-
ter was found to perform very well. The high transmission together
with the deflecting magnetic field allowed for the detection of very
low-abundance negative ions, confirming its capabilities in negative
ion/positive ion coincidence experiments.
Both anions and cations were found and identified at core-
level photon energies. Positive fragments were singly and doubly-
charged; triply-charged cations (CO2 +++ ) are expected to form but
due to their dissociative nature they fragment further much before
they reach the detector.

41
REFERENCES

1. G. Wight and C. Brion, “K-shell energy loss spectra of 2.5 kev

electrons in co 2 and n 2 o,” Journal of Electron Spectroscopy and
Related Phenomena, vol. 3, no. 3, pp. 191–205, 1974.
2. K. Okada, H. Yoshida, Y. Senba, K. Kamimori, Y. Tamenori,
H. Ohashi, K. Ueda, and T. Ibuki, “Angle-resolved ion-yield mea-
surements of co 2 in the o 1 s to rydberg excitation region,”
Physical Review A, vol. 66, no. 3, p. 032503, 2002.
3. D. Harris, “The pioneer in the hygiene of ventilation.,” The
Lancet, vol. 176, no. 4542, pp. 906–908, 1910.
4. E. Almqvist, History of Industrial Gases. Springer US, 2003.
5. B. Bransden and C. Joachain, Physics of Atoms and Molecules.
Prentice Hall, 2003.
6. P. W. Ubachs, Notes on Molecular Physics. Vrije Universitet,
Amsterdam, 2004.
7. J. Hollas, Modern Spectroscopy. Wiley, 2004.
8. K.-H. Gericke, Laserchemistry course. Institut für Physikalische
und Theoretische Chemie der Technischen Universität Carolo-
Wilhelmina zu Braunschweig, 2007.
9. P. Atkins and R. Friedman, Molecular Quantum Mechanics. OUP
Oxford, 2011.
10. G. Miessler, P. Fischer, and D. Tarr, Inorganic Chemistry. Pearson
Education, 2013.
11. A. Kivimäki, C. Stråhlman, T. J. Wasowicz, J. A. Kettunen, and
R. Richter, “Yields and time-of-flight spectra of neutral high-
rydberg fragments at the k edges of the co2 molecule,” The
Journal of Physical Chemistry A, 2016.
12. E. Burhop and W. Asaad, “The auger effect,” vol. 8 of Advances
in Atomic and Molecular Physics, pp. 163 – 284, Academic Press,
1972.
13. L. Meitner, “Über die Entstehung der β -Strahl-Spektren radioak-

43
References

tiver Substanzen,” Zeitschrift fur Physik, vol. 9, pp. 131–144, Dec.

1922.
14. T. Åberg, “Theory of the radiative auger effect,” Phys. Rev. A,
vol. 4, pp. 1735–1740, Nov 1971.
15. T. Åberg and J. Utriainen, “Evidence for a "radiative auger effect"
in x-ray photon emission,” Phys. Rev. Lett., vol. 22, pp. 1346–1348,
Jun 1969.
16. T. A. Carlson and M. O. Krause, “Relative abundances and recoil
energies of fragment ions formed from the x-ray photoionization
of n2, o2, co, no, co2, and cf4,” The Journal of Chemical Physics,
vol. 56, no. 7, pp. 3206–3209, 1972.
17. G. Öhrwall, M. Sant’Anna, W. C. Stolte, I. Dominguez-Lopez,
L. Dang, A. S. Schlachter, and D. W. Lindle, “Anion and cation for-
mation following core-level photoexcitation of co2,” Journal of
Physics B: Atomic, Molecular and Optical Physics, vol. 35, no. 21,
p. 4543, 2002.
18. E. Rühl and R. Flesch, “Mechanism of anion formation in c 1s
→ pi*-excited carbon dioxide.,” The Journal of chemical physics,
vol. 121, no. 11, pp. 5322–5327, 2004.
19. A. Hitchcock, C. Brion, and M. Van der Wiel, “Ionic fragmen-
tation of inner shell excited states of co2 and n2o,” Chemical
Physics Letters, vol. 66, no. 2, pp. 213–217, 1979.
20. S. H. Southworth, R. Wehlitz, A. Picón, C. S. Lehmann, L. Cheng,
and J. F. Stanton, “Inner-shell photoionization and core-hole de-
cay of xe and xef2,” The Journal of chemical physics, vol. 142,
no. 22, p. 224302, 2015.
21. M. Kivimäki, Alagia, R. Richter, et al., “Metastable fragment pro-
duction at the c 1s and o 1s edges of the co2 molecule,” Journal of
Physics B: Atomic, Molecular and Optical Physics, vol. 47, no. 15,
p. 155101, 2014.
22. E. Rühl and H.-W. Jochims, “Core-excitation in molecules:
Formation of cations and anions,” Zeitschrift für Physikalische
Chemie, vol. 195, no. Part_1_2, pp. 137–152, 1996.
23. C. Stråhlman, R. Sankari, A. Kivimäki, R. Richter, M. Coreno, and
R. Nyholm, “A tandem time–of–flight spectrometer for negative–
ion/positive–ion coincidence measurements with soft x-ray ex-
citation,” Review of Scientific Instruments, 2016.
24. H. Wiedemann, Synchrotron Radiation. Advanced Texts in
Physics, Springer, 2003.
25. K. Wille, The Physics of Particle Accelerators: An Introduction.
Oxford University Press, 2000.
26. E. S. Trieste, “Elettra and FERMI lightsources.” http:

44
 References

//[Link]/lightsources/elettra/
elettra-beamlines/gas-phase/specifications/all.
html/, 2016. [Online; accessed May-2016].
27. R. Blyth, R. Delaunay, M. Zitnik, J. Krempasky, R. Krempaska,
J. Slezak, K. Prince, R. Richter, M. Vondracek, R. Camilloni, et al.,
“The high resolution gas phase photoemission beamline, elet-
tra,” Journal of electron spectroscopy and related phenomena,
vol. 101, pp. 959–964, 1999.
28. K. Prince, R. Blyth, R. Delaunay, M. Zitnik, J. Krempasky, J. Slezak,
R. Camilloni, L. Avaldi, M. Coreno, G. Stefani, et al., “The
gas-phase photoemission beamline at elettra,” Journal of syn-
chrotron radiation, vol. 5, no. 3, pp. 565–568, 1998.
29. W. Wiley and I. H. McLaren, “Time-of-flight mass spectrome-
ter with improved resolution,” Review of Scientific Instruments,
vol. 26, no. 12, pp. 1150–1157, 1955.
30. E. Kukk, Electron and ion spectroscopy, Lecture Notes, XFYS4275.
University of Turku, 2005.
31. O. Plekan, M. Coreno, V. Feyer, A. Moise, R. Richter,
M. De Simone, R. Sankari, and K. Prince, “Electronic state
resolved pepico spectroscopy of pyrimidine,” Physica Scripta,
vol. 78, no. 5, p. 058105, 2008.
32. C. Truesdale, D. Lindle, P. Kobrin, U. Becker, H. Kerkhoff,
P. Heimann, T. Ferrett, and D. Shirley, “Core-level photoelectron
and auger shape-resonance phenomena in co, co2, cf4, and ocs,”
The Journal of chemical physics, vol. 80, no. 6, pp. 2319–2331,
1984.
33. E. Itälä, E. Kukk, D. Ha, S. Granroth, A. Caló, L. Partanen,
H. Aksela, and S. Aksela, “Fragmentation patterns of doubly
charged acrylonitrile molecule following carbon core ioniza-
tion,” The Journal of chemical physics, vol. 131, no. 11, p. 114314,
2009.
34. T. Märk and E. Hille, “Cross section for single and double ioniza-
tion of carbon dioxide by electron impact from threshold up to
180 ev,” The Journal of Chemical Physics, vol. 69, no. 6, pp. 2492–
2496, 1978.

45