Taking Snapshots

of Atomic Wave Functions

with a Photoionization Microscope

Aneta S. Stodolna
Cover: Pictures on the cover show three photoelectron images of the hydrogen atom,
which were recorded with the use of a photoionization microscope, for slightly dif-
ferent excitation energies (increasing from left to right). The central image diverges
significantly in its shape and size as in this particular case a Stark state is excited
and an atomic wave function dominates the observed structure. This image, which
directly shows an electronic orbital in hydrogen is the most important result of the
thesis and has received a significant attention in the scientific and social media all
over the world (e.g. 70.000 ’likes’ on Facebook).

ISBN: 978-90-77209-80-6
Copyright © 2014 by Aneta S. Stodolna.
An electronic copy of this thesis is available at www.amolf.nl/publications.
 Taking Snapshots
of Atomic Wave Functions
with a Photoionization Microscope

Proefschrift

ter verkrijging van de graad van doctor

aan de Radboud Universiteit Nijmegen
op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann,
volgens besluit van het college van decanen
in het openbaar te verdedigen op vrijdag 20 juni 2014
om 10.30 uur precies

door

Aneta Sylwia Stodolna

geboren op 20 maart 1984
te Ostroda, Polen
 Promotor: Prof. dr. M.J.J. Vrakking

Manuscriptcommissie: Prof. dr. D.H. Parker

Prof. dr. H. Helm (Freiburg University, DE)
Prof. dr. H.J. Bakker (FOM Institute AMOLF)

The work described in this thesis was performed at the FOM Institute for Atomic and
Molecular Physics (AMOLF), Science Park 104, 1098 XG Amsterdam, The Netherlands.
This work is part of the research programme of the ‘Stichting voor Fundamenteel Onderzoek
der Materie (FOM)’, which is financially supported by the ‘Nederlandse Organisatie voor
Wetenschappelijk Onderzoek (NWO)’.
To My Husband
Contents

1 Introduction 1
1.1 Old Quantum Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Schrödinger equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Attempts towards measuring ψ . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Scope of this thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2 Stark effect 13
2.1 History of the discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 The Stark effect in the point of view of the Quantum Mechanics . . . . . . 15
2.2.1 The Schrödinger equation in parabolic coordinates . . . . . . . . . . 16
2.2.2 Linear and quadratic Stark effect by means of perturbation theory . 19
2.3 Energy levels in a presence of electric fields – Stark maps . . . . . . . . . . 21
2.3.1 Hydrogen atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.3.2 Nonhydrogenic atoms . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.4 Field ionization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.4.1 Hydrogen atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.4.2 Nonhydrogenic atoms . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5 Theoretical calculations of the Stark problem . . . . . . . . . . . . . . . . . 31
2.5.1 Perturbation theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.5.2 Wentzel-Kramers-Brillouin – Quantum Defect theory (WKB-QD) . . 33
2.5.3 Wave packet calculations . . . . . . . . . . . . . . . . . . . . . . . . 34

3 Experimental facilities 39
3.1 Photoionization microscopy of hydrogen atoms . . . . . . . . . . . . . . . . 39
3.1.1 Characteristics of the atomic hydrogen source . . . . . . . . . . . . . 41
3.1.2 Transport and density of H atoms in the interaction region . . . . . 43
3.1.3 Two-photon excitation into the mixture of 2 s– and p–states . . . . 44
3.1.4 Photoionization of excited hydrogen atoms – Pulsed Dye Amplifier . 47
3.1.5 Detection of photoelectrons – velocity map imagining (VMI) spec-
trometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2 Photoionization microscopy of helium atoms . . . . . . . . . . . . . . . . . . 53
3.2.1 Metastable helium source . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2.2 Photoionization of He (21 S) atoms with UV light . . . . . . . . . . . 58
3.2.3 Detection of photoelectrons – Penning ionization problem . . . . . . 60

4 Semiclassical interpretation of the interference patterns in the photoion-

ization microscopy experiments of hydrogen atoms 61
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2 Classical theory of the interference phenomenon . . . . . . . . . . . . . . . . 62
4.2.1 Motion along the ξ coordinate . . . . . . . . . . . . . . . . . . . . . 63
4.2.2 Motion along the η coordinate . . . . . . . . . . . . . . . . . . . . . 64
 4.2.3 Direct and indirect photoelectron trajectories . . . . . . . . . . . . . 65
4.2.4 Interference patterns – semiclassical wave function . . . . . . . . . . 66
4.3 Experimental results and discussion . . . . . . . . . . . . . . . . . . . . . . 68
4.3.1 Semiclassical interpretation of the interference patterns . . . . . . . 69
4.3.2 Direct and indirect ionization processes . . . . . . . . . . . . . . . . 70
4.3.3 Beating effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4 Resonance effects in photoionization microscopy . . . . . . . . . . . . . . . . 76
4.5 Resonance effects in experiments and theory to date . . . . . . . . . . . . . 78

5 Hydrogen atoms under magnification: direct observation of the nodal

structure of Stark states 81
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.2 Resonance phenomenon in photoionization microscopy . . . . . . . . . . . . 82
5.3 Experimental results and discussion . . . . . . . . . . . . . . . . . . . . . . 84
5.3.1 m = 0 Stark states in high electric fields . . . . . . . . . . . . . . . . 85
5.3.2 m = 1 Stark states in low electric fields . . . . . . . . . . . . . . . . 88
5.3.3 Differences between resonant and non-resonant excitations . . . . . . 91
5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

6 Visualizing the coupling between red and blue Stark states using pho-
toionization microscopy 93
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.2 Ionization process in hydrogenic and nonhydrogenic . . . . . . . . . . . . . 94
6.3 Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.4 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.4.1 Above the V (η) barrier ionization of helium atoms . . . . . . . . . . 98
6.4.2 Interference narrowing effects in helium . . . . . . . . . . . . . . . . 100
6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

References 107

Summary 121

Samenvatting 125

Acknowledgements 129

List of publications 133

About the author 135

1 Introduction

1.1 Old Quantum Mechanics

Towards the end of the 19th century various branches of physics had been inte-
grated in a general and coherent theory – Classical P hysics. The classical program
divided objects into two categories: matter – perfectly localizable corpuscles obey-
ing Newton’s laws of classical mechanics and radiation – exhibiting a wave-like
behavior, manifested in the phenomena of interference and diffraction, and following
Maxwell’s laws of electromagnetism. The undeniable success of classical physics lies
in the fact that almost all known phenomena at that very time could be explained
by the theory of matter and radiation. Whenever the explanation could not be found
the failure was attributed to the mathematical difficulties in solving the problem –
but the classical program was never questioned.
Physicists of the classical period, proceeding towards greater unity of various
branches of science, attributed more unity to the classical theory than it essentially
had. The most striking example comes from a consideration of the wave propagation
phenomenon. First observed for matter, e.g. surface waves of a liquid, provided
physicists with a clue that similarly to surface waves on the water also light needs
a medium to transmit its waves. This medium, the aether, was assumed to have
a set of highly unusual properties since light and material bodies pass through it
without any drag or friction [1]. In the 19th century numerous experiments did
not reveal the nature of the expected medium. Moreover, the famous Michelson -
Morley experiment in 1887 [2], designed to measure the speed of the aether with
high precision, gave a negative result that was later explained by Einstein [3] in
1905 within the framework of the special theory of relativity. Finally the concept of
the aether was abandoned.
At the turn of the 20th century various experiments exploring the structure of
matter and its interaction with electromagnetic radiation brought unexpected results
which could not be understood in a classical manner. It became evident that the
classical physics is not as general as physicists believed.
The first disagreement emerged by studying radiation of the black body. Accord-
ing to the classical physics this is a body that absorbs all incident electromagnetic
radiation and while being at thermal equilibrium emits radiation proportional to
frequency squared [4]. Consequently, with an increase of frequency the emitted ra-

1
Chapter 1. Introduction

Figure 1.1: Spectral radiance distribution of the black body radiation as a function of wave-
length calculated using a classical formula of the Rayleigh-Jeans law (blue curve) and the
Planck’s law (red, green and black curves) for different temperatures. As the black body
temperature increases the intensity maximum shifts to shorter wavelengths.

diation goes to infinity leading to the so-called ”ultraviolet catastrophe” (see blue
curve in Fig.1.1). However, experiments showed that black body radiation has a
continuous frequency spectrum with a characteristic frequency peak, which shifts
with the black body’s temperature (red, green and black curves in Fig.1.1). In 1900
Planck succeeded in calculating the radiation curve by postulating that the energy
exchange between matter and radiation takes place by discrete and indivisible por-
tions of energies - quanta [5, 6]. Planck postulated that the quantum of energy must
be proportional to the frequency ν of the radiation via:

E = hν, (1.1)

where h is the proportionality factor known as the Planck’s constant. The hypoth-
esis of quanta seemed unacceptable to most of the contemporary physicists and the
Planck’s success was perceived as a lucky mathematical trick. However, Planck’s the-
ory was to be confirmed by numerous experiments investigating matter and radiation
on the microscopic scale.
With the discovery of radioactivity in 1896 [7] first properties of atomic nuclei

2
 1.2 The Schrödinger equation

were investigated. One year later Thomson while working with cathode rays discov-
ered the electron [8] and in 1904 proposed the ”plum pudding model” of the atom
[9]. According to this model the atom is composed of electrons surrounded by a soup
of positive charge that balances negative charge of electrons. In 1911 Rutherford di-
rected a beam of alpha particles on a thin gold foil and by studying the scattered
particles proposed a first planetary model of the atom [10]. The Rutherford atom
comprises a positively charged, tiny and heavy nucleus and surrounding electrons.
However, this model did not explain the atom’s stability and predicted a continuous
emission/absorption spectrum of light by the atom. The latter was contradictory
with experiments showing that emission and absorption spectra are composed of
narrow spectral lines. In 1913 Bohr [11] succeeded in explaining the atomic spectra
by postulating that the energy levels of an atom are quantized. Bohr introduced
the idea that electrons do not travel freely around the nucleus but move along a
certain number of orbits (stationary states) with a well-defined energy. Within the
framework of Bohr’s theory an electron can drop from a high-energy state E i to a
lower state E k by emitting a photon hν, which frequency satisfies the relation:

hν = Ei − Ek . (1.2)
Similarly, in the presence of light the atom in a low-energy state E k may absorb
a photon hν and undergo a transition to a higher energy state E j provided that the
total energy is conserved:

hν = Ej − Ek . (1.3)
In 1914 the experiment by Franck and Hertz [12] on inelastic collisions between
electrons and atoms brought another confirmation of the quantization of atomic
energy levels. Two years later Sommerfeld [13] extended the Bohr’s theory by in-
troducing elliptical orbits, the orientation of the orbital plane (spatial quantization)
and a relativistic correction for the mass of the electron giving birth to what is now
known as the Old Quantum M echanics. The Bohr-Sommerfeld theory provided an
explanation of the Stark effect [14, 15], the ordinary Zeeman effect [16] and the
fine structure of the hydrogen spectrum. Nevertheless, this theory contained errors,
ambiguities, contradictions and failed with aperiodic systems [17]. Furthermore, the
quantization rules were determined in an empirical manner.

1.2 The Schrödinger equation

In 1923 de Broglie [18] stated the hypothesis that the wave-corpuscle duality is
a general property of microscopic objects. Hence, as light waves are known to have
corpuscular properties, shown by Einstein in the photoelectric effect [19], moving
particles should as well have wave properties given by de Broglie’s equation:

h
λ= , (1.4)
p

3
Chapter 1. Introduction

where λ is the wavelength of the particle and p is its momentum. At about the
same time Heisenberg [20] introduced a new representation of quantum mechan-
ics – M atrix M echanics. In this new theory only physically observable quantities
are represented by matrices which, in contrast to the classical mechanics, obey a
non-commutative algebra. In the meantime, Schrödinger [21], while working on de
Broglie’s theory of W ave M echanics, obtained the equation of propagation of a
wave function ψ representing a quantum system:

∂
i~ ψ = Hψ,
b (1.5)
∂t

where i is the imaginary unit, ~ is the reduced Planck constant and H b is the
Hamiltonian operator. The undeniable power of the Schrödinger equation lies in
the fact that by knowing the wave function ψ one could predict and describe all
dynamical properties of the quantum system at a given instant of time t. Moreover,
Schrödinger has shown that both matrix mechanics and wave mechanics, although
mathematically different, lead to the same result and are equivalent.
Soon after Schrödinger’s publication, Born interpreted the wave function ψ in
the wave equation by stating that since ψ has a certain spatial extension one cannot
attribute to a quantum particle a precise position (or momentum) but can only
define a probability of finding the particle in a given region of space (or momentum
space) [22]. The probability density P (r), describing a probability of finding the
particle in the volume element (r, r + dr), is related to the wave function ψ(r) via:

P (r) = ψ ∗ (r)ψ(r) = |ψ(r)|2 , (1.6)

where ψ ∗ (r) denotes the complex conjugate of ψ(r). In other words, relation (1.6)
says that the larger the probability of the presence of the particle the more intense
the wave at that point is.
Few months later Heisenberg formulated a general consequence of the statistical
interpretation of the wave-corpuscle nature of quantum objects – the uncertainty
principle – which says that it is impossible, under any circumstances, to attribute
simultaneously to the quantum particle a precise position and a precise momentum
[23]. Finally Bohr completed the theory of Quantum Mechanics with the comple-
mentarity principle, which indicates that it is not possible to observe both the wave
and corpuscle aspects of quantum objects simultaneously, but together they present
a full description of the microscopic objects [24].
Consequently, a new theoretical model of the hydrogen atom was presented,
where the probability of finding the electron in an element volume dV is calculated
by means of the Schrödinger equation (1.5). With boundary conditions imposed upon
the wave function ψ, the allowed solutions of the wave equation create a discrete
set of solutions, so-called eigenfunctions, where each of them represents an energy
state of the atom. There exists an analytical solution of the Schrödinger equation
for the hydrogen atom if the electron position is given in spherical coordinates r, ϕ

4
 1.2 The Schrödinger equation

and θ. Consequently, the wave function ψ might be replaced by a product of three

functions along each coordinate, i.e. by:

ψ(r, ϕ, θ) = R(r) · Φ(ϕ) · Θ(θ), (1.7)

leading to separation of the wave equation into three total differential equations.
As a result, three separation constants are introduced, which are linked to three
quantum numbers n, m and l. These numbers can be associated with quantum
numbers introduced by the Bohr-Sommerfeld model, i.e. n is the total quantum
number that characterizes the basic energy level for the electron and describes the
size of the orbital, l - the azimuthal quantum number characterizes the electron’s
angular momentum and indicates the shape of the orbital, m - the magnetic quantum
number specifies the orientation in space of the orbital. In Quantum Mechanics these
numbers have values:

n = 1, 2, 3, . . . , ∞,
l = 0, 1, 2, . . . , n − 1,
m = 0, ±1, ±2, . . . , ±l.
(1.8)

For a given set of the quantum numbers n, l and m the complete eigenfunction
ψnlm , which satisfies the wave equation, is of the form [25]:

s
eimϕ (2l + 1)(l − |m|)! |m|
ψnlm = √ (sin|m| θ)Pl (cosθ)·
2π 2(l + |m|)!
s l
4(n − l − 1)!Z 3 2Zr
  
−Zr
2l+1 2Zr
· e na0
Ln+a , (1.9)
[(n + l)!]3 n4 a30 na0 na0

|m|
where Z is the atomic number,
 a0 
is the Bohr radius, Pl (cosθ) is the associated
Legendre polynomial and L2l+1
n+a
2Zr
na0 is the associated Laguerre polynomial.
Fig.1.2 shows the probability density, calculated using Eq. (1.9), for the hydrogen
atom in the states: n = 6, l = 0 → 5 and m = 0. One can see that for different
values of the azimuthal quantum number l orbitals have different shapes, which are
best described as spherical (l = 0), polar (l = 1), or cloverleaf (l = 2). As the value
of the angular quantum number becomes larger more complex shapes are observed.
Moreover, the azimuthal quantum number gives the number of planar nodes going
through the nucleus placed at the center of the coordinate system. Closer inspection
of plotted charge distributions reveals that the nodal lines of the eigenfunctions are
seen to be concentric circles and radii.

5
Chapter 1. Introduction

(6,0,0) (6,1,0)

z
x

(6,2,0) (6,3,0)

(6,4,0) (6,5,0)

Figure 1.2: Charge distributions for the hydrogen atom in the spherical states (n,l,m) =
(6,0,0), (6,1,0), (6,2,0), (6,3,0), (6,4,0) and (6,5,0). Note that the probability-density dis-
tribution has been multiplied by r for a better visibility.

6
 1.3 Attempts towards measuring ψ

1.3 Attempts towards measuring ψ

Although the wave function ψ, also known as a ”quantum state”, is a central
concept of Quantum Mechanics its observation remains elusive. By using the wave
function to calculate the expectation value of experimental observables, one can
obtain properties like the position of an atom or the momentum of the electron.
However, according to the Heisenberg uncertainty principle these projections only
reveal a fraction of the information encoded in the wave function and destroy other
intriguing features as superposition or entanglement. It has been shown [26] that it
is impossible to determine a completely unknown wave function of a single state.
Nevertheless, in the last decades different direct and indirect experimental methods
for imaging the wave function of photons, nuclei, atoms and molecules have been
developed.
The oldest method of mapping wave functions of atoms, molecules and nuclei em-
ploys collision experiments. A possible experimental setup requires a beam of fast
particles (electrons or protons) and a target material M . When the impact particle
has high kinetic energy it can kick out a particle from the target at large polar angles
ensuring that the momentum transfer in the inelastic collision is maximized and very
large. The recoil momentum of the ion M + is equal and opposite in sign to the mo-
mentum of the target particle immediately prior to the ”knock-out”. By varying the
azimuthal angle between the two outgoing particles for each characteristic binding
energy a measurement of the cross-section directly gives the momentum distribution
of the target particle |ψ(p)|2 prior to knock-out. This technique of electron momen-
tum spectroscopy (EMS) (i.e., binary (e, 2e) spectroscopy) allowed measurement
of the momentum distribution of electrons in the ground state of atomic hydrogen
[27] as well as in helium and other many-electron atoms [28], and small molecules
[29]. More recently a spherically averaged Dyson orbital electron momentum density
distribution of the highest occupied molecular orbital (HOMO) has been observed
for different molecules [30]. Similarly to the (e, 2e) reaction (p, 2p), (p, pn), (e, e’p),
(γ, p) and (p, pα) reactions have been used to map wave functions of nuclei [31, 32].
Photons are of particular interest to quantum information processing (QIP) as
owing to their mobility they are irreplaceable carriers of information in quantum
communication networks [33]. A successfully implementation of photons in QIP re-
quires inter alia a detail characterization of a quantum state. This can be accom-
plished by means of tomographic methods, where the wave function of a quantum
system is estimated by employing reconstruction techniques. Quantum-state tomog-
raphy (QST) by taking many sets of measurements allows obtaining a quorum -
a set of observables providing tomographically complete information about a quan-
tum state. The first experimental demonstration of optical QST in 1993 [34] allowed
the reconstruction of the Wigner function and the density matrix of a squeezed
state of light. Consecutive tomographic experiments provided complete information
of different quantum states e.g. the single-photon Fock states [35] or the squeezed
”Schrödinger cats” [36]. Alternatively, one can measure the wave function ψ of a
single photon by means of so-called weak-measurements [37]. In this method pho-
tons with the same initial wave function ψ(x) first undergo a weak measurement of

7
Chapter 1. Introduction

3
a bb 0.8
2 2 0.5

1 0.6
1

Y (Å)
Y (Å)

Y (Å)

2
|y|
0 0 0.4
0
-1
0.2
-2 -1 -0.5
0
-3
-3 -2 -1 0 1 2 3 -1 -0.5 0 0.5 1
X (Å) X (Å)

Figure 1.3: Examples of a molecular and an atomic orbital measured by means of orbital
tomography. (a) Reconstructed wave function of the HOMO of N2 obtained from a tomo-
graphic inversion of the high-harmonic spectra [38]. Note that this wave function has both
positive and negative values meaning that this image shows ψ(x) up to an arbitrary phase.
(b) Retrieved square of the wave function |ψ(r)|2 of the 2p orbital in Ne reconstructed from
the high-harmonic spectrum [39].

a transverse position, which does not disturb the state, and then their momenta are
measured normally (a strong measurement). Although this method is called by the
authors as ”direct”, it only means that this method is free from complicated sets of
measurements and computations.
Tomographic methods found also applications in mapping the highest-lying or-
bitals of molecules, which are of particular interest as they are responsible for chem-
ical properties. For example the HOMO and LUMO (lowest unoccupied molecular
orbital) can be reconstructed for molecules, placed on a thin film, by means of the
ultraviolet angle-resolved photoelectron spectroscopy (ARPES) [40]. The advantage
of ARPES is that it works for low photon energies, minimizing damage to the sam-
ple, and that it does not require a tunable photon source. Another approach has
been shown by the group of Corkum et al. [38] who used high-harmonics, generated
from intense femtosecond laser pulses focused on aligned N2 molecules, to obtain the
full three-dimensional shape of the HOMO. A crucial element of this tomographic
technique is that it makes use of the high-harmonic emission process from aligned
molecules, which is known to be sensitive to the spatial structure of the electronic
wave function [41]. In brief, an intense low-frequency laser ionizes an aligned molecule
and consequently a fraction of electron wave function tunnels from the HOMO. When
the laser field direction reverses this electron wave packet ψc is driven back to the
molecule where it overlaps with the remaining portion of the initial wave function
ψg . By coherent addition of ψc and ψg a dipole is induced, seen as the asymmetric
electron density distribution, which oscillations result in emission of high harmonic
radiation. The high harmonic spectra, measured for different angles between the
molecular axis and the re-colliding electron, are tomographically inverted and the
initial wave function ψg is reconstructed (see Fig. 1.3a). However, there are two main
limitations of this method: (1) the orbital must be fixed in the laboratory frame and
(2) the coupling between ionization in the field and the recollision process must be
disentangled. Therefore, the orbital tomography in this form is only applicable for

8
 1.3 Attempts towards measuring ψ

a b

HOMO LUMO

Figure 1.4: (a) Spatial image of a quantum corral, built by placing 48 Fe atoms into a
ring on a metal surface, with visible eigenstates obtained by means of STM [45]. (b) Direct
comparison between the experimental STM images of the highest-lying orbitals of a molecule
placed on an insulating layer (top panel) with the calculated electron densities |ψ(r)|2 for a
free molecule (bottom panel) [46].

systems that can be naturally aligned and only for the simplest molecular orbitals
as the HOMO and HOMO-1[42]. Recently, these limitations have been overcome
by adding a second-harmonic field polarized orthogonally to the fundamental field
[39]. By varying the time delay between the two harmonic fields the angle between
ionization and recollision is controlled and resolved. This extended version of the
orbital tomography allowed reconstruction of the 2p orbital for neon (see Fig. 1.3b).
It is also possible to expand this method to unaligned molecules if there exists a
characteristic angular distribution of the tunneling ionization probability, which can
be related to a specific orbital, such as in CO2 [43] or in C2 H6 [44].
Wave functions might also be observed in a more direct way in a sense that
already raw images, without any reconstruction procedures, reveal elements of the
quantum state under examination. One of such a direct method is scanning - tun-
neling microscopy (STM), which have allowed the observation of eigenstates of the
so-called ”quantum corral”, created by placing 48 Fe atoms into a ring on a Cu
surface [45] (see Fig. 1.4a). STM imaging techniques rely on the nonzero conduc-
tance of its tunneling junction meaning that atoms/molecules must be positioned
on metals or semiconductors. As the experiment on the quantum corral showed,
the electronic structure of a single atom is strongly perturbed by the surface elec-
trons. However, this limitation might be overcome by using insulating films of a
thickness of few atomic layers, what ensures a sufficient electronic decoupling and
consequently observation of the HOMO and LUMO of an individual molecule [46]
(see Fig. 1.4b). Alternatively, one can use the molecular-frame photoelectron angular
distribution (MFPAD) method to observe nodal planes of the HOMO and HOMO-1
in polar molecules [47]. In this technique molecules with a permanent dipole mo-
ment are three-dimensionally oriented by combined laser and electrostatic field and
subsequently ionized with intense, circularly polarized laser pulses, which prevents
recollision of the free electron with its parent ion. The presence of the nodal plane

9
Chapter 1. Introduction

in the HOMO leads to a left-right splitting of the momentum distribution along

the x-axis (e.g. parallel to the detector plane), what can be directly observed in
photoelectron images. Very recently it was shown that in experiments on aligned
molecules, which are photoionized by extreme ultraviolet (XUV) laser light polar-
ized perpendicularly to the internuclear axis, the recorded MFPADs directly reveal
the nodal structure of molecular wave functions [48].

1.4 Scope of this thesis

Additional possibilities for imaging the nodal planes of atomic orbitals can be
created by placing atoms into a DC electric field, which introduces a quantization
axis and aligns the orbitals prior to being measured. When atoms are photoex-
cited into quasi-bound Stark states with narrowband laser pulses, projection of the
launched photoelectrons on a position-sensitive detector, placed perpendicularly to
the electric field direction directly reveals the nodal structure of the atomic orbitals
prepared in Stark states prior to ionization. This is the principle of an imaging tech-
nique called photoionization microscopy (PM), proposed more than 30 years ago
[49–51]. A hundred years after Bohr introduced the model of quantized energy levels
in hydrogen atoms it was possible, for the first time, to directly measure atomic
orbitals of hydrogen by means of photoionization microscopy [52], as presented in
Chapter 5.
This thesis deals with experiments employing the photoionization microscopy
technique and making use of the simplest atoms in nature: hydrogen and helium. It
may convey the impression that the helium atom is slightly more complex than hy-
drogen since it only has one additional electron and its ionic core contains one extra
proton and two neutrons. However, from the point of view of quantum mechanics
these two atoms are fundamentally different and represent two separate groups of
atoms. Namely, for hydrogenic atoms the quantum mechanical calculations can be
performed exactly and with high accuracy while for nonhydrogenic atoms various
approximation methods need to be used to carry out calculations to a high degree
of accuracy. In Chapter 2 it is shown that for the hydrogen atom the Schrödinger
equation (1.5) in parabolic coordinates has exact analytical solutions and in the
presence of an external electric field it remains separable. On the contrary, the finite
size of the ionic core in helium leads to the situation when the wave equation be-
comes nonseparable. Consequently, both atoms behave differently while exposed to
an external electric field, a key element of photoionization microscopy, and ionize in
a distinct manner. Chapters 5 and 6 illustrate that due to these dissimilarities the
nodal structure of the electronic wave function is always observable in the experi-
ments on hydrogen atoms (Chapter 5), whereas in the case of helium the observation
is possible owing to the interference narrowing effect (Chapter 6).
This thesis is organized as follows. Chapter 2 contains a detailed theory of the
Stark effect in the hydrogen atom and demonstrates distinct properties of hydrogenic
and nonhydrogenic atoms exposed to external electric fields. Particularly, atomic en-
ergy levels, with a special interest on crossings between states, and field ionization

10
 1.4 Scope of this thesis

mechanisms are considered. Moreover, in the last part of Chapter 2 different the-
oretical approaches to the Stark problem, which are used in calculations presented
in this thesis, are discussed. Chapter 3 provides a detailed description of the ex-
periments on hydrogen and helium atoms, with the emphasis on crucial elements
in the experimental setups, the efficiency of the overall processes and the data ac-
quisition procedure. Prior to the discussion of the experimental results, a theory
of photoionization microscopy, within the framework of semiclassical and quantum
mechanical theory, is presented in Chapter 4. The same chapter introduces exper-
imental results on interference effects in hydrogen, which provide a deeper insight
into the interactions between the core and the photoelectron since they are reflected
in various classical trajectories by which the photoelectron moves en route the de-
tector. Moreover, the interference images reveal, for the first time, beating patterns
between so-called direct and indirect photoelectron trajectories. Resonant phenom-
ena in hydrogen are treated extensively in Chapter 5 and the presented experimental
results show in a convincing way that the nodal structure of electronic wave func-
tions is directly observable in these experiments. Finally, Chapter 6 demonstrates
the importance of the interference narrowing effect in the visualization of nodes of
wave functions in helium atoms and by that very fact indicates the fundamental
differences between hydrogenic and nonhydrogenic atoms.

11
2 Stark effect

2.1 History of the discovery

In 1896 the Dutch experimental physicist Zeeman [53] discovered that a mag-
netic field splits spectral lines of atoms into doublets or triplets. Motivated by this
discovery many attempts were made to find an analogous effect caused by an electric
field. In 1901 Voigt [54] analyzed this problem within the framework of classical elec-
trodynamics and calculated that two sodium D lines in an electric field of 300 V/cm
should be split by about 3×10−4 Å, which suggested that such an effect could not be
observed experimentally. Regardless of Voigt’s negative result Stark, who was one
of the earliest supporters of the quantum theory, decided to experimentally study
the effect of an electric field on spectral lines. By using hydrogen and a transverse
electric field of the order of 104 V/cm he observed in 1913 a splitting of the hydrogen
H β and H γ lines into several components [55]. Simultaneously and independently
the Italian physicists Lo Surdo observed splitting of the same hydrogen lines while
studying the spectra emitted by so-called ”retrograde rays” [56, 57]. The advantage
of Lo Surdo’s experiment was that the gradient of the electric field present in a dis-
charge tube, directly revealed the relationship between the strength of the electric
field and the splitting (see Fig. 2.1a-b).
The first satisfactory theoretical interpretation of the Stark effect was already
obtained in 1916 independently by Epstein [14] and Schwarzschild [15] within the
framework of the Old Quantum Mechanics. To examine the influence of an external
electric field on an electron they used a system of parabolic coordinates, which is
defined as follows:

ξ = r + z,
η = r − z,
y
ϕ = tan−1 .
x
(2.1)

and

13
Chapter 2. Stark effect

a b c

Figure 2.1: Longitudinal (a) and transverse (b) Stark effects observed for the hydrogen Hγ
line in the experiment by J.S. Foster [58]. (c) A schematic representation of the electron
motion (black thick curve) in parabolic coordinates according to the old quantum mechanics
interpretation. Confined to this region the electron has three periodic motions, i.e. around
the electric field direction given by m, along the ξ and η coordinates given by nξ and nη ,
respectively [25]

p
x= ξηcosϕ,
p
y= ξηsinϕ,
1
z= (ξ − η),
2
1
r = (ξ + η),
2
(2.2)

and it is drawn schematically in Fig. 2.1c. The surfaces with constant ξ or η are
paraboloids obtained by revolving about the electric field axis. In this coordinate
system the nucleus is placed at the focus (x = y = z = 0). Note that atomic units
are used, unless specified otherwise.
According to their interpretation the electron follows an elliptical orbit confined
to the space between the parabolas (ξmin , ξmax ) and (ηmin , ηmax ) while the plane
of the orbit processes about the azimuthal angle ϕ (see Fig .2.1c). The total energy
of the electron is given by [59]:

E = E0 + ∆E, (2.3)

where E0 is the field-free energy of the electron expressed by E0 = 1/2n2 and ∆E

represents the shift in the energy levels due to the presence of the external electric

14
 2.2 The Stark effect in the point of view of the Quantum Mechanics

field and can be expanded as a power series in the electric field strength F as:

∆E = AF + BF 2 + CF 3 + · · · . (2.4)

In parabolic coordinates the total principal quantum number is given by n =

nξ + nη + nϕ , where nξ and nη are the so-called parabolic quantum numbers and nϕ
is the azimuthal quantum number. By analogy to the azimuthal quantum number l
in the Kepler problem nϕ can be expressed as nϕ = m + 1, where m = 0, 1, 2, . . .
[59]. The linear term AF and the quadratic term BF 2 represent the first– and the
second–order Stark effect, respectively, with A and B given by:

3
A= n(nξ − nη ), (2.5)
2

1 4
B=− n {17n2 − 3(nξ − nη )2 − 9m2 + 19}. (2.6)
16
The quadratic Stark effect was experimentally observed in 1918 by Takamine
and Kokubu [60]. One year latter Kramers [61] estimated the relative intensities of
spectral lines and determined the polarizations of the hydrogen-line components,
ipso facto obtaining an excellent agreement with Stark’s measurements.

2.2 The Stark effect in the point of view of the

Quantum Mechanics
The Old Quantum Mechanics, despite being successful in giving a quantitative
explanation of the Stark effect, was actually full of ambiguities, as has been em-
phasized at the end of Section 1.1. Only a few months after Schrödinger’s break-
through publication Epstein approached the Stark effect problem again, this time
from Schrödinger’s quantum theory point of view [62]. Once more an excellent agree-
ment between the calculated positions of lines with the experimental observation was
obtained and the new expressions for the linear and quadratic Stark effect were iden-
tical with the ones attained by means of the Old Quantum Mechanics. Furthermore,
within the framework of the new theory of Quantum Mechanics the restriction for
the azimuthal quantum number nϕ > 0 appeared automatically. That was a clear
advantage over the Bohr-Sommerfeld theory, where this restriction was artificially
introduced to eliminate plane orbits in which the electron would sooner or later
undergo a collision with the nucleus.

15
Chapter 2. Stark effect

2.2.1 The Schrödinger equation in parabolic coordinates

A single positively charged proton, which constitutes a hydrogen nucleus, pro-
duces a Coulomb potential, i.e. a potential depending only on the distance r of the
particle from the center of force, ensuring that the Hamiltonian operator has spher-
ical symmetry and the wave equation (1.5) is separable in spherical coordinates, as
shown in Section 1.2. While dealing with perturbation problems, where an external
force distinguishes a particular direction in space (e.g. Stark effect, photoelectron ef-
fect, etc.), one can use parabolic coordinates ξ, η, and ϕ to separate the Schrödinger
equation [63–65].
The Hamiltonian operator for the hydrogen atom is given by:

2
b = − ∇ + V (r),
H (2.7)
2
where ∇2 is the Laplacian operator and V (r) = −1/r represents the Coulomb
potential. Thus, the time independent Schrödinger equation in parabolic coordinates
may be written as:

" #
∇2 2
− − ψ = Eψ, (2.8)
2 ξ+η

with the Laplacian operator in parabolic coordinates expressed by:

! !
2 4 ∂ ∂ 4 ∂ ∂ 1 ∂2
∇ = ξ + η + . (2.9)
ξ + η ∂ξ ∂ξ ξ + η ∂η ∂η ξη ∂ϕ2

To find solutions of the wave equation (2.8) one can assume that the wave function
is of the form:

ψ(ξ, η, ϕ) = u1 (ξ)u2 (η)eimϕ , (2.10)

where m is the magnetic quantum number. Substituting Eq. (2.9) and (2.10) into
Eq. (2.8) and separating it with constants Z1 and Z2 related by Z1 + Z2 = 1, one
obtains two independent equations for functions u1 (ξ) and u2 (η):

! !
d du1 1 m2
ξ + Eξ + Z1 − u1 = 0, (2.11a)
dξ dξ 2 4ξ
! !
d du2 1 m2
η + Eη + Z2 − u2 = 0. (2.11b)
dη dη 2 4η

16
 2.2 The Stark effect in the point of view of the Quantum Mechanics

The solutions of these equations

√ can be found by assuming that u1 (ξ) behaves as
1
1 |m|
exp(− 2 εξ) for large ξ, where ε = −2E, and as ξ 2 for small ξ. Accordingly, the
same assumptions should be made for u2 (η). As a result the normalized eigenfunction
for the hydrogen atom is given in terms of associated Laguerre polynomials [63]:

1 1 3
eimϕ n1 ! 2 n2 ! 2 ε|m|+ 2
ψnn1 n2 m = √ · ·
πn (n1 + |m|)! 32 (n2 + |m|)! 32
1 1 |m| |m|
· e− 2 ε(ξ+η) (ξη) 2 |m| Ln1 +|m| (εξ)Ln2 +|m| (εη), (2.12)

where, n is the principal quantum number, n1 and n2 are nonnegative integers

called parabolic quantum numbers. Additionally, n1 and n2 are the numbers of nodes
in the u1 (ξ) and u2 (η) wave functions and relate to n via:

n = n1 + n2 + |m| + 1, (2.13)

and

Z1 1
n1 = − (|m| + 1), (2.14a)
ε 2
Z2 1
n2 = − (|m| + 1). (2.14b)
ε 2
By solving equations (2.14a) or (2.14b) for ε one obtains the energy for the
hydrogen atom:

1
E=− . (2.15)
2n2
The above expression is identical with the one obtained by means of the old
quantum mechanics, as shown in Section 2.1.
The electron probability distribution in parabolic coordinates can be calculated
by taking the squared absolute value of Eq. (2.12). Fig. 2.2 shows eigenstates cal-
culated for n = 6 and m = 0, which are described by a set of quantum numbers
(n, n1 , n2 , m). Closer examination reveals that the nodes lie on parabolas and one
can observe a visible asymmetry with respect to the plane z = 0. For n1 > n2 the
charge distribution of the electron is localized along the +z axis whereas for n1 < n2
it is localized on the negative side of z. Therefore, the displacement of the center
of mass of the charge distribution results in a permanent electric dipole moment,
which is conspicuous for all states.

17
Chapter 2. Stark effect

(6,0,5,0) (6,1,4,0)

x
z

(6,2,3,0) (6,3,2,0)

(6,4,1,0) (6,5,0,0)

Figure 2.2: Charge distribution of the electron for parabolic eigenstates: (n, n1 , n2 , m)
=(6, 0, 5, 0), (6, 1, 4, 0), (6, 2, 3, 0), (6, 3, 2, 0), (6, 4, 1, 0), and (6, 5, 0, 0). The
dipole moments, which give rise to the first order Stark effect, are conspicuous. Note that
the nucleus is at the center of the coordinate system.

18
 2.2 The Stark effect in the point of view of the Quantum Mechanics

The hydrogenic parabolic nn1 n2 m states and the spherical nlm states can be
related to each other via a unitary transformation [66]:

n−1
!
X n − 1 n − 1 m + n1 − n2 m − n1 + n2 n − 1 n − 1
ψnn1 n2 m = , , , , , l, m ψnlm
2 2 2 2 2 2
l=|m|

(2.16)

where (. . . | . . .) is the Clebsh-Gordan coefficient. For example, the (n, n1 , n2 , m)

= (6, 0, 5, 0) parabolic state shown in the top-left corner of the Fig .2.2 is a linear
combination of spherical states with weights given by the Clebsh-Gordan coefficients,
i.e.:

r √
1 5 5 5 1 1
ψ6050 = − √ ψ600 + ψ610 − √ ψ620 + ψ630 − √ ψ640 + √ ψ650 . (2.17)
6 14 2 21 6 2 7 6 7

2.2.2 Linear and quadratic Stark effect by means of pertur-

bation theory
A homogenous electric field of strength F in the z direction introduces a pertur-
bation potential that can be expressed in parabolic coordinates as:

1
Fz = F (ξ − η). (2.18)
2
Consequently the Schrödinger equation for the hydrogen atom in a uniform elec-
tric field is given by [63–65]:

" #
∇2 2 F (ξ − η)
− − + ψ = Eψ. (2.19)
2 ξ+η 2

The above equation can again be separated by introducing three functions, each
along one parabolic coordinate, as in Eq. (2.10). The one-dimensional equations for
ξ and η are in analogy to equations (2.11a) and (2.11b):

! !
d du1 1 m2 1 2
ξ + Eξ + Z1 − − F ξ u1 = 0 (2.20a)
dξ dξ 2 4ξ 4
! !
d du2 1 m2 1 2
η + Eη + Z2 − + F η u2 = 0 (2.20b)
dη dη 2 4η 4

with the so-called ”effective” charges subject to the condition Z1 + Z2 = 1. The

above differential equations are too complex to be solved exactly. The simplest way

19
Chapter 2. Stark effect

to calculate approximate energy levels for the hydrogen atom in an electric field is to
use perturbation theory. Namely, first solve the zero field problem and then calculate
the electric field effect by means of a perturbation series.
(0)
In the absence of the electric field the effective charge Z1 is given by Eq. (2.14a).
The first order correction of this quantity is calculated as the integral of the per-
turbation potential evaluated over the unperturbed eigenfuction u1 (ξ) given by Eq.
(2.12) [63]:

Z ∞
(1) 1 1 −2
Z1 = F ξ 2 u21 (ξ)dξ = F ε (6n21 + 6n1 |m| + m2 + 6n1 + 3|m| + 2). (2.21)
4 0 4
Altogether the effective charge Z1 is given by:

!
(0) (1) |m| + 1 1
Z1 = Z1 +Z1 = ε n1 + + F ε−2 (6n21 +6n1 |m|+m2 +6n1 +3|m|+2).
2 4
(2.22)

The expression for Z2 is obtained by replacing n1 with n2 . By substituting the

expressions of the effective charges in the relation Z1 + Z2 = 1 and solving it with
respect to ε one obtains the energy of hydrogen levels to the first approximation:

1 1 3
E = − ε2 = − 2 + F n(n1 − n2 ). (2.23)
2 2n 2
The second term on the right side is the linear Stark effect, which is identical
with the expression calculated by Schwarzschild and Epstein given by Eq. (2.5). The
first order Stark effect results in a symmetrical splitting of the energy levels about
their field-free positions. The energetically highest Stark component corresponds to
n1 = n − 1, n2 = 0 and refers to the situation when the electron is predominately
located on the positive side of the z-axis, i.e. located uphill with respect to the
Coulomb potential (e.g. the (6, 5, 0, 0) state shown in Fig. 2.2). This state is a
so-called blue state as the transition energy is shifted towards shorter wavelengths
while compared to the field-free case. By analogy the lowest Stark level, a so called
red state, corresponds to n1 = 0, n2 = n − 1 and has the electron located downhill
with respect to the Coulomb potential, i.e. on the negative side of the z-axis as
shown for the (6, 0, 5, 0) state in Fig. 2.2. The energy separation between the two
extreme components of a Stark manifold is given by:

∆E = 3F n(n − 1). (2.24)

To obtain the energy expression including the quadratic Stark effect one needs
to calculate the second order perturbations for Z1 and Z2 using the general formulae
of perturbation theory [64]:

20
 2.3 Energy levels in a presence of electric fields – Stark maps

(2) F2 X |(ξ 2 )n1 n0 |2

1
Z1 = (0) (0) 0
16 0 Z (n ) − Z
n 6=n 1
1 1 1 (n1 )
1

1
= − F 2 ε−5 (|m| + 2n1 + 1)[4m2 + 17(2|m|n1 + 2n21 + |m| + 2n1 ) + 18].
16
(2.25)

As a result, the energy is given to second order by:

1 1 3 1
E = − ε2 = − 2 + F n(n1 −n2 )− F 2 n4 [17n2 −3(n1 −n2 )2 −9m2 +19]. (2.26)
2 2n 2 16
The third term on the right side is the quadratic Stark effect and is identical
with the Old Quantum Mechanical expression in Eq. (2.6). In contrast to the linear
Stark effect it depends also on the magnetic quantum number m. By having a closer
look at Eq. (2.26) one can realize that the quadratic Stark effect always results in a
lowering of the levels.

2.3 Energy levels in a presence of electric fields –

Stark maps
2.3.1 Hydrogen atom
The uniqueness of the hydrogen atom comes from the fact that it has only one
electron moving in the Coulomb central potential and consequently its energy de-
pends only on the principal quantum number n, as shown in Eq. (2.15). This implies
that in the absence of external perturbations energy levels are degenerate with re-
spect to the azimuthal and magnetic quantum numbers l and m. When the hydrogen
atom is placed in an external electric field the degeneracy of the lm states of a par-
ticular n is lifted. Nevertheless, states that differ only in the sign of the magnetic
quantum number remain degenerate. Thus the energy level specified by the principal
quantum number n is split into n(n + 1)/2 sublevels, which can be characterized by
parabolic quantum numbers n1 ,n2 and m.
A peculiar consequence of the symmetrical Coulomb potential is that the hy-
drogen atom has a permanent electric dipole moment, except for the ground state
(n = 1), which is clearly visible in Fig. 2.2. Therefore, while being placed in an ex-
ternal electric field hydrogen exhibits a linear Stark effect. The second-order Stark
effect is always present as the electric dipole moment is modified by the electric field.
Fig. 2.3a shows a Stark map calculated for the hydrogen atom in the vicinity of the
n = 15 manifold. This figure points out that the linear Stark effect is the dominant
feature, whereas curvature of the levels due to the second– and higher– order effects
is hardly visible for the given energy and field range. However, the quadratic Stark
effect becomes conspicuous for higher electric fields and higher states.

21
Chapter 2. Stark effect

n=15
n=15

Figure 2.3: Calculated Stark maps for (a) hydrogen atom m = 0 states (solid lines) and
m = 1 (dashed lines) states in the vicinity of the n = 15 manifold, obtained by means of
seventeenth order perturbation theory [67]. Stark structures of the lithium atom for (b) m
= 1 and (c) m = 0 states (reproduced from Ref. [68]).

22
 2.3 Energy levels in a presence of electric fields – Stark maps

Another peculiarity observed in the hydrogen Stark map is that states of different
principal quantum number n actually cross, in spite of the fact that they have the
same m [69]. However, relativistic effects can produce small coupling between these
states, that otherwise would cross, leading to occurrence of so-called anticrossings
[70]. Nevertheless, this effect is beyond the experimental resolution of experiments
presented in this thesis and therefore will not be further discussed. In hydrogen the
even and the odd m levels are interleaved, as can be seen in Fig. 2.3a by comparing
the positions of energy levels for m = 0 (solid lines) and for m = 1 (dashed lines).

2.3.2 Nonhydrogenic atoms

Atoms other than hydrogen have a complex nucleus, comprising protons and
neutrons, and more than one electron, resulting in a finite size of the ionic core.
Consequently, the core potential experienced by a valence electron is no longer purely
Coulombic, hence the Hamiltonian is no longer separable and there is no analytical
solution to the wave equation (Eq. (1.5)). To calculate the wave function of a ”one-
electron” atom, i.e. an atom with one electron excited into a state with high principal
quantum number n, one can use the methods of Quantum Defect theory (QD) [[71]
and references therein]. In QD the energy of atoms is given by:

Z2
E= , (2.27)
2(n − δl )2
where δl is an empirically observed quantum defect for a series with orbital an-
gular momentum l and provides a measure of the difference between the ionic core
potential and the pure Coulomb potential. In Rydberg atoms, a highly excited elec-
tron spends most of its time at large distance (r ∼ n2 ), where its wave function can
be accurately represented by Coulomb functions. However, there is a significant dif-
ference between states with high and low orbital angular momentum, as an electron
in a high l state never comes close to the ionic core, preventing mutual interaction
– its quantum defect is small and the atom behaves hydrogenically. By contrast, a
Rydberg electron in a low l state comes close to the core and it can both polarize and
penetrate it. This leads to significant changes of the wave function and of the atomic
energy level with respect to its hydrogenic counterpart – for such states quantum
defects are large.
In zero electric field the non-Coulombic potential acting upon the electron already
removes the n2 degeneracy of the n shell and thus the energy levels, especially for
states with low l, are depressed. Since the energy levels are not degenerate with
respect to the orbital quantum number l, in the presence of an electric field the
first order perturbation to the energy vanishes. Therefore, nonhydrogenic atoms
exhibit a quadratic Stark effect for low values of the electric field and when the
Stark interaction becomes comparable to the core interactions responsible for the
quantum defect, the Stark shift becomes linear – similar to the hydrogen case.
Fig. 2.3 shows calculated Stark maps for lithium in (b) m = 1 and (c) m = 0
states, which at low field strength are quasidegenerate and therefore not interleaved

23
Chapter 2. Stark effect

like hydrogen [72]. When the quantum defect is small, as for l = 1 states (δp ≈ 0.05
[68]), the system shows largely hydrogenic behavior, similar to the H Stark map
shown in Fig. 2.3a. The only noticeable deviation is near zero field, where the 15p
state is significantly depressed, experiencing initially a quadratic Stark effect in
contrast to the linear behavior of the remaining levels. For large quantum defects,
like in the case of l = 0 states (δs ≈ 0.4 [68]) that contribute to the Stark maps
calculated for m = 0, the structure differs dramatically from the hydrogenic one. A
conspicuous result is that the 15s and 16s levels never display a linear Stark effect
and are repelled by levels above and below them. When they are squeezed between
the adjacent manifolds the entire Stark structure is altered and the resemblance to
the hydrogenic structure is lost, in contrast to the m = 1 case.
Another consequence of the non-Coulombic potential is that red and blue states
are coupled by their slight overlap at the core what results in avoided crossings. A
break of a dynamical symmetry of the Coulomb potential leads to a situation when
states with the same m cannot cross. The first anticrossing between red and blue
states of adjacent n manifolds occurs for fields:

1
F > . (2.28)
3(n − δl )5

Most of the avoided crossings are too small to be observed in Fig. 2.3b, although
this effect has its significance as will be explained in the next section. The m = 0
states exhibit larger anticrossings. The maximum possible repulsion between levels is
half of the energy separation, thus ∼ 1/(n − δl )3 . Therefore, larger quantum defects
produce larger repulsions between levels, but the repulsion can suddenly diminish
creating an unexpected pseudocrossing as shown in Fig. 2.3c at F = 4.8 kV/cm and
E = 507 cm−1 . In this particular case, this is caused by the fact that the interacting
levels from the n = 16 and n = 14 manifolds have the largest and the smallest
quantum detects, respectively.

2.4 Field ionization

2.4.1 Hydrogen atom
When a hydrogen atom is placed in an electric field it experiences the combined
Coulomb-Stark potential V (r) = − 1r + F z . This potential lowers the ionization
threshold (from E = 0, which is the field-free ionization limit) to
√ a saddle point on
the z–axis where the potential energy has the value Esp = −2 F (see Fig. 2.4a).
If one ignores all quantum effects and the Stark shift of energy levels, an electron
bound by energy E can be ionized by a field with strength:

1
F = . (2.29)
16n4

24
 2.4 Field ionization

Figure 2.4: (a) Comparison between a pure Coulombic potential (dashed lines) and the
combined Coulomb-Stark potential (solid lines) along the z–axis for a field of F = 0.01
(a.u.) in the z direction. Plots of the potentials (b) V (ξ) and (c) V (η) for a red state with
Z2 = 0.99 (dashed lines) and a blue state with Z2 = 0.10 (dotted lines) in an electric field
of F = 0.01 (a.u.) for m = 1.The solid line is a pure Stark potential with values (b)+ 41 F ξ
and (c) − 14 F η.

However, in 1930 Gebauer and Rausch van Traubenberg [73], while studying the
ionization thresholds of Stark states, observed that red states, which are located at
lower energies are less stable with respect to ionization compared to blue states.
This counter-intuitive phenomenon was explained by Lanczos [74] owing to later
developments in quantum mechanics.
To qualitatively explain the field ionization process for the hydrogen atom one
can replace functions u1 (ξ) and u2 (η) in Eqs. (2.20a) and (2.20b) with [64]:

χ1 (ξ)
u1 (ξ) = √ , (2.30a)
ξ
χ2 (η)
u1 (η) = √ . (2.30b)
η
√
By introducing a normalization factor of 2π the wave function is given by:

1
ψ(ξ, η, ϕ) = √ χ1 (ξ)χ2 (η)eimϕ . (2.31)
2πξη

25
Chapter 2. Stark effect

Functions χ1 (ξ) and χ2 (η) describe the electron motion along the ξ and η co-
ordinates, respectively. For each coordinate there exists a certain effective potential
acting on the electron, which is given by:

Z1 m2 − 1 F ξ
V (ξ) = − + + , (2.32a)
2ξ 8ξ 2 8
Z2 m2 − 1 F η
V (η) = − + − . (2.32b)
2η 8η 2 8
Both effective potentials are plotted in Fig. 2.4 for a red (Z2 = 0.99) and a
blue (Z2 = 0.10) state in the Stark manifold. The different shape of the V (ξ) and
V (η) potentials leads to qualitatively different wave functions. For vanishing electric
fields and E < 0 both χ1 (ξ) and χ2 (η) have a discrete eigenvalue spectrum shape
characterized by a set of quantum numbers (n, n1 , n2 , m). When the electric field
does not vanish the spectrum of the ξ equation remains discrete, while the spectrum
of the η equation becomes continuous. Closer examination of Eq. (2.32a) and Fig.
2.4b shows that the motion in the ξ direction is always bound since V (ξ) → F ξ as
ξ → ∞. Therefore, the χ1 (ξ) function has a well-defined number of nodes which is
equal to the parabolic quantum number n1 . Thus, for ionization to occur the electron
must escape to n = ∞. For states with energy bigger than the peak in the potential
V (η) the entire region is classically accessible for n > 0 and ionization occurs for
fields:

E2
F > . (2.33)
4Z2
The effective charge Z2 for a blue state n1 − n2 ≈ n has a value Z2 ≈ 1/n and
for the red state Z2 ≈ 1, thus if blue and red states have the same energy the red
one will be the most ionized (see Fig. 2.4c). It has been shown [75] that for m = 0
states belonging to the same Stark manifold the ionization threshold electric field
for the bluest state is by a factor of three larger than the threshold electric field of
the reddest state.
If the energy is less than the peak of the V (η) potential, the electron can tunnel
through the barrier. For energies far below the top of the potential the transmission
through the barrier is vanishingly small and the resonances are so sharp as to be
bound states. With increase of the energy the transmission through the barrier
increases, hence the width of resonances increases and the amplitude of χ2 (η) in the
inner well decreases. Nevertheless, at resonance, when n2 is a good quantum number
fitting the number of nodes of the χ2 (η) wave function, the amplitude of the χ2 (η)
function in the inner potential well of V (η) becomes much larger than for any energy
away from a resonance and below the peak of the potential V (η). Finally, for energies
above the barrier, the wave function has an amplitude which hardly depends on the
energy and smoothly varies.
In a moderate electric field each energy level of the hydrogen atom ionizes at some
uniform rate Γ, which is a rapidly increasing function of the field. In this regime each

26
 2.4 Field ionization

state undergoes a simple exponential decay with a lifetime τ = 1/Γ. For energies
below the top of the V (η) potential barrier a tunneling process occurs and accurate
calculations of ionization rates have been made [64, 76–78]. Numerical solutions of
the Stark problem for the hydrogen atom are very efficient, but for many practical
purposes it is more convenient to have analytical formulae. Damburg and Kolosov
obtained a formula for the level width Γ, associated with the tunneling effect, by
employing an asymptotic method, which gives reliable results for a wide range of
electric fields for an arbitrary n [79, 80]:

" !#
(4R)2n2 +|m|+1 2 1 3 2 2 53
Γ= 3 exp − R− n F 34n2 +34n2 |m|+46n2 +7m +23|m|+ ,
n n2 !(n2 + |m|)! 3 4 3
(2.34)

where R = (−2E)3/2 /F and E is the Stark energy. The validity of this asymp-
totic formula has been experimentally confirmed by Koch and Mariani, who made
the first precise measurement of absolute ionization rates for individual sublevels of
hydrogen [75]. With an increase of the field strength Stark states start to broaden
and create a ”sea” of continuous Stark levels superimposed on other higher-lying
discrete states. As the energy levels do not mix for hydrogen, the degenerate ”sea”
of levels has no effect on the dynamics of the hydrogen ionization process. Indeed,
photoionization spectra obtained by Rottke and Welge [81] showed sharp line struc-
tures of quasistable field-ionizing states superimposed on continua for energies below
the field-free ionization limit E = 0 and above the classical saddle point energy Esp .

2.4.2 Nonhydrogenic atoms

As has been pointed out in Section 2.3.2 the finite size of the ionic core signif-
icantly alters Stark maps for nonhydrogenic atoms. Additionally it influences the
process by which multi-electron atoms ionize in the presence of a static electric field.
Actually there are two different forms of field ionization observed for nonhydrogenic
atoms for energies below the top of the V (η) potential. The first type of field ion-
ization is a pure tunneling effect occurring exactly the same way as in the hydrogen
atom – a state with an approximate quantum number n1 ionizes itself and keeps the
same value of n1 if its energy is enough to penetrate the potential barrier in V (η).
Ionization rates for this form of field ionization increase rapidly in the vicinity of
the classical ionization threshold meaning that states broaden until they disappear
to form a continuum at the top of the V (η) potential. The second type of field ion-
ization is due to coupling of slowly ionizing states to rapidly ionizing states and is
similar to the autoionization process in multi-electron atoms or to predissociation
in molecules [84]. A simple classical picture explaining this process comprises an
electron in a blue orbit, which once per revolution comes close the core and has a
non-zero probability of being scattered from the stable blue orbit into the degener-
ate red continuum. Thus this process takes place for all nonhydrogenic Stark states
except the reddest state with n1 = 0, which ionizes like hydrogen at the classical

27
Chapter 2. Stark effect

ionization limit given by Eq. (2.33) [83]. The ionization rate for the autoionization
process is not exponentially dependent on the field, as for the hydrogenic ionization
process, but is rather a highly nonmonotonic function of the field strength [83, 85].
As a consequence, levels retain their character for a considerable range of fields. Due
to the coupling between states, there are no stable states (i.e. long-lived blue states)
above the classical ionization limit and they decay more rapidly by ionization than
by radiative decay, as shown experimentally by Littman [84].

a1 G(sec )-1

b 109
Normalized signal intenisty

108

107

6
10

0
0 1 15.4 15.6 15.8 16.0
Frequency (GHz) E(kV/cm)

Figure 2.5: (a) Example of a characteristically asymmetric line shape observed in the pho-
toionization spectrum of Rb in a F = 158 V/cm static electric field for an excitation energy
located between the 40p and 41p levels. Dots represent the best fit to a Fano profile [82].
(b) Ionization rates for the (n, n1 , n2 , m) = (12, 6, 3, 2) state of Na in the vicinity of
an avoided crossing with the rapidly ionizing (14, 0, 11, 2) state [83]. (c) Photoionization
spectra of lithium measured in the region of an avoided crossing between the (18, 16, 0, 1)
and (19, 1, 16, 1) states [68].

28
 2.4 Field ionization

The coupling between the blue states and the red continua leads to interferences
in the excitation amplitudes and gives rise to asymmetric line shapes resembling
Fano profiles, which are commonly observed in the excitation spectra of autionizating
states [86]. The transition strength to the mixed discrete-continuum eigenfunction
exhibits constructive interference on one side of the peak and destructive interference
on the other side leading to the asymmetric profile. Broadening and asymmetry
introduced by core effects are greatest for m = 0 resonances as they penetrate the
core the most. The first observation of Fano profiles in the photoionization spectrum
of rubidium by Feneuille [82] (see Fig. 2.5a) confirmed that the field ionization of
nonhydrogenic atoms can have the form of an autoionization process.

Field ionization process in the vicinity of an avoided-crossing

In multi-electron atoms the non-Coulombic core potential acting on the outer
electron mixes states with the same m value but different n1 or n2 quantum numbers,
causing anticrossings between levels. In the vicinity of an avoided crossing the state
follows an adiabatic path when the field increases slowly meaning that none of the
levels crosses. If the field increases rapidly, an atom traverses the avoided crossing
diabatically – it ”jumps” across the anticrossing. Zimmerman et al. [68] investigated
the level anticrossing in lithium for interacting (n, n1 , n2 , m)= (18, 16, 0, 1) and
(19, 1, 16, 1) states (see Fig. 2.5c). They observed a disappearance of the oscillator
strength in the upper level and an increase in the intensity of the other level. The
eigenstates at the anticrossing are symmetric
√ and asymmetric combination of the
interacting states A and B, i.e. (1/ 2)(ψA ± ψB ). Thus when two Stark states
have approximately equal oscillator strengths away from the crossing, the oscillator
strength will vanish in one of the eigenstates and double in the other one at the
avoided crossing.
Not only the oscillator strengths are altered in the vicinity of an anticrossing,
but also the ionization rates are strongly perturbed. In 1976 Littman [83] observed
for sodium atoms a strong increase of the ionization rate for the (12, 6, 3, 2) state
while mixing with the rapidly ionizing (14, 0, 11, 2) state (see Fig. 2.5b). They found
good agreement with a simple model based on a two-level system stating that at the
avoided crossing the eigenstates are linear combinations of the two Stark states.
In reality the situation is more complicated as there is a simultaneous interaction
between the crossing levels and the continua to which there are coupled. Therefore,
near the avoided crossing the ionization rate may be estimated from the properly
weighted sum of the couplings between the components of the mixture and each of
the accessible continuum states [87]. If one considers unperturbed discrete states |φi i
coupled by the core interaction (Vc ) to continua |ψjE i indicated by matrix elements
VijE then each state has a core-induced ionization rate given by:

X X
Γi = 2π | hφi | Vc |ψjE i |2 = 2π |VijE |2 . (2.35)
j j

These states are also mixed with each other by the core interaction and near a

29
Chapter 2. Stark effect

crossing of two states A and B the normalized eigenstates are given by:

|φ+ i = cosθ |φA i + sinθ |φB i , (2.36a)

|φ− i = −sinθ |φA i + cosθ |φB i . (2.36b)

The angle θ has values from 0 to 2π, with θ = π/4 at the center of the avoided
crossing. Finally, the ionization rate of the mixed states are given by:

" #
X Γ A + Γ B Γ A − ΓB
p
Γ± = 2π | hφ± | Vc |ψjE i |2 = ± cos2θ+ ΓA ΓB cosγ ·sin2θ ,
j
2 2
(2.37)

where γ is an angle representing the decay-channel overlap and it is defined by:

VAjE VBjE
P
j
cosγ = rP rP . (2.38)
(VAjE )2 (VBjE )2
j j

It is interesting to calculate the minimum width of the mixed states if ΓA 6= ΓB :

v !2
u
ΓA + Γ B u Γ +Γ
A B
Γmin = − t − ΓA ΓB sin2 γ. (2.39)
2 2

For γ = 0 a maximum overlap between the decay channels is achieved, hence,

there is only one continuum and consequently one of the mixed states is completely
decoupled from the continuum leading to a decrease of its ionization rate and corre-
sponding spectral width. Stated in other words: there are circumstances when terms
in Eq. (2.37) have opposite signs and will interfere, leading to a reduction of the ion-
ization rate. In some cases the core-induced width is canceled to such a degree that
the dominant decay mechanism is ionization via hydrogenic tunneling. However, the
tunneling rate exceeds the core-induced ionization rate only in a very limited field-
energy region, when the resonance is located very near the top of the V (η) potential
barrier.
Such a line-narrowing or interference narrowing effect has been first observed by
Feneuille [88] for rubidium atoms in a narrow energy and field range (∆E = 0.03
cm−1 and ∆V = 2 V/cm), for the ionization of the (23, 0, 21, 0) state in the electric
field of 2150 V/cm. The authors interpreted the observed decrease in the ionization
rate as a strong perturbation introduced by the spin-orbit interaction, which coupled
states with different values of m. The incorrectness of this interpretation had been
pointed out by Liu [87], who observed the same effect for sodium atoms in the vicin-
ity of an avoided crossing between two energy levels: (20, 19, 0, 0) and (21, 17, 3, 0).

30
 2.5 Theoretical calculations of the Stark problem

a (21,7,12,1) b (21,7,12,1)

(20,9,9,1)
(18,16,0,1)
(1
9,
(20,9,9,1) 12
,5
,1
)

(19,12,5,1)

(18,16,0,1)

Figure 2.6: (a) Sodium Stark map in the vicinity of a four-level crossing with minima in
the ionization rate indicated by circles. (b) Calculated decay rates of the four states plotted
in (a) showing six regions of line-narrowing [92].

As confirmed by calculations based on the Wentzel-Kramers-Brillouin – Quantum

Defect theory (WKB-QD) [89–91], the observed phenomenon was due to interference
between discrete and continuum coupling amplitudes of both states. Consequently,
the (20, 19, 0, 0) state was partially decoupled from the continua resulting in a
decrease of its ionization width by two orders of magnitude. Further study by McNi-
choll [92] brought more details about the interference-narrowed Stark resonances in
sodium. Fig. 2.6 shows the examined case of a four-level crossing indicating six re-
gions of line-narrowing, the number of distinct pairs of levels, each positioned slightly
outside the center of the anticrossing. For each line-narrowing region the calculated
ionization rates show a decrease by circa three orders of magnitude. In the vicin-
ity of another avoided crossing the authors observed a reversal of the asymmetry
of the Fano profile caused by the sign change of transition amplitude. It has been
pointed out by Yang [93] that there are much more interference-narrowing regions
to be found in sodium than in rubidium. In the former case the ion-core potential
is smaller, hence, Stark levels preserve the hydrogenic structure to a much greater
degree and there are many regions where 2, 3 or 4 hydrogenic levels cross. Thus, it is
a bit ironic that the effect was first observed for rubidium atoms. The line-narrowing
phenomenon has also been observed for helium [94], lithium [95] and barium [96].

2.5 Theoretical calculations of the Stark problem

As has been pointed out in Section 2.2.2 differential equations (2.20a) and (2.20b)
cannot be exactly solved with the help of elementary functions. However, different
theoretical approaches allow the calculation of the approximate energies of Stark
states by means of: perturbation theory [97], the Wentzel - Kramers - Brillouin
(WKB) approximation [76, 78], WKB - Quantum Defect theory (WKB-QD) [89–
91, 98] or numerical methods [99–101]. All these procedures are suitable for distinct
field intensities, energies, state characters (red or blue) and types of atoms under

31
Chapter 2. Stark effect

investigation. In this thesis three different methods are employed for calculations of
the Stark effect in hydrogen and helium and only these approaches will be elaborated
in more detail below.

2.5.1 Perturbation theory

The first quantum-mechanical problem treated by Rayleigh-Schrödinger pertur-
bation theory was the Stark effect in the hydrogen atom [21], for which the first
two orders of perturbed energies were immediately obtained. A formula for the third
order, which required some effort, was delivered a few years later [102]. However, the
next order was so tedious that it took almost 50 years to obtain the correct formula
[103]. In 1978 Silverstone proposed a solution of the Stark effect in hydrogen to
arbitrarily high orders [67]. In this approach the derived formula for the N th order
energy E (N ) is a polynomial in the parabolic and magnetic quantum numbers:

(N )
XXX
E (N ) = 4−N n3N −2 Erij M r k1i k2j , (2.40)
r­0 i­0 j­0

(N )
where ki = ni + 12 |m| + 12 , M = 14 (m2 − 1) and Erij are polynomial coefficients
computed to seventeenth order and available from the Physics Auxiliary Publication
Service (PAPS). To calculate the perturbed energy of a specific state (n, n1 , n2 , m)
one can use the N th order energy obtained with Eq. (2.40) and insert it in the power-
series expansions in the field:

E = E (0) + E (1) F + E (2) F 2 + · · · . (2.41)

Despite the fact that for the Stark effect perturbation theory yields a noncon-
vergent, asymptotic expansion of the energy [97, 104], it can still be used as a good
estimation for low fields if the expansion is not carried too far [105]. For a given
maximum error  one can calculate the value of the field strength FN (), for which
the N th term has a magnitude :

1/N

FN () = . (2.42)
E (N )
Perturbation theory is more accurate for the n1 − n2  0 resonances (red states)
than for the n1 − n2  0 (blue states), as has been shown experimentally by Koch
[105], who got an accurate value of the energy for the reddest state (30, 0, 29, 0).
The discrepancy between the theory and the experiment was less than the exper-
imental uncertainty. In contrast, the energy of the blue state (25, 21, 2, 1) could
not be given with the same high accuracy due to the oscillatory behavior of the
perturbation series about the experimental result. Moreover, this theory is poorly
suited for describing anticrossings in multi-electron atoms [94]. Nevertheless, when
used in a pure Coulomb potential, where levels actually cross, it works rather well
[75].

32
 2.5 Theoretical calculations of the Stark problem

2.5.2 Wentzel-Kramers-Brillouin – Quantum Defect theory

(WKB-QD)
For the hydrogen atom in a very strong electric field it is more convenient to solve
the differential equations (2.20a) and (2.20b) by means of an approximation method
than by using perturbation theory [63]. The WKB method is especially suitable for
states with a high principal quantum number n and can be used for arbitrary field
strength F practically over the entire energy domain. The first successful imple-
mentation of this method was made by Lanczos already in 1930 [106]. Using the
expressions for wave functions χ1 (ξ) and χ2 (η) from Eq. (2.30a) and (2.30b) the
differential equations can be re-written as:

d2 χ1
+ φ1 (ξ)χ1 = 0, (2.43a)
dχ21
d2 χ2
+ φ2 (η)χ2 = 0, (2.43b)
dχ22

where the functions φ1 (ξ) and φ2 (η) are essentially the ”local kinetic energies” of
the electron at the position ξ and η, respectively. The energy function φ1 (ξ) = 12 E +
Z1 m2 −1 1
ξ − 4ξ 2 − 4 F ξ has two classical turning points ξ1 and ξ2 defining a region where
it has positive values. The wave function χ1 (ξ) has the unnormalized approximate
form [63, 89, 101]:

!
Z ξ2
−1/4
p π
χ1 (ξ) = aφ1 (ξ)cos φ1 (ξ)dx − , (2.44)
ξ1 4

where a is the normalization constant. This function must be bound and decrease
exponentially on both sides of the turning points, thus the following quantization
condition can be written:

!
Z ξ2 p 1
φ1 (ξ)dx = n1 + π. (2.45)
ξ1 2

The approximation 2.45 is very accurate for large n1 , fairly accurate for small
values of n1 and in the region ξ ∼ 0 it breaks down. However, this limitation can be
2
bypassed through the Langer correction by replacing the centrifugal term m4ξ−1 2
 2
m
with 4ξ 2 , as shown in Ref. [89]. For fixed values of E, m and F there exists
many solutions to Eq. (2.45), which correspond to different values of Z1 and n1 .
Consequently, allowed values of Z2 are obtained from the condition Z1 + Z2 = 1 and
serve as an input to the characterization of the χ2 (η) wave function.
To determine the asymptotic form of χ2 (η → ∞) in the Coulomb + Stark po-
tential Harmin proposed a theory based on a separation of physical space into an

33
Chapter 2. Stark effect

inner and outer region. In the former region the interaction of the electron with the
electric field is negligible compare to that with the core, hence the potential is purely
m2 Z2
Coulombic + centrifugal, i.e. V1 (η) = 4η 2 − η . In the latter region the Coulomb

field is negligible and the potential reduces to V2 (η) = − 14 F η [89]. For all values of
η, except η ∼ 0, the χ2 (η) wave function is approximated by the WKB solution.
By matching forms of χ2 (η) from the inner and outer regions the asymptotic phase
and amplitude of χ2 (η) is obtained for fixed n1 , m, E, F and Z2 . An extension
of this approach to nonhydrogenic atoms allowed obtaining an excellent agreement
with an experimental photoionization spectrum of Na [90, 107] as well as with Li
photoionization spectra [91].

2.5.3 Wave packet calculations

During a photoionization process in a DC electric field time-dependent electron
wave packets are created, which rapidly evolve in the combined Coulomb + Stark
potential. The dynamics of such wave packets is determined by the static electric
field, the laser polarization, the wavelength of the laser light and the pulse duration.
It has been shown that precisely calculated wave packets allow for a deeper insight
into dynamical processes of photoionization in Rydberg atoms [108, 109]. In this
thesis a vast amount of calculations are based on a wave packet propagation approach
in which the time dependent Schrödinger equation is solved by using a split-operator
technique. The principle of this method is to separate the time evolution operator
U (t, t0 ) (see Eq. (2.48)) into parts, which are easy to calculate for each calculation
step. Presented below, a theoretical description of this method is given based on the
PhD thesis of Antoine M. Ollagnier [110].
The wave function ψ(~r, t) at a given time delay t and at the position ~r obeys the
time dependent Schrödinger equation:
∂
i ψ(~r, t) = Hψ(~r, t), (2.46)
∂t
which can be re-written in terms of the time evolution operator U (t, t0 ) as:

ψ(~r, t) = U (t, t0 )ψ(~r, t0 ), (2.47)

where ψ(~r, t0 ) is the initial wave function at t0 , and U (t, t0 ) is defined by:

Z t
U (t, t0 ) = 1 − i HU (t0 , t0 )dt0 . (2.48)
t0

For a time-independent Hamiltonian, U (t, t0 ) is simply:

U (t, t0 ) = e−iH(t−t0 ) . (2.49)

34
 2.5 Theoretical calculations of the Stark problem

The following Hamiltonian is considered:

H = Hat + HF + Hint , (2.50)

where Hat is the atomic part corresponding to:

1 ∂2 L2
Hat = − + + V (r), (2.51)
2 ∂r2 2r2
where L is the angular momentum operator and V (r) is the atomic potential
describing the interaction between the valence electron and the ionic core.
The second term on the right-hand side of Eq. (2.50), i.e. HF corresponds to the
Stark Hamiltonian associated with the static electric field F . It is given by:

HF = F z = F rcosθ, (2.52)

where F is the strength of the electric field, and θ the polar angle between the
position vector ~r and the axis of the electric field.
Finally, the last term on the RHS of Eq. (2.50) Hint defines the interaction
between the atom and the laser field.
Using first order perturbation theory, the wave function ψ(~r, t) describing exci-
tation from the initial state at energy E0 to the final state of E is given by:

ψ(~r, t) = Φ(~r)e−iE0 t + Ψ(~r, t)e−iEt , (2.53)

where Φ is the initial wave function of the bound electron at time t0 at the energy
E0 which satisfies the equation:

Hat Φ = E0 Φ. (2.54)

Combining equations (2.46), (2.50) and (2.53) leads to:

∂Ψ
i − (Hat − E + F z)Ψ = S(~r, t). (2.55)
∂t
The excitation process is taken into account via the source term S(~r, t) defined
as:

!
t2
S(~r, t) = exp S(r)Ylm (2.56)
∆t2pulse

35
Chapter 2. Stark effect

where Ylm is the spherical harmonic describing the initial state characterized by
quantum numbers l and m, and S(r) describes the radial dependence of the initial
wave function. In practice, this term is defined by:

S(r) = r2 e−2r . (2.57)

The width of the laser pulse ∆tpulse is taken into account in the source term by
a Gaussian dependence.
Integration of equation (2.55) leads to the following propagator:

" #
1+ i ∆t
2 (Hat + F z − E)
Ψ(t + ∆t) = " # Ψ(t)+
1 − i ∆t
2 (Hat + F z − E)
" #
∆t2 ∆t
+ − i∆t + (Hat + F z − E) S(~r, t + ). (2.58)
2 2

The first term corresponds to the first order Padé approximation of the operator
exp(−i(Hat + F z − E)∆t). It can be split as follows:

" #
1+ i ∆t
2 (Hat + F z − E)
" # Ψ(t) = e−i(Hat +F z−E)∆t Ψ(t) =
1 − i ∆t
2 (Hat + F z − E)

∆t ∆t
= e−iF z 2 e−i(Hat −E)∆t e−iF z 2 Ψ(t). (2.59)
The first part only acts on the angular part of the wave function, while the
second part only acts on the radial part. In order to be able to apply the exponential
terms, these operators have to be linearized. For the angular part, a [2,2] order Padé
approximant is used and a canonical linearization leads to:

" #" #
−iF z ∆t 1 − iF za ∆t 1 − iF zb ∆t
e[2,2] 2 = 2 2
+ o(∆t4 ) (2.60)
1 + iF za ∆t
2 1 + iF zb ∆t
2
√ √
with a = (1/2 + i 3/6) and b = (1/2 − i 3/6).
For the radial part, a [1,1] order Padé approximant is used:

" #
−i(H−E)∆t 1 − i(Hat − E) ∆t
e[1,1] = 2
+ o(∆t2 ). (2.61)
1 + i(Hat − E) ∆t
2

36
 2.5 Theoretical calculations of the Stark problem

Finally, equation (2.58) can be re-written as follows:

" #" #
1 − iF za ∆t
2 1 − iF zb ∆t
2
Ψ(t + ∆t) = ·
1 + iF za ∆t
2 1 + iF zb ∆t
2
" #" #" #
1 − i(Hat − E) ∆t
2 1 − iF za ∆t
2 1 − iF zb ∆t
2
· Ψ(t)+
1 + i(Hat − E) ∆t
2 1 + iF za ∆t
2 1 + iF zb ∆t
2
" # !
∆t2 ∆t
+ − i∆t + (Hat + F z − E) S ~r, t + (2.62)
2 2

Because of the cylindrical symmetry of the Hamiltonian the dependence on the

azimuthal angle ϕ is not included. As a consequence, the propagation occurs in 2
dimensions using a (k, l) grid, where k is the radial coordinate and l is orbital angular
momentum.
The Stark operator F z = F rcos(θ) operates on the angular part of Ψ according
to:

s s
(l + m + 1)(l − m + 1) m (l + m)(l − m) m
cos(θ)Ylm = Yl+1 + Y . (2.63)
(2l + 3)(2l + 1) (2l + 1)(2l − 1) l−1

The radial part of the atomic Hamiltonian is given by (63):

" !2 #
d2 l(l + 1) Zl αd −(r/rc )3
− 2+ + − − 4 1−e , (2.64)
dr 2r2 r 2r

where αd is the dipole polarizability of the ionic core and rc is its radius. In the
region r ¬ rc the potential experienced by the valence electron is no longer −1/r.
The effective charge Zl (r) is given by:

(1) (3)
(2)
Zl = 1 + (Z − 1)e−αl r
+ αl re−αl r
. (2.65)

The values of the coefficients αl can be found in e.g.[98, 111–113].

At a fixed time delay t, the wave packet is written as:

lX
max

ψ(ρ, z) = Ψ(k, l)Ylm (cosθ). (2.66)

l=0

In calculations, the total wave function is integrated over time and the quan-
2
tity ρ|ψ(~r, t))|
√ is computed, which describes the probability of finding an electron,
where ρ = r − z 2 . Also the time-dependent flux of ejected electrons is calculated
2

providing an insight into the dynamics of the photoionization process.

37
3 Experimental facilities

This chapter deals with all technical aspects of photoionization microscopy ex-
periments, on hydrogen and helium atoms, whose results are presented and discussed
in Chapters 4, 5 and 6. The chapter is divided into two parts, each dedicated to one
experimental setup. At the beginning of each part the general idea of the experi-
ment is described and the following sections present crucial steps of the experiment
in greater detail. This includes the design of an atomic or metastable source, de-
scription of lasers systems, of an imaging spectrometer and of the data acquisition
procedure. In these sections also efficiency calculations of the overall processes are
provided.

3.1 Photoionization microscopy of hydrogen atoms

The experiment on hydrogen atoms was performed making use of an experi-
mental setup shown in Fig. 3.1, which comprised two differentially pumped vacuum
chambers. Atomic hydrogen was produced by photodissociating H2 S gas, introduced
into a first vacuum chamber by a pulsed valve (a) that was facing upwards (along the
x-axis), using 213 nm light from a Nd:YAG laser (Quanta-Ray, Spectra Physics).
The polarization of the 213 nm laser was vertical (along the x-axis), maximizing
production of H-atoms with a recoil velocity in the yz-plane. The atomic hydrogen
beam was formed by a 3 mm aperture (b) separating the first vacuum chamber
from the second vacuum chamber, which contained a velocity map imaging (VMI)
spectrometer [114]. The hydrogen atoms entered the VMI through a 1.5 mm hole in
the repeller electrode (c). In the VMI, a static electric field in the range from 200
to 850 V/cm was created by applying a voltage difference across the repeller and
extractor (d) electrodes. A third electrode (e) in the VMI contained deflection plates
and was kept at ground potential.
The field-free ionization potential of the hydrogen atom is 13.6 eV [115], mean-
ing that to photoionize it from the 2 S1/2 ground state one can choose a one-photon
absorption scheme employing light at 91.2 nm. Alternatively, hydrogen can be firstly
excited to an n = 2 state by means of the Lyman-alpha line and secondly photoion-
ized at 364.7 nm. In our experiments the hydrogen atoms were excited to a mixture
of n = 2 s– and p–states by means of a two-photon transition, making use of 3 mJ,
243 nm, 8 ns pulses from a tunable dye laser (PDL-3, Spectra Physics). From there,

39
Chapter 3. Experimental facilities

e j
d
c

b
h
i

g
a
f

Figure 3.1: Schematic overview of the hydrogen experiments. H2 S gas was introduced into
a first vacuum chamber by a pulsed valve (a) and subsequently it was photodissociated with
laser light at 213 nm. A 3 mm aperture (b) placed between the first and second vacuum
chambers served as a skimmer forming the atomic hydrogen beam. Hydrogen atoms passing
the velocity map imagining spectrometer were ionized by means of a two-color process.
By applying a voltage difference across the repeller (c) and extractor (d) electrodes and
keeping the third electrode (e) at ground potential the photoelectrons were accelerated in a
DC electric field towards a position-sensitive detector comprising micro-channel plates (f ),
a phosphor screen (g) and a CCD camera (h). The whole spectrometer was surrounded by a
µ-metal shield (i), which suppressed the earth magnetic field. One-third of the way toward
the detector the photoelectrons passed an Einzel lens (j), which enabled magnification of
photoelectron images by up to one order of magnitude.

they were ionized using 600 µJ, 8 ns, narrow-band, tunable laser pulses with wave-
lengths between 365 and 367 nm, obtained by frequency doubling the output of a
Fourier-limited, home-built three-stage, four-pass pulsed dye amplifier, using a cw
(continuous wave) ring-dye laser as its seed (899-29 Autoscan II, Coherent) and a
Nd:YAG laser (GCR-290, Spectra Physics) as its pump. By doing so, the emphasis
was on ionization processes resulting from the excitation of high-lying, auto-ionizing
Stark states. The polarization of the 243 nm laser was perpendicular to the static
electric field axis (i.e. parallel to the two-dimensional detector), whereas the polar-

40
 3.1 Photoionization microscopy of hydrogen atoms

ization of the 365-367 nm laser was either parallel or perpendicular to the static
electric field axis. The resulting photoelectrons were projected onto a micro channel
plate (MCP) detector (f) followed by a phosphor screen (g), and recorded with a
CCD (charged coupled device) camera (h). The magnetic field inside the VMI was
suppressed by a µ-metal shield (i). An electrostatic zoom lens (j) was used [116]
to magnify the images by about one order of magnitude without deteriorating the
image quality. Images were typically acquired for 5000 and 10000 laser shots for
non-resonant and resonant excitation of a Stark state, respectively.

3.1.1 Characteristics of the atomic hydrogen source

To produce atomic hydrogen H2 S gas has been chosen due to its continuous
absorption spectrum in the region of 160-270 nm with a maximum at about 195 nm
[117]. Hydrogen sulfide is a toxic, flammable, colorless, liquefied gas, which inhaled
at high concentrations can result in unconsciousness, coma, and death. Therefore
extra safety precautions were taken during the experiments, e.g. all gas connections
were contained within housings connected to an external gas exhaust and personal
safety was guarded by H2 S detectors. The hydrogen sulfide gas was injected into
the first vacuum chamber by a pulsed valve mounted perpendicularly to the VMI
(along the x-axis shown in Fig. 3.1). The gas flow was controlled by changing the
distance between a sealing viton disc (φ = 1 mm) and a 1 mm nozzle, making use
of a piezo element [118]. The sealing disc was mounted on a stainless steel pillar
(10 mm long), which was connected to a piezo disc (φ = 50 mm) as shown in Fig.
3.2b. It turned out that the piezo discs (P-286.23, PI) break down rather quickly
while being exposed to a highly corrosive gas, as shown in Fig. 3.2 where a picture
of a brand new piezo disc (a) is compared with one taken after 10 hours exposure to
H2 S gas (b). To overcome this limitation a vacuum grease (Fomblin Grease VAC3)
was used, which was smeared on the piezo element creating a 1 mm thick protective

a b

(1) (2)

Figure 3.2: Comparison between a brand new piezo disc (a) and one exposed for 10 hours
to H2 S gas with noticeable cracks on the disc surface (b). In the latter image the stainless
steel pillar (1) with a viton disk (2) mounted on the piezo disc are visible.

41
Chapter 3. Experimental facilities

layer and providing a trouble-free operation for a period of 6 months. In experiments

the pulsed valve was operated at 10 Hz by a home-build piezo valve voltage driver.
As photolysis light source, a Nd:YAG laser (Quanta-Ray, Spectra Physics) was
used, which delivered 7 ns light pulses at 213 nm by means of frequency mixing 1064
and 266 nm pulses in a 12×12×2 mm BBO (barium borate) crystal. An average
output energy of 1.5 mJ per pulse was achieved at 10 Hz. Within one laser pulse the
number of H2 S molecules undergoing dissociation can be calculated as:

h  σabs · E · λ i
Ndiss = N 1 − exp − 5.03 × 1015 , (3.1)
l02

where N is the number of H2 S molecules, σabs is the absorption cross section at

the photolysis wavelength λ[nm] (σabs = 2×10−18 cm−2 at 213 nm [117]), E is the
energy of the laser pulse [J/pulse] and l0 is the photolysis width [cm]. The number
of molecules N can be expressed as:

N = n0 · l03 , (3.2)

where n0 is the density of molecules in the photolysis volume of l03 . A reasonable H

atom beam quality, i.e. where inelastic collisions with residual H2 S or HS fragments
do not significantly change the H atom velocity, is obtained as long as n0 ·l0 ¬ 5×1015
cm−2 . Hence, equation 3.1 can be re-written as:

h  σabs · E · λ i
Ndiss = 5 × 1015 · l02 1 − exp − 5.03 × 1015 . (3.3)
l02

Fig. 3.3a shows the calculated number of H atoms arising from photodissociation
of H2 S gas by making use of 1.5 mJ at 213 nm as a function of the photolysis width
l0 . One can see that this process is highly inefficient for tightly focused laser beams,
whereas the focal spot of 0.1 cm ensures 90% production of H atoms. The process
starts to saturate around l0 = 0.2 cm thus the beam size should not be bigger than
this value. In the experiment this requirement was fulfilled by placing a f = 1 m
lens 80 cm in front of the pulsed valve ensuring a 213 nm beam size of 0.1×0.2 cm
in the photodissociation region. The efficiency of the atomic hydrogen source was
further optimized by setting the time of the H2 S gas injection to 40 µs prior the
laser pulse and keeping the pressure in the source chamber with the loaded gas at
3.8×10−4 Torr. At higher pressures the condition n0 · l0 ¬ 5 × 1015 cm−2 was not
optimum and the quenching of hydrogen atoms, due to collisions with other atoms
and molecules, was significant.
To ensure the highest density of H2 S molecules in the photolysis region the
213 nm laser beam was positioned approximately 1 mm above the pulsed valve.
During experiments H2 S gas was injected through the 1 mm nozzle with a speed of
460 m/s providing a gas pressure of 0.6 Torr at the exit of the nozzle. By making a
crude approximation that the pressure in the photolysis region was the same as at

42
 3.1 Photoionization microscopy of hydrogen atoms

a 1.0 b
Hydrogen atoms production

Normalized signal intensity

3x1012
0.8

2x1012 0.6

0.4
12
1x10
0.2

0.0
0.0 0.1 0.2 0.3 0.4 0.5 16000 17000 18000 19000
Photolysis width l0 (cm) Velocity (m/s)

Figure 3.3: (a) Dependence of the number of hydrogen atoms produced as a function of the
photolysis width calculated for an average energy of 1.5 mJ/pulse at 213 nm, subject to the
conditions that the product of n0 and l0 was 5×1015 cm−2 . (b) Measured velocity profile
of the hydrogen atomic beam (squares) fitted with a Gaussian profile (solid line) indicating
that the mean velocity of the H fragments was 18000 m/s and the FWHM (hull width at
half maximum) was 700 m/s.

the nozzle exit the average density of H2 S molecules had a value of 2×1016 H2 S/cm3
and the amount of produced hydrogen atoms by making use of 1.5 mJ at 213 nm
was 5.5×1012 H atoms per pulse.

3.1.2 Transport and density of H atoms in the interaction

region
After photodissociation the atomic hydrogen beam was formed by a 3 mm hole
separating the two vacuum chambers (approximately 6.5 cm away from the pulsed
valve) followed by a 1.5 mm hole in the repeller plate, providing a well-collimated
beam. The number of hydrogen atoms transported into the interaction region inside
the VMI can be calculated as:

Ntrans = Ndiss · I(θ) · ∆Ω, (3.4)

where I(θ) is the angular distribution of H fragments and ∆Ω is the solid angle.

When a molecule is directly dissociated using linearly polarized light the angular
distribution of fragments is given by [119]:

" #
1
I(θ) = 1 + 2βP2 (cosθ) , (3.5)
4π

with P2 (cosθ) being the second-order Legendre polynomial, θ the angle between
the polarization vector and the detection direction and β the anisotropy parameter.

43
Chapter 3. Experimental facilities

At 222 nm the β parameter for H2 S has a value of -0.42±0.01 [119] meaning that
most of the H atoms have their velocity vector in a plane perpendicular to the laser
polarization. Therefore, in the experiment the laser polarization was chosen to be
vertical, i.e. along the x-axis, to maximize the production of H atoms in the direction
of the VMI.
The solid angle of a cone with an apex angle 2ϕ can be calculated
  as ∆Ω =
r
2π(1 − cosϕ). Considering the experimental geometry ϕ = arctg d , where r is the
radius of the hole in the repeller plate and d is the distance between the interaction
region inside the VMI and the photolysis region (d = 9.5 cm), hence ∆Ω = 1.96 ×
10−4 . Taking this into account the percentage of hydrogen atoms resulting from the
photodissociation of H2 S molecules and being transported into the interaction region
is 0.002% and corresponds to a value of Ntrans = 1.2 × 108 H atoms/pulse.
In the interaction region H atoms may be photoionized by absorbing three pho-
tons at 243 nm if the laser pulse energy is high enough. By detecting the cre-
ated H+ ions and changing the time delay between the photolysis and photoion-
ization laser pulses the velocity spread of the atomic hydrogen beam was measured.
Fig. 3.2b shows experimentally measured velocities (black squares) with a fitted
Gaussian profile (solid line) indicating the mean velocity of H fragments to be
ϑ = (1.800 ± 0.035) × 104 m/s. Due to the velocity spread the hydrogen atomic
beam occupied a certain volume that can be calculated as ϑbeam = πr2 lz , where lz
is the longitudinal spread in the z-axis given as: lz = (ϑmax − ϑmin ) · d/ϑ = 0.37 cm.
Thus the beam volume is Vbeam = 0.0065 cm3 and therefore, the density of H atoms
in the interaction region is:

Ntrans Hatoms
n= = 1.9 × 1010 . (3.6)
Vbeam pulse

3.1.3 Two-photon excitation into the mixture of 2 s– and p–

states
As has been mentioned at the beginning of Section 3.1 atomic hydrogen has a
field-free ionization potential of 13.6 eV and therefore, in our experiments H atoms
were photoionized with three photons including a two-photon resonant transition
at 243.06 nm to a mixture of n = 2 s− and p−states. The photoexcitation step
was accomplished by making use of UV light obtained by frequency doubling the
output of a tunable pulsed dye laser comprising an external amplification stage. In
this section first the laser setup is described followed by calculations of the efficiency
of two-photon excitation into the n = 2 state.

Source of 243 nm laser beam

The pulsed dye laser (PDL-3, Spectra Physics) delivered tunable light in the vis-
ible range with an average bandwidth of 0.1 cm−1 . It was pumped with 110 mJ
at 355 nm provided by a Nd:YAG laser (GCR-290, Spectra Physics). By using

44
 3.1 Photoionization microscopy of hydrogen atoms

a b
1
x Dye flow 2
3
jq
y
z 4

Figure 3.4: (a) Side view of a Bethune cell with a circular capillary tube through which the
dye flows. (b) End view of the cell with a pump beam divided into four equal segments (1-4)
that uniformly excite the dye from the top, front, back and the bottom.

Coumarin 480 dye the laser delivered 8 ns pulses with average energy of 2 mJ at
486 nm. The pulses were further amplified in an external dye cell filled with the
same dye at a concentration of 2.5×10−4 mol/l. The cell was pumped with 220 mJ
at 355 nm from the Nd:YAG laser ensuring further amplification of the pulse ener-
gies by approximately 10 times. After the amplification stage the second harmonic
was generated with a BBO crystal (6×6×5 mm), which led to the production of
8 ns pulses with an average energy of 3 mJ at 243 nm. Prior to being sent into the
interaction region the beam size was shaped to 2.5 mm in diameter by passing a 3:2
telescope.
The external dye cell was of the Bethune type [120], i.e. a standard 45 deg right
angle prism with a drilled hole for dye solution as shown in Fig. 3.4a. In this geometry
the pump beam must have a vertical width four times bigger than the bore diameter
to ensure homogenous pumping of the dye. As shown in Fig. 3.4b when pump light
enters the cell though an input window it undergoes internal reflections and four
equal segments (1-4) homogeneously pump the top, front, back and the bottom of
the dye, respectively. The advantage of using a large pumping area is that it prevents
damaging of the input window. Even though the amplified beam spatial profile is
decoupled from that of the pump beam the temporal profile is affected as a dye
amplifier mimics the intensity fluctuations of the pump pulses. Also it is important
that the seed beam must have a size slightly bigger than the bore as a complete
filling ensures more efficient energy extraction, reduces saturation in the dye cell
and helps to deplete amplified spontaneous emission (ASE). In the experiments, the
Bethune cell had a bore diameter of 4 mm and a length of 35 mm. It was mounted on
translation stages enabling its precise positioning in the xy direction and rotation
along ϕ and θ angles (see Fig. 3.4). To shape the Nd:YAG beam to the size of
16×35 mm a f = −100 mm plano-concave lens was used.

45
Chapter 3. Experimental facilities

Efficiency of the photoexcitation step

For a two-photon excitation process with a single light beam the population in
the excited state can be written as [121]:

  σ I 2 ∆T 
2γ
Nexcited = Nground 1 − exp . (3.7)
hν
with Nground being the population in the ground states, σ2γ the two-photon
absorption cross-section, I the intensity of the light beam, ∆T the pulse duration
and hν the photon energy. The amount of hydrogen atoms in the 2 S1/2 ground state
can be expressed as:

Nground = n · Vinter , (3.8)

where n is the density of H atoms per cm3 and Vinter the volume of interaction
between the atomic hydrogen beam (φH = 0.15 cm) and the 243 nm laser beam
2
(φ243 = 0.25 cm), i.e. Vinter = πrH l243 = 0.0044 cm3 .
The 1s-2s two-photon cross-section for the hydrogen atom is given by the ex-
pression [122]:

h cm4 s i
σ2γ = 2.75 × 10−17 g(∆Ω)G(2) , (3.9)
W
where G(2) is the two-photon statistic factor, which for dye and excimer lasers
has a value of 2 [121], and g(∆Ω) is the spectral linewidth function given by [122]:

 1/2
4ln(2)
π
g(∆Ω) = 2 + 2∆ω 2 )1/2
, (3.10)
(∆ωD L

with ∆ωD and ∆ωL being the Doppler and the laser linewidth, respectively. For
the experimental conditions ∆ωD = 9.6 × 1010 s−1 and ∆ωL = 1.9 × 1010 s−1 , hence
g(∆Ω) = 9.4 × 1012 s−1 and the two-photon cross-section for hydrogen is:

h cm4 i
σ2γ = 5.18 × 10−28 . (3.11)
W
During the experiments the average energy at 243 nm was 3 mJ/pulse, and the
pulse duration was 8 ns giving the light beam intensity of I = 6 MW/cm2 . Hence,
the population of H atoms in the mixture of n = 2 s– and p–states was:

46
 3.1 Photoionization microscopy of hydrogen atoms

  σ I 2 ∆T 
2γ H∗
Nexcited = n · Ninter 1 − exp − = 1.5 × 104 atoms . (3.12)
hν pulse

3.1.4 Photoionization of excited hydrogen atoms – Pulsed

Dye Amplifier
The final and the most crucial step in the experiments on hydrogen atoms was
photoexcitation into Stark states, whose energy levels were located above the sad-
dle point energy and below the field-free ionization potential. This step required a
narrow-band, tunable in the rage of 365-367 nm and a pulsed laser system operated
at the same frequency as the pulsed valve and the PDL-3.
Pulse amplification of a continuous single mode dye laser has found wide appli-
cation in experiments where tunable, high power and near Fourier-limited pulses are
required [123–126]. This technique was first demonstrated by Salour [127], who am-
plified the output of a cw dye laser in a chain of three dye cells pumped by a N2 laser.
By pumping an organic dye in a cell, molecules are populated into the first excited
(electronic) singlet manifold. In a very short time they decay into low-lying substates
of the manifold through non-radiative transitions. Subsequently, molecules can radi-
ate their energies and drop to one of substates in the lowest singlet vibration-rotation
manifold (spontaneous emission). If light is present during that process molecules
can be stimulated to emit coherent photons with the same frequency as the input
photons (stimulated emission). In this way a pulsed dye amplifier (PDA) generates
coherent light pulses at a wavelength that is controlled by a seed laser. The design
of Salour resulted in pulses with a peak power of 100 kW and a linewidth of 60 MHz
in the visible. Since then many variations have been introduced leading to a higher
peak power [123], greater gain [128] and significant reduction of ASE [129]. The
laser system used in the hydrogen experiments was a three-stage, four-pass pulsed
dye amplifier based on a design by Cromwell [130]. It is schematically shown in Fig.
3.5 with a general description of its operation given in the figure caption. In the text
below all elements of the PDA are explained in detail.

Lasers
A ring dye laser (899-29 Autoscan II, Coherent) [131] pumped by a cw Nd:YVO4
(Verdi V-12, Coherent) laser was exploited as the seed beam for the PDA. This dye
laser delivered a continuous, well-characterized and tunable output over the whole
visible spectrum with a typical bandwidth of 500 kHz. An external wavelength meter
combined with a software package ensured automatic selection of wavelengths with
an absolute accuracy of ±200 MHz and a reproducibility of ±50 MHz over the 450 to
900 nm operating range. In experiments on hydrogen atoms Pyridine 2 dye ensured
operation in the range of 725-745 nm providing an average output power of 190 mW
at 725 nm. As the pump beam for the PDL, a Q-switch Nd:YAG laser (Quanta-Ray
GCR-3, Spectra Physics) was used. This laser was chosen due to its high peak power
at 532 nm – a wavelength readily absorbed by many dyes. At 10 Hz it delivered 8 ns

47
Chapter 3. Experimental facilities

ø 6mm

HR

FR T2

WP2
20%
BS ø 3mm

98% GL
BS
Nd:YAG WP1

T1
SI

Verdi 899-29 FI
ø 1mm

Figure 3.5: Schematic representation of a home-built pulsed dye amplifier (PDA) laser sys-
tem, which comprised a ring-dye laser (899-29, Coherent), pumped by a Nd:YVO4 laser
(Verdi, Coherent). The vertically polarized output beam passed first through a Faraday iso-
lator (FI) and then through a spatial isolator (SI) composed of two achromatic doubles and
a pin-hole. The latter shaped the beam before entering a φ = 1 mm dye cell, pumped with
2% of the output energy of a Nd:YAG (GCR-3, Spectra Physics) laser, provided by a 98%
beam splitter (BS). After amplification in the first dye cell the beam passed through a 1:3
telescope (T1), a half wave plate (WP1), which rotated the polarization by 45 deg, and a
Glan-Laser (GL) polarizer. In the second φ = 3 mm dye cell the beam was amplified twice.
The dye cell was pumped with 20% of the output power delivered by the Nd:YAG laser beam.
After the first amplification the beam was reflected back to the cell by using a high reflecting
mirror (HR). The beam polarization was rotated by 90 deg due to a double pass through a
Fresnel rhomb (FR) ensuring its reflection by the GL polarizer. After passing a wave plate
(WP2) the polarization was rotated back to 0 deg and the beam diameter was doubled by a
1:2 telescope (T2) before it entered the last dye cell φ = 6 mm, where the final amplification
took place.

pulses with the average energy of 400 mJ per pulse. By making use of beam splitters
the pump energy was distributed in the ratio of 2:20:78 over the three dye cells, as
shown in Fig. 3.5.

Optical and spatial isolators

Before entering a chain of amplifiers the seed beam passed through two isolators.
The first one, a Faraday isolator, prevented amplified spontaneous emission from

48
 3.1 Photoionization microscopy of hydrogen atoms

the amplifier chain to go back to the 899-29 and cause mode-hopping. This optical
isolator consisted of two Glan-Laser calcite polarizers, a Faraday rotator and a half-
wave plate as shown in Fig. 3.6. The transmission through the Faraday isolator
was independent of wavelength, due to broadband antireflection coatings, and had
a maximum value of 70%. The Faraday isolator worked as follows: the laser beam
passed through the first polarizer (GL1), which was set to pass vertically polarized
light – the polarization of the seed beam. Subsequently, the beam passed through
a Faraday rotator which rotated the polarization of light by +45 deg due to the
Faraday effect. Then the beam was filtered by the second polarizer (GL2), which was
set for this polarization. At the end light traveled through a half-wave plate (WP)
which rotated the polarization plain back to a vertical direction. If unpolarized light
(e.g. ASE) was coming in the opposite direction, namely from the amplifier chain to
the 899-29 laser, it was first filtered by the +45 deg polarizer (GL2) and then the
Faraday rotator flipped polarization to 90 deg (horizontally polarized). Consequently,
the light was completely blocked by the GL1 polarizer.
The second isolator consisted of two achromatic doublet lenses and a pinhole.
This spatial filter played a twofold role: it controlled the growth of ASE and shaped
the seed beam profile for the first dye cell. The ASE reduction was possible as
the spot size of the laser beam was much bigger than the pinhole diameter. Under
these conditions only a well-collimated beam, i.e. the beam from the 899-29 laser,
passed through the pinhole whereas an uncollimated beam (e.g. ASE) was efficiently
suppressed. For a lens of focal length f and a beam radius w, the pinhole diameter
d can be determined as:

2f λ
d < 2w0 = (3.13)
πw
where w0 is the focal radius of the beam and λ is its wavelength. The seed beam
had w = 0.5 mm at λ = 730 nm meaning that for a f = 100 mm lens the focal spot
radius was 46 µm. Therefore, the pinhole diameter was chosen to be d = 75 µm. For
such a combination of lens and pinhole the measured transmission was 97%.
When light is diffracting on a circular aperture, like a pinhole, it creates an Airy
pattern with a bright spot in the center known as the Airy disk. This fact was used
to calculate the focal length of the second lens, which re-collimated the beam. The
radius of the Airy disk is given by [132]:

fλ
r = 1.22 . (3.14)
d
For the beam expanding from the 75 µm pinhole the radius of the first minimum
is 1 mm at a distance of 84 mm. Therefore, by using a f = 75 mm lens the seed
beam had a FWHM approximately equal to the first dye cell diameter, i.e. 1 mm.

49
Chapter 3. Experimental facilities

GL 1 Faraday rotator GL 2 WP

Figure 3.6: Faraday optical isolator comprising two Glan-Laser calcite polarizers (GL1 and
GL2), a Faraday rotator and a half-wave plate (WP). The double arrows represent the light
polarization and B is the magnetic field direction inside the Faraday rotator.

Amplification stages
For the amplification stages, dye cells of the Bethune type were used, which are
described in Section 3.1.3 and shown in Fig. 3.4. The cells had sizes of 1 mm φ
×20 mm long, 3 mm φ ×35 mm long and 6 mm φ ×35 mm long for the first, second
and third dye cell, respectively. The cells were mounted on translation stages which
allowed their positioning with respect to the seed beam. Before each cell a pair of two
achromatic lenses served as telescopes with magnifying powers of ×1, ×3, and ×2
to shape the beam profile for the first, second and third dye cell, respectively. The
size of the pump beam was four times wider than the bores owing to plano-concave
lenses placed in the front of the Bethune cells.
A polarization rotation technique was used to accomplish a double pass, double
amplification through the second dye cell (see Fig. 3.5). Firstly, a half-wave plate
(WP1) rotated the beam polarization by 45 deg. Secondly, the beam was filtered
through a polarizer (GL) and entered the second dye cell. After passing the cell the
polarization was rotated by 45 deg in a λ/4 Fresnel rhomb (FR) and the beam was
reflected back by a high reflecting mirror (HR) and made a second pass through the
rhomb which rotated its polarization by another 45 deg. Finally, the beam passed
through the cell the second time and was completely reflected by the polarizer (GL)
in the direction of the third Bethune cell. A second half-wave plate (WP2) rotated
the polarization plane back to the vertical direction.

Characteristics of PDA
The pulsed dye amplifier system was operated in two different wavelength ranges.
In the first one Kiton Red dye was used for the 899-29 laser and Sulforhodamine 640
(SR640) dye for the Bethune cells. The ring dye laser provided tunable light in the
range of 580-650 nm with an average output power of 300 mW at 640 nm. SR640
used in the amplifier chain allowed pulse generation in the range from 605 to 628 nm.
The dye concentration was 2.1×10−4 , 9.2×10−5 and 7.9×10−5 moles/liter for the
first, second and third dye cell, respectively. The energy output curve measured for
these conditions is shown in Fig. 3.7a. The maximum output energy of 29.5 mJ was

50
 3.1 Photoionization microscopy of hydrogen atoms

a b

Figure 3.7: (a) Energy output curve of the PDA for Sulforhodamine 640. (b) Schematic
view of the PDA output beam profile, where light pink indicates the seed beam profile and
dark pink a fraction of the beam, which was pulsed dye amplified.

achieved at 616 nm. The measured portion of the ASE was 33% over the whole
tunable range.
In the second tunable wavelength range, suitable for experiments on hydrogen
atoms, Pyridine 2 dye was used for the 899-29 laser, providing an average output
power of 190 mW at 725 nm. Bethune cells contained Styryl 7 dye at concentrations
of 3.8×10−4 , 1.3×10−4 and 5.6×10−5 moles/liter for the first, second and third dye
cell, respectively. Due to instability of the 899-29 laser in the case of wavelengths
shorter than 725 nm it was impossible to measure the energy output curve as was
done for SR640. The maximum output energy of 3.5 mJ was obtained at 729.4
nm. This low output energy could be explained by the moon-like beam profile of
the amplified beam (see Fig. 3.7b). This uneven beam profile showed that only
approximately 30% of the seed beam was successfully amplified. This may have been
caused by micro-damages of the second dye cell that prevented efficient pumping the
dye from the top, back and the bottom. The other cause may be reflected fluorescence
from the Fresnel rhomb into the dye cell, which led to gain competition between the
amplified and the reflected light. Nevertheless, the amplified output beam after being
frequency doubled in a KDP (potassium dihydrogen phosphate) crystal (10×10×30
mm) provided 650 µJ/pulse at 365 nm, which was sufficient to photoionize hydrogen
atoms from the n = 2 states into quasi-bound Stark states.

51
Chapter 3. Experimental facilities

3.1.5 Detection of photoelectrons – velocity map imagining

(VMI) spectrometer
Proposed more than a decade ago, the velocity map imaging technique, combin-
ing a three-plate electrostatic lens with a position-sensitive detector, revolutionized
the imaging of ions and electrons resulting from photoionization and photodisso-
ciation experiments [114]. By changing the commonly electrode configuration used
in imaging techniques, i.e. a positive plate (repeller) and a negative grid, to three
plane electrodes, a better image resolution was achieved. Even more importantly
all charged particles with the same initial velocity vector, regardless of their origi-
nal position, were mapped onto the same point in the focal plane. In recent years
VMI was successfully implemented in experiments exploring atomic and molecular
dynamics on femtosecond [133, 134] and attosecond [135, 136] time-scales, in mass
spectroscopy [137, 138], slow photoelectron spectroscopy [139] and for the first time
in the observation of holographic structures in atoms [140].
The velocity map imaging spectrometer used in the hydrogen experiments com-
prised three plates (repeller, extractor and ground) mounted inside a 50 cm flight
tube. Each electrode had a diameter of 70 mm and a thickness of 1.25 mm. The spac-
ing between the repeller and extractor was 17 mm, whereas the spacing between the
extractor and ground was 30 mm. The repeller plate had a hole of 1.5 mm in diame-
ter, through which hydrogen atoms entered the VMI. The holes in the extractor and
ground electrodes had a 20 mm diameter. The advantage of using open electrodes is
that the equipotential surfaces are bent and can be adjusted simply by changing the
voltage. Thus such an electrode configuration acts as an electrostatic lens enabling
to focus the image sharply on the detector. In our VMI the ratio between the re-
peller and extractor voltages was always 5/4 ensuring that all photoelectrons with
the same velocities perpendicular to the flight tube ended up at the same position
on the detector. It should be pointed out that the VMI worked perfectly only when
the interaction region between the hydrogen atoms and laser lights was exactly in
the middle of the repeller and extractor electrodes.
At the other end of the flight tube a position-sensitive detector was mounted,
which comprised a set of two MCPs (4 cm in diameter), a phosphor screen (4 cm in
diameter) and a CCD camera (F-100 B/C, Pike). When photoelectrons hit the MCPs
they ejected electrons, which were subsequently multiplied through the channels of
the MCPs. The created shower of electrons reached the phosphor screen causing
fluorescence and the produced image was recorded by the CCD camera. Knowing
the photoelectron time-of-flight, the camera was triggered allowing efficiently sup-
pression of background signals.
The size of the photoelectron images obtained with the VMI is proportional to the
square root of the initial kinetic energy. For example photoelectron images recorded
slightly above the saddle point energy for an electric field of 800 V/cm would have
a maximum radius of ∼100 µm, meaning that it would be impossible to distinguish
individual interference rings. Therefore, we used an electrostatic Einzel lens [116],
composed of three stainless steel plates with a hole size of 10 mm and spaced by
5 mm plastic separators, which was mounted inside the flight tube approximately

52
 3.2 Photoionization microscopy of helium atoms

a b

Figure 3.8: Experimental images recorded for photoelectrons with the kinetic energy of 10
meV at an electric field of 170 V/cm (a) without the zoom lens on (voltages at the electrodes:
repeller -1.44 kV, extractor -1.16 kV), and (b) with the Einzel lens on (repeller at -1.96 kV,
extractor at -1.68 kV and lens at -1.68 kV), allowing magnification of image (a) by a factor
of 5. Note that in both images the signal intensities have been normalized: red corresponds
to 1 whereas dark blue to 0.

one-third of the way toward the detector (see Fig. 3.1). To ensure that the zoom lens
would not change the energy of the photoelectrons the potential before and after the
lens had to be the same. This was realized by setting the two outer electrodes at the
ground potential. By applying a negative voltage to the central plate photoelectron
images were magnified up to 1 order of magnitude without deteriorating the image
quality. Fig. 3.8 shows two photoelectron images acquired for the same electric field
of 170 V/cm in the interaction region and for photoelectrons with the same initial
energy of 10 meV, with the zoom lens off (a) and on (b). By applying -1.68 kV to
the lens and adjusting the voltage at the repeller plate for an optimum focusing, the
initial image (without the lens) was enlarged by a factor of 5.
In the experiments, photoionization spectra of the hydrogen atom were measured
for different electric fields. This was accomplished by scanning the 899-29 laser con-
tinuously over a range of 10 GHz (with a speed of 100 MHz/s) with synchronized
acquisition of photoelectron images over a small region on the detector (100×100
pixels), where the photoelectrons resulting from the experimental events were ex-
pected. Each image was recorded for approximately 20 laser shots. By using an
external program the total signal intensity per image was calculated and plotted as
a function of the photoionization laser wavelength (e.g. see Fig. 5.4).

3.2 Photoionization microscopy of helium atoms

The field-free ionization potential of helium is almost twice as large as that of
hydrogen, i.e. 24.59 eV [141], but unlike the latter it has metastable states, which
remain stable in the presence of an external DC electric field. This fact has been
used in designing the photoionization microscopy experiments on helium atoms.
Namely, an ionization scheme was chosen, where helium atoms are first excited into

53
Chapter 3. Experimental facilities

the 21 S metastable state (20.62 eV) by means of a DC electric discharge and then
photoionized by absorbing one photon of UV light.
To produce (21 S) metastable helium atoms a discharge source was designed,
which is shown in greater detail in the inset of Fig. 3.9. In the experiments, helium
gas was injected via a gas inlet (a) into a quartz tube (b), which surrounded a
tantalum needle (c). A negative voltage applied to the needle caused a DC electric
discharge between the needle tip and a grounded skimmer (d). Resulting metastable
helium atoms were formed in a beam by the skimmer. The He∗ beam intensity was
constantly measured by a Faraday cup (e). A 1 mm × 300 µm slit in a copper
plate (f), which was placed 20 cm after the source chamber, shaped the beam before
it entered a velocity map imaging spectrometer (g). Inside the VMI a DC electric
field was produced by applying a voltage difference across the electrodes and He∗
atoms were excited into quasi-bound Stark states by making use of a tunable UV
laser with a wavelength in the range of 312–314 nm. UV light was obtained by
means of frequency doubling the output of a ring dye laser (899-29, Coherent) with
a commercial frequency doubler (WaveTrain, Spectra Physics) providing an average
output power of 40 mW. The polarization of the UV light was chosen to be parallel
to the electric field enabling the population of m = 0 Stark states. The launched
photoelectrons were imaged on a position-sensitive detector comprising two MCPs
(h), a phosphor screen (i) and a CCD camera (j). Due to low kinetic energies of
the photoelectrons it was necessary to magnify the photoelectron images making
use of an Einzel lens (k). In order to suppress contributions resulting from Penning
ionization of background gas, the velocity map imaging spectrometer was surrounded
by a liquid nitrogen cold trap (l). Photoelectron images were obtained by measuring
the difference between acquisitions with and without the UV laser.

3.2.1 Metastable helium source

The first metastable helium source combining a corona discharge with a sonic
nozzle was proposed by Searcy in 1974 [142]. This device was remarkable for its
simple construction as it consisted of a nozzle made of a dielectric material and
an ordinary steel sewing needle, which served as the anode. By applying a positive
voltage to the anode the corona discharge between the needle and a skimmer was
ignited. The discharge was maintained across a pressure gradient and provided a
beam flux of 1010 metastable helium atoms per second per steradian. The beam
was dominated by triplet metastables, as 21 S metastables rapidly converted to 23 S
by collisions with thermal electrons [143]. One year later Leasure [144] presented a
modified source, which increased the beam flux by two orders of magnitude. This
was achieved by improving the source geometry, i.e. introducing a glass tube and
a floating bias supply. Further modification of the source, made by Fahey [145],
resulting in an enhancement of the beam flux to 3.5×1014 He∗ atoms s−1 sr−1 . In
this type of metastable helium source the needle served as the cathode of a low
voltage discharge, which was maintained at a negative potential with respect to a
grounded skimmer. A slight variation of this device was introduced by Rooijakkers
et al. [146], who added liquid nitrogen cooling to reduce the longitudinal velocity of

54
 3.2 Photoionization microscopy of helium atoms

j
h i
l

e k

g
f
m

d
c a
b

Figure 3.9: Schematics of the experimental setup used in the photoionization microscopy
experiments on helium atoms. The inset shows a metastable helium source, where He gas
was loaded via an inlet (a) into a quartz tube (b). By applying a negative voltage to the
tantalum needle (c) a discharge took place between the needle tip and the grounded skimmer
(d) resulting in the production of (21 S) metastable helium atoms. A He∗ beam was shaped
by the skimmer and its intensity was monitored by a Faraday cup (e). The beam was further
shaped by a slit in a copper plate (f ) before it entered a main chamber containing a velocity
map imaging spectrometer (g). Inside the VMI He∗ atoms were photoionized with UV light
and the photoelectrons were projected on a detector comprising two MCPs (h), a phosphor
screen (i) and a CCD camera (j). Halfway in the detector flight tube an Einzel lens (k) was
mounted. To reduce the effects of Penning ionization, a cold trap (l) was employed and to
suppress the Earth magnetic field the spectrometer was surrounded by a µ-metal shield (m).

55
Chapter 3. Experimental facilities

helium atoms, which was essential in their experiments.

The metastable helium source, used for experiments presented in this thesis, was
based on the latter design and its detail scheme is shown in Fig. 3.10 a. A tantalum
needle (1) served as the cathode and was located inside a 6 mm diameter quartz
tube (2), which was filled with helium gas introduced by a gas inlet (3). Tantalum
was chosen due to its high melting point and good thermal conductivity. As a nozzle
(4) a piece of boron-nitride (BN) was used because of its softness, workability and
poor electrical conductivity. A boron-nitride cup had a 0.25 mm diameter hole, was
attached to the quartz tube by using a GE varnish and pressed into a copper block
(5), which was cooled with water supplied by bellows. To avoid arcing between the
needle and the block (which was grounded) a 2 mm thick ceramic ring (6) was
placed on top of the cup. The skimmer (7) was a cone-shape piece made of stainless
steel with a 1 mm hole at the apex and kept at a ground potential. The apparatus
comprising elements (1)-(6) was mounted on a xyz translation stage, which enabled
its precise positioning with respect to the skimmer. Owing to that it was possible to
optimize the beam flux, which is dependent on the nozzle - skimmer position. The
source was mounted inside a vacuum chamber (shown in Fig. 3.9) that was pumped
by a 540 l/s turbo pump, assuring a base pressure of 2.3×10−8 Torr (without gas).
Helium gas was supplied into the source chamber by a needle valve, which allowed
precise adjustment of the gas pressure. Typically, the discharge started for a pressure
of 1.5×10−4 Torr in the source chamber and after the ignition the pressure was
reduced to 6×10−5 Torr. The source chamber had two windows through which it
was possible to observe the discharge glow by a naked eye (see Fig. 3.10c).
The needle in the source was the cold-cathode type, meaning it was not inde-
pendently heated and its operation was as follows. By applying a sufficiently high
negative voltage to the cathode electrons were emitted by means of field electron
emission. These electrons ionized the gas creating ions, which, due to the presence
of a negative potential, were accelerated toward the needle. By gaining sufficient
energy as a result of the acceleration the ions kicked out electrons from the cath-
ode surface. These secondary electrons sustained the discharge and consequently the
voltage could be lowered. Therefore, to ignite and sustain the DC electric discharge
in the source two power supplies were used, which were connected in a simple elec-
trical network as shown Fig. 3.10b. When helium gas was loaded a negative voltage
of -3 kV was applied to the needle by a high-voltage power supply (0-10 kV, 0-3 mA,
Wallis Hivolt). The ignited discharge resulted in a current increase and the current
flow across the high-voltage power supply was blocked by a 1 kΩ resistor while a 6
kV diode allowed for it to pass through a high-current power supply (0-1.5 kV, 0-100
mA, Heinzinger), which was set to 700 V. The latter power supply maintained the
discharge whereas the former was switched off. The current through the discharge
was limited to 20 mA.
The intensity of the metastable helium beam was measured with a Faraday cup.
In this simple detector He∗ atoms struck a stainless steel target and were de-excited.
As the atoms had an internal energy higher than the work function of the stainless
steel, secondary electrons were emitted, which induced a current measured with a
picoammeter (6487, Keithley). The relation between the current I and the beam

56
 3.2 Photoionization microscopy of helium atoms

5
a 3
7
1 4

6
2

b c

Figure 3.10: (a) Metastable helium source comprising a tantalum needle (1), quartz tube
(2), gas inlet (3), boron-nitride nozzle (4), cooling block (5), ceramic ring (6) and skimmer
(7). The needle is connected to high-voltage (HV) power supplies. (b) Electrical circuit con-
necting two power supplies, which deliver high-voltage to initiate and sustain the discharge.
(c) View through a window showing the discharge glow between the nozzle and the skimmer.

flux F is given by [147]:

I = eγF ∆Ω (3.15)
where e is the elementary charge, γ the secondary electron emission coefficient
and ∆Ω the solid angle (for the experimental geometry ∆Ω = 7.5×10−6 ). During
the discharge helium atoms were promoted into 23 S and 21 S metastable states.
However, in the used source design the atomic beam is dominated up to 97% by
triplet metastables [146] since singlet metastables rapidly convert to 23 S due to
collisions with thermal electrons. Therefore, we assume that the measured current
comes from the impact of triplet metastables. For 23 S He atoms hitting a stainless
steel target γ has a value of 0.69 [148]. Thus the flux was determined to be 2.5×1014
He (23 S) atoms per second per steradian at the discharge current of 20 mA. Hence,
the intensity of singlet metastables was 7.5×1012 He (21 S) atoms s−1 sr−1 . The

57
Chapter 3. Experimental facilities

beam intensity increased linearly with the discharge current and slowly decreased
with the gas pressure. It was strongly depended on the nozzle-skimmer position in
the xy-plane and hardly on the z direction. The source ran stably for many hours
of operation without any adjustments.
The design of the metastable source allows operation for gasses different than
helium. The source was tested for argon gas, where the discharge was ignited at
-1.7 kV and at 2.4 ×10−3 Torr in the source chamber. At a discharge current of 22
mA the beam intensity was 2.8×1014 metastable argon atoms s−1 sr−1 . For neon the
discharge started at -2 kV and a dark red discharge glow was observed. However, after
few moments the discharge quenched and the beam flux could not been measured.

3.2.2 Photoionization of He (21 S) atoms with UV light

Inside the VMI metastable helium atoms interacted with tunable UV light, which
was obtained by means of frequency doubling cw light provided by the ring dye laser.
Kiton Red dye used in the ring dye laser ensured its tunable operation in the range
of 610-630 nm with an average output power of 400 mW at 625 nm. Prior being
sent into the VMI the laser beam passed through an external ring cavity frequency
doubler (WaveTrain, Spectra Physics). This commercial device, designed for single
frequency cw pump lasers, guaranteed 10% efficiency for second harmonic generation
(SHG).

Frequency doubling of continuous wave laser light

The technology for SHG of cw lasers is substantially different compared to ar-
rangements used for pulsed laser systems. The reason is that the power of the second
harmonic is proportional to the square of the fundamental wave power, i.e. in the
first approximation:

P2 = cP12 , (3.16)

where c is the so-called conversion coefficient dependent on e.g. the refractive

index, the fundamental light wavelength, phase matching, etc. Usually the output
power for cw lasers is about 1 W, which makes it challenging to efficiently generate
the second harmonic in a single pass through a nonlinear crystal. Therefore, light
should be locked in an external resonator containing a nonlinear crystal. In such a
cavity the energy of many round-trips is stored and by constructive superposition
the initial light power is increased [149]. The total beam path in the cavity is a
multiple of the beam’s wavelength to ensure its enhancement. As vibrations are
always present the resonator length is actively stabilized by a piezo element driven
in a servo loop.
The WaveTrain consisted of a resonator based on the DeltaConceptT M , i.e. com-
prising a prism mounted on a piezo element, a nonlinear crystal (BBO) and two
mirrors. As the main limiting factors for efficient SHG are the losses in the res-
onator, both the prism and the crystal were cut at Brewster’s angle. To fulfill the

58
 3.2 Photoionization microscopy of helium atoms

phase matching condition, for a maximum conversion efficiency, critical phase match-
ing was used. This method takes advantage of the angle dependence of the refractive
index and by tilting the crystal phase matching was achieved. A corrective signal
needed for the servo loop was generated exploiting the Pound-Drever method [150].
Namely, the incoming beam first passes through an electro-optic phase modulator,
which adds sidebands to the spectrum of the laser line, and then it interferes with
the unchanged beam reflected from the cavity. The superposition of both beams pro-
vides a radio-frequency modulation signal, which is monitored with a photo-diode.
Due to active stabilization the 899-29 laser could be continuously scanned as fast as
10 GHz/s, although in experiments the scanning speed was chosen to be 100 MHz/s.

Photoabsorption cross-sections for resonant and non-resonant transitions

The efficiency of a non-resonant transition from the metastable 21 S state to the
field-free ionization potential E = 0 can be calculated by means of the Lambert-
Beer law, i.e. p = nσl with p being the photoabsorption probability, n the density of
He (21 S) atoms in the interaction region, σ the absorption cross-section, which for
312.18 nm has a value of 9.7 ×1018 cm−2 [151], and l the length of the interaction
region. The density of 21 S metastables can be calculated as:

F ∆Ω
n= , (3.17)
Aϑ
where A is the area of the beam shaped/defined by the slit in the copper ring, i.e.
A = 300 µm ×1 mm and ϑ is the velocity of the beam, which had an approximate
value of 1500 m/s. The density of metastables was 1.3×105 He (21 S) atoms/cm3 .
The He beam had a diameter of 1 mm, hence the photoabsorption probability was
p = 1.2×10−13 . As the average output power of UV light was 40 mW the number of
photons per second was 6.3×1016 . Thus the expected number of events per second
for this non-resonant transition was 7.6×103 .
A photoabsorption cross-section for a resonant excitation into a Stark state is at
least one order of magnitude higher compared to the non-resonant transition. This
is clearly demonstrated in Fig. 6.2b, where the calculated photoabsorption spectrum
of helium atoms in the electric field of 468 V/cm is shown. Unfortunately, the cal-
culation did not provided exact values of photoabsorption cross-sections, however,
based on the signal to noise ratio it is expected that the most intense resonances in
the experimental spectrum (Fig. 6.2a) correspond to the situation when almost all
hydrogen atoms are ionized giving approximately 105 events per second.

59
Chapter 3. Experimental facilities

3.2.3 Detection of photoelectrons – Penning ionization prob-

lem
In the experiments 21 S metastable helium atoms were photoionized in a DC
electric field of 468 V/cm. The laser polarization was chosen to be parallel to the
electric field, which resulted in the population of m = 0 Stark states. To detect
photoelectrons resulting from resonant photoexcitation of Stark states, located at
energies between the saddle point energy (Esp = -132.4 cm−1 ) and the field free
ionization potential (E = 0), the velocity map spectrometer was used. The VMI
comprised three electrodes with dimensions and spacing as given in Section 4.1.5.
The flight tube had a length of 33 cm and included an Einzel lens mounted 18 cm
from the repeller plate. The lens magnified the photoelectron images by one order of
magnitude. A position-sensitive detector comprised MCPs, a phosphor screen and a
CCD camera.
The MCPs can be operated at pressures as high as 10−5 Torr. Thus, in the original
setup design, the He∗ source was directly connected to the main chamber. Under
typical operation conditions, when helium gas was loaded into the source chamber,
the pressure in the main chamber was 7×10−7 Torr. However, the background signal
was so high that it was impossible to operate the MCPs. This signal was caused by
Penning ionization, i.e. due to collisions of metastable He atoms with residual gas
atoms or with themselves resulting in the production of ions and electrons [152]:

He∗ + A → He + A+ + e− , (3.18a)
∗ ∗ − −
He + He → He + He + e (or+
He+
2 + e ). (3.18b)

As the process given by Eq. (3.18a) is linearly dependent on the residual gas
pressure a liquid nitrogen cold trap was installed in the main chamber. It provided
a decrease of the pressure by one order of magnitude. Moreover, a 20 cm buffer
chamber was introduced between the source and the main chamber to reduce the
number of ions and electrons resulted from the reaction given by Eq. (3.18b). At the
connection with the main chamber a copper ring with a 300 µm × 1 mm hole was
used, which also played the role of a gasket. These two elements decreased the base
pressure in the interaction region to 4×10−9 Torr and significant suppression of the
background signal was achieved.
Nevertheless, images observed on the detector were not free from background
signals. As the metastable helium beam and UV light were continuous it was impos-
sible to trigger the camera. To suppress the background signals the photoelectron
images were recorded with and without UV light and the final images were obtained
by subtracting both images. Typical acquisition times were 20 minutes per image.

60
 Semiclassical interpretation of

4 the interference patterns in the

photoionization microscopy
experiments of hydrogen atoms

4.1 Introduction
The hydrogen atom is the most abundant chemical element in the universe, which
due to its simplicity is the most studied and best known particle, yet still attracting
attention of theoretical and experimental scientists [153, 154]. The effect of an electric
field on this atom, discovered a hundred years ago, was the first problem successfully
treated within the framework of quantum mechanics [62]. Implementation of laser
spectroscopy to study hydrogen atoms in DC electric fields brought more information
concerning properties of quasi-bound Stark states [70, 75, 81, 155, 156] and boosted
theoretical treatments of this subject [89, 157].
The problem of electron motion in a combined Stark + Coulomb potential can
also be treated by means of classical mechanics. In a series of famous papers, Kon-
dratovich and Ostrovsky developed a semiclassical theory describing photoionization
of the hydrogen atom in a uniform electric field [50, 51, 158, 159]. The advantage
of the semiclassical approach is that the Stark effect solved in parabolic coordinates
has, in contrast to the quantum mechanical approach, exact analytical solutions,
whose interpretation is much more intuitive. For example, the photoionization cross
section, calculated in the dipole approximation, is known to be caused by resonance
and interference effects. Relative contributions of these phenomena depend on the
photoelectron energy E and therefore, the energy range can be divided into charac-
teristic regions. Additionally, in the presence of a uniform electric field far from the
atomic core, the photoelectrons move asymptotically and parallel to the field axis.
Hence, one can measure the spatial distribution of the photoelectron’s density on
a plane placed perpendicularly to the field axis. The recorded photocurrent should
show interference patterns resulting from the interference among an infinite number
of classical trajectories along which the photoelectrons move towards the screen. Fur-
thermore, if the photoionization proceeds from a particular Stark state the measured
photocurrent represents the squared wave function describing the electron motion
along the ξ coordinate.
Following suggestions of Kondratovich and Ostrovsky [51] an apparatus allowing
observation and measurement of photoionization interferences of electrons is drawn
schematically in Fig. 4.1. It comprises two plane plates procuring a uniform electric
field of strength F , laser light focused halfway between the plates, a beam of neutral

61
Chapter 4. Photoionization microscopy

hn

Figure 4.1: Schematic representation of a ”photoionization microscope”, i.e. an apparatus

where atoms are ionized by focused laser light in a presence of a DC electric field, created
by two plane plates, and the resulting photoelectrons are projected on a position-sensitive
detector, which in this configuration is one of the electrodes. The created interference patters
are strongly correlated to the processes occurring in the vicinity of the atom and in special
case, when the photoionization proceeds from a Stark state, the observed dark fringes reflect
the nodal structure of an electronic wave function.

atoms passing through the focus of the laser and a position-sensitive detector po-
sitioned perpendicular to the field axis. The apparatus is called a ”photoionization
microscope” [49] as electronic wave functions and interference fringes are visible on
a microscopic scale.

4.2 Classical theory of the interference phenomenon

The aforementioned interference effect can be exclusively explained by studying
the peculiarities of the classical motion of the photoelectrons in the presence of a
combined Stark + Coulomb potential. Mathematically this problem becomes equiv-
alent to the description of an artificial satellite motion in a planet’s gravitational
field with a constant acceleration [160]. Here the gravitational field is an analogy
to the Coulomb potential and the acceleration is an analogy to the homogeneous
electric field. By introducing a reduced time variable τ , according to the relation
dτ = dt/r, the equations of the classical motion for an electron with the total energy
E in a DC electric field are given by [158]:

62
 4.2 Classical theory of the interference phenomenon

!
dξ 1 Z1 p2φ 1
= 2ξ E+ + 2 − Fξ , (4.1a)
dτ 2 ξ 4ξ 4
!
dη 1 Z2 p2φ 1
= 2η E+ + 2 + Fη . (4.1b)
dτ 2 η 4η 4

In these equations pφ is the orbital angular momentum and, as before (see Section
2.2), the separation constants Z1 and Z2 are related in the following way: Z1 +Z2 = 1.
These equations can be simplified by assuming a plane motion, i.e. (pφ = 0) as
the Coulomb field is the dominant force affecting the photoelectron trajectories at
macroscopic distances. The separation constants Z1 and Z2 can be associated with
an ejection angle β between the initial velocity direction of the photoelectron and
the field axis as follows:

Z1 = cos2 (β/2), (4.2a)

2
Z2 = sin (β/2). (4.2b)

Integration over the equations of motion gives the parabolic coordinates ξ and
η as Jacobi elliptic functions of the reduced time variable τ . General solutions can
be found in Ref. [158]. Here only the energy region between the saddle point energy
and the field free ionization potential i.e. Esp ¬ E ¬ 0 is elaborated further as
the experimental results presented in this thesis were measured solely in this energy
range.

4.2.1 Motion along the ξ coordinate

As previously discussed in Section 2.4 (see Fig. 2.4), the electron motion along
the ξ coordinate is periodic in τ and confined in the region of 0 ¬ ξ ¬ ξmax , where

" !1/2 #
|E| Z1
ξmax = 1+ +1 , (4.3)
F Zc

with Zc being the so-called critical separation constant Zc = E 2 /4F . The ex-
pression for the ξ coordinate of the launched electron in the photoionization process
is given by Jacobi elliptic functions:

!
sn2 (ϕ|mξ )
ξ(τ ) = ξmax , (4.4)
m−1 2
ξ − sn (ϕ|mξ )

with the elliptic modulus and the Jacobi amplitude given by:

63
Chapter 4. Photoionization microscopy

!
1 1
mξ = 1− p , (4.5a)
2 1 + Z1 /Zc
!
1 Z1 √
ϕ= √ 1+ −Eτ . (4.5b)
2 Zc

The period of the photoelectron motion along the ξ coordinate is equal to

!−1/4
√ Z1
Tξ = 2 2K(mξ ) 1 + (−E)−1/2 , (4.6)
Zc

where K(mξ ) is the complete elliptic integral of the first kind.

4.2.2 Motion along the η coordinate

The motion along the η coordinate depends on the angle β with respect to the
electric field at which the photoelectron is launched. If β is smaller than the critical
angle, i.e.:

!
−E
β < βc = 2arcsin √ , (4.7)
2 F

the electron’s motion √is bound, even though its energy lies above the saddle
point (i.e. Esp < E < −2 Z2 F ). This situation corresponds to the case described
in Section 2.4 where the electron’s energy lies below the top of the potential barrier
V (η). However, in the latter case the photoelectron also has a nonzero probability
to tunnel through the barrier, whereas the classical description does not include any
quantum effects. √
For β ­ βc (i.e. −2 Z2 F ¬ E ¬ 0 ) the expression for the time-dependence of
the η-coordinate has the form:

!2
1 − cn(Θ|mη )
η(τ ) = η+ , (4.8a)
sn(Θ|mη )
!1/2
|E| Z2
η+ = . (4.8b)
F Zc

Elliptic functions in Eq. (4.8a) contain parameter mη and argument Θ given as:

64
 4.2 Classical theory of the interference phenomenon

" !1/2 #
1 Zc
mη = 1+ , (4.9a)
2 Z2
!1/4
Z2 √
Θ= 1+ −Eτ. (4.9b)
Zc

Examination of Eq. (4.8) shows that the photoelectron escapes from the atom to
infinity when sn(Θ|mη ) vanishes, i.e. Θ = 2K(mη ), which corresponds to time:

!−1/4
Z1
Tη = 2K(mη ) 1 + (−E)−1/2 . (4.10)
Zc

4.2.3 Direct and indirect photoelectron trajectories

By placing a position-sensitive detector perpendicular to the electric field one
can measure electrons launched during photoionization processes. At any point on
the detector, in the classically accessible region, there exist an infinite number of
classical trajectories that connect this point with the photoionization region (see
Fig. 4.2a). These trajectories can be labeled using the number N of oscillations
in the ξ coordinate, i.e. the number of turning points that occur in the trajectory
between the atom and the detector. For a given N there exist two trajectories labeled
”+”or ”-” depending on whether the number of minima along ξ(τ ) is equal to N or
N − 1.
Photoelectron trajectories can be further classified into two categories repre-
senting different types of ionization processes. In the so-called ”direct” ionization
process the photoelectron is ejected downfield towards the hole in the Coulomb +
Stark potential. In this situation the photoelectron is launched approximately in the
direction of the detector, hence it does not interact with the parent ion and only
follows simple trajectories without crossing the z-axis (direct trajectories). During
ionization the photoelectron may be ejected up field, i.e. away from the hole in the
potential, and it ionizes due to further interactions with the core. In this so-called
”indirect” ionization process the photoelectron after being scattered by the core
follows more complex trajectories along which it crosses the z-axis at least once
(indirect trajectories). Both processes depend on the electric field strength and the
energy of photoelectrons. The indirect process dominates for energies close to the
saddle point, where direct trajectories are forbidden for any ejection angle β. The
onset of direct trajectories corresponds to the situation when Tξ = Tη , and it is given
by a threshold energy [161]:

Edir ≈ 0.775Esp . (4.11)

65
Chapter 4. Photoionization microscopy

a b
180

Emission angle b (deg)

5
135
x(mm)

0 90 Simple trajectories N≤2

bc
45
b0
-5 Bound
Complex N≥2
0
-10 -5 0 5 -1 0 1 2 3 4
z(mm) Reduced Energy e

Figure 4.2: (a) Examples of classical trajectories leading to the same point M on a detector
placed 10 µm from an atom, which is photoionized in a presence of an electric field F =
1 V/cm and at energy E = -2 cm−1 . The thin dashed line indicates the field axis and allows
distinguishing the direct trajectories (labeled 0+ and 1− ), which do not cross the z-axis, from
the indirect trajectories (labeled 1+ and 2+ ), which cross the z-axis once. The thick dashed
line specifies the classically allowed region (on the left). The gray shaded area shows launched
angles smaller than the critical angle βc , for which the electron motion is bound [162]. (b)
Relation between critical angles βc and β0 and the reduced energy ε = E/|Esp |. Depending
on the ejection angle β the electron motion is bound (β ¬ βc ), complex (βc ¬ β ¬ β0 ) or
simple (β ­ β0 ) (figure adopted from Ref. [163]).

Direct and indirect trajectories can also be distinguished by introducing a critical

angle β0 such that the electron launched at angles smaller than β0 follows a complex
motion (indirect ionization), whereas for angles larger than β0 the electron’s motion
is simple (direct ionization). Both critical angles βc and β0 divide the energy space
into three regions as shown in Fig. 4.2b.

4.2.4 Interference patterns – semiclassical wave function

The fact that distinct photoelectron trajectories pass through the same point on
the detector leads to the appearance of an interference pattern, which may be ob-
served experimentally. When the photoionization process does not involve resonant
excitation into a Stark state the observed patterns can be interpreted exclusively
by making use of the semiclassical approximation developed for this problem that
considers the phase accumulated by the electron on a way between the atom and
the detector [161].
The wave function of a hydrogen atom placed in a DC electric field expressed
in parabolic coordinates (see Eq. (2.31)) is a product of functions χ1 (ξ) and χ2 (η),
which satisfy the ordinary differential equations:

66
 4.2 Classical theory of the interference phenomenon

! !
d2 χ1 E Z1 m2 − 1 F ξ
+ + − − χ1 = 0, (4.12a)
dξ 2 2 ξ 4ξ 2 4
! !
d2 χ2 E Z2 m2 − 1 F η
+ + − + χ2 = 0. (4.12b)
dη 2 2 η 4η 2 4

Similar to the discussion regarding the classical electron motion, here only the
planar motion is further considered i.e. m = 0. Hence, modified momenta along both
parabolic coordinates have the form [76]:

!1/2
E Z1 Fξ
p̃ξ = + − , (4.13a)
2 ξ 4
!1/2
E Z2 Fη
p̃η = + + . (4.13b)
2 η 4

Fig. 4.2a shows only four examples from the infinite number of photoelectron
trajectories, which pass via the point M (ξ0 , η0 ) in the classically accessible region. In
the semiclassical approximation the wave function at the point M can be represented
as a sum of all trajectories connecting this point with the photoionization region:

∞
X ∞
X
ψ(M ) = XN + exp(iSN + ) + XN − exp(iSN − ), (4.14)
N =0 N =1

with XN ± being a slowly varying real function depending on the ejection angle
β and the angle θ between the x-axis and the plane of trajectory, and SN ± being
the action along the classical trajectories N ± . The expression for the total action in
terms of modified momenta along the parabolic coordinates ξ and η is given by:

Z η0 Z ξmax Z ξ0
SN ± = p̃η (Z2 , E)dη + 2N p̃ξ (Z1 , E)dξ ± p̃ξ (Z1 , E)dξ. (4.15)
0 0 0

Thus, the total action takes into account the number of complete oscillations of
the photoelectron along the ξ coordinate and the accumulated phase from the last
axis crossing of the trajectory before reaching the point M . The explicit expressions
of the phase integrals can be found in Ref. [50, 89, 90, 107].
Finally, in the laboratory polar coordinates (R, θ) the wave function measured
at the point M (R, θ) on the plane detector can be written as:

67
Chapter 4. Photoionization microscopy

X
ψ(M (R, θ)) = [X(βi , θ)(sinβi dβi )/(RdR)]1/2 exp{i[Sξ (βi ) + Sη (βi )]}, (4.16)
i

where the function X(βi , θ) = XN ± represents the weight of a given classical

trajectory, βi denotes all ejection angles for which photoelectron trajectories are
detected at the radius R; Sξ (βi ) and Sη (βi ) are the phase integrals for the motion
along the ξ and η coordinates [161].
The first experimental observation of the interference pattern resulting from in-
terference among various classical trajectories was presented by the group of Blon-
del in experiments employing a photodetachment microscope [164]. The observed
oscillatory patterns were interpreted as interferences between only two possible tra-
jectories by which electrons move when ejected from negative ions in a presence of a
DC electric field. On the contrary, in photoionization of a neutral atom the launched
electron experiences a combined DC + Coulomb potential, where the latter cannot
be neglected and leads to existence of an infinite number of photoelectron trajec-
tories. These interference patterns were first observed by photoionizing metastable
(6s[3/2]J=2 ) Xe atoms and also first revealed the presence of direct and indirect tra-
jectories [162, 165]. The observed photoelectron density distributions were in good
agreement with calculations based on the semiclassical theory described above.

4.3 Experimental results and discussion

In this section the aforementioned interference effects will be illustrated by pre-
senting a series of new experimental results, which were obtained by performing
photoionization microscopy experiments on hydrogen atoms. A detailed description
of these experiments is given in Chapter 3 and here only the general concept is
explained.
In the experiments a beam of atomic hydrogen was produced by photodissociating
H2 S gas making use of 213 nm light delivered by a Nd:YAG laser. The atomic beam
was formed by a skimmer separating a source vacuum chamber from a main vacuum
chamber, which contained a velocity map imaging (VMI) spectrometer [114]. The
hydrogen atoms entered the VMI through a hole in the repeller plate. In the VMI
a DC electric field was imposed by applying a voltage difference across the repeller
and extractor electrodes. The hydrogen atoms were excited to a mixture of n = 2
s– and p–states by means of a two–photon transition, making use of 243 nm pulses
from a tunable dye laser. From there, they were ionized using narrowband, tunable
laser light in the range from 365 to 367 nm, obtained from a Fourier-limited, home-
built pulsed dye amplifier. At this step the emphasis was on ionization processes
resulting from the excitation into a continuum of Stark states in the energy region
between the saddle point energy Esp and the field free ionization potential (E = 0).
The polarization of the 243 nm laser was chosen to be perpendicular to the static
electric field (i.e. parallel to the two-dimensional detector), whereas the polarization
of the 365-367 nm laser was parallel or perpendicular to the static electric field. The

68
 4.3 Experimental results and discussion

resulting photoelectrons were projected onto a position sensitive detector comprising

a stack of two microchannel plates, a phosphor screen and a CCD camera. In order
to distinguish individual interference fringes in the recorded photoelectron images,
an electrostatic zoom lens was used [116], which magnified the images by one order
of magnitude without deteriorating their quality.

4.3.1 Semiclassical interpretation of the interference patterns

In the experiments on hydrogen atoms distinct interference patterns were inves-
tigated for different values of the electric field ranging from 200 to 850 V/cm and for
the photoionization laser polarization chosen to be parallel (π–polarized) or perpen-
dicular (σ–polarized) to the field axis. Observed patterns were independent of the
laser polarization, as shown in Fig. 4.3, where two images recorded for the photo-
electrons excited 10 cm−1 above the field-free ionization limit with π–polarized (a)
and σ–polarized (b) laser light are compared. From these R experimental images the
radial probability distributions, defined as Pexp (R) = R P (R, θ)dθ, where P (R, θ)
denotes the intensity distribution in polar coordinates R and θ, are extracted and
compared in Fig. 4.3c. Regardless the laser polarization the same number of fringes
is observed and the interference patterns are similar. The inner part of the image,
i.e. for the image radius 0 < R ¬ 4 mm, is dominated by direct trajectories whereas
in the outer part (4 < R ¬ 5.3 mm) the indirect trajectories are noticeable. The
classical photoelectron trajectories are described in detail in the next section. Note
that the radial distribution for the π–polarization (black line) corresponds to the
photoelectron image acquired for 25000 laser shots whereas in the case of the σ–
polarization (gray line) the image was recorded only for 10000 events. Therefore, in
the former case the signal to noise ratio is better and the interference fringes are
more conspicuous.
All photoelectron images discussed in this chapter resulted from photoexcitation
of hydrogen atoms into the continuum of Stark states. Recorded images showed a
smooth evolution of interference patterns with the increasing excess energy, con-
firming the lack of resonance effects. Therefore, observed oscillatory structures can
be explained exclusively as a result of constructive and deconstructive interferences
between classical photoelectron trajectories. To validate this interpretation exper-
imentally measured radial probability distributions Pexp (R) were compared with
calculated ones, by means of the semiclassical theory developed R in Ref. [161], where
radial probability distributions were defined as Pcalc (R) = R |ψ(M (R, θ))|2 dθ, with
ψ(M (R, θ)) given by Eq. (4.16). The left column in Fig. 4.4 shows four photoelectron
images recorded by making use of σ-polarized laser light and setting the electric field
in the interaction region to 822 V/cm. The signal intensity was normalized per im-
age: red corresponds to the highest intensity whereas black corresponds to the lowest
signal intensity. In the right column the experimental radial probability distributions
(solid lines) are compared with the results of semiclassical calculations (dashed lines)
for a given value of the reduced energy ε = E/|Esp |, where Esp = 175.4 cm−1 . The
calculated distributions were scaled radially to get the best fit with the experimental
result at ε = -0.825. Even though experimentally observed fringes for R < 0.5 mm

69
Chapter 4. Photoionization microscopy

a 1.0 c
5

Normalized radial distribution

0 0.8

-5 0.6
-5 0 5
5
b 0.4
Ry (mm)

0 0.2

-5
0.0
-5 0 5 0 1 2 3 4 5 6
Rx (mm) R (mm)

Figure 4.3: Raw photoelectron images resulting from the excitation of electrons 10 cm−1
above the field-free ionization limit with the laser polarization chosen to be parallel (a) and
perpendicular (b) to the electric field axis. Note that images ware acquired for 25000 and
10000 laser shots for the π–polarized (a) and σ–polarized (b) laser light, respectively. (c)
Normalized radial probability distributions extracted from images shown in (a) - black line,
and in (b) - gray line. Despite the laser polarization the same interference patterns are
observed, which in the inner and outer part of the images corresponds to direct and indirect
trajectories, respectively.

are not very well resolved the agreement between Pcalc (R) and Pexp (R) is satisfy-
ing, validating the interpretation of the observed oscillatory structures as a result
of interferences between the infinite number of classical trajectories by which pho-
toelectrons move towards the detector. Visible singularities in the calculated radial
probability distributions (dashed lines in Fig. 4.4) are due to classical singularities
(caustics) in the used semiclassical approximation. As shown in Ref. [166] incorpo-
ration of the uniform Ariy approximation removes the divergences at caustics and
implementation of the Bessel-function approximation fixes the divergent behavior of
the wave function near R = 0 (apparent in the image for ε = −0.797).

4.3.2 Direct and indirect ionization processes

Fig. 4.5 shows examples of photoelectron images resulting from photoionization of
hydrogen atoms in a DC electric field F = 804 V/cm. By scanning the laser frequency,
atoms were excited into the continuum located in the energy range: Esp ¬ E ¬ 0.
The polarization of the laser was chosen to be parallel to the electric field and each
image was recorded for 10000 laser shots. The photoelectron image presented in the
top left corner corresponds to the case when the excitation energy was equal to the
field-free ionization potential. The subsequent images were recorded by decreasing
the excess energy in steps of approximately 5 cm−1 . The last image, shown in the

70
 4.3 Experimental results and discussion

4 e = -0.882

-4
-4 0 4

4 e = -0.825

-4
-4 0 4

4 e = -0.797

-4
-4 0 4

= -0.737
4 ee=-0.882
Ry (mm)

-4
-4 0 4
Rx (mm) R(mm)

Figure 4.4: The left column shows four photoelectron images recorded for electrons ejected
at reduced energies ε = −0.882, ε = −0.825, ε = −0.797 and ε = −0.737. Hydrogen atoms
were placed in an electric field F = 822 V/cm and the laser polarization was perpendicular
to the electric field. In the right column the experimentally measured radial probability distri-
butions Pexp (R) (solid lines) are compared with the calculated probability densities Pcalc (R)
(dashed lines) showing a good agreement in the number of fringes. Note that singularities
in Pcalc (R) are due to limitations of the used semiclassical theory.

bottom left corner, was acquired for an excitation energy circa 3.7 cm−1 above the
saddle point energy, which for the given electric field was located at energy Esp =
-173.5 cm−1 .
The major observation in Fig. 4.5 is the presence of two distinct concentric struc-

71
Chapter 4. Photoionization microscopy

tures, which are clearly visible in the top three rows. The inner structure can be cor-
related with direct trajectories i.e., with the photoelectrons ejected downfield during
the ionization process (direct ionization) and the outer structure with the photo-
electrons ejected upfield, which interacted with the parent ion and crossed the field
axis at least once (indirect ionization). This is confirmed by the fact that the in-
ner structure becomes visible in photoelectron images recorded at reduced energies:
ε > −0.777. This observation agrees well with the classically predicted threshold en-
ergy, given by Eq. (4.11), above which direct trajectories are allowed. By analyzing
photoelectron images in the bottom two rows one can see that the outer concentric
structure, corresponding to the indirect ionization process, dominates in the excess
energy region close to the saddle point energy. As shown in Fig. 4.2b, when the
reduced energy is smaller than -0.775 Esp the angle β is always smaller than β0
meaning that photoelectrons have to interact with the parent ion in order to go
through a bottleneck in the Coulomb + Stark potential. With increase of the excess
energy the opening in the effective potential becomes larger and when E > Edir
the opening is large enough for photoelectrons to directly escape through it, i.e.
without interaction with the ion. With increase of the photoelectron energy the size
of photoelectron images increases and the radius of the outermost ring, built up by
indirect trajectories, can be calculated with [161]:

!1/2
√ E 2
Rmax = 2L +√ , (4.17)
F F

where L is the distance between the photoionization region and the detector.
Another observation is the appearance of an intense peak in the center of the
photoelectron image corresponding to a radius of R = 0. If the Coulomb potential
is neglected this structure can be built up only with electrons launched at β = 0
or β = 180◦ . However, when the Coulomb force is taken into account there exist
many trajectories with β 6= 0 or β = 180◦ which form this intense and narrow peak
at R = 0 [163]. Since the number of classical trajectories is proportional to sin(β)
the peak at R = 0 is observable for all energies above the saddle point, though, it
becomes especially prominent with the appearance of direct trajectories.

4.3.3 Beating effect

Direct and indirect trajectories differ not only in their origin but they also cre-
ate distinctively different interference structures. The oscillatory pattern resulting
from an indirect ionization process shows smaller spacings between the fringes com-
pared to the structure resulting from direct trajectories (see Fig. 8 in Ref. [161]).
Therefore, in the inner part of a photoelectron image direct and indirect oscillatory
patterns give rise to a beating pattern, which is visible in the photoelectron images
shown in Fig. 4.5. Particularly it is noticeable in the image recorded at ε = -0.287,
which is presented in Fig. 4.6a with the experimental radial probability distribution
(solid line in (b)). At radii close to the center of the image, i.e. (R ¬ 1.5 mm) the

72
 4.3 Experimental results and discussion

 =0  = -0.051  = -0.085  = -0.114  = -0.143

 = -0.203  = -0.229  = -0.258  = -0.287  = -0.316

 = -0.342  = -0.374  = -0.408  = -0.431  = -0.460

 = -0.489  = -0.515  = -0.546  = -0.604  = -0.659

 = -0.719  = -0.746  = -0.777  = -0.806  = -0.834

6  = -0.863  = -0.892  = -0.921  = -0.950  = -0.979

Ry (mm)

-6
-6 Rx0(mm) 6

Figure 4.5: Examples of photoelectron images recorded at an electric field of 804 V/cm for
different values of the reduced energy ε. The photoelectron energy decreases from the field-
free ionization potential (ε = 0 – top left image) to the energy close to the saddle point
energy (ε = -0.979 – bottom right image).

interference is enhanced and distinct rings can be distinguished whereas for larger
R the two contributions, direct and indirect, partially cancel each other leading
to a structureless area. This situation is very well reproduced by the semiclassical

73
Chapter 4. Photoionization microscopy

Normalized radial distribution

6 a 1.0 b
0.8
Ry (mm)

0.6
0
0.4

0.2
-6 0.0
-6 0 6 0 1 2 3 4 5 6
Rx (mm) R (mm)

Figure 4.6: (a) Photoelectron image recorded at the scaled energy ε = −0.287 with a con-
spicuous beating patter, in the inner part of the image, resulting from constructive and
deconstructive interferences between direct and indirect photoelectron trajectories. (b) Ex-
perimental radial distribution (solid line) indicating a presence of individual ring for R ¬
1.5 mm and a blurred area for larger R, what agrees well with the semiclassical calculations
(dashed line).

calculations (dashed line in (b)), which for R ¬1.5 mm demonstrate two groups of
fringes: intense ones, which are interleaved with small fringes and consequently dom-
inate the observed interference pattern. For larger R the intensities in both groups
of fringes become comparable resulting in a blurred region, where individual rings
are indistinguishable.
To investigate the beating effect in detail an experiment was performed where
hydrogen atoms were placed in an electric field F = 796 V/cm and irradiated with
π–polarized laser light. Photoelectron images were recorded in the energy range
from -80 to 0 cm−1 by changing the excess energy in steps of only 2 cm−1 and
keeping the exposure time at 10000 laser shots per image. The experimental results
are presented in the form of a contourplot shown in Fig. 4.7a, where measured
intensity distributions for a given excess energy are plotted as a function of the image
radius R. For a better visualization of the beating effect the individual measurements
were normalized to the total intensity of the image. Two intense branches, which at
energy E = -80 cm−1 appear at R = 1.1 mm and Rmax = 4 mm and are marked
with white dashed lines, correspond to the outer rings created by direct and indirect
trajectories, respectively. The most striking observation in this plot is the presence of
a sequence of interference structures in the inner part of the photoelectron images, i.e.
below the first outer ring where both direct and indirect trajectories interfere. This
beating pattern is particularly conspicuous in the energy range from -80 to -40 cm−1 ,
where at least five beating branches can be distinguished with an average spacing
between the fringes equal to 4 cm−1 , as indicated with a white arrow. These beating
structures are very well reproduced by semiclassical calculations shown in Fig. 4.7b.
The calculations reveal more branches of beating patterns, which are present in the
energy range where direct and indirect trajectories interfere, i.e. Edir < E. Most
likely limitations in the spatial resolution of the recorded images is the reason why

74
 4.3 Experimental results and discussion

a b
5 5 10
0.4

4 4 8
0.3
3 3 6
R (mm)

R (mm)
0.2
2 2 4

0.1
1 1 2

0 0 0 0
-80 -70 -60 -50 -40 -30 -20 -10 0 -80 -70 -60 -50 -40 -30 -20 -10 0
Energy (cm-1) Energy (cm-1)

Figure 4.7: Measured (a) and calculated (b) contourplots of the radial intensity distribu-
tions in photoelectron images, which resulted from ionization of hydrogen atoms in a F =
796 V/cm DC electric field with π–polarized laser light. White dashed lines in (a) indi-
cates intense branches corresponding to the outermost rings created by direct and indirect
trajectories. In both plots the beating patterns created by constructive and deconstructive
interference between direct and indirect photoelectron trajectories are visible. The measured
and the calculated spacing between the beating fringes are 4 cm−1 and 4.3 cm−1 , respectively,
as marked in white arrows.

only a limited number of beating structures are experimentally observed in the

energy range close to Edir . The calculated spatial frequency of the beating fringes
is 4.3 cm−1 , which agrees very well with the experimentally determined spacing
between the patterns. However, the calculations are not free from computational
problems, which lead to the appearance an artifact at E = -39 cm−1 .
Furthermore, the recorded data show a periodic modulation in the intensity of
the outermost ring, which is particularly conspicuous in the energy range between -
40 and 0 cm−1 . This effect is emphasized in Fig. 4.8 by presenting four photoelectron
images (a-d), which were recorded at energies close to each other, and where one can
see that the outermost ring is visible only in two images recorded at E = -2 cm−1 and
E = -6 cm−1 . The measured intensities of the outermost rings for different excess
energies were normalized to the total intensity per image and are shown in (e) as
black squares. To exclude that the observed oscillations are caused by fluctuations
in the signal intensity, the total intensity of photoelectron images, normalized to the
intensity of the image recorded at E = -32 cm−1 , is plotted in (f). One can see that
the periodic modulation observed in (e), with an average energy spacing between
extrema ∆E = 5.3 cm−1 , does not show any correlations with fluctuations in the
total signal intensity shown in (f). The semiclassical calculations, partially shown in
Fig. 4.7b, reveal a periodic oscillation in the intensity of the outermost ring with an
average spacing between two extrema equal to ∆E = 5.2 cm−1 , what agrees very
well with the experimentally measured value.
At the moment when this thesis is written many semiclassical calculations are
running in order to understand the origin of this periodic modulation. One possible

75
Chapter 4. Photoionization microscopy

explaination is that the observed intensity modulation may be created as a result

of the interaction between the indirect trajectories with the so-called closed orbits
[167–169]. As explained in Section 3.1, when the ejection angle is smaller than the
critical angle βc the electron stays forever in the vicinity of the atom. These types
of photoelectron trajectories are called bound orbits, in contrast to open orbits, i.e.
trajectories along which the electron escapes from the atom. Direct and indirect
trajectories are both open orbits. There exists a special set of trajectories belonging
to the bound orbits – so-called closed orbits, which are related to the occurrence
of blue-shifted Stark states [166, 168]. Hence, the classical picture would be that
along these trajectories the electron first goes out from the atom and then it is re-
turned back by the electric field. When the electron propagates back towards the
atom it starts interfering with outgoing waves and produces oscillations, which can
be observed in the absorption spectrum [81, 156, 170]. For E > 0 there is only one
closed orbit, which is along the electric field, thus oscillations will be visible only
for laser light polarized along the electric field axis (π–polarized). One experimental
observation which confirms this hypothesis of interference between indirect trajec-
tories and closed orbits is the fact the modulation in the intensity of the outer ring
was not observed for σ–polarized laser light. However, this problem requires further
theoretical and experimental studies before any conclusions are drawn.

4.4 Resonance effects in photoionization microscopy

In the discussed energy range i.e., Esp ¬ E ¬ 0, an electron may be resonantly
excited into a quasi-bound Stark state. In this situation the electron motion (given
by Eq. (4.12)) is always bound along the ξ coordinate and classically bound along
the η coordinate, since its energy lies below the top of the V (η) potential. However,
from a quantum mechanical point of view the electron has a nonzero probability
to tunnel through the V (η) potential barrier and to escape from the atom. Such a
quantum effect may be incorporated into the discussed semiclassical approximation
by the Bohr-Sommerfeld quantization condition [50]:

Z ξmax h 1 i
p̃ξ dξ = π n1 + |m| + 1 , (4.18)
ξ0 2

which relates the separation constant Z1 with the resonance energy E. With such
a condition for a given value of E the separation constant Z2 can be determined,
and thus by integrating Eq. (4.12b) the outgoing wave function can be expressed by
[51]:

1 h 1 i
χ2 (η) ≈ Cη −1/4 exp i F 1/2 η 3/2 + δ , (4.19)
2i 3
with a constant C and some phase δ . When a position-sensitive detector is placed
far enough from the photoionization region the η = const coordinate crossing the

76
 4.4 Resonance effects in photoionization microscopy

E = -2 cm
-1
a E = -4 cm-1 b e

E = -6 cm-1 c E = -8 cm-1 d f

Figure 4.8: (a-d) Photoelectron images recorded in the vicinity of the field free ionization
potential, which show the change in the intensity of the outermost ring over a short interval
of the excess energy. (e) Periodic oscillation in the intensity of the outermost ring, which is
particularly prominent in the energy range from 0 to -40 cm−1 . Note that here the intensity
of the outermost ring was normalized to the total signal in the image. (f ) The total signal
intensity of a series of recorded photoelectron images normalized to the intensity of the
image recorded at E = -32 cm−1 .

plane of the detector can be approximated by a plane surface, which is perpendicu-

lar to the z-axis. The component of the photocurrent density perpendicular to the
surface η = const is denoted by:

 dψ ∗ dψ 
jη ≈ i ψ − ψ∗ , (4.20)
dη dη

where ψ represents the final state electron wave function given by Eq. (2.31).
Combination of Eq. (2.31), (4.19) and (4.16) gives the photoelectron current density
measured on the detector:

1 2 1/2 1
jη = |C| F |χ1 (ξ)|2 . (4.21)
4 ξ
Thus, when the hydrogen atom is excited into a Stark state prior to being ionized,
the observed interference pattern represents the squared modulus of the electronic
wave function χ1 (ξ), which describes the electron motion along the ξ coordinate.

77
Chapter 4. Photoionization microscopy

The radius of the resonant interference pattern is given by:

" !1/2 #1/2

c
√ E E2 4Z1
Rmac = 2L + 2
+ , (4.22)
F F F

and can be compared with the expression given by Eq. (4.17), which describes the
classically allowed region along the ξ coordinate for the non-resonant case. One sees
c
that Rmax > Rmax meaning that the interference pattern resulting from resonant
excitation into a Stark state extends radially further compared to the non-resonant
case. Moreover,
√ the positions of the minima in the photoelectron image are given by
Rn = 2Lξn , where ξn is the nth zero of the electronic wave function χ1 (ξ). Hence,
the experimentally observed minima are a direct projection of the nodal structure
of the electronic wave function along the ξ coordinate. Therefore, photoionization
microscopy is a technique, which allows direct observation of the electronic wave
function on a microscopic scale.

4.5 Resonance effects in experiments and theory

to date
In the first experimental implementation of the photoionization microscopy tech-
nique, Xe atoms were used, which were photoionized in a DC electric field of 320 V/cm
with one UV photon, revealing the presence of direct and indirect ionization pro-
cesses [165]. Photoelectron images recorded for resonant excitation to Stark states
showed interference patterns, for which the number of dark fringes increased smoothly
with the excess energy, hence showing no sign of the transverse component of the elec-
tronic wave function. The results were qualitatively explained within the framework
of the semiclassical approximation and it was shown that the continuum contribu-
tion dominated the interference structure in Xe atoms [161]. In a following paper
[162], which treated the experimental results with greater detail, it was suggested
that information regarding the localization of the wave function could be extracted
from relative intensities between direct and indirect contributions.
In the last few years new semiclassical and quantum mechanical theories on
photoionization microscopy were developed. The already mentioned work of Zhao
and Delos [166] extended the semiclassical approximation of Bordas [161] by means
of Airy-function and Bessel-function approximations and Maslov indices. By using
the Bohr-Sommerfeld quantization condition, the tunneling effect into the classically
forbidden regions was included. To analyze effects of Stark resonances on spatial
interference patterns the authors developed a quantum mechanical theory [171].
In their
√ approach mixed parabolic and semiparabolic coordinates were used, i.e.
ξ = r + z and η = r − z, and a new expression for the outgoing wave function was
derived. By numerically solving the Schrödinger equation for H atoms in an electric
field F = 5714 V/cm with energies near and at a resonance, the authors calculated
the probability densities shown in Fig. 4.9. The image in the middle corresponds to

78
 4.5 Resonance effects in experiments and theory to date

Figure 4.9: Calculated probability densities for H atoms places in an electric field F = 5714
V/cm and excited with π-polarized laser light to three energies: E = -160 and -164 cm−1
correspond to non-resonant transitions and show a moderate change in the shape of the
interference patterns, whereas E = -162.36 cm−1 corresponds to a resonant excitation into
a Stark state. In the latter case the shape and the intensity of the interference pattern
changes dramatically and the number of dark fringes equals the number of nodes of the
electronic wave function χ1 (ξ) [171].

resonant excitation into a narrow Stark state (n, n1 , n2 , m)=(21, 17, 3, 0) and shows
that the total probability distribution is one order of magnitude higher compared to
the other non-resonant images. Additionally, the shape of the interference pattern is
different – it shows more rings and the number of dark fringes equals the parabolic
quantum number n1 , which is the number of nodes of the electronic wave function
χ1 (ξ). On top of that the size of the resonant image is larger than the other two,
although this is hardly visible in this figure. The authors showed that the (21, 17, 3, 0)
Stark state contributes to the total densities at the other two energies E = -160 and
-164 cm−1 , however, its contribution is very small compared to the situation at E
= -162.36 cm−1 . Thus the differences in the spatial distributions of the probability
densities at E = -162.36 cm−1 are attributed to the presence of the Stark state and
are associated with tunneling through the V (η) potential barrier. Experimentally
the resonance effect was observed in hydrogen [52], helium and lithium atoms [172],
with first two cases being further discussed in this thesis (Chapters 5 & 6).

79
 Hydrogen atoms under

5 magnification: direct
observation of the nodal
structure of Stark states

To describe the microscopic properties of matter, quantum mechanics uses wave

functions, whose structure and time dependence is governed by the Schrödinger
equation. In atoms the charge distributions described by the wave function are rarely
observed. The hydrogen atom is unique, since it only has one electron and, in a DC
electric field, the Stark Hamiltonian is exactly separable in terms of parabolic coor-
dinates (η, ξ, ϕ). As a result, the microscopic wave function along the ξ coordinate
that exists in the vicinity of the atom, and the projection of the continuum wave
function measured at a macroscopic distance share the same nodal structure. In
this chapter photoionization microscopy experiments where this nodal structure is
directly observed are reported. The experiments provide a validation of theoretical
predictions that have been made over the last three decades.

5.1 Introduction
The development of quantum mechanics in the early part of the last century has
had a profound influence on the way that scientists understand the world. Central
to quantum mechanics is the concept of a wave function that satisfies the time-
dependent Schrödinger equation [21]. According to the Copenhagen interpretation,
the wave function describes the probability of observing the outcome of measure-
ments on a quantum mechanical system, such as measurements of the energy or
the position or momenta of constituents [173]. The Copenhagen interpretation thus
allows reconciling the occurrence of nonclassical phenomena on the nano-scale with
manifestations and observations made on the macro-scale, which correspond to view-
ing one of a number of possible realizations allowed for by the wave function.
Despite the overwhelming impact on modern electronics and photonics, under-
standing quantum mechanics and the many possibilities that it describes continues
to be intellectually challenging, and has motivated numerous experiments that illus-
trate the intriguing predictions contained in the theory [174]. Using ultrafast lasers,
Rydberg wave packet experiments have been performed illustrating how coherent su-
perpositions of quantum mechanical stationary states describe electrons that move
on periodic orbits around nuclei [175]. The wave function of each of these electronic
stationary states is a standing wave, with a nodal pattern that reflects the quantum

81
Chapter 5. Hydrogen atoms under magnification

numbers of the state. Mapping of atomic and molecular momentum wave functions
has been extensively explored by means of (e, 2e) spectroscopy, using coincident
detection of the momentum of both an ejected and a scattered electron to retrieve
the momentum distribution of the former prior to ionization [28]. In the spirit of
scanning tunneling methods, orbital tomography based on high harmonic genera-
tion was developed as a method allowing the determination of atomic and molecular
orbitals [38, 39]. In this chapter experiments are presented where the nodal struc-
ture of electronic wave functions of hydrogen atoms is measured, making use of a
photoionization microscopy experiment.
The hydrogen is a unique atom, since it only has one electron and, in a DC electric
field, the Stark Hamiltonian is exactly separable in terms of parabolic coordinates.
For this reason, an experimental method was proposed about thirty years ago, when
it was suggested that experiments ought to be performed projecting low-energy
photoelectrons resulting from the ionization of hydrogen atoms onto a position-
sensitive two-dimensional detector placed perpendicularly to the static electric field,
thereby allowing the experimental measurement of interference patterns directly
reflecting the nodal structure of the quasi-bound atomic wave function [49, 51, 158].

5.2 Resonance phenomenon in photoionization mi-

croscopy
In a static electric field F the wave function of atomic hydrogen can be separated
in terms of the parabolic coordinates (η, ξ, ϕ) and consequently, the wave function
may be written as a product of functions χ1 (ξ) and χ2 (η) that separately describe
the dependence along ξ and η , i.e. ψ(η, ξ, ϕ) = √ 1 χ1 (ξ)χ2 (η)eimϕ . The functions
2πξη
χ1 (ξ) and χ2 (η) satisfy the ordinary differential equations:

!
d2 χ1 E Z1 m2 − 1 1
+ + − − F ξ χ1 = 0, (5.1a)
dξ 2 2 ξ 4ξ 2 4
!
d2 χ2 E Z2 m2 − 1 1
+ + − + F η χ2 = 0. (5.1b)
dη 2 2 η 4η 2 4

In these expressions m is the magnetic quantum number and Z1 and Z2 are

separation constants subject to the condition Z1 + Z2 = 1. The parabolic quantum
numbers n1 and n2 are related to the principal quantum number via n = n1 +
n2 + |m| + 1. Functions χ1 (ξ) and χ2 (η) have n1 and n2 nodes along the ξ and
η coordinates, respectively. Fig. 5.1a shows the potential energy landscape for the
hydrogen atom in an 808 V/cm electric field. The electron motion is always bound in
the ξ coordinate whereas the motion along the η coordinate depends on the energy
available for its motion. Figures in the bottom panel illustrate the potential energy
curves:

82
 5.2 Resonance phenomenon in photoionization microscopy

=5000 a.u. =3000 a.u. =1000 a.u.

2000
a

E=-163.3 cm-1
1000

E=-172.8 cm-1
y (a.u.)

=100 a.u.
-1000

=300 a.u.

-2000
=500 a.u.

-4000 -2000 0 2000

z (a.u)
z (a.u.)

0.000

b 0.002 c
0.000

-0.002

-0.002

0 10000 20000 0 10000 20000

 (a.u.)  (a.u.)

Figure 5.1: (a) Potential energy landscape and relevant coordinate system for hydrogen
atom photoionization microscopy in an 808 V/cm electric field. The hydrogen atom sits at
the origin of the (z, y) coordinate system and the electric field is along the z-axis (a.u. =
atomic units). The boundary between the classically allowed and the classically forbidden
region is plotted (solid lines) at the excitation energies of the four measurements that are
shown in Fig. 5.3, i.e. ranging from E = -172.8 cm−1 to E = -163.3 cm−1 (thick outer
solid line). Close to the saddle-point, the electron can only escape through a very narrow gap
in the Coulomb + dc field potential. The parabolic coordinates η = r − z and ξ = r + z are
illustrated by plotting a series of contours at constant η (dashed lines) and ξ (dotted lines).
The electron motion is always bound in the ξ coordinate whereas the motion along the η
coordinate depends on the energy available for the η-motion. (b) and (c) Potential energy
2 2
curves V (η) = − Z2η2 + m8η−12 − F8η and V (ξ) = − Z2ξ1 + m8ξ−12 + F8ξ , describing the motion
along the η and ξ coordinates [65], where Z1 = (n1 + |m|+1
2
)/n and Z2 = (n2 + |m|+1
2
)/n.
V (η) and V (ξ) are shown for the (n, n1 , n2 , m)=(30, 3, 26, 0) quasi-bound state at E =
-163.3 cm−1 .

83
Chapter 5. Hydrogen atoms under magnification

Z2 m2 − 1 F η
V (η) = − + − , (5.2a)
2η 8η 2 8
Z1 m2 − 1 F ξ
V (ξ) = − + + , (5.2b)
2ξ 8ξ 2 8
defining the motion along the η and ξ coordinates, respectively. Ref. [49] con-
tained a remarkable prediction for the special case where the atomic hydrogen pho-
toionization involves the excitation of quasi-bound Stark states. In this case, where
both n1 and n2 are good quantum numbers and the electron tunnels through the
barrier in the potential energy curve associated with the η coordinate, the mea-
surements should show a total of n1 dark fringes, directly revealing an important
signature of the Stark state involved. However, to date, this experiment was never
performed.
Motivated by the theoretical predictions for the configuration of the above-
mentioned ”photoionization microscope” [49], a photodetachment microscope for
negative ions was first constructed by Blondel et al. [164]. Their experiments clearly
revealed interferences between the photoelectrons en route to the detector, in agree-
ment with simple semiclassical considerations by Du [176]. In photodetachment,
the photoelectrons follow one of two possible parabolic trajectories to the two-
dimensional detector. By contrast, in a photoionization experiment that starts from
a neutral sample, the photoelectrons move in a combined static electric + Coulomb
field, significantly complicating the dynamics and leading to the existence of an in-
finite number of trajectories that the photoelectron can follow to the detector [163].
Given the considerable challenges connected to the experimental use of atomic
hydrogen, first attempts to implement photoionization microscopy were performed
on Xe atoms by Nicole et al. [165]. Observed interference rings were interpreted in the
framework of a semiclassical treatment [161], excluding the possibility of resonant
excitation of a Stark state. The experiments themselves were performed both with
and without the resonant excitation of quasi-bound Stark states, and no significant
differences were observed [162]. More recently, photoionization microscopy experi-
ments were performed for Li atoms [172], revealing first indications of differences in
the radial distributions for on- and off-resonance excitation. This work provides the
motivation for the hydrogen experiments reported in this chapter, where results for
both resonant and non-resonant ionization are presented, and where the long-time
predictions of Demkov et al. [49, 158] is convincingly validated.

5.3 Experimental results and discussion

In the experiments an atomic hydrogen beam was formed by collimating hydrogen
atoms resulting from the photodissociation of H2 S gas in a first vacuum chamber (see
Fig. 5.2 and Chapter 3 for details). The hydrogen atoms were ionized in the active
region of a velocity map imaging (VMI) spectrometer [114], where a DC electric
field ranging between 250 and 850 V/cm was applied. The atoms were resonantly

84
 5.3 Experimental results and discussion

e
c
b
a

243 nm

213 nm 365-367 nm

Figure 5.2: Schematic overview of the experiment. An atomic hydrogen beam was formed
by photodissociating H2 S and placing a 3 mm aperture (a) 65 mm downstream. In the
active region of a velocity map imaging (VMI) spectrometer, the ground state hydrogen
atoms were first excited to a mixture of n = 2 s− and p−states by a two-photon transition
using a pulsed 243 nm laser. Next, they were ionized by a Fourier-limited, tunable (365-
367 nm), UV laser. By applying a voltage difference across the repeller (b) and extractor
(c) electrodes, the photoelectrons were accelerated towards a two-dimensional detector (d),
consisting of a set of microchannel plates (MCPs), a phosphor screen and a CCD camera.
En route to the MCP detector, the photoelectrons passed through a three-element Einzel
lens (e), allowing an increase of the diameter of the recorded image by about one order of
magnitude.

excited to a mixture of n = 2 s– and p–states by a two-photon transition (λlaser =

243 nm) and were ionized using narrowband, tunable laser pulses ( λlaser = 365-367
nm, τlaser = 8 ns) from a Fourier-limited, home-built pulsed dye amplifier [130].
The polarization of the 365-367 nm laser was chose to be parallel or perpendicular
the static electric field (i.e. π–polarized or σ–polarized) to populate m = 0 or m =
1 Stark states, respectively. A dual micro-channel plate (MCP) detector followed
by a phosphor screen and a CCD camera was used to record the photoelectrons.
An electrostatic zoom lens magnified the images by about one order of magnitude
compared to the size that would have been measured without this lens [116].

5.3.1 m = 0 Stark states in high electric fields

In one set of the performed experiments the DC electric field in the interaction
region was set to 808 V/cm and the photoionization laser was polarized along the
electric field axis ensuring excitation of the m = 0 Stark states. The main results of
the experiments are shown in Fig. 5.3. In this figure calculated and experimental re-
sults are shown for four experiments, where the hydrogen atoms were excited to the

85
Chapter 5. Hydrogen atoms under magnification

2 (30,0,29,0)
0.1 10
0.0 0
5
-0.1
-2 0
0.0 0.4 -2 0 2
2 (30,1,28,0) 10
0.1

0.0 0 5

-0.1
-2
Ry (mm)

0
y (mm)

0.0 0.4 -2 0 2
2 (30,2,27,0)
0.1 6
0.0 0 4
2
-0.1
-2 0
0.0 0.4 -2 0 2
2 (30,3,26,0) 5
0.1

0.0 0 3

-0.1 1
-2
0.0 0.4 -2 0 2
z (mm) Rx (mm)

Figure 5.3: Experimental observation of the transverse nodal structure of four atomic
hydrogen Stark states. The images in the middle show experimental measurements for
(n, n1 , n2 , m)= (30, 0, 29, 0), (30, 1, 28, 0), (30, 2, 27, 0) and (30, 3, 26, 0). In-
terference patterns are clearly observed where the number of nodes corresponds to the value
of n1 . The results may be compared to TDSE calculations shown to the left (for details
see text), revealing that the experimentally observed nodal structures originate from the
transverse nodal structure of the initial state that is formed upon laser excitation. A com-
parison of the experimentally measured (solid lines) and calculated radial (dashed lines)
probability distributions P(R) is shown to the right of the experimental results. In order to
make this comparison, the computational results were Rscaled to the macroscopic dimensions
of the experiment. Please note that since, P (R) = P (R, α)Rdα, the radial probability
distributions,P (R) have a zero at R = 0, even if the two-dimensional images P (R, α) do
not.

(n, n1 , n2 , m)= (30, 0, 29, 0), (30, 1, 28, 0), (30, 2, 27, 0) and (30, 3, 26, 0) quasi-
bound Stark states. As indicated in Fig. 5.3, the states lie at energies of -172.82 cm−1 ,
-169.67 cm−1 , -166.45 cm−1 and -163.30 cm−1 with respect to the field-free ion-
ization limit, i.e. just above the saddle-point in the Coulomb + DC field poten-

86
 5.3 Experimental results and discussion

tial, which lies at -174.00 cm−1 . According to Eq. (2.35), the validity of which was
checked experimentally [177], the ionization rate of these states covered a range from
Γ=2.2×1010 s−1 to Γ=7.25×109 s−1 , which (using δE(cm−1 ) = 5.3 × 10−12 Γ(s−1 ))
implies line widths comparable to the effective bandwidth of our excitation laser
(0.005 cm−1 ). These states could readily be identified in wavelength scans, since the
ionization is complete before the hydrogen atoms leave the interaction region of the
VMI [75]. By contrast, in the same energy range Stark states in the n = 31 manifold
(Γ > 1012 s−1 ) lead to very broad resonances, while states in the n = 29 manifold
((Γ < 106 s−1 ) undergo insufficient ionization before the atoms fly out of the in-
teraction region. Total ionization spectra as a function of excitation energy in the
given static electric field were successfully reproduced by means of the semiclassical
Stark theory of Harmin (see [89, 107] and Section 2.5.2). The parabolic quantum
number n1 was identified by comparing the experimental spectra with separate the-
oretical excitation curves for each n1 -channel. Given the value of n1 , the value of n2
was subsequently determined by applying the WKB quantization rule along the η
coordinate.
The main result of the experiment, which is directly visible in the four images
shown in the middle of Fig. 5.3, is the observation of an interference pattern with a
number of dark fringes corresponding to the value of n1 . This observation validates
the prediction by Demkov and co-workers [8, 9] and illustrates that photoionization
microscopy can be used to visualize the nodal structure of χ1 (ξ) for quasi-bound
Stark states of the hydrogen atom. A rationalization for this behavior can be found
in the calculations shown to the left of the experimental images. Here, results of
propagating the time-dependent Schrödinger equation (TDSE) following excitation
of the hydrogen atom at the energies used in the experiments are shown (see [109]
and Section 2.5.3). The graphs represent time-integrated
√ plots of the two-dimensional
electron density ρ|ψ(~r, t))|2 , where ρ = r2 − z 2 , evaluated from the time of exci-
tation (t = 0) to a time delay of 600 ps. The nodal structure that is observable at
a large distance from the proton (here: 0.4 µm) clearly has its origin in the trans-
verse nodal structure of the initial state that is formed upon laser excitation. The
calculation was also carried beyond a distance of 0.4 µm and showed no significant
changes.
A direct comparison of the experimental (solid line) and calculated (dashed line)
results that is obtained by scaling the radial coordinate in the calculation, is shown
to the right of the experimental results.
R Here a comparison of the measured radial
probability distribution Pexp (R) = P (R, α)Rdα (where P (R, α) represents the
intensity distribution in the experimental image in polar coordinates R and α), and
the calculated radial probability distribution Pcalc = R|ψ(R, t)|2 is given, showing
very satisfactory agreement and validating the assignment of the number of dark
fringes as the parabolic quantum n1 . Observed differences between the experimental
and calculated results may be due to differences in the experimental conditions and
the assumptions made in the calculations (where the Stark states were excited using
a 250 ps excitation pulse), imperfections in the experimental images, and possible
smearing effects due to the finite lifetime of the excited Stark states.

87
Chapter 5. Hydrogen atoms under magnification

c (38,0,36,1) (38,1,35,1)

(38,2,34,1) (38,3,33,1)

b
2 (38,4,32,1) (38,5,31,1)

Ry (mm)
0

-2
-2 0 2
Rx (mm)

Figure 5.4: (a) Photoionization spectrum for hydrogen atoms ionized in a DC electric field of
295.5 V/cm calculated, by means of Harmin’s WKB-QD theory, where resonances associated
with quasi-bound Stark states are indicated by arrows. For this plot a different notation of
the Stark states is used, i.e. the numbers in the brackets correspond to the parabolic quantum
numbers (n1 , n2 ) and colors reflect different manifolds: n = 38 (black), n = 39 (red) and
n = 40 (blue). (b) An experimental photoionization spectrum in the same energy range as
in (a), with prominent resonances originating from photoexcitation of hydrogen atoms into
the n = 38 Stark manifold. (c) Photoelectron images recorded for quasi-bound Stark states
(n, n1 , n2 , m)=(38, 0, 36, 1), (38, 1, 35, 1), (38, 2, 34, 1), (38, 3, 33, 1), (38, 4, 32, 1)
and (38, 5, 31, 1), with the number of observed dark fringes equal to the parabolic quantum
number n1 , i.e. the number of nodes of the electronic wave function χ1 (ξ), meaning that
the transverse component of the wave function along the ξ coordinate is directly projected
on the detector in the experiments.

5.3.2 m = 1 Stark states in low electric fields

The transverse component of the electronic wave function χ1 (ξ) was also observed
at lower electric fields as well as in Stark manifolds containing m = 1 states, which
were populated by making use of σ–polarized laser light, i.e. polarized perpendicular
to the electric field axis. The main results of the experiments, where the hydrogen
atoms were ionized in a DC electric field of 295.5 V/cm, are shown in Fig. 5.4. In
the top panel (a) a calculated, by means of the WKB-QD theory (see Section 2.5.2),
photoionization spectrum for the given electric field strength indicates the presence
of three overlapping Stark manifolds, i.e. n = 38, n = 39 and n = 40 for excess
energies just above the saddle point (Esp = −105.17 cm−1 ). It should be noted that
due to a limited energy gap between the resonances a different notation for the Stark
states is used in this plot, namely numbers in brackets correspond to the parabolic
quantum numbers (n1 , n2 ). The experimental photoionization spectrum (b), mea-

88
 5.3 Experimental results and discussion

sured by scanning the photoexcitation laser frequency, demonstrates prominent res-

onances in the vicinity of the saddle point energy. However, their number suggests
that only one Stark manifold was populated during the experiments due to the afore-
mentioned limitation of the experimental setup. Calculated ionization rates for the
reddest states in adjacent manifolds have values of Γ=2.5×107 s−1 , Γ=1.3×1011 s−1
and Γ=9.9×1012 s−1 for the (n, n1 , n2 , m)=(38, 0, 36, 1), (39, 0, 37, 1) and
(40, 0, 38, 1) state, respectively, indicating that the observed resonances originated
from the excitation of the hydrogen atoms into the n = 38 Stark manifold. Therefore,
for the resonances lying at energies of -104.55 cm−1 , -103.02 cm−1 , -101.46 cm−1 ,
-99.95 cm−1 , -98.51 cm−1 and -96.98 cm−1 the following quasi-bound Stark states
are assigned: (38, 0, 36, 1), (38, 1, 35, 1), (38, 2, 34, 1), (38, 3, 33, 1), (38, 4, 32, 1)
and (38, 5, 31, 1), respectively. The photoelectron images recorded for these states
are shown in Fig. 5.4c. The main observation here is that the agreement between the
number of observed dark fringes and the parabolic quantum number n1 is satisfying,
which validates the proper assignment of the quantum numbers.
To confirm that the measured oscillatory patterns share the same nodal structure
as the electronic wave function χ1 (ξ) TDSE calculations were performed. The Stark
states measured at F = 295.5 V/cm have lifetimes three orders of magnitude longer
compared to the Stark states excited at 808 V/cm, hence, the propagation time of the
wave packets needed to be extended from 600 ps to 10 ns to ensure a sufficient decay
of these states in the calculations. Additionally, the increased stability of the Stark
states results in narrower widths of the resonances and consequently, the excitation
pulse in the calculations must be long enough to resolve such spectral lines. In Fig.
5.5b the calculated radial distributions using a pulse duration of 250 ps (red dotted
line) and 4000 ps (blue dashed line) are compared with the experimentally measured
radial distribution (black solid line) of the (38, 3, 33, 1) state, which is shown in
the upper panel (a). When the excitation pulse is too short the resonant state is
populated along with a fraction of the continuum and as a result the calculated
radial distribution contains both resonant and non-resonant contributions, which
lead to the appearance of overlapping fringes as shown in the calculations using
250 ps pulses. For a pulse duration of 4000 ps only the resonant contribution is
calculated providing a good agreement between the experimental and calculated
radial probability distributions. The number of observed dark fringes is identical to
n1 , which confirms that the transverse component of the χ1 (ξ) function is directly
observed also in the experiments employing σ–polarized laser light and low electric
fields.
As the photoelectron image size scales with F −1/2 (see Eq. (4.17) and (4.22))
it is advantageous to use lower electric fields, which ensure larger images and ipso
facto higher spatial resolution. Consequently, the electrostatic lens can be operated at
lower voltages, which significantly reduces the effects of finite aberrations in the zoom
lens and therefore the recorded images appeared to be sharper. This improvement in
the image quality is noticeable by comparing the photoelectron images recorded at F
= 808 V/cm (Fig. 5.3) with the ones obtained at F = 295.5 V/cm (Fig. 5.4). However,
the Stark splitting is linearly proportional to the field strength, hence the electric
field cannot be set too low as individual Stark states need to be resolved with the laser

89
Chapter 5. Hydrogen atoms under magnification

a (38,3,33,1)

-1.5 -1 -0.5 0 0.5 1 1.5

Rx (mm)
Normalized radial distribution

1.0 b
0.8

0.6

0.4

0.2

0.0
-1.5 -1 -0.5 0 0.5 1 1.5
R (mm)

Figure 5.5: (a) 3D representation of the photoelectron image measured for the (38, 3, 33, 1)
state obtained by smoothing the raw data with a Legendre fit and plotting the signal intensity
as the z-axis. (b) Direct comparison between the measured (black solid line) and calculated
(dashed and dotted lines) radial probability distributions for the (38, 3, 33, 1) Stark state.
TDSE calculations shown here used a 250 ps (red dotted line) and 4000 ps (blue dashed
line) excitation pulse to emphasize that the long lifetime of the Stark state (τ = 0.23 µs)
results in a narrow resonance that can be excited along with a fraction of the continuum if
the excitation pulse is too short. As a result the calculated radial probability contains both
resonant and non-resonant contributions that conceal the wave function of the Stark state
and the agreement between the calculation and experiment is diminished. Note that for 250
ps two fringes at R = 0.5 mm overlap whereas with the excitation pulse of 4000 ps only the
Stark state is populated in the calculations and all fringes are resolved.

bandwidth, which in the experiments was 0.005 cm−1 . Furthermore, with a decrease
of the electric field the saddle point energy increases and experimentally available
quasi-bound states belong to manifolds with high principal quantum numbers, i.e.
are located at energies where the density of Stark states is higher and many adjacent
manifolds overlap. The latter effect may be troublesome in the calculations and in
experiments making use of broad-band laser light.

90
 5.3 Experimental results and discussion

3 3
2 a 2 b 3 2 c

Ry (mm)
2 2
2
0 0 0
1 1 1

-2 (30,2,27,0)
0 -2 0 -2 0
-2 0 2 -2 0 2 -2 0 2
Rx (mm)
d

(a)
(b)
(c)

Figure 5.6: Evidence for resonant ionization by tunneling through the Coulomb + DC elec-
tric field potential. A comparison is shown between a measurement carried out for the (n,
n1 , n2 , m) = (30, 2, 27, 0) resonance (b) and two non-resonant measurements performed
1.8 cm−1 below (a) and 1.1 cm−1 above (c) this resonance. The normalized radial distri-
bution of the resonant measurement extends significantly further outwards than the non-
resonant measurements (d). The inset in (d) shows a comparison of the radial extension of
the experimental images, defined as the position of the outer maximum (color triangles) and
the theoretical radial extension Rmax (blue line) according to the classical formula (exclud-
ing tunneling contributions) given by Eq. (4.17). The experimental and theoretical radial
extensions were matched for the measurement at E = -165.37 cm−1 .

5.3.3 Differences between resonant and non-resonant excita-

tions
A striking observation in the experiments is the pronounced difference between
images recorded following resonant excitation of a quasi-bound Stark state and im-
ages recorded following non-resonant excitation of the ionization continuum. This
is illustrated in Fig. 5.6, where a comparison is shown between the image for the
(n, n1 , n2 , m)=(30, 2, 27, 0) state and two non-resonant images recorded 1.8 cm−1
below and 1.1 cm−1 above this resonance. Remarkably, the outermost ring in the
resonant image extends significantly further radially than in the two surrounding
images. In fact, when comparing the experimental results (color triangles in the in-
set) to calculations using the classical method used in Ref. [163] (solid blue line), one

91
Chapter 5. Hydrogen atoms under magnification

sees that the position of the outer ring in the image for the (30, 2, 27, 0) resonance
extends further outwards by about 70 %. This is in line with recent theoretical work
by Zhao and Delos, who developed both a semiclassical and a quantum mechanical
theory for the hydrogen atom photoionization microscopy problem [166, 171]. They
predicted a ”remarkable tunneling effect” that applies only in the case of resonant
excitation of quasi-bound Stark states. Classically, the electron is trapped by the
potential barrier V (η)
 (see Fig. 5.1b) if the emission angle is smaller than a critical
−E
angle βc = 2arcsin √
2 F
. However, in case of excitation to a quasi-bound state,
electrons with an emission angle smaller than βc may tunnel through the V (η) po-
tential barrier, leading to a situation where the electron can reach a position on the
detector that is not classically accessible.
Generally, the image measured at a resonance corresponds to a coherent superpo-
sition of resonant and non-resonant contributions, the latter corresponding to direct
excitation into the ionization continuum. As a consequence a beating between these
two contributions is expected. In the hydrogen measurements that are presented
here, the resonant contribution strongly dominates. For example, the signal (i.e. the
total number of detected electrons per acquisition) at the (30, 2, 27, 0) resonance
in Fig. 5.6 (case b) was stronger than the signal at the adjacent non-resonant po-
sitions (cases a and c) by a factor 10. Therefore, the image essentially represents a
direct macroscopic projection of the microscopic electronic quantum state. In other
atoms electron-electron interactions (as manifested by quantum defects) that couple
the initial state (n, n1 , n2 , m) to other states have a major influence on the elec-
tronic wave that is observed. For example, the above-mentioning tunneling in the
η coordinate is largely absent in non-hydrogenic systems, because (n, n1 , n2 , m)
initial states that require tunneling couple to states that do not. In Xe, the coupling
among parabolic states led to a complete disappearance of the resonant effects [165],
whereas in Li it led to a substantial reduction of the contrast between resonant and
non-resonant excitation [172].

5.4 Summary
In conclusion it was shown that the concept of photoionization microscopy, as
theoretically proposed more than 30 years ago and the subject of recent theoretical
work predicting the possibility to experimentally observe non-classical photoioniza-
tion dynamics involving tunneling through the V (η) potential barrier, can be ex-
perimentally realized, providing both a beautiful demonstration of the intricacies of
quantum mechanics and a fruitful playground, where the fundamental implications
of this theory can be further explored. For example, predictions have already been
made for the case where both electric and magnetic fields are present [178]. The
experimental observations of the nodal structures of the wave functions presented
in this chapter are not specific to the field strengths and laser excitation conditions
(i.e. the polarization direction of the exciting laser) used, but are a general phe-
nomenon that is observable and can be exploited over a wide range of experimental
conditions.

92
 Visualizing the coupling

6 between red and blue Stark

states using photoionization
microscopy

In nonhydrogenic atoms placed in a DC electric field, the finite size of the ionic
core introduces a coupling between quasi-bound Stark states that leads to the oc-
currence of avoided crossings between states that otherwise would cross. Near an
avoided crossing, the interacting states may have decay amplitudes that cancel each
other, leading to a decoupling of one of the states from the ionization continuum.
This well-known interference narrowing effect, observed as a strongly electric field-
dependent decrease in the ionization rate, has previously been observed in several
atoms [87, 88, 94–96]. Here the photoionization microscopy technique is used to vi-
sualize the interference narrowing effect, thereby explicitly revealing the mechanism
by which Stark states decay. The interference narrowing allows measurements of the
nodal patterns of red Stark states, which are otherwise not observable due to their
intrinsic short lifetime.

6.1 Introduction
In the previous chapter we have seen that hydrogen atoms placed in an exter-
nal field are conveniently described using parabolic coordinates (ξ, η, ϕ), since the
Schrödinger equation becomes exactly separable and the Stark wave function can
be written as ψ(η, ξ, ϕ) = √ 1 χ1 (ξ)χ2 (η)eimϕ . For a given value of the principal
2πξη
quantum number n, the Rydberg states fan out into Stark manifolds that are char-
acterized by parabolic quantum numbers n1 and n2 . The energy of the Stark states
is, to first order in the electric field strength F , given by E = − 2n1 2 + 32 F n(n1 − n2 ).
Red Stark states (n1 < n2 ) are localized downhill on the Coulomb + DC field poten-
tial and display higher ionization rates than blue Stark states (n1 > n2 ), where the
electron is located uphill. For field strengths F > 1/3n5 adjacent Stark manifolds
overlap, and blue and red Stark states cross each other without coupling.
Substantial differences occur in the case of nonhydrogenic atoms, where, due to
the finite size of the ionic core, the interaction of the Rydberg electron with the core
deviates from being purely Coulombic. In the absence of an electric field, the energy
levels of the Rydberg states are now given by a Rydberg-like formula, where the
principal quantum number n is replaced by n∗ = n − δl , where δl is the quantum
defect [71]. The existence of non-zero quantum defects leads to couplings among the

93
Chapter 6. Visualizing the coupling

Stark states. Avoided crossings arise when two Stark states cross each other [68] and
Stark states ionize by a process similar to autoionization [84], which significantly
decreases the lifetimes of the states. Moreover, the coupling between the blue states
and the red continua leads to interferences in the excitation amplitudes and gives
rise to asymmetric line shapes resembling Fano profiles [86].
Studies of Stark states around an avoided crossing have included observations
of interference narrowing [87, 88, 92, 94–96], i.e. a localized and strongly electric
field-dependent increase in the lifetime of some of the Stark states. As explained in
Ref. [87]and in Section 2.4.2, this interference narrowing occurs because the matrix
element describing the ionization rate can be written as a sum of individual contri-
butions from uncoupled Stark states that may have opposite signs, and that may
therefore cancel each other. For example, a two orders of magnitude reduction in the
line width of the (n, n1 , n2 , m)=(20, 19, 0, 0) Stark state of sodium was observed
in Ref. [87].

6.2 Ionization process in hydrogenic and nonhy-

drogenic
In this chapter, the photoionization microscopy technique [52, 165, 172] is ap-
plied to observe changes in the nature of the Stark wave function at a point where a
blue and a red Stark state cross, and where interference narrowing occurs. Photoion-
ization microscopy, proposed about thirty years ago [49, 51, 158], predicts that the
projection of very slow (typically meV) photoelectrons that are formed by photoion-
ization in the presence of a DC electric field onto a position-sensitive two-dimensional
detector would lead to the observation of interference patterns, resulting either from
the nodal structure of Stark states prepared in the experiment, or from quantum-
mechanical path length differences experienced by the photoelectron along multiple
interfering pathways connecting the ionized atom to a particular position on the
detector. In a first experimental realization, Nicole and co-workers observed inter-
ference patterns for the ionization of xenon atoms in a DC electric field [165], which
could be fully explained by the latter mechanism [161]. The same non-resonant effect
was observed in hydrogen atoms and is described in detail in Chapter 4. Observa-
tion of the nodal structure of quasi-bound Stark states was only accomplished very
recently in photoionization microscopy experiments on hydrogen ( Ref. [52, 179] and
Chapter 5) and lithium [172].
The nature of the ionization process leaves its signature in the interference pat-
terns that can be recorded in a photoionization microscopy experiment. If the ion-
ization of the hydrogen atom takes place from a quasi-bound Stark state, the elec-
tron must tunnel through an effective potential along the η coordinate, V (η) =
2
− Z2η2 − m8η−1
2 + F8η (see e.g. Fig. 2.4). This tunneling only affects the electronic wave
function χ2 (η), and therefore experimentally observed dark fringes directly reveal
the nodal structure of χ1 (ξ). If the ionization takes place by direct excitation into the
continuum (in other words, when the electron is excited above the barrier in V (η)),

94
 6.3 Experimental methods

interferences due to semiclassical path length differences along trajectories are ob-
served [161]. A special case arises in nonhydrogenic systems, where the Schrödinger
equation is no longer separable in parabolic coordinates, n1 and n2 are no longer
good quantum numbers and ionization from a quasi-bound Stark state (i.e. a state
that is bound in the V (η) potential) can occur when the electron is scattered by the
core into parabolic states lying above the barrier in the V (η) potential. The latter
process, similar to autoionization, is the dominant mechanism of field ionization in
nonhydrogenic atoms.

6.3 Experimental methods

Here, the case of a singlet state in helium is considered, where δs = 0.14,
δp = −0.012, δd = 0.002 and δl­3 ≈ 0. In the experiments (see Fig. 6.1 and for
more detail Chapter 3), metastable helium atoms (21 S) were generated in a water-
cooled source, where a cw discharge was maintained between a Tantalum needle,
surrounded by a 6 mm diameter quartz tube filled with He gas, and a grounded, stain-
less steel skimmer [146]. The intensity of the metastable helium beam was 7.5×1012
He∗ /sec/steradian. The He∗ atoms were photoexcited using the frequency-doubled
output of a cw ring dye laser (899-29, Coherent) that was tunable between 625 and
630 nm. Second harmonic generation was accomplished making use of an external
ring cavity frequency doubler (WaveTrain, Spectra Physics), which ensured 10% ef-
ficiency. The resulting photoelectrons were detected using a velocity map imaging
(VMI) spectrometer [114], with a 33 cm flight length and an Einzel lens placed
half-way to the two-dimensional detector, which enabled a ten-fold magnification of
the images observed on the position-sensitive detector [116]. The detector consisted
of a stack of 4 cm diameter microchannel plates (MCPs), followed by a phosphor
screen and a CCD camera. To suppress contributions from Penning ionization of
background gas, the VMI was surrounded by a LN2 cold trap, and photoelectron
images were recorded by measuring the difference between acquisitions with and
without the UV laser.

6.4 Results and discussion

Throughout the experiments the DC field strength was maintained at approxi-
mately 468 V/cm and the polarization of the UV laser was set parallel to the DC
electric field axis, resulting in the population of m = 0 Stark states (δs = 0.14). By
scanning the dye laser frequency with a simultaneous acquisition of the signal on the
detector a photoionization spectrum was acquired, which revealed a large number of
quasi-bound Stark states lying between
√ the saddle-point in the Coulomb + DC field
potential at an energy Esp = −2 F = −132.4 cm−1 and the field-free ionization
limit (E = 0). A part of this spectrum, for excitation energies in the vicinity of the
saddle point, is shown in Fig. 6.2a, where distinct Stark states are observed. The
hydrogenic quantum numbers (n, n1 , n2 , m) were assigned to most of the spec-
tral lines using the numerical calculations based on Harmin’s WKB-QD theory (see

95
Chapter 6. Visualizing the coupling

312-315 nm
h

f
e
g

d
a

b
c

Figure 6.1: Schematic representation of the experimental setup. Metastable (21 S) helium
atoms are created by means of a DC electric discharge maintained between a Tantalum
needle (a), surrounded by a quartz tube (b) filled with He gas provided by a gas inlet (c),
and a grounded skimmer (d). The ignition starts at -2 kV and is sustained at -0.7 kV
and 20 mA. The He∗ beam, formed by the skimmer, crosses a tunable UV laser inside a
velocity map imaging spectrometer, permitting the excitation of quasi-bound Stark states that
subsequently ionize. By applying a voltage difference across the repeller (e) and extractor (f )
plates the photoelectrons are accelerated towards a position-sensitive detector (g) comprising
4 cm diameter MCP plates, a phosphor screen and a CCD camera. Halfway towards the
detector, the photoelectrons pass through an Einzel lens (h), enabling magnification of the
recorded photoelectron images by one order of magnitude.

Section 2.5.2). For example the most prominent spectra lines, marked with black
arrows, belong to the n = 28 Stark manifold and are equally spaced by E = 1.70
cm−1 . The spectral line at E = −118.07 cm−1 corresponds to the bluest state in
the manifold, i.e. (28, 27, 0, 0), whereas the line at E = −129.99 cm−1 is identified
as the (28, 20, 7, 0) Stark state. The n = 28 is one of the lowest (in terms of en-
ergy) manifolds observed in the spectrum and due to its high stability the observed
spectral lines are intense and very narrow, e.g. the experimental full width at half
maximum (FWHM) of the (28, 27, 0, 0) state is 0.0060±0.0003 cm−1 . By contrast,
the red state (33, 5, 27, 0), marked with a gray arrow, has a width of 0.026±0.002
cm−1 , meaning that its lifetime is one order of magnitude shorter than the lifetime
of the (28, 27, 0, 0) state. In hydrogen, blue and red states differ in their lifetimes by

96
 6.4 Results and discussion

0.9 a

*
0.6
Normalized signal intensity

0.3

* * *
* **
0.0
b
0.9

0.6

0.3

0.0
-130 -128 -126 -124 -122 -120 -118
-1
Energy (cm )

Figure 6.2: (a) Experimental photoionization spectrum of helium atoms placed in a DC

electric field of 468 V/cm and ionized in the vicinity of the saddle point (Esp = −132.4
cm−1 ). The numerous quasi-bound Stark states observed here belong to different manifolds
ranging from n = 27 to n = 34. Black arrows indicate intense and narrow spectral lines
from the n = 28 manifold with the bluest state, i.e. (n, n1 , n2 , m)=(28, 27, 0, 0) at
E = −118.07 cm−1 . A gray arrow points at a red state (33, 5, 27, 0), whose lifetime is
one order of magnitude shorter compared to the bluest state in the n = 28 manifold. Black
asterixes (star symbols) indicate resonances for which the interference narrowing effect was
observed. The spectral lines show asymmetric Fano profiles, which are characteristic for
autoionization processes. (b) Calculated photoabsorption spectrum, by means of numerical
calculations, displaying a very good agreement, in terms of intensity and position, with the
experimental spectral lines shown in (a).

many orders of magnitude [75], whereas in helium, or other nonhydrogenic atoms,

the blue states are coupled to the red continua and decay much faster. This type of
field ionization is a dominant process in nonhydrogenic atoms and it manifests itself
by asymmetric profiles of the spectral lines. Such Fano profiles, resulting from con-
structive and destructive interference between the transition strengths to the mixed
discrete-continuum eigenfunctions, are conspicuous in the experimental spectrum.
Fig. 6.2b shows the calculated photoabsorption spectrum for helium atoms in
an electric field of 468 V/cm. The photoabsorption cross section was computed by
Fourier transforming the correlation function hψ(t = 0)|ψ(t)i as implemented in Ref.
[98]. This calculation is similar to the wave packet code (see Section 2.5.3) except
ψ(t = 0) = zψg , where z = rcos(θ) and ψg is a ground state wave function. The
calculation reproduces very well the experimentally observed spectral lines including
their positions and relative intensities. The Fourier transform is performed in all
calculations meaning that the calculated line widths are inversely related to the
computational time.

97
Chapter 6. Visualizing the coupling

1 1
2 a 2 b e

0 0

-2 0 -2 0
-2 0 2 -2 0 2
1 1
2 c 2 d
Ry (mm)

0 0

-2 0 -2 0
-2 0 2 -2 0 2
Rx (mm)

Figure 6.3: Photoelectron images recorded for Stark states that decay via core-induced cou-
pling to fast-ionizing red states. The examples shown in (a)-(d) are the (28, 22, 5, 0),
(28, 25, 2, 0), (29, 21, 7, 0) and (21, 23, 5, 0) states, acquired at energies of -126.59 cm−1 ,
-121.46 cm−1 , -119.42 cm−1 and -115.87 cm−1 , respectively, in the presence of a 468 V/cm
DC electric field. In all images, the intensities are normalized to the maximum occurring in-
tensity. The main observation is that the number of interference minima increases smoothly
as a function of the energy. The interferences can be explained by considering the phase
difference between different trajectories by which electrons can move from the ion to the
detector. This is further illustrated in (e), where the positions of experimentally observed
interference minima (+) are compared with semiclassically calculated positions (solid lines,
using the method described in [161]).

6.4.1 Above the V (η) barrier ionization of helium atoms

In the experiments photoelectron images were acquired for all Stark resonances in
the energy range between the saddle point and E = -75 cm−1 , resulting in more than
120 images. The available signal levels did not permit measurements away from the
pronounced Stark resonances. In Figure 6.3 (a)-(d), a series of representative images
are displayed, revealing the appearance of an interference structure. Images (a)-(d)
were recorded for the (28, 22, 5, 0), (28, 25, 2, 0), (29, 21, 7, 0) and (21, 23, 5, 0)
Stark states, respectively. However, there is no correlation between a number of nodes
and the parabolic quantum number n1 . On the contrary, the number of dark fringes
increases smoothly with the photoelectron kinetic energy (defined with respect to the
saddle-point). As illustrated in Fig. 6.3e, the energy dependence of this interference
structure can be understood semiclassically by considering the phase accumulation
along different trajectories that allow the electron to reach the same position on
the two-dimensional detector, as discussed in Section 4.2 and in Ref. [161]. In other
words these results, which account for about 95% of the images that were recorded,
are fully consistent with the results previously reported for xenon in Ref. [165] and
for hydrogen in Chapter 4. Consequently, these results will not be further discussed.

98
 6.4 Results and discussion

a 1 b 1 c 1
2 2 2
Ry (mm)
0 0 0

-2 0 -2 0 -2 0
-2 0 2 -2 0 2 -2 0 2
Rx (mm)
1.0 d 1.0 e 1.0 f
Normalized radial
distribution

0.5 0.5 0.5

0.0 0.0 0.0

0 1 2 0 1 2 0 1 2
R (mm)
-130.0
g h
9E-3

-130.2 (27,26,0,0)
(34,1,32,0)

6E-3
-130.4
Energy (cm-1)

FWHM (cm-1)

(34,1,32,0)
-130.6

-130.8
3E-3
(31,7,23,0)
-131.0
464 466 468 470 472 464 466 468 470 472
Electric field (V/cm) Electric field (V/cm)

Figure 6.4: (a)-(c) Three raw photoelectron images recorded at energies of -130.95 cm−1 ,
-130.43 cm−1 and -130.19 cm−1 corresponding to adjacent Stark states identified as (n,
n1 , n2 , m)=(31, 7, 23, 0), (34, 1, 32, 0) and (27, 26, 0, 0) respectively. (d)-(f ) Comparison
between the experimental radial distributions (solid line), extracted from the images in the
top panel, and calculated radial distributions (dashed lines) obtained by means of TDSE
calculations. The main result is the appearance of a minimum in the radial distribution
for the (34, 1, 32, 0) state, which corresponds to the value of n1 and directly reveals the
nodal structure of the χ1 (ξ) wave function. N.B. The computational results shown here and
in Fig. 6.6 were scaled to the dimensions of the experiment in order that the best fit is
achieved for the (34, 1, 32, 0) state. (g) Stark map for helium obtained from the calculated
photoabsorption cross sections (squares); super-imposed dashed lines schematically indicate
the field dependence of the related diabatic, uncoupled Stark states. The gray bar indicating
the experimental field strength illustrates that the photoelectron image for the (34, 1, 32, 0)
state was recorded in the vicinity of an avoided crossing of this red state with the blue (27,
26, 0, 0) state. (h) Field-dependent reduction in the line width of the (34, 1, 32, 0) state,
extracted from the calculated photo-absorption spectra and confirming the occurrence of the
interference narrowing effect.

99
Chapter 6. Visualizing the coupling

6.4.2 Interference narrowing effects in helium

For eight Stark states the observed nodal pattern differed markedly from the
semiclassical predictions. These images were recorded in the vicinity of avoided cross-
ings between red and blue Stark states, and reveal the nodal structure along the ξ
coordinate of the red Stark states involved. To illustrate this, Fig. 6.4 (a)-(f) show
raw images and extracted radial distributions for three measurements performed
at energies just above the saddle-point, involving the (31, 7, 23, 0), (34, 1, 32, 0)
and (27, 26, 0, 0) states. The (34, 1, 32, 0) state is an extreme red Stark state,
whereas the other two states are a state in the middle of the Stark manifold and
an extreme blue state. Remarkably, the radial distribution of the red (34, 1, 32, 0)
state extends further than that of the two adjacent states, and – unlike the adja-
cent images – has a minimum. Both of these observations point towards hydrogenic
behavior [52]. The observed minimum in the radial distribution reflects the value of
the parabolic quantum number n1 , which describes the nodes of the electronic wave
function χ1 (ξ). Furthermore, the increased radial extension of the image reflects the
occurrence of a tunneling process, which allows the electron to reach positions on
the detector that are not accessible when the electron ionizes by passing over the
saddle-point in the Coulomb + DC field potential [166, 171]. This interpretation is
further confirmed by comparing the experimental radial distributions shown in Fig.
6.4 (d)-(f) with results from wave packet calculations based on numerical solution
of the time-dependent Schrödinger equation (TDSE) [98]. In these calculations, the
He∗ atoms were ionized using a 1.5 ns long bandwidth-limited laser pulse, and the
propagation was continued up to a time delay of 3 ns. To extract a two-dimensional
image from the calculations, the flux through a screen placed at a distance of 0.8 µm
was evaluated. It was checked that the radial distributions do not change anymore
qualitatively beyond this distance. The numerical radial distributions (scaled to the
size of the experimentally measured distributions) show satisfactory agreement and
reproduce the essential features of the experiment, i.e. the occurrence of a minimum
in the radial distribution for the (34, 1, 32, 0) state and the larger radial extension
of this image.
From the calculated photoabsorption cross sections for helium in different energy
and field ranges Stark maps were obtained by reading off local maxima in the com-
puted absorption strength as a function of laser wavelength. The Stark map shown
in Fig. 6.4g reveals that the measurement for the (34, 1, 32, 0) state was performed
in the vicinity of an avoided crossing with the (27, 26, 0, 0) state. The approximate
hydrogenic assignments of the diabatic, uncoupled Stark states indicated in Fig. 6.4g
were obtained by exploiting the similarities between the calculated Stark map and
a hydrogenic calculation (using the formulas given in Section 2.5.1). Monitoring the
line widths of the Stark states in the photoabsorption spectra (by fitting a Lorentz
profile to the observed resonances), it was discovered that the condition under which
the experiment for the (34, 1, 32, 0) state was performed, coincided with a minimum
in the line width of this state (Fig. 6.4h). In other words, the observation of the nodal
pattern of the (34, 1, 32, 0) state was made possible by a lifetime enhancement (i.e.
interference narrowing) of this state, due to the coupling of this state with the blue

100
 6.4 Results and discussion

(27, 26, 0, 0) state near a field strength of 468 V/cm.

Fig. 6.5a presents the calculated photoabsorption spectra in the vicinity of the
avoided crossing between the (34, 1, 32, 0) and (27, 26, 0, 0) Stark states. The spectra
were calculated by increasing the electric field from 464 V/cm (top line) in steps of
0.2 V/cm till 468 V/cm (bottom line). The narrowing effect is admittedly visible
here, but it is not very apparent since in this particular case the line-narrowing
occurs over a wide range of the electric field with a minimum width at 468 V/cm,
as shown in Fig. 6.4h. Nevertheless, the main observation here is the disappearance
of the oscillator strength of one of the interfering components at about 466.6 V/cm.
The same effect was observed experimentally by McNichol et al. in sodium [92]
and was explained as interference between the dipole transition matrix elements.
Thus it is an analogues effect to the line-narrowing effect, but the cancelation in
the amplitude is for other process, which in this case is an electric dipole transition
[93, 180]. The energy-field region where this effect occurs, marked with a box, is
shown in 6.5b in greater detail. Closer examination of this figure shows that the
asymmetry of the line shape is reversed when the electric field changes from 466.4
to 466.6 V/cm. As stated at the beginning of this chapter and in Section 2.4.2, the
spectral lines of nonhydrogenic atoms resemble Fano profiles that are characterized
by the line shape parameter q, which is proportional to the ratio of the transition
amplitude to a bound state and the amplitude of the direct photoionization (to
the continuum) [86]. The q reversal observed in 6.5b indicates a sign change of the
transition amplitude related to the direct ionization amplitude as shown in Ref.
[92]. By that very fact the conditions for the interference narrowing effect to occur
are created, i.e. the interfering states have transition amplitudes of opposite signs
that can cancel each other leading to a decoupling of one of the interfering states
from the continuum. Indeed, the experimental photoionization spectrum (shown in
Fig. 6.2a) revealed that the states, where the line-narrowing effect was observed,
showed reversed Fano profiles compared to the states with which they interfered in
the vicinity of an avoided crossing.
A very similar reasoning underlies the measurement of seven other photoelectron
images, listed in Table 6.1 and marked by black stars in Fig. 6.2a, where the exper-
imentally observed interference patterns disagree with the semiclassical prediction.
In each of these cases, an avoided crossing between a red and a blue Stark state can
be identified, and the nodal structure observed on the detector was that of the red
Stark state involved. Furthermore, all these states showed a strongly field-dependent
increase in their lifetimes and the photoionization spectrum revealed a reversal of
Fano profiles between the interfering states. Here, only the case of the (35, 5, 29, 0)
state is further discussed as this measurement was taken a strongly perturbed region
due to the vicinity of another avoided crossing, i.e. between the (31, 13, 17, 0) and
(28, 26, 1, 0) states.
In Fig. 6.6a a raw photoelectron image of the (35, 5, 29, 0) state measured at an
energy of -119.73 cm−1 and an electric field of 467.4 V/cm is presented. The ex-
tracted radial probability distribution, shown in (b) as a red line, differs significantly
from the radial distributions of the adjacent states, which were measured at E =
-119.83 cm−1 (black line) and E = -119.42 cm−1 (green line). The red line shows

101
Chapter 6. Visualizing the coupling

a
6000
464 V/cm

b
Signal intenity (arb. units)

465 V/cm
4000

466 V/cm

2000 466 V/cm

467 V/cm

467 V/cm

468 V/cm
0 -1
-130.3 cm

-130.5 -130.4 -130.3 -130.2 -130.1

Energy (cm-1)

Figure 6.5: (a) Photoabsorption spectra in the region of an avoided crossing between the (34,
1, 32, 0) and (27, 26, 0, 0) states calculated with a step of 0.2 V/cm for an electric field
ranging between 464 V/cm (top line) to 468 V/cm (bottom line). The main observation
here is the disappearance of the oscillator strength in the line on the left side due to the
interference of electric dipole transition matrix elements. (b) Magnification of the region
marked with a box in (a), indicating a reversal of the asymmetry of the line shape, resulting
from a sign change in the transition amplitude related to the ionization into the continuum.

five minima reflecting the nodes of the electronic wave function χ1 (ξ) since n1 = 5,
which is further confirmed in (c), where the experimental radial probability distri-
bution for the (35, 5, 29, 0) state (solid line) is compared with TDSE calculations
(dashed line), displaying a very good agreement. Moreover, the radial expansion of
the photoelectron image for the (35, 5, 29, 0) state is larger compared to the adja-
cent states, which validates the involvement of the tunneling effect. According to the
semiclassical formula given by Eq. (4.17), the image at -119.83 cm−1 should have
the smallest radius as the photoelectrons’ kinetic energy in this case is the lowest.
However, the image size at -119.83 cm−1 , given by the black line in (b), exceeds the
radius of the image recorded at -119.42 cm−1 (green line), what suggests that this
state, i.e. (28, 26, 1, 0) is also perturbed due to coupling to the (35, 5, 29, 0) state.
This is further confirmed by investigating the plotted Stark map in (d), where the
experimental electric field (467.4 ± 0.5 V/cm) is marked by a grey bar showing that
measurements at 119.42 cm−1 and -119.73 cm−1 were taken in a strongly perturbed
region coinciding with crossings between the (35, 5, 29, 0) and (31, 13, 17, 0) and
(28, 26, 1, 0) hydrogenic states. In the calculated photoabsorption spectra a line

102
 6.4 Results and discussion

Normalized radial distribution

Normalized radial distribution
1
2 a 1.0 b 1.0 c
Ry (mm)

0
0.5 0.5

-2
0 0.0 0.0
-2 0 2 0 1 2 0 1 2
Rx (mm) R (mm) R (mm)
-119.5 1E10
d e
A
-119.6

Ionization rate (s )
-1
1E9
Energy (cm-1)

-119.7
B
1E8
-119.8 C

-119.9
1E7

-120.0
464 466 468 470 472 466 467 468 469
Electric field (V/cm) Electric field (V/cm)

Figure 6.6: (a) Raw photoelectron image for the (35, 5, 29, 0) state measured at E = -
119.73 cm−1 in the vicinity of an avoided crossing. (b) Experimental radial distributions
of three adjacent resonances: E = -119.83 cm−1 (black line), E = -119.73 cm−1 (red line)
and E = -119.49 cm−1 (green line), showing that the (35, 5, 29, 0) state has five minima
and extends radially further than the other two states, which suggests the involvement of
tunneling in the ionization process. (c) Comparison between the experimental (solid line)
and theoretical (dashed line, obtained from TDSE calculations) radial distribution showing
an excellent agreement in the position of the nodes. A calculated Stark map (d) indicates
that the measurement was taken for a DC electric field of approximately 467.4 V/cm, where
the (35, 5, 29, 0) state (A) has an avoided crossing with the (31, 13, 17, 0) (B) and (28,
26, 1, 0) (C) states. Note that the absence of black squares implies a disappearance of the
oscillator strength in the photoabsorption spectra for a given state. The calculated ionization
rate of the (35, 5, 29, 0) state, shown in (e), decreases by three orders of magnitude within
a field range of 0.5 V/cm, due to the occurrence of interference narrowing.

width of the (35, 5, 29, 0) state was too broad to fit any Lorentz profile and conse-
quently, it was impossible to determine the line-narrowing effect in this way. On the
other hand, in TDSE calculations the (35, 5, 29, 0) state was very distinct due to its
unique radial distribution with five minima. Therefore, for this state photocurrent
fluxes as a function of time were computed for different values of the electric field,
providing information about its lifetime and revealing a three orders of magnitude
drop in the ionization rate over a field range of 0.5 V/cm as shown in (e), ipso facto
confirming the occurrence of the interference narrowing effect.

103
Chapter 6. Visualizing the coupling

Table 6.1: List of quasi-bound Stark states for which the nodal structure of the electronic
wave function χ1 (ξ) was directly observed in the experiments due to the interference nar-
rowing effect.
Energy Electric field Stark state Avoided crossing with:
( cm−1 ) (V/cm) (n, n1 , n2 , m) (n, n1 , n2 , m)
-130.43 468.0 (34, 1, 32, 0) (27, 26, 0, 0)
-128.39 468.0 (35, 1, 33, 0) (28, 21, 6, 0)
-128.23 468.0 (34, 2, 31, 0) (29, 16, 12, 0)
-126.35 468.0 (35, 2, 32, 0) (29, 17, 11, 0)
-124.28 468.0 (35, 3, 31, 0) (30, 14, 15, 0)
(31, 13, 17, 0)
-119.73 467.4 (35, 5, 29, 0)
(28, 26, 1, 0)
-117.42 467.4 (35, 6, 28, 0) (29, 22, 6, 0)
-108.80 468.0 (35, 10, 24, 0) (29, 27, 1, 0)

6.5 Conclusions
Hence, the observations encountered in the present experiments can be fully ex-
plained. Given the Stark lifetime (­ 100 ps) needed for the observation of sharply
defined resonances in the photoexcitation efficiency as a function of laser wavelength,
most helium photoionization microscopy measurements involved the excitation of
quasi-bound Stark states in the middle or on the blue side of the Stark manifold.
These states are strongly bound along both the ξ and η coordinates, and can only
ionize as a result of core-induced coupling to red continuum states, i.e. states where
the motion along the η coordinate is unbound. Consequently, as shown in Fig. 6.3,
the interference patterns that are observed for these states are governed by the phase
differences that the photoelectrons accumulate along multiple classical paths that
they can follow between the ion and the detector. Moreover, a number of experi-
ments were performed, where quasi-bound Rydberg states were excited on the red
side of a Stark manifold, and in the vicinity of an avoided crossing with a blue state
that induced an enhancement of the Rydberg state lifetime as a result of interfer-
ence narrowing (listed in Table 6.1 and shown in Fig 6.4 and 6.6). All of these red
states furthermore had in common that they occurred at energies just below the
parabolic critical energy, which marks the top of the V (η) potential. Under these
circumstances, the electron may leave the atom by means of a tunneling process
that leaves the quantization along the ξ coordinate intact. This ionization mecha-
nism occurs in competition with the core-induced coupling mechanism and is more
prominent for atoms with very small quantum defects (δl  1), where the hydrogenic
structure of interfering Stark states is preserved to a greater degree [93]. Interfer-
ence narrowing is essential, because it causes a destructive interference between the
bound-continuum coupling amplitudes of the red and blue states that cross, decou-
pling one of these states from the ionization continuum. The lack of coupling causes

104
 6.5 Conclusions

the hydrogenic tunneling to be the dominant ionization mechanism, thereby allowing

the observation of the n1 quantum number of the red Stark state on the detector,
similar to the recent work on the hydrogen atom (Chapter 5 and [52]). At the same
time the presented results are different from the photoionization microscopy exper-
iments on lithium [172], where the hydrogenic behavior was forced by exciting only
m = 1 Stark states, which are characterized by small quantum defects (δp = 0.0048).
For the similarities between hydrogen and lithium in m = 1 states see e.g. Fig. 2.3.
The utility of the photoionization microscopy technique in the present experi-
ments on helium is that it visualizes, through an observation of the nodal structure
of the red Stark state, the DC electric field-induced coupling between the red and
blue Stark states in the vicinity of the avoided crossing. In future, it will be of great
interest to perform experiments such as the ones presented here in a field-dependent
manner, thereby allowing a complete mapping out the couplings that occur in Stark
manifolds around field-induced avoided crossings.

105
Bibliography

[1] E. T. Whittaker. A History of the theories of aether and electricity. Green

and Co., Dublin (1910).

[2] A. A. Michelson and E. W. Morley. On the Relative Motion of the Earth and
the Luminiferous Ether. American Journal of Science, 34, 333–345 (1887).
[3] A. Einstein. Zur Elektrodynamik bewegter Körper. Annalen der Physik, 322,
891–921 (1905).
[4] R. C. Douga. The presentation of the Planck radiation formula (tutorial).
Physics Education, 11, 438–443 (1976).
[5] M. Planck. Verh. Dt. Phys. Ges., 2, 202–204 (1900).
[6] M. Planck. Verh. Dt. Phys. Ges., 2, 237–245 (1900).

[7] H. Becquerel. Sur les radiations émises par phosphorescence. Comptes Rendus,
122, 420–421 (1896).
[8] J. Thomson. Cathode rays. Philosophical Magazine, 44, 293 (1897).
[9] J. J. Thomson. XXIV. On the structure of the atom: an investigation of the
stability and periods of oscillation of a number of corpuscles arranged at equal
intervals around the circumference of a circle; with application of the results
to the theory of atomic structure. Philosophical Magazine Series 6, 7, 237–265
(1904).
[10] L. M. Rutherford. Philosophical Magazine, 21, 669 (1911).

[11] N. Bohr. Philosophical Magazine, 26, 476 (1913).

[12] G. H. J. Franck. Über Zusammenstösse zwischen Elektronen und Molekülen
des Quecksilberdampfes und die Ionisierungsspannung desselben. Verh. Dtsch.
Phys. Ges., 16, 457–467 (1914).
[13] A. Sommerfeld. Annalen der Physik, 56, 1 (1916).

[14] P. S. Epstein. Zur Theorie des Starkeffektes. Annalen der Physik, 50, 489–520
(1916).
[15] K. Schwarzschild. Zur Quantenhypothese. Sitzungsberichten der Kgl. Preuss.
Akad. d. Wiss., 26, 548–568 (1916).

107
References

[16] A. Sommerfeld. Zur Theorie des Zeeman-Efekts der Wasserstofflinien, mit

einem Anhang uber den Stark-Efekt. Physikalische Zeitschrift, 17, 491–507
(1916).

[17] A. Messiah. QUANTUM MECHANICS ; VOL. 1. Amsterdam, North-holland

(1964).
[18] L. de Brogile. Nature, 112, 540 (1923).
[19] A. Einstein. Über einen die Erzeugung und Verwandlung des Lichtes betref-
fenden heuristischen Gesichtspunkt. Annalen der Physik, 17, 132–148 (1905).
[20] W. Heisenberg. Zeitschrift für Physik, 33, 879 (1925).
[21] E. Schrödinger. An Undulatory Theory of the Mechanics of Atoms and
Molecules. Physical Review, 28, 1049–1070 (1926).

[22] M. Born. Zur Quantenmechanik der Stossvorgänge. Zeitschrift für Physik, 37,
863–867 (1926).
[23] W. Heisenberg. Über den anschaulichen Inhalt der quantentheoretischen Kine-
matik und Mechanik. Zeitschrift für Physik, 43, 172–198 (1927).
[24] N. Bohr. The Quantum Postulate and the Recent Development of Atomic
Theory. Nature, 121, 580–590 (1928).
[25] H. White. Introduction to atomic spectra. McGraw-Hill book company, inc.
(1934).
[26] W. K. Wootters and W. H. Zurek. A single quantum cannot be cloned. Nature,
299, 802–803 (1982).
[27] B. Lohmann and E. Weigold. Direct measurement of the electron momentum
probability distribution in atomic hydrogen. Physics Letters A, 86, 139–141
(1981).
[28] I. E. Mccarthy and E. Weigold. Wavefunction Mapping in Collision Experi-
ments. Reports on Progress in Physics, 51, 299–392 (1988).
[29] C. E. Brion. Looking at orbitals in the laboratory: The experimental investiga-
tion of molecular wavefunctions and binding energies by electron momentum
spectroscopy. International Journal of Quantum Chemistry, 29, 1397–1428
(1986).

[30] C. E. Brion, G. Cooper, Y. Zheng, I. V. Litvinyuk and I. E. McCarthy. Imaging

of orbital electron densities by electron momentum spectroscopy - a chemical
interpretation of the binary (e,2e) reaction. Chemical Physics, 270, 13–30
(2001).

108
 References

[31] J. W. Watson, M. Ahmad, D. W. Devins, B. S. Flanders, D. L. Friesel, N. S.

Chant, P. G. Roos and J. Wastell. 40 Ca(p, pn)39 Ca and 48 Ca(p, pn)47 Ca
neutron knockout reactions at 149.5 MeV. Phys. Rev. C, 26, 961–964 (1982).

[32] J. Mougey, M. Bernheim, A. Bussire, A. Gillebert, P. X. H, M. Priou, D. Royer,

I. Sick and G. Wagner. Quasi-free (e, e’p) scattering on 12C, 28Si, 40Ca and
58Ni. Nuclear Physics A, 262, 461 – 492 (1976).
[33] C. Monroe. Quantum information processing with atoms and photons. Nature,
416, 238–246 (2002).

[34] D. T. Smithey, M. Beck, M. G. Raymer and A. Faridani. Measurement of the

Wigner distribution and the density matrix of a light mode using optical ho-
modyne tomography: Application to squeezed states and the vacuum. Physical
Review Letters, 70, 1244–1247 (1993).

[35] A. I. Lvovsky, H. Hansen, T. Aichele, O. Benson, J. Mlynek and S. Schiller.

Quantum State Reconstruction of the Single-Photon Fock State. Physical Re-
view Letters, 87, 050402 (2001).
[36] A. Ourjoumtsev, H. Jeong, R. Tualle-Brouri and P. Grangier. Generation of
optical ‘Schrödinger cats’ from photon number states. Nature, 448, 784–786
(2007).
[37] J. S. Lundeen, B. Sutherland, A. Patel, C. Stewart and C. Bamber. Direct
measurement of the quantum wavefunction. Nature, 474, 188–191 (2011).
[38] J. Itatani, J. Levesque, D. Zeidler, H. Niikura, H. Pepin, J. C. Kieffer, P. B.
Corkum and D. M. Villeneuve. Tomographic imaging of molecular orbitals.
Nature, 432, 867–871 (2004).
[39] D. Shafir, Y. Mairesse, D. M. Villeneuve, P. B. Corkum and N. Dudovich.
Atomic wavefunctions probed through strong-field light-matter interaction. Na-
ture Physics, 5, 412–416 (2009).

[40] P. Puschnig, S. Berkebile, A. J. Fleming, G. Koller, K. Emtsev, T. Seyller, J. D.

Riley, C. Ambrosch-Draxl, F. P. Netzer and M. G. Ramsey. Reconstruction of
Molecular Orbital Densities from Photoemission Data. Science, 326, 702–706
(2009).
[41] N. Hay, M. Lein, R. Velotta, R. D. Nalda, E. Heesel, M. Castillejo, P. L.
Knight and J. P. Marangos. Investigations of electron wave-packet dynamics
and high-order harmonic generation in laser-aligned molecules. Journal of
Modern Optics, 50, 561–577 (2003).
[42] S. Haessler et al. Attosecond imaging of molecular electronic wavepackets.
Nature Physics, 6, 200–206 (2010).

109
References

[43] H. Niikura. Mapping Molecular Orbital Symmetry on High-Order Harmonic

Generation Spectrum Using Two-Color Laser Fields. Physical Review Letters,
105(5), p. 053003 (2010).

[44] H. Niikura. Probing the Spatial Structure of a Molecular Attosecond Electron

Wave Packet Using Shaped Recollision Trajectories. Physical Review Letters,
107(9), p. 093004 (2011).
[45] M. F. Crommie, C. P. Lutz and D. M. Eigler. Confinement of Electrons to
Quantum Corrals on a Metal-Surface. Science, 262, 218–220 (1993).

[46] J. Repp, G. Meyer, S. M. Stojkovi, A. Gourdon and C. Joachim. Molecules

on Insulating Films: Scanning-Tunneling Microscopy Imaging of Individual
Molecular Orbitals. Physical Review Letters, 94, 026803 (2005).
[47] L. Holmegaard et al. Photoelectron angular distributions from strong-field ion-
ization of oriented molecules. Nature Physics, 6, 428–432 (2010).
[48] A. Rouzée, A. Harvey, F. Kelkensberg, W. K. Siu, G. Gademann, O. Smirnova
and M. J. Vrakking. Imaging the electronic structure of bonding orbitals in the
XUV ionization of aligned molecules. (submitted for publication) (2013).
[49] Y. N. Demkov, V. D. Kondratovich and V. N. Ostrovskii. Interference of
electrons resulting from the photoionization of an atom in an electric field.
JETP Letters, 34, 403– (1981).
[50] V. D. Kondratovich and V. N. Ostrovsky. Resonance and interference phenom-
ena in the photoionisation of a hydrogen atom in a uniform electric field. I.
Resonances below and above the potential barrier. Journal of Physics B-Atomic
Molecular and Optical Physics, 17, 1981–2010 (1984).
[51] V. D. Kondratovich and V. N. Ostrovsky. Resonance and interference phe-
nomena in the photoionization of a hydrogen atom in a uniform electric field.
IV. Differential cross sections. Journal of Physics B-Atomic Molecular and
Optical Physics, 23, 3785–3809 (1990).

[52] A. S. Stodolna, A. Rouzée, F. Lépine, S. Cohen, F. Robicheaux, A. Gijsbert-

sen, J. H. Jungmann, C. Bordas and M. J. J. Vrakking. Hydrogen Atoms un-
der Magnification: Direct Observation of the Nodal Structure of Stark States.
Physical Review Letters, 110, 213001 (2013).
[53] P. Zeeman. The Effect of Magnetisation on the Nature of Light Emitted by a
Substance. Nature, 55, 347 (1897).
[54] W. Voigt. Ueber das elektrische Analogon des Zeemaneffectes. Annalen der
Physik, 309, 197–208 (1901).
[55] J. Stark. Observation of the Separation of Spectral Lines by an Electric Field.
Nature, 92, 401 (1913).

110
 References

[56] A. Surdo. Sul fenomeno analogo a quello di Zeeman nel campo elettrico. Ren-
diconti della Reale Accademia dei Lincei, 22, 664–666 (1913).
[57] A. Surdo. Über das elektrische Analogon des Zeeman-Phnomens. Zeitschrift
für Physik, 15, 122 (1914).
[58] J. S. Foster. Observations on the Stark Effect in Hydrogen and Helium. Phys-
ical Review, 23, 667–684 (1924).
[59] A. Sommerfeld. Atombau und Spektrallinien. Fiedr. Vieweg & Sohn (1960).

[60] T. Takamine and N. Kokubu. The Effect of an Electric Field on the Spectrum
Lines of Hydrogen Part III. Memoirs of the College of Science, Kyoto Imperial
University, 3, 271–273 (1918).
[61] H. Kramers. Intensities of Spectral Lines. Andr. Fred. Host & Son (1919).

[62] P. S. Epstein. The Stark Effect from the Point of View of Schroedinger’s
Quantum Theory. Physical Review, 28, 695–710 (1926).
[63] H. Bethe and E. Salpeter. Quantum mechanics of one- and two- electron
atoms. Springer (1957).
[64] E. M. Landau, L. D. & Lifshitz. Quantum Mechanics Non-Relativistic Theory.
Oxford, Pergamon Press (1965).
[65] T. F. Gallagher. Rydberg atoms. Cambridge University Press (1994).
[66] D. Park. Relation between the parabolic and spherical eigenfunctions of hydro-
gen. Zeitschrift für Physik, 159, 155–157 (1960).

[67] H. J. Silverstone. Perturbation theory of the Stark effect in hydrogen to arbi-

trarily high order. Physical Review A, 18, 1853–1864 (1978).
[68] M. L. Zimmerman, M. G. Littman and D. Kleppner. Stark structure of the
Rydberg states of alkali-metal atoms. Physical Review A, 20, 2251–2275 (1979).

[69] G. J. Hatton. Avoided crossings of resonance energy curves of one-electron

atoms in an external electric field. Physical Review A, 16, 1347–1351 (1977).
[70] T. Bergeman. Relativistically Enhanced Ionization Rates at Stark-Effect Level
Crossings in Hydrogen. Physical Review Letters, 52, 1685–1688 (1984).
[71] M. J. Seaton. Quantum defect theory. Reports on Progress in Physics, 46, 167
(1983).
[72] C. Fabre, Y. Kaluzny, R. Calabrese, L. Jun, P. Goy and S. Haroche. Study
of non-hydrogenic near degeneracies between m sublevels in the linear Stark
effect of sodium Rydberg states. Journal of Physics B: Atomic and Molecular
Physics, 17, 3217 (1984).

111
References

[73] R. Gebauer and H. R. van Traubenberg. Über den Starkeffekt dritter Ordnung
bei den Serienlinien Hβ und Hγ des Wasserstof. Zeitschrift für Physik, 62,
289 (1930).
[74] C. Lanczos. Zeitschrift für Physik, 68, 204 (1931).
[75] P. M. Koch and D. R. Mariani. Precise Measurement of the Static Electric-
Field Ionization Rate for Resolved Hydrogen Stark Substates. Physical Review
Letters, 46, 1275–1278 (1981).
[76] M. H. Rice and J. R. H. Good. Stark Effect in Hydrogen. Journal of Optical
Society of America, 52, 239–246 (1962).
[77] J. O. Hirschfelder and L. A. Curtiss. Spontaneous Ionization of a Hydrogen
Atom in an Electric Field. I. The Journal of Chemical Physics, 55, 1395–1402
(1971).
[78] D. S. Bailey, J. R. Hiskes and A. Riviere. Electric field ionization probabilities
for the hydrogen atom. Nuclear Fusion, 5, 41 (1965).
[79] R. J. Damburg and V. V. Kolosov. An asymptotic approach to the Stark effect
for the hydrogen atom. Journal of Physics B: Atomic and Molecular Physics,
11, 1921 (1978).
[80] R. J. Damburg and V. V. Kolosov. A hydrogen atom in a uniform electric field.
III. Journal of Physics B: Atomic and Molecular Physics, 12, 2637 (1979).
[81] H. Rottke and K. H. Welge. Photoionization of the hydrogen atom near the ion-
ization limit in strong electric fields. Physical Review A, 33, 301–311 (1986).
[82] S. Feneuille, S. Liberman, J. Pinard and A. Taleb. Observation of Fano Profiles
in Photoionization of Rubidium in the Presence of a dc Field. Physical Review
Letters, 42, 1404–1406 (1979).
[83] M. G. Littman, M. L. Zimmerman and D. Kleppner. Tunneling Rates for
Excited States of Sodium in a Static Electric Field. Physical Review Letters,
37, 486–489 (1976).
[84] M. G. Littman, M. M. Kash and D. Kleppner. Field-Ionization Processes in
Excited Atoms. Physical Review Letters, 41, 103–107 (1978).
[85] W. van de Water, D. R. Mariani and P. M. Koch. Ionization of highly excited
helium atoms in an electric field. Physical Review A, 30, 2399–2412 (1984).
[86] U. Fano. Effects of Configuration Interaction on Intensities and Phase Shifts.
Physical Review, 124, 1866–1878 (1961).
[87] J.-Y. Liu, P. McNicholl, D. A. Harmin, J. Ivri, T. Bergeman and H. J. Met-
calf. Interference narrowing at crossings of sodium Stark resonances. Physical
Review Letters, 55, 189–192 (1985).

112
 References

[88] S. Feneuille, S. Liberman, E. Luc-Koenig, J. Pinard and A. Taleb. Field-

induced stabilisation of Stark states in the rubidium atom. Journal of Physics
B: Atomic and Molecular Physics, 15, 1205 (1982).
[89] D. A. Harmin. Hydrogenic Stark effect: Properties of the wave functions. Phys-
ical Review A, 24, 2491–2512 (1981).
[90] D. A. Harmin. Theory of the Stark effect. Physical Review A, 26, 2656–2681
(1982).
[91] D. A. Harmin. Analytical study of quasidiscrete Stark levels in Rydberg atoms.
Physical Review A, 30, 2413–2428 (1984).
[92] P. McNicholl, T. Bergeman and H. J. Metcalf. Lifetime measurements of
interference-narrowed sodium Stark resonances. Physical Review A, 37,
33023312 (1988).
[93] D.-H. Yang, D. Lieberman, P. van der Straten, T. Bergeman and H. Metcalf.
Precision mapping of electrostatic fields using interference-narrowed Stark res-
onances. Physical Review A, 40, 5026–5039 (1989).
[94] A. Nussenzweig, E. E. Eyler, T. Bergeman and E. Pollack. Line shapes of
ionizing Stark resonances in helium. Physical Review A, 41, 4944–4957 (1990).
[95] G. D. Stevens, C.-H. Iu, T. Bergeman, H. J. Metcalf, I. Seipp, K. T. Taylor
and D. Delande. Precision measurements on lithium atoms in an electric field
compared with R -matrix and other Stark theories. Physical Review A, 53,
1349–1366 (1996).
[96] A. König, J. Neukammer, H. Hieronymus and H. Rinneberg. Field-ionization
Stark spectra of Ba Rydberg atoms at high density of states. Physical Review
A, 43, 2402–2415 (1991).
[97] E. Titchmarsh. Elgenfunction Expansions Associated With Second Order Dif-
ferential Equations Part II. Oxford Univ. Press (1958).
[98] F. Robicheaux and J. Shaw. Calculated electron dynamics in an electric field.
Physical Review Letters, 56, 278–289 (1997).
[99] R. J. Damburg and V. V. Kolosov. A hydrogen atom in a uniform electric
field. Journal of Physics B: Atomic and Molecular Physics, 9, 3149 (1976).
[100] E. Luc-Koenig and A. Bachelier. Interpretation of Electric-Field-Induced Os-
cillations in the Cross Section of a One-Electron Atom above the Ionization
Limit. Physical Review Letters, 43, 921–924 (1979).
[101] E. Luc-Koenig and A. Bachelier. Systematic theoretical study of the Stark
spectrum of atomic hydrogen. II. Density of oscillator strengths. Comparison
with experimental absorption spectra in solid-state and atomic physics. Journal
of Physics B: Atomic and Molecular Physics, 13, 1769 (1980).

113
References

[102] Y. Ishida and S. Hiyama. Scientific Papers of the Institute of Physical and
Chemical Research, 9, 1 (1928).
[103] S. Alliluev and I. Malkin. Calculations of the Stark effect in hydrogen atoms
by using the dynamical symmetry O(2,2) × O(2). Journal of Experimental
and Theoretical Physics, 39, 627 (1974).
[104] J. Killingbeck. Quantum-mechanical perturbation theory. Reports on Progress
in Physics, 40, 963 (1977).
[105] P. M. Koch. Resonant States in the Nonperturbative Regime: The Hydrogen
Atom in an Intense Electric Field. Physical Review Letters, 41, 99–103 (1978).
[106] C. Lanczos. Zeitschrift für Physik, 65, 431 (1930).
[107] D. A. Harmin. Theory of the Nonhydrogenic Stark Effect. Physical Review
Letters, 49, 128–131 (1982).
[108] G. M. Lankhuijzen and L. D. Noordam. Atomic streak camera. Optics Com-
munications, 129, 361–368 (1996).
[109] F. Robicheaux and J. Shaw. Calculated electron dynamics in a strong electric
field. Physical Review Letters, 77, 4154–4157 (1996).
[110] A. M. Ollagnier. Microscopie de Photoïonisation :une étude classique, semi-
classique et quantiqu. Ph.D. thesis, lUNIVERSITÉ CLAUDE BERNARD -
LYON 1 (2007).
[111] M. J. Amoruso and W. R. Johnson. Relativistic One-Electron Calculations of
Shielded Atomic Hyperfine Constants. Physical Review A, 3, 6–12 (1971).
[112] W. Johnson, D. Kolb and K.-N. Huang. Electric-dipole, quadrupole, and
magnetic-dipole susceptibilities and shielding factors for closed-shell ions of
the He, Ne, Ar, Ni (Cu+), Kr, Pb, and Xe isoelectronic sequences. Atomic
Data and Nuclear Data Tables, 28, 333 – 340 (1983).
[113] D. Kolb, W. R. Johnson and P. Shorer. Electric and magnetic susceptibilities
and shielding factors for closed-shell atoms and ions of high nuclear charge.
Physical Review A, 26, 19–31 (1982).
[114] A. Eppink and D. Parker. Velocity Map Imaging of ions and electron using
electostatic lenses: Application in photoelectron and photofragment ion imaging
of molecular oxygen. Review of Scientific Instruments, 68, 3447–3484 (1997).
[115] W. Johnson and G. Soff. The lamb shift in hydrogen-like atoms, 1¬ Z ¬ 110.
Atomic Data and Nuclear Data Tables, 33, 405 – 446 (1985).
[116] H. L. Offerhaus, C. Nicole, F. Lepine, C. Bordas, F. Rosca-Pruna and M. J. J.
Vrakking. A magnifying lens for velocity map imaging of electrons and ions.
Review of Scientific Instruments, 72, 3245–3248 (2001).

114
 References

[117] L. C. Lee, X. Wang and M. Suto. Quantitative photoabsorption and fluores-

cence spectroscopy of H2S and D2S at 49-240 nm. Journal of Chemical Physics,
86, 4353–4361 (1987).
[118] D. Proch and T. Trickl. A High-Intensity Multi-Purpose Piezoelectric Pulsed
Molecular-Beam Source. Review of Scientific Instruments, 60, 713–716 (1989).
[119] G. van Veen, K. Mohamed, T. Baller and A. de Vries. Photofragmentation of
H2S in the first continuum. Chemical Physics, 74, 261 – 271 (1983).
[120] D. S. Bethune. Dye cell design for high-power low-divergence excimer-pumped
dye lasers. Applied Optics, 20, 1897–1899 (1981).
[121] U. Czarnetzki, K. Miyazaki, T. Kajiwara, K. Muraoka, M. Maeda and H. F.
Döbele. Comparison of various two-photon excitation schemes for laser-
induced fluorescence spectroscopy in atomic hydrogen. Journal of Optical So-
ciety of America Bbickel, 11, 2155–2162 (1994).
[122] G. A. Bickel and G. A. McRae. Determination of the 1s-2s two-photon exci-
tation cross-section in atomic hydrogen. Spectrochimica Acta Part B: Atomic
Spectroscopy, 55, 57 – 63 (2000).
[123] P. Drell and S. Chu. A megawatt dye laser oscillator-amplifier system for high
resolution spectroscopy. Optics Communications, 28, 343 – 348 (1979).
[124] M. Ashfold, R. Dixon, K. Rosser, R. Stickland and C. Western. Sub-doppler
spectroscopy of the C’ v2 = 2 Rydberg level of ND3 at vacuum ultraviolet
energies. Chemical Physics, 101, 467 – 482 (1986).
[125] L. A. Rahn, R. L. Farrow and G. J. Rosasco. Measurement of the self-
broadening of the H2 Q (0˘5) Raman transitions from 295 to 1000 K. Phys.
Rev. A, 43, 6075–6088 (1991).
[126] M. D. D. Rosa and R. L. Farrow. Cross sections of photoionization and ac
Stark shift measured from Doppler-free B ¬ X(0, 0) excitation spectra of CO.
Journal of Optical Society of America B, 16, 861–870 (1991).
[127] M. Salour. Powerful dye laser oscillator-amplifier system for high resolution
and coherent pulse spectroscopy. Optics Communications, 22, 202 – 206 (1977).
[128] E. S. Lee and J. W. Hahn. Four-pass amplifier for the pulsed amplification of
a narrow-bandwidth continuous-wave dye laser. Optics Letters, 21, 1836–1838
(1996).
[129] J. W. Hahn and Y. S. Yoo. Suppression of Amplified Spontaneous Emission
from a Four-Pass Dye Laser Amplifier. Applied Op, 37, 4867–4870 (1998).
[130] E. Cromwell, T. Trickl, Y. T. Lee and A. H. Kung. Ultranarrow Band-
width Vuv-Xuv Laser System. Review of Scientific Instruments, 60, 2888–2892
(1989).

115
References

[131] J. T. F. Johnston, R. H. Brady and W. Proffitt. Powerful single-frequency

ring dye laser spanning the visible spectrum. Applied Optics, 21, 2307–2316
(1982).
[132] E. Hecht. Optics. Addison Wesley (2002).
[133] V. Blanchet. Femtosecond resolved dynamics in small polyatomic molecules by
velocity map imaging. AIP Conference Proceedings, 1501, 1355–1364 (2012).
[134] M. E. Corrales, G. Gitzinger, J. Gonząlez-Vązquez, V. Loriot, R. de Nalda
and L. Baśares. Velocity Map Imaging and Theoretical Study of the Coulomb
Explosion of CH3I under Intense Femtosecond IR Pulses. The Journal of
Physical Chemistry A, 116, 2669–2677 (2012).
[135] G. Sansone et al. Electron Localization following Attosecond Molecular Pho-
toionization. Nature, 465, 763–766 (2010).
[136] F. Calegari et al. Observation of ultrafast electron dynamics in N2 molecules
induced by attosecond pulses. In CLEO: 2013. Optical Society of America
(2013), QF2C.5.
[137] J. H. Jungmann, A. Gijsbertsen, J. Visser, J. Visschers, R. M. A. Heeren
and M. J. J. Vrakking. A new imaging method for understanding chemical
dynamics: Efficient slice imaging using an in-vacuum pixel detector. Review
of Scientific Instruments, 81, 103112 (2010).
[138] A. T. Clark et al. Multimass Velocity-Map Imaging with the Pixel Imaging
Mass Spectrometry (PImMS) Sensor: An Ultra-Fast Event-Triggered Camera
for Particle Imaging. The Journal of Physical Chemistry A, 116, 10897–10903
(2012).
[139] C. Hock, J. B. Kim, M. L. Weichman, T. I. Yacovitch and D. M. Neumark.
Slow photoelectron velocity-map imaging spectroscopy of cold negative ions.
The Journal of Chemical Physics, 137, 244201 (2012).
[140] Y. Huismans et al. Time-Resolved Holography with Photoelectrons. Science,
331, 61–64 (2011).
[141] K. S. E. Eikema, W. Ubachs, W. Vassen and W. Hogervorst. Lamb shift
measurement in the 1t1 S ground state of helium. Phys. Rev. A, 55, 1866–1884
(1997).
[142] J. Q. Searcy. A supersonic molecular beam metastable atom source initiated
by direct discharge. Review of Scientific Instruments, 45, 589–590 (1974).
[143] A. V. Phelps. Absorption Studies of Helium Metastable Atoms and Molecules.
Physical Review, 99, 1307–1313 (1955).
[144] E. L. Leasure, C. R. Mueller and T. Y. Ridley. ”Hot” metastable atom, molec-
ular beam source. Review of Scientific Instruments, 46, 635–637 (1975).

116
 References

[145] D. W. Fahey, W. F. Parks and L. D. Schearer. High flux beam source of thermal
rare-gas metastable atoms. Journal of Physics E: Scientific Instruments, 13,
381 (1980).

[146] W. Rooijakkers, W. Hogervorst and W. Vassen. An intense collimated beam

of metastable helium atoms by two-dimensional laser cooling. Optics Commu-
nications, 123, 321–330 (1996).
[147] J. Ashmore and R. Sang. Cathode design for a low-velocity metastable neon
cold cathode discharge source. MEASUREMENT SCIENCE & TECHNOL-
OGY, 12, N17–N21 (2001).
[148] F. B. Dunning, R. D. Rundel and R. F. Stebbings. Determination of secondary
electron ejection coefficients for rare gas metastable atoms. Review of Scientific
Instruments, 46, 697–701 (1975).

[149] M. Brieger, H. Bsener, A. Hese, F. v.Moers and A. Renn. Enhancement of

single frequency {SGH} in a passive ring resonator. Optics Communications,
38, 423 – 426 (1981).
[150] R. Drever, J. Hall, F. Kowalski, J. Hough, G. Ford, A. Munley and H. Ward.
Laser phase and frequency stabilization using an optical resonator. Applied
Physics B, 31, 97–105 (1983).
[151] S. Hussain, M. Saleem and M. A. Baig. Measurement of the oscillator strength
distribution in helium. Physical Review A, 76, – (2007).
[152] P. J. J. Tol, N. Herschbach, E. A. Hessels, W. Hogervorst and W. Vassen. Large
numbers of cold metastable helium atoms in a magneto-optical trap. Phys. Rev.
A, 60, R761–R764 (1999).
[153] W. Ubachs, W. Vassen, E. J. Salumbides and K. S. E. Eikema. Precision
metrology on the hydrogen atom in search for new physics. Annalen der Physik,
525, A113–A115 (2013).

[154] J. C.-H. Chen, B. L. Hanson, S. Z. Fisher, P. Langan and A. Y. Kovalevsky. Di-

rect observation of hydrogen atom dynamics and interactions by ultrahigh res-
olution neutron protein crystallography. Proceedings of the National Academy
of Sciences, 109, 15301–15306 (2012).
[155] P. M. Koch and D. R. Mariani. Double-resonance Stark spectroscopy as a probe
of hydrogen substate production in fast atomic collisions. Journal of Physics
B: Atomic and Molecular Physics, 13, L645 (1980).
[156] W. L. Glab and M. H. Nayfeh. Stark-induced resonances in the photoionization
of hydrogen. Physical Review A, 31, 530–532 (1985).
[157] G. Alvarez, R. J. Damburg and H. J. Silverstone. Photoionization of atomic
hydrogen in an electric field. Physical Review A, 44, 3060–3082 (1991).

117
References

[158] V. D. Kondratovich and V. N. Ostrovsky. Resonance and interference phe-

nomena in the photoionisation of a hydrogen atom in a uniform electric field.
II. Overlapping resonances and interference. Journal of Physics B-Atomic
Molecular and Optical Physics, 17, 2011–2038 (1984).
[159] V. D. Kondratovich and V. N. Ostrovsky. Resonance and interference phe-
nomena in the photoionisation of a hydrogen atom in a uniform electric field.
III. Comparison with recent experimental and theoretical results. Journal of
Physics B-Atomic Molecular and Optical Physics, 23, 21–43 (1990).

[160] V. V. Beletsky. Essays on the motion of celestial bodies. Birkhäuser Verlag

(2001).
[161] C. Bordas, F. Lepine, C. Nicole and M. J. J. Vrakking. Semiclassical descrip-
tion of photoionization microscopy. Physical Review A, 68, – (2003).

[162] F. Lepine, C. Bordas, C. Nicole and M. J. J. Vrakking. Atomic photoionization

processes under magnification. Physical Review A, 70, 033417 (2004).
[163] C. Bordas. Classical motion of a photoelectron interacting with its ionic core:
Slow photoelectron imaging. Physical Review A, 58, 400–410 (1998).
[164] C. Blondel, C. Delsart and F. Dulieu. The photodetachment microscope. Phys-
ical Review Letters, 77, 3755–3758 (1996).
[165] C. Nicole, H. L. Offerhaus, M. J. J. Vrakking, F. Lépine and C. Bordas. Pho-
toionization Microscopy. Phys. Rev. Lett., 88, 133001 (2002).
[166] L. B. Zhao and J. B. Delos. Dynamics of electron wave propagation in pho-
toionization microscopy. I. Semiclassical open-orbit theory. Physical Review
A, 81, 053417 (2010).
[167] M. L. Du and J. B. Delos. Effect of Closed Classical Orbits on Quantum
Spectra - Ionization of Atoms in a Magnetic-Field .1. Physical Picture and
Calculations. Physical Review A, 38, 1896–1912 (1988).

[168] J. Gao, J. B. Delos and M. Baruch. Closed-Orbit Theory of Oscillations in

Atomic Photoabsorption Cross-Sections in a Strong Electric-Field .1. Compar-
ison between Theory and Experiments on Hydrogen and Sodium above Thresh-
old. Physical Review A, 46, 1449–1454 (1992).
[169] J. Gao and J. B. Delos. Closed-orbit theory of oscillations in atomic photoab-
sorption cross sections in a strong electric field. II. Derivation of formulas.
Physical Review A, 46, 1455–1467 (1992).
[170] T. S. Luk, L. DiMauro, T. Bergeman and H. Metcalf. Continuum Stark Spec-
troscopy. Physical Review Letters, 47, 83–86 (1981).

118
 References

[171] L. B. Zhao and J. B. Delos. Dynamics of electron wave propagation in photoion-

ization microscopy. II. Quantum-mechanical formulation. Physical Review A,
81, 053418 (2010).

[172] S. Cohen, M. M. Harb, A. Ollagnier, F. Robicheaux, M. J. J. Vrakking, T. Bar-

illot, F. Lépine and C. Bordas. Wave Function Microscopy of Quasibound
Atomic States. Physical Review Letters, 110, 183001 (2013).
[173] H. Wimmel. Quantum Physics & Observed Reality: A Critical Interpretation
of Quantum Mechanics. World Scientific, Singapore (1992).

[174] C. Jönsson. Elektroneninterferenzen an mehreren künstlich hergestellten

Feinspalten. Zeitschrift für Physik, 161, 454–474 (1961).
[175] L. D. Noordam and R. R. Jones. Probing Rydberg electron dynamics. Journal
of Modern Optics, 44, 2515–2532 (1997).

[176] M. L. Du. Oscillations of electron flux in photodetachment of H− in an electric

field. Physical Review A, 40, 4983–4987 (1989).
[177] C. Delsart, L. Cabaret, C. Blondel and R. J. Champeau. Two-step high-
resolution laser spectroscopy of the Stark substates of the n=44 level in atomic
hydrogen. Journal of Physics B: Atomic and Molecular Physics, 20, 4699
(1987).
[178] H. Wang, M. Chini, S. Chen, C.-H. Zhang, F. He, Y. Cheng, Y. Wu, U. Thumm
and Z. Chang. Attosecond Time-Resolved Autoionization of Argon. Physical
Review Letters, 105, 143002 (2010).

[179] C. T. L. Smeenk. A New Look at the Hydrogen Wave Function. Physics, 6, 58

(2013).
[180] C. A. Kocher. Vanishing of Dipole Matrix Elements at Level Crossings. Phys-
ical Review A, 6, 35–37 (1972).

119
Summary

The development of QuantumM echanics in the early 20th century provided

a new mathematical tool to describe phenomena occurring at the nano-scale and
changed the way that scientists understand nature. Although the theory gives a
qualitative explanation of the effects taking place on the nano-scale, it contains nu-
merous elements that cannot be classically intuited, such as wave-particle duality,
the Heisenberg uncertainty principle, the Pauli exclusion principle etc. A fundamen-
tal role in quantum mechanics plays the wave function ψ, which satisfies the time-
dependent Schrödinger equation and provides all information about the dynamical
properties of a quantum system at a given instant of time t. From the theoretical
point of view it is possible to predict what a wave function is, but a measurement
of a wave function remains tough due to its exquisite fragility. Strictly speaking,
one of the peculiarities of quantum mechanics is that most attempts to directly ob-
serve a wave function actually destroy it in a process called collapse. Despite this
difficulty, in the last decades many direct and indirect experimental methods were
proposed resulting in wave function imaging of such quantum objects as photons,
nuclei, atoms and molecules (Chapter 1).
In this thesis a novel technique – photoionization microscopy – is exploited to
directly observe electronic wave functions in atoms like hydrogen and helium. This
alternative imaging technique, proposed more than 30 years ago, makes use of a
unique property of the hydrogen atom placed in a static electric field, namely, the
symmetry of its Hamiltonian operator, which ensures that the Schrödinger equa-
tion is separable in parabolic coordinates (ξ, η, ϕ). Consequently, the wave function
may be written as: ψ(η, ξ, ϕ) = √ 1 χ1 (ξ)χ2 (η)eimϕ , where χ1 (ξ) and χ2 (η) are
2πξη
functions describing the electron motion along the ξ and η coordinate, respectively,
and have n1 , n2 nodes, where n1 , n2 are so-called parabolic quantum numbers. As
shown explicitly in Chapter 2, for each coordinate there exists a certain effective po-
tential acting on the electron and resulting in motion that is always finite along the
ξ coordinate. Hence, for the ionization to occur the electron must escape to η = ∞,
which is possible in two situations: (1) if it has enough energy to surmount the top
of the V (η) potential or (2) when the energy lies below the top of the V (η) potential,
and when the electron has a nonzero probability of tunneling through the barrier.
In the latter case the ideal conditions for photoionization microscopy are created
since the tunneling process leaves the quantization along the ξ coordinate intact.
The theory predicts that a projection of such photoelectrons onto a two-dimensional
detector placed perpendicularly to the static electric field would allow the experi-
mental measurement of interference patterns directly reflecting the nodal structure
of the electronic wave function χ1 (ξ).
In the framework of this thesis the validity of photoionization microscopy as a

121
Summary

technique suitable for imaging electronic wave functions is tested for two, fundamen-
tally different (from the point of view of quantum mechanics) atoms: hydrogen and
helium. The main purpose of Chapter 2 is to present theoretical and experimental
consequences resulting from the finite size of the ionic core in helium and other non-
hydrogenic atoms in contrast to hydrogen, whose core contains only a single proton.
It is shown that the non–Coulombic potential of nonhydrogenic atoms leads to cou-
pling among Stark states. Consequently, so-called avoided crossings arise between
levels that in hydrogen would cross and an additional type of field ionization mecha-
nism occurs: (3) when the electron’s energy lies below the top of the V (η) potential
the electron can be scattered by the core into another parabolic state, which lies
above the barrier in V (η), resulting in a process similar to autoionization. However,
this type of field ionization, which is a dominant decay mechanism in nonhydro-
genic atoms, removes the quantization along the ξ coordinate and by that very fact
precludes a measurement of the electronic wave function in nonhydrogenic atoms.
Nevertheless, as shown in Chapter 2, a peculiar effect takes place in the vicinity of
avoided crossings, so-called interference narrowing, which alters the ionization pro-
cess and in principle may permit a direct observation of the nodal structure in χ1 (ξ),
as explored in Chapter 6.
To turn the idea of mapping wave functions with photoionization microscopy
into experimental reality was by no means trivial. According to the theoretical de-
scription a ”photoionization microscope” should comprise two plates producing a
homogeneous electric field, a beam of neutral atoms, a focused laser beam that pho-
toionizes the atoms, and a position-sensitive detector placed perpendicular to the
electric field axis, on which the resulting photoelectrons are projected. Chapter 3
provides a detailed description of the experimental realization of this concept, sep-
arately for hydrogen and for helium, since the intrinsic differences between these
atoms resulted in designing two distinct experimental setups. The hydrogen atom,
in contrast to helium, does not exist as a chemically stable species in nature and
thus in the experiments it had to be produced by means of photodissociation of a
H2 S molecule. High ionization thresholds in both atoms excluded the original idea
of one-photon excitation from the ground state into high lying Stark states. There-
fore, alternative photoionization schemes were used, which took advantage of atoms
prepared in metastable states prior to being photoexcited. For helium, which as a
noble gas has metastable states resistant to a DC electric field, a discharge source
providing a high-flux beam of (21 S) metastable helium atoms was built. Since hy-
drogen does not have the same property it was photoexcited into n = 2 states,
by absorbing two photons at 243 nm, in the same interaction region where it was
subsequently photoexcited into Stark states by absorbing one photon of UV light
delivered by a home-built, tunable and narrow-band pulsed dye amplifier. The final
photoexcitation step in the helium experiments was accomplished by making use of
one photon of UV light produced by frequency doubling the output of a cw ring dye
laser. To fulfill the requirements regarding the homogeneous electric field and the
position-sensitive detector, a velocity map imaging spectrometer was used. Inside
this device a DC electric field was created by applying a voltage difference across
the electrodes, which acted as an electrostatic lens enabling to focus the created pho-

122
 Summary

toelectrons sharply on the detector. In our experiments the detector consisted of two
micro-channel plates, a phosphor screen and a CCD camera. The last crucial element
in both setups was an Einzel lens, which ensured magnification of recorded photo-
electron images by one order of magnitude. The calculated efficiencies of the overall
process in both experiments are given in Chapter 3 together with the description of
the main limiting factors in both systems.
During the photoionization microscopy experiments hydrogen and helium atoms
were photoionized in the presence of DC electric fields and the resulting photoelec-
trons were projected on a detector, creating characteristics oscillatory structures.
In the experiments where atoms were photoexcited into the continuum of Stark
states, i.e. when the field ionization process was via mechanism (1), the observed
interference patterns are fully understood considering the phase accumulation along
different trajectories that allow the photoelectrons to reach the same position on the
two-dimensional detector. In Chapter 4 the experimental results for hydrogen atoms
are presented, where the aforementioned interference patterns are observed and ex-
plained within the framework of semiclassical theory. In the recorded photoelectron
images distinct oscillatory structures are distinguished, related to so-called ”direct”
and ”indirect” ionization processes, which differ in the circumstances under which
the photoelectron is launched, i.e. whether it interacted with the parent ion or not.
For the first time it was observed that direct and indirect trajectories lead to beating
patterns in the inner part of the photoelectron images. This effect was studied in
greater detail and was very well reproduced by the semiclassical calculations.
According to the theory of photoionization microscopy, in the special case when
the photoionization follows a resonant excitation into a Stark state, the observed
interference patterns directly reflect the nodal structure of the Stark state prepared
in the experiment. This resonant effect is extensively investigated in Chapter 5,
where the photoelectron images recorded for hydrogen atoms excited into various
Stark states are presented. Remarkably, the number of observed dark fringes equals
to n1 , i.e. the number of nodes of the electronic wave function χ1 (ξ), irrespective
of the laser polarization or the electric field strength. To confirm that the observed
interference structures are caused by the resonant effect, photoelectron images were
recorded also for energies lying below and above the Stark state. For these non-
resonant cases the interference patterns have less rings and their origin can be fully
understood considering quantum-mechanical path length differences experienced by
the photoelectron along multiple interfering pathways, as presented in Chapter 4.
Moreover, the image recorded at the resonance extends radially further indicating an
involvement of a quantum effect since the photoelectrons reach classically forbidden
positions on the detector. This observation confirms that the ionization mechanism
was via tunneling through the potential barrier in V (η) keeping the quantization
along the ξ coordinate intact. Finally, an excellent agreement between the exper-
imental results and the quantum mechanical calculations was achieved, ipso facto
validating the theoretical predictions made more than three decades ago and showing
a new technique enabling the direct observations of electronic wave functions.
The nature of the ionization process governs whether the microscopic wave func-
tion χ1 (ξ), that exists in the vicinity of the atom, and the projection of the continuum

123
Summary

wave function, measured at a macroscopic distance, share the same nodal structure.
In nonhydrogenic atoms the ionization occurs predominantly via the process (3),
which prevents a successful measurement of electronic wave functions. However, this
limitation can be overcome by enforcing a hydrogenic behavior either by populating
states with small quantum defects or by taking a measurement in the vicinity of an
avoided crossing, where the interference narrowing effect occurs. The latter approach
was used in the experiments on helium atoms, whose results are presented and dis-
cussed in Chapter 6. 95% of the recorded images showed interference patterns similar
to those presented in Chapter 4, hence without any imprint of the wave function,
and in several cases, coinciding with the occurrence of the interference narrowing
effect, the presence of the nodal structure of χ1 (ξ) was revealed. By that very fact
it was demonstrated that photoionization microscopy may be a technique suitable
for imaging wave functions in nonhydrogenic atoms. In Chapter 6 it is also shown
that a ”photoionization microscope” can be applied as a tool in the investigations
of couplings between Stark states.

124
Samenvatting

De ontwikkeling van de kwantummechanica in het begin van de twintigste eeuw

zorgde voor nieuw wiskundig gereedschap om natuurlijke fenomenen te beschrijven
die zich op de nano-schaal afspelen en het veranderde de manier waarop wetenschap-
pers de natuur begrijpen. Hoewel de theorie een kwalitatieve beschrijving geeft voor
effecten die op de nano-schaal plaatsvinden, bevat het ook verscheidene elementen
die klassiek niet intuı̈tief af te leiden zijn, zoals de golf-deeltjes dualiteit, het onzeker-
heidsprincipe van Heisenberg, het Pauli uitsluitingsprincipe enzovoorts. Een funda-
mentele rol in de kwantummechanica is weggelegd voor de golffunctie ψ, welke moet
voldoen aan de tijdafhankelijke Schrödingervergelijking en alle informatie bevat over
de dynamische eigenschappen van een kwantumsysteem op elk gegeven moment in
tijd t. Vanuit theoretisch oogpunt is het mogelijk een golffunctie te beschrijven, maar
een golffunctie experimenteel te observeren blijft uitdagend door zijn extreme gevoe-
ligheid. Één van de eigenaardigheden van de kwantummechanica is dat de meeste
pogingen om een golffunctie direct te observeren deze daadwerkelijk vernietigen in
een proces dat collapse (ineenstorting) wordt genoemd. Toch zijn er in de afgelopen
decennia vele directe en indirecte experimentele methodes voorgesteld, wat geresul-
teerd heeft in het afbeelden van de golffunctie van kwantumobjecten zoals het foton,
atoomkernen, atomen en molekulen (Hoofdstuk 1).
In dit proefschrift wordt een nieuwe techniek – fotoionisatie-microscopie - aan-
gewend om de elektrongolffunctie in atomen zoals waterstof en helium direct te
observeren. Deze alternatieve techniek, die meer dan 30 jaar geleden is voorgesteld,
maakt gebruik van een unieke eigenschap van het waterstof atoom dat zich in een
statisch elektrisch veld bevindt. De symmetrie van zijn Hamiltoniaanse operator
maakt het namelijk mogelijk dat de Schrödingervergelijking geschreven kan worden
in parabolische coördinaten (ξ, η, ϕ). Hierdoor kan de golffunctie geschreven worden
als: ψ(η, ξ, ϕ) = √ 1 χ1 (ξ)χ2 (η)eimϕ , waar χ1 (ξ) en χ2 (η) functies zijn die de elek-
2πξη
tronenbeweging langs de ξ en η coördinaat beschrijven en respectievelijk n1 en n2
knopen hebben. De getallen n1 en n2 zijn de zogenoemde parabolische kwantumge-
tallen. Zoals expliciet is aangetoond in hoofdstuk 2, bestaat er voor elke cordinaat
een zekere effectieve potentiaal die op de elektronen werkt en resulteert in beweging
langs de ξ coördinaat die altijd eindig is. Om ionisatie plaats te laten vinden moet
het elektron daarom ontsnappen naar η = ∞. Dit is in twee gevallen mogelijk: (1) als
het elektron genoeg energie heeft om de top van de V (η) potentiaal te overbruggen
of (2) wanneer de energie lager is dan het maximum van de V (η) potentiaal, maar
er een eindige waarschijnlijkheid is om door de barriére te tunnelen. In het laat-
ste geval worden de ideale omstandigheden gecreëerd voor fotoionisatie-microscopie,
omdat het tunnelproces de kwantisatie langs de ξ-coördinaat intact laat. De theorie
voorspelt dat een afbeelding van zulke getunnelde foto-elektronen, op een tweedi-

125
Samenvatting

mensionale detector die loodrecht op de richting van het statisch elektrisch veld is
geplaatst, experimenteel een interferentiepatroon zou vertonen dat direct de knoop-
structuur van de elektronengolffunctie χ1 (ξ) laat zien.
In dit proefschrift wordt de geldigheid van fotoionisatie-microscopie, als geschikte
techniek voor het afbeelden van elektrongolffuncties, getest op twee fundamenteel
verschillende (vanuit het oogpunt van de kwantummechanica) atomen: waterstof en
helium. Het voornaamste doel van hoofdstuk 2 is om theoretische en experimen-
tele gevolgen te presenteren van het eindige formaat van de ionenkern in helium
en andere niet-waterstofachtige atomen. In vergelijking met waterstof, waar de kern
slechts één enkel proton bevat. Het wordt aangetoond dat de niet–Coulombische
potentiaal van niet-waterstofachtige atomen leidt tot koppelingen tussen Stark toe-
standen. Dit heeft tot gevolg dat zogenoemde vermeden kruisingen opkomen tussen
niveaus die in waterstof zouden kruisen en een nieuw soort veldionisatiemechanisme
plaatsvindt. Mechanisme (3): wanneer de energie van het elektron lager is dan het
maximum van de V (η) potentiaal, kan het elektron door de kern verstrooid wor-
den in een andere parabolische toestand. Wanneer de energie van deze parabolische
toestand boven de V (η) barriére ligt, resulteert dit in een proces dat vergelijkbaar
is met auto-ionisatie. Echter, dit type veldionisatie, welke een dominant vervalme-
chanisme is in niet-waterstofachtige atomen, verwijdert de kwantisatie langs de ξ-
coördinaat en verzet zich daardoor tegen een meting van de elektronen golffunc-
tie in niet-waterstofachtige atomen. Toch, zoals getoond in hoofdstuk 2, vindt er
een bijzonder effect plaats in de buurt van vermeden kruisingen. Dit zogenoemde
interferentieversmalling-effect, verandert het ionisatieproces en staat in principe een
directe observatie van de knoopstructuur in χ1 (ξ) toe, zoals bestudeerd wordt in
hoofdstuk 6.
Het idee van het afbeelden van golffuncties met fotoionisatie-microscopie in expe-
rimentele realiteit omzetten was zeker geen triviale opgave. Volgens de theoretische
beschrijving zou een ”fotoionisatie-microscoop”moeten bestaan uit twee platen die
een homogeen elektrisch veld produceren, een straal van neutrale atomen, een ge-
focusseerde laserstraal die de atomen fotoioniseert, en een positiegevoelige detector
die loodrecht op de elektrische veldrichting staat waar de fotoelektronen op kun-
nen worden geprojecteerd. Hoofdstuk 3 biedt een gedetailleerde beschrijving van de
experimentele realisatie van dit concept, voor zowel waterstof als voor helium, aan-
gezien de intrinsieke verschillen tussen deze atomen resulteerden in het ontwerpen
van twee afzonderlijke experimentele opstellingen. Het waterstofatoom, in tegenstel-
ling tot helium, bestaat niet als een chemisch stabiele soort in de natuur en daarom
moest het in de experimenten geproduceerd worden door fotodissociatie van een H2 S
molekuul. Hoge ionisatiebarriéres in de beide atomen sloten het originele idee van
één-foton-excitatie vanuit de grondtoestand naar een hooggelegen Stark-toestand uit.
Daarom zijn alternatieve foto-ionisatie concepten aangewend, welke gebruik maken
van geprepareerde atomen die zich in metastabiele toestanden bevinden voordat ze
foto-gexciteerd worden. Voor helium, wat als edelgas metastabiele toestanden heeft
die immuun zijn voor een DC elektrisch veld, werd een ontladingsbron gebouwd die
een straal van metastabiele (21 S) heliumatomen leverde. Aangezien waterstof niet
dezelfde eigenschap heeft, werd het gexciteerd naar een n = 2 toestand door het ab-

126
 Samenvatting

sorberen van twee fotonen met een golflengte van 243 nm. In dezelfde regio werd het
vervolgens gexciteerd naar een Stark-toestand door het absorberen van één UV fo-
ton dat geleverd werd door een huisgemaakte, verstelbare, smalbandige, gepulseerde
kleurstof versterker. De laatste foto-excitatie stap in de experimenten aan helium
werd gedaan door gebruik te maken van een UV foton dat geproduceerd wordt door
frequentieverdubbeling van licht uit een CW ring kleurstof laser. Om aan de eisen
voor de homogeniteit van het elektrisch veld en de positiegevoelige detector te vol-
doen, werd gebruik gemaakt van een velocity map imaging spectrometer (VMIS).
In dit apparaat werd een DC veld opgewekt door het aanleggen van een voltage-
verschil tussen de beide elektrodes, welke fungeerden als een elektrostatische lens
die het mogelijk maakte om de gecreëerde fotoelektronen scherp te focusseren op de
detector. In onze experimenten bestond de detector uit twee micro-channel plates
(MCPs), een fosfor scherm en een CCD camera. Het laatste cruciale element in beide
opstellingen was een Einzel-lens, welke zorgde voor een vergroting van de opgeno-
men fotoelektronen afbeeldingen van één ordegrootte. De berekende efficiëntie van
het gehele proces in beide experimenten is gegeven in hoofdstuk 3, samen met de
beschrijving van de voornaamste limiterende factoren in beide systemen.
Tijdens de fotoionisatie-microscopie-experimenten werden waterstof- en helium-
atomen foto-geı̈oniseerd in de aanwezigheid van DC elektrische velden en de resulte-
rende fotoelektronen werden op een detector geprojecteerd, wat karakteristiek oscil-
lerende structuren creëerde. In de experimenten waar atomen door foto-excitatie in
het continum van Stark-toestanden werden gexciteerd, oftewel wanneer het veldioni-
satieproces verliep via mechanisme (1), zijn de geobserveerde interferentiepatronen
volledig begrepen. Hiervoor dient namelijk de faseaccumulatie, langs de verschillende
banen die de fotoelektronen toestaan de tweedimensionale detector op dezelfde plaats
te bereiken, in acht worden genomen. In hoofdstuk 4 worden de experimentele resul-
taten voor de waterstofatomen gepresenteerd, waar de genoemde interferentiepatro-
nen worden geobserveerd en uitgelegd binnen het kader van de semiklassieke theorie.
In de opgenomen afbeeldingen van de fotoelektronen worden verschillende oscille-
rende structuren onderscheiden die gerelateerd zijn aan de zogenoemde ”directeën
ı̈ndirecteı̈onisatieprocessen. Deze verschillen in de omstandigheden waaronder het
fotoelektron wordt gelanceerd, met andere woorden of het interageerde met het ion
of niet. Hiermee is voor de eerste keer waargenomen dat directe en indirecte banen
leiden tot golvende patronen in het centrale gedeelte van de afbeeldingen van de
fotoelektronen. Dit effect is verder in detail bestudeerd en zeer goed gereproduceerd
door semiklassieke berekeningen.
Volgens de theorie achter fotoionisatie-microscopie toont het geobserveerde inter-
ferentiepatroon, in het bijzondere geval dat de fotoionisatie volgt op een resonante
excitatie naar een Stark-toestand, direct de knoopstructuur van de Stark-toestand
die in het experiment wordt voorbereid. Dit resonante effect is uitgebreid onder-
zocht in hoofdstuk 5, waar de opgenomen afbeeldingen van de fotoelektron voor
waterstofatomen, die geëxciteerd zijn naar verschillende Stark-toestanden, worden
gepresenteerd. Opvallend genoeg is het aantal geobserveerde donkere ringen gelijk
aan n1 , oftewel aan het aantal knopen van de elektrongolffunctie χ1 (ξ), ongeacht de
laserpolarisatie of de sterkte van het elektrisch veld. Om te bevestigen dat de geob-

127
Samenvatting

serveerde interferentiestructuren worden veroorzaakt door het resonante effect, zijn

afbeeldingen van de fotoelektronen opgenomen voor energien die onder en boven de
Stark-toestand lagen. In deze niet-resonante gevallen hebben de interferentiepatro-
nen minder ringen en kan hun herkomst volledig worden begrepen. Hiervoor dient
men kwantummechanische padlengteverschillen in acht te nemen die de het fotoelek-
tron ervaart langs de verschillende interfererende paden, zoals gepresenteerd wordt
in hoofdstuk 4. Verder moet vermeld worden dat de opgenomen afbeeldingen zich
radiaal verder uitbreiden, wat lijkt te wijzen op betrokkenheid van een kwantum ef-
fect, aangezien de fotoelektronen klassiek verboden posities op de detector bereiken.
Deze waarneming bevestigt dat het ionisatie mechanisme verliep via het tunnelen
door de potentiaal barriére in V (η), waardoor de kwantisatie langs de ξ-coördinaat
behouden blijft. Tot slot is er een uitstekende overeenkomst bereikt tussen de expe-
rimentele resultaten en de kwantummechanische berekeningen. Dit bekrachtigt ipso
facto de theoretische voorspellingen van drie decennia geleden en toont aan dat deze
techniek een directe waarneming van elektrongolffuncties mogelijk maakt.
De aard van het ionisatieproces bepaalt of de microscopische golffunctie χ1 (ξ),
die bestaat in de nabijheid van het atoom, en de afbeelding van de continumgolf-
functie, gemeten op een macroscopische afstand, dezelfde knoopstructuur hebben.
In niet-waterstofachtige atomen vindt ionisatie voornamelijk plaats via mechanisme
(3), wat een succesvolle meting van de golffunctie voorkomt. Echter, deze beperking
kan overwonnen worden door een waterstofachtig gedrag op te leggen. Dit kan ofwel
door het vullen van toestanden met kleine kwantumdefecten of door een meting te
doen in de buurt van een vermeden kruising, waar het interferentie versmalling effect
plaatsvindt. De laatstgenoemde aanpak is gebruikt in de experimenten aan helium-
atomen, waarvan de resultaten worden gepresenteerd en bediscussieerd in hoofdstuk
6. Ongeveer 95% van de opgenomen afbeeldingen toont interferentiepatronen verge-
lijkbaar met die uit hoofdstuk 4, maar dan zonder enige afdruk van de golffunctie. In
verschillende gevallen vindt ook het interferentieversmalling-effect plaats, waarmee
de aanwezigheid van de knoopstructuur van χ1 (ξ) was ontdekt. Door dat simpele
feit is het aangetoond dat fotoionisatie-microscopie mogelijk een geschikte techniek
is voor het afbeelden van golffuncties in niet-waterstofachtige atomen. In hoofdstuk
6 wordt ook getoond dat een ”fotoionisatie-microscoop”kan worden aangewend als
techniek in het onderzoek naar koppelingen tussen Stark-toestanden.

128
Acknowledgements

In my life I experienced a variety of exotic and astonishing travels, but my PhD

was the journey that cost me the most energy, led to countless ups and downs and
brought all kinds of emotions in return: from total hopelessness to entire satisfaction
and happiness. During these few years of my doctorate I met numerous inspiring
and helpful people without whom this journey would have never ended successfully.
On top of that a sincere support of my family and friends helped me immensely
during some rough times. Therefore, the last pages of this book I devote to these
essential people and hopefully in a few simple thank words I will express my honest
gratefulness to all of them.
First of all, I would like to thank my supervisor Marc Vrakking for giving me
an opportunity to start a PhD in his group at AMOLF despite our ”disagreement”
during the job interview. Marc, I am grateful for your confidence in entrusting me
with the photoionization microscopy project, which turned out to be a true chal-
lenge and at a certain point almost an impossible task, but with your motivation
and support it became a world-wide success. I was really lucky to had you as my
supervisor since you always had time for discussion, ridiculously even more when
you moved to Berlin, had always plan ”B” or even ”C” when things were getting off
the trajectory and have such a versatile knowledge of science and technology that
there was always a solution to my problems. During my PhD I have learnt a lot
from you and I appreciate the chance of climbing M ount Everest together – it was
a pleasure and honor to work with you.
My PhD project would have never been successful without the supporting staff of
AMOLF, which is one of its kind and ensures smooth and pleasant research. I would
like to sincerely thank Rob for being the best technician a PhD student could wish
for. Rob, I am thankful not only for your exceptional assistance in the lab, which was
always quick and efficient, but also for your kindness, life-wisdom and useful advices.
You are a wonderful person and I wish you all the best during your retirement. I want
to thank Hinco for being our unofficial group technician on Fridays and for proving a
helping hand despite lack of time. Moreover, I am thankful to Ad for teaching me the
first steps with the ring-dye laser and for his sincere interest in my research even after
he left the group. My experiments would not have been successful without state-of-
the-art machines, which were designed and built at AMOLF. Therefore, I wish to
thank the Design Department, especially Iliya for his accessibility and creativity in
providing emergency solutions to my numerous disasters in the lab. Many thanks
to Henk-Jan for his patience with creating beautiful figures of experimental setups
for our publications and this thesis. The Mechanical Workshop in AMOLF is simply
the best in manufacturing high quality parts thanks to people working there, in
particular Henk, Menno and Niels, and their great manager Wim. Thank you all

129
Acknowledgements

for an incredible cooperation in the last few years when you never said NO to my
sophisticated requests and always delivered adequate parts in no time. From the
Electronic Engineering I would like to thank Duncan for his assistance in many
projects, Ronald for countless repairs of the control box and making kilometers of
cables for my experiments, Idsart for his support with high voltage elements and
Henk for solving the biggest problem during my PhD. Many thanks to the Software
Engineering, in particular to Marco Konijnenburg for making my work with Lisa
Cluster completely effortless and to Marco Seynen for ensuring connection to my
very old laser control computer. Sincere thanks to the ICT department, especially
to Carl, Jan, Rutger and Wiebe for their excellent assistance during all these years.
Finally, I would like to acknowledge people from other supportive departments in
AMOLF: Anuok, Arnelli, Erny, Grace, Tatiana, Silvia, Willem, Wouter for their
support, kind words and smile in AMOLF’s corridors.
The work presented in this thesis is a result of collaboration with several brilliant
scientists and researchers with whom I had the pleasure to work over last few years.
Franck Lépine deserves honest thanks for his great contribution to the hydrogen
experiment and assistance afterwards. Franck, I am grateful for the time spent to-
gether in the lab, your wise suggestions and advises, many theoretical discussions
and black humor that made me laugh despite tiredness and late hours. I would like
to acknowledge Francis Robicheaux for developing the wave packet code, which I
used for many calculations presented in this thesis, as well as for explaining its com-
plicated theory. Many thanks to Christian Bordas for his input to our publications.
Moreover, I would like to express my gratitude to Tom Bergeman for his indispen-
sable assistance in the interpretation of He photoionization spectra, especially for
bringing back to life his old computer code, the final version of which got lost many
years ago. Tom, I wish I could meet you in person to thank you for all those do-
zens of e-mails that we exchanged over last few years. Finally, I would like to thank
Wim Vassen for his precious hints on designing and building the metastable helium
source.
After my first year of PhD I became a part of the ICONIC network where I had
a chance to learn a lot about different applications of the velocity map imaging tech-
nique. I would like to acknowledge the funders for giving me the opportunity to be
a member of this organization and for numerous meetings and schools during which
I could share my knowledge with other PhD students and postdocs. In particular I
would like to thank my mentor Benjamin Whitaker for his interest in my research
and stimulating discussions. I am thankful to Dave Parker for his warm welcome and
to Wim van der Zande for his challenging questions which motivated me to broaden
my knowledge. Many thanks to André Eppink for arranging all these great meetings,
schools and dinners. Lastly, I wish to thank all other members of ICONIC for great
times we had together, for their support and friendship.
People with whom I worked every day – my terrific colleagues – had a huge
impact on my PhD project. From the XUV group in AMOLF I wish to thank Julia
for being an amazing companion who truly shared all my ups and downs during all
these years. Julia, I am thankful for your compassion, helping hand in the lab, hours
of discussions, wise advises and suggestions, and of course our morning gossip sessions

130
 Acknowledgements

and passion for chocolate ;-) You turned out to be not only the best office mate,
super-helpful colleague, but also a true friend I could always count on. Many thanks
to Arjan who taught me useful tricks in the lab boosting my confidence in performing
experiments and gave me the best advice during my whole PhD. Arnaud, thank you
for coming from Berlin and fixing my experiment within 48 hours as you promised!
Ymkje, the Felice campaign under your supervision was simply the best as three of us
created an unbeatable team of girls that performed a superb experiment despite all
doubts of some of our male colleagues. Thank you for that great experience and for
always being super helpful, friendly and very funny ;-) Many thanks to Georg, Freek
and Wing Kiu for all inspiring discussions, exciting group uitjes, delicious dinners,
and great parties – especially the one with terrible sangria! I wish to express my
gratitude to Jesse for translating my summary into Dutch and also for going out
with me for countless dinners in Berlin. The latter also applies to my other amazing
colleagues from Berlin: Axel, Christian, ChungHsin, Faruk, Fede and Martin. Thank
you all for being fantastic hosts and your efforts at showing me the city – especially
by night! Moreover, Fede thank you one more time for all your suggestions and tips
regarding our holidays in Argentina and Faruk for advising us Cay Caulker which
turned out to be one of the best places we visited during our RTW trip!
Halfway through my PhD the whole XUV group moved to Berlin while I stayed
in Amsterdam for the next couple of years. I am grateful to Huib Bakker, who
became my official supervisor, for always finding time to discuss progress in my PhD
project. Moreover, I wish to thank Ron Heeren for ”adopting” me into his group
so I had never felt redundant or alone. Moreover, many thanks to people from the
BIMS group for always being enjoyable and friendly. In particular I am thankful
to Andras, Don, Julia, Gert and Marc for all those fantastic lunches that we had
together. Finally, to all my colleagues I forgot to mention – thanks a lot!
The final year of my PhD was really tough for me due to some health problems.
However, I was lucky to meet great specialists who spent months on massaging and
mobilizing my arm so I was able to work with a computer again. I would like to
acknowledge Gerard Vierbergen for his patience in fixing all my new injuries, for
his optimism and great stories which were always cheering me up. Many thanks
to Jolanda van der Peet for very painful dry needling therapy which helped me
enormously. I am thankful to Marty Cook for his alternative treatments which were
restoring a balance in my body and mind.
The time of my PhD would not be so enjoyable without my awesome friends - old
ones from Poland and new ones that I have met in the Netherlands. Many thanks to
absolutely wonderful friends from Alkmaar: Claire and Marc, Ingrid and Hans, Ni-
colas, Gaga and Ben, Ellen and Tom, Raimo, Anna and Annelies, Lydia and Pichon,
Virginie and Alain, Soledad, Maria and Goulven for all delicious dinners, BBQs,
drinks in Bier M useum, litters of La Chouffe in P rovadja, poker nights, Sinter-
klaas celebrations, Gondelvaarten in Koedijk and mostly for our crazy parties in the
infamous bake-shed! Gorące podziękowania dla naszej wspaniałej, aczkolwiek nielic-
znej Polonii z południowo-centralnej Holandii: Magdy i Alicji. Dziewczyny, dzięki
wielkie za wszystkie szalone weekendy spędzone razem, zapoznanie nas z elitarnym
klubem International, a przede wszystkim za Wasze zrozumienie oraz wsparcie w

131
Acknowledgements

kwestiach polsko-holenderskich. Chciałabym również podziękować wyjątkowej gru-

pie przyjaciół z Gdańska: Ani, Mai, Stachowi, Jackowi, Agacie, Michałowi, Maćkowi,
Soni, Martynce i Piotrkowi za wszystkie ciepłe powitania i pożegnania w Stacji De
Luxe, Wasze wizyty w Alkmaar, wspólne wyjazdy na narty bądź do Ekwadoru lub
Malezji oraz za fakt pozostawania w bardzo dobrych kontaktach pomimo setek ki-
lometrów które nas dzielą! Moim szalonym i cudownym przyjaciółkom z podwórka:
Ani, Agnes, Kasi i Judycie jestem wdzięczna za ich wsparcie oraz wszystkie cudowne
chwile spędzone razem.
Z całego serca pragnę podziękować moim dwóm wspaniałym rodzinom: Smół-
kowskim i Stodólnym za ich nieustającą pomoc, wyjątkowe wsparcie oraz gorący
doping w trakcie mojego doktoratu. Mamie Danusi i tacie Leszkowi jestem ogromnie
wdzięczna za ich trud w wychowaniu mnie na osobę ambitną, która nieprzerwanie
dąży do wyznaczonego celu, ponieważ bez tej silnej cechy charakteru (częściowo od-
ziedziczonej po Was) nigdy nie osiągnęłabym tak wiele. Ta książka jest poniekąd owo-
cem Waszej ciężkiej pracy w zapewnieniu mi najlepszego wykształcenia. Ogromne
podziękowania dla mamy Ewy i taty Krzysia za ich gorące przyjęcie do rodziny, zain-
teresowanie w postępach mojego doktoratu, liczne toasty za wszelkie sukcesy jak i
również za bardzo ciekawe dysputy naukowe oraz gorliwe pytania odnoście ”moich
atomów”. Mojej kochanej siostrzyczce Inusi chciałabym podziękować za jej cudowne
wsparcie w trudnych chwilach, szczerą radość z moich sukcesów oraz za bardzo do-
bre porady techniczno-naukowe jakich nie spodziewałabym się po tak młodej osobie
- jednym słowem jest z Ciebie bardzo dobry materiał na naukowca! Ogromny bu-
ziak dla Moni za nadanie ostatnich szlifów angielskiej wersji podziękowań. Arkowi
i Asi dziękuję za liczne odwiedziny w Holandii. Wszystkim pozostałym Smółkom i
Stodólniakom dziękuje również gorąco!
Last but not the least I wish to thank my husband Maciej for being the best
travel buddy, a wonderful life partner and an outstanding scientist with whom I had
many stimulating discussions regarding my research. Maciej, the most I am thankful
for your remarkable support in the last year of my PhD, when I had to stop working
for a few months and you overtook all my household duties (every wife dreams of
that). Moreover, you became my personal secretary who typed dozens of e-mails and
to spare me many working hours with a computer you edited figures for this thesis
and put the whole text into LATEX. Without your help, advice, patience and love I
would have never finished this book and therefore I dedicate it to you.

Aneta S. Stodolna, Southeast Asia, April 2014

132
List of publications

ˆ Visualizing the coupling between red and blue Stark states using photoioniza-
tion microscopy, A.S. Stodolna, F. Lépine, T. Bergeman, F. Robicheaux, A.
Gijsbertsen, J.H. Jungmann, C. Bordas and M.J.J. Vrakking, (submitted for
publication).
ˆ Photoelectron holography in strong optical and dc electric fields, A.S. Stodolna,
Y. Huismans, A. Rouzée, F. Lépine and M.J.J. Vrakking, (to be published in
the Journal of Physics: Conference Series).

ˆ Hydrogen atoms under magnification: Direct Observation of the Nodal Struc-

ture of Stark States, A.S. Stodolna, A. Rouzée, F. Lépine, S. Cohen, F. Ro-
bicheaux, A. Gijsbertsen, J.H. Jungmann, M.J.J. Vrakking, Phys. Rev. Lett.
110, 213001 (2013).
ˆ Photoelectron angular distributions from the ionization of xenon Rydberg states
by midinfrared radiation, Y. Huismans, A. Rouzée, A. Gijsbertsen, P. S. W. M.
Logman, F. Lépine, C. Cauchy, S. Zamith, A. S. Stodolna, J. H. Jungmann, J.
M. Bakker, G. Berden, B. Redlich, A. F. G. van der Meer, K. J. Schafer, and
M. J. J. Vrakking, Phys. Rev. A 87, 033413 (2013).
ˆ Scaling laws for photoelectron holography in the midinfrared wavelength regime,
Y. Huismans, A. Gijsbertsen, A.S. Smolkowska (Stodolna), J.H. Jungmann, A.
Rouzée, P.S.W.M. Logman, F. Lépine, C. Cauchy, S. Zamith, T. Marchenko,
J.M. Bakker, G. Berden, B. Redlich, A.F.G. van der Meer, M.Yu. Ivanov, T.-
M. Yan, D. Bauer, O. Smirnova and M.J.J. Vrakking, Phys. Rev. Lett. 109,
013002 1-5 (2012).

ˆ Time-resolved holography with photoelectrons, Y. Huismans, A. Rouzée, A.

Gijsbertsen, J.H. Jungmann, A.S. Smolkowska (Stodolna), P.S.W.M. Logman,
F. Lépine, C. Cauchy, S. Zamith, T. Marchenko, J.M. Bakker, G. Berden,
B. Redlich, A.F.G. van der Meer, H.G. Muller, W. Vermin, K.J. Schafer, M.
Spanner, M.Yu. Ivanov, O. Smirnova, D. Bauer, S.V. Popruzhenko and M.J.J.
Vrakking, Science 331, 61-64 (2011).

ˆ Yrast structure of the two-proton- and threeneutron-hole nucleus 203 Hg from

the decay of a 53/2+ isomer, B. Szpak, K.H. Maier, A.S. Smolkowska (Sto-
dolna), B. Fornal, R. Broda, M.P. Carpenter, N. Cieplicka, R.V.F. Janssens,
W. Krolas, T. Pawlat. J. Wrzesinski and S. Zhu, Phys. Rev. C 83. 064315
(2011).

133
About the author

Aneta Sylwia Stodólna (Smółkowska)

was born in Ostróda, North-East Po-
land in 1984. She spent her childhood
and attended primary school in Sam-
borowo. In 2003 Aneta graduated with
honors from Secondary School no.1 in
Olsztyn in biological-chemistry class.
She bet on physics and studied at
Gdansk University of Technology on the
Faculty of Applied Physics and Mathe-
matics for 5 years. During that period
she did several scientific internships. The first was at the Henryk Niewodniczanski
Institute of Nuclear Physics, Department of the Structure of Nucleus, Cracow, Po-
land. The second was at the National Institute of Nuclear Physics, Legnaro, Italy. She
has done a half-year schooling at AGH University of Science and Technology, Faculty
of Physics and Applied Computer Science, Cracow, Poland with special emphasis
on nuclear physics. She worked there in the field of ”Yrast structure in neutron–rich
nuclei from the vicinity of doubly magic 208 Pb”, which was the main topic of her
Master’s thesis. In June 2008 Aneta received a M.Sc. degree cum laude in Applied
Physics. Between July 2008 and March 2009 she worked at Katholieke Universiteit
Leuven, Institute for Nuclear and Radiation Physics, Belgium, designing and buil-
ding a new facility for Collinear Resonant Ionization Spectroscopy at the ISOLDE
hall at CERN, Switzerland to investigate the structure of exotic nuclei. In March
2009 Aneta started her PhD research at FOM Institute for Atomic and Molecular
Physics (AMOLF), Amsterdam, the Netherlands in the group of Extreme-Ultraviolet
Physics. Aneta’s research – photoionization microscopy of H and He atoms – brought
an experimental proof of a quantum theory developed in the 1980’s and allowed di-
rect observation of one of the most elusive quantum objects – the wave function. Her
recent paper in P hysics Review Letter on ”Hydrogen Atoms under Magnification:
Direct Observation of the Nodal Structure of Stark States” received a significant
attention in the scientific media all over the world and was highlighted in the top
scientific journals as N ature, N ature N anotechnology, Science, European P hysics
Society, N ew Scientists and P hysics W orld. Furthermore, this work was chosen as
one of the TOP 10 Breakthroughs in Physics in 2013 awarded by P hysics W orld.

135