Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article)
A little more than a decade ago our team extended the field of defect spectroscopy in
 cryocrystals to solid ⁴He matrices, in both their body-centered cubic (bcc) and hexagonally
 close-packed (hcp) configurations. In this review paper we survey our pioneering activities in the&#...
Gespeichert in:
| Veröffentlicht in: | Физика низких температур |
|---|---|
| Datum: | 2006 |
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | Englisch |
| Veröffentlicht: |
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
2006
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/120885 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) / P. Moroshkin, A. Hofer, S. Ulzega, A. Weis // Физика низких температур. — 2006. — Т. 32, № 11. — С. 1297–1319. — Бібліогр.: 82 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1860027815446970368 |
|---|---|
| author | Moroshkin, P. Hofer, A. Ulzega, S. Weis, A. |
| author_facet | Moroshkin, P. Hofer, A. Ulzega, S. Weis, A. |
| citation_txt | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) / P. Moroshkin, A. Hofer, S. Ulzega, A. Weis // Физика низких температур. — 2006. — Т. 32, № 11. — С. 1297–1319. — Бібліогр.: 82 назв. — англ. |
| collection | DSpace DC |
| container_title | Физика низких температур |
| description | A little more than a decade ago our team extended the field of defect spectroscopy in
cryocrystals to solid ⁴He matrices, in both their body-centered cubic (bcc) and hexagonally
close-packed (hcp) configurations. In this review paper we survey our pioneering activities in the
field and compare our results to those obtained in the related fields of doped superfluid helium and
doped helium nanodroplets, domains developed in parallel to our own efforts. We present experimental
details of the sample preparation and the different spectroscopic techniques. Experimental
results of purely optical spectroscopic studies in atoms, exciplexes, and dimers and their interpretation
in terms of the so-called bubble model will be discussed. A large part of the paper is devoted
to optically detected magnetic resonance, ODMR, processes in alkali atoms. The quantum nature
of the helium matrix and the highly isotropic shape of the local trapping sites in the bcc phase
make solid helium crystals ideal matrices for high resolution spin physics experiments. We have investigated
the matrix effects on both Zeeman and hyperfine magnetic resonance transitions and
used ODMR to measure the forbidden electric tensor polarizability in the ground state of cesium.
Several unexpected changes of the optical and spin properties during the bcc—hcp phase transition
can be explained in terms of small bubble deformations.
|
| first_indexed | 2025-12-07T16:50:44Z |
| format | Article |
| fulltext |
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11, p. 1297–1319
Spectroscopy of atomic and molecular defects in solid
4He using optical, microwave, radio frequency,
magnetic and electric fields
(Review Article)
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
Physics Department, Universit� de Fribourg, Chemin du Mus�e 3, 1700 Fribourg, Switzerland
E-mail: antoine.weis@unifr.ch
Received July 7, 2006, revised July 28, 2006
A little more than a decade ago our team extended the field of defect spectroscopy in
cryocrystals to solid 4He matrices, in both their body-centered cubic (bcc) and hexagonally
close-packed (hcp) configurations. In this review paper we survey our pioneering activities in the
field and compare our results to those obtained in the related fields of doped superfluid helium and
doped helium nanodroplets, domains developed in parallel to our own efforts. We present experi-
mental details of the sample preparation and the different spectroscopic techniques. Experimental
results of purely optical spectroscopic studies in atoms, exciplexes, and dimers and their interpre-
tation in terms of the so-called bubble model will be discussed. A large part of the paper is devoted
to optically detected magnetic resonance, ODMR, processes in alkali atoms. The quantum nature
of the helium matrix and the highly isotropic shape of the local trapping sites in the bcc phase
make solid helium crystals ideal matrices for high resolution spin physics experiments. We have in-
vestigated the matrix effects on both Zeeman and hyperfine magnetic resonance transitions and
used ODMR to measure the forbidden electric tensor polarizability in the ground state of cesium.
Several unexpected changes of the optical and spin properties during the bcc—hcp phase transition
can be explained in terms of small bubble deformations.
PACS: 32.30.–r, 32.60.+i, 33.35.+r, 33.50.–j, 67.80.–s, 76.70.Hb
Keywords: cryocrystals, quantum solids, matrix isolation spectroscopy, solid He, exiplexes,
atomic bubbles, dimers, optical detected magnetic resonance, Stark effect.
1. Introduction
Since the 1950’s, chemists have used solid noble
gases or solid nitrogen matrices, so-called cryocrystals
for the study of trapped unstable or reactive species,
such as free radicals. The technique has become known
under the name of matrix isolation spectroscopy
(MIS). Solid helium matrix isolation spectroscopy
(sHeMIS) described here, is an extension of that field
to solid matrices of 4He. This field has emerged from
the spectroscopy of point defects in superfluid 4He
(He II), studied first with electrons and He ions, and
later with atoms and other atomic ions. Historical as-
pects of defect spectroscopy in liquid and solid helium
were reviewed in [1,2]. Several prominent aspects of
sHeMIS are intimately related to dopant spectroscopy
in other quantum solids/fluids, such as solid hydro-
gen or helium nanodroplets. S. Kanorsky and one of us
(AW) have pioneered and developed the original field
of sHeMIS research at the Max-Planck-Institute for
Quantum Optics in Garching (Germany) more than a
decade ago. The studies were later continued and ex-
tended at the Institute for Applied Physics of the Uni-
versity of Bonn and are being pursued by our team
since 2000 at the Physics Department of the Univer-
sity of Fribourg (Switzerland).
Helium solidifies only under pressure. This is the
main obstacle for the preparation of doped helium
crystals and the primary reason why it has taken many
decades to extend standard MIS methods to helium
© P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis, 2006
matrices. In conventional MIS, the doped samples are
prepared by condensing the host and the guest (dop-
ant) substances on a cold finger, with the dopant com-
ing from a thermal jet, a discharge plasma, or a laser
ablation target. With solid helium one has first to pre-
pare the host solid and then dope it with the guest
substance using a procedure described herein.
In this invited paper we review our research on de-
fect structures formed by atoms and molecules in solid
4He matrices. Optical, radio frequency (rf), and mi-
crowave resonance spectroscopy was used to study the
samples with and without external magnetic and elec-
tric fields. It will not pass unnoticed that this review
has a strong focus on our own activities in the field.
This simply reflects the fact that, aside from a single
publication, there has been, to our knowledge, no
other report on optical experiments performed on
atomic/molecular defect structures in solid helium.
In this review we will not cover a related field of
research originally developed by a group in Cherno-
golovka, who studies so-called impurity—helium so-
lids formed by condensing a jet of impurity—helium
gas mixture into liquid helium [3]. In those studies a
highly porous structure composed of frozen impurity
(H2, D2, N2, Ne, Kr) clusters is formed inside of bulk
liquid helium. X-ray scattering, ultrasound propaga-
tion, and magnetic resonance (ESR) studies of such
samples have been reported. A review of some recent
achievements in this field can be found in [4]. Re-
cently the doping technique was modified and first re-
sults on doped solid helium were reported [5,6].
Among all the known cryocrystals solid helium
plays an outstanding role due to its macroscopic quan-
tum nature. This quantum aspect of solid helium also
plays a prominent role in the studies described below.
The softness of solid He, and, to a lesser extent that
of solid hydrogen, is due to the large amplitude
zero-point fluctuations of the constituents which al-
lows any entrapped or implanted species, hereafter
called defects or dopants, to conserve their symmetry
to a large extent. As a rule the spectra of the defects
have a simpler structure than the one encountered for
defects held in the more rigid conventional matrices.
Moreover, the defect can impose its own symmetry on
the local environment of the trapping site. Solid he-
lium has the further advantage that it has a symmetric
and a uniaxial crystalline phase (at low pressures) be-
tween which one can easily switch by changing the
temperature or the pressure of the matrix. A compari-
son of the dopants’ properties in both phases has al-
lowed the detailed study of small changes in the trap-
ping-sites symmetry when the phase boundary is
crossed.
The paper is organized as follows. In Sec. 2 we will
review the main properties of solid helium which are
relevant for the experiments, and in Sec. 3 we describe
the structure of the atomic defects in solid helium and
introduce the so-called atomic bubble model, origi-
nally developed for trapped electrons, but which has
been successfully extended for the description of opti-
cal properties of the trapped species. The next four
sections are devoted to experimental studies. In Sec. 4
we describe the details of the various experimental ap-
paratus used for the different studies and in Sec. 5 we
present the results of the purely optical studies of at-
oms, exciplexes and the recently-observed diatomic
molecules. In Sec. 6 we introduce the optically de-
tected magnetic resonance (ODMR) technique, a
powerful method for studying spin physics in dilute
samples of paramagnetic atoms, which combines mag-
netic resonance with optical preparation and detec-
tion. ODMR has been the center of our activities for
many years and we will cover the topics of optical
pumping, relaxation times, hyperfine transitions, and
Stark interactions. Section 7 finally discusses the ef-
fects of small bubble deformations on the optical, mi-
crowave, and rf spectra. The paper ends with remarks
regarding the use of He isolated atoms to search for a
permanent electric dipole moment of the electron
(Sec. 8) and with a concluding Section 9.
2. Solid 4He
Helium, either 3He or 4He, has the unique prop-
erty of being the only natural substance which stays
liquid under its vapor pressure down to the absolute
zero of temperature. Details of the p–T phase diagram
of 4He are shown in Fig. 1 from which it can be seen
that the solid phase can only be reached by applying
pressure p in excess of 25 bar. Under that condition,
1298 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
1.4 1.5 1.6 1.7 1.8
T, K
26
28
30
He II
0 1
2.17
3 15
He II
He I
T, K
25.3
50
1000
He I
�-line
W
of experiments
orking region
p,
ba
r
p,
ba
r
bcc
bcc
hcp
fcc
Fig. 1. 4He phase diagram. He I and He II denote the
normal fluid and the superfluid phase, respectively. The
solid phases are labeled according to their crystalline
structures: bcc (body-centered cubic), hcp (hexagonally
close-packed), and fcc (face-centered cubic). The p–T re-
gion studied in our experiments is shown on the right
panel.
4He crystallizes into three different crystalline struc-
tures depending on T and p [7]. The (isotropic)
body-centered cubic (bcc) and the (uniaxial) hexa-
gonally close-packed (hcp) structures can be obtained
at moderate pressures. Above 1000 bar and 15 K, 4He
solidifies in a face-centered cubic (fcc) structure as do
all the heavier rare gases.
The exceptional behavior of He can be understood
as a manifestation of macroscopic quantum properties.
The interaction between neighboring He atoms at
small internuclear separations is dominated by the
Pauli repulsion between their closed electronic
S-shells. At intermediate distances, an attractive in-
duced dipole—dipole interaction dominates, however,
due to the small electric polarizability, � = 3.33·10 �4
Hz/(V/cm)2 the attraction is very weak and cannot
produce a bound state. Even at T = 0, due to the
Heisenberg uncertainty principle � �x p � �, the local-
ization of a helium atom requires a zero-point kinetic
energy E0 = �p2/2M which is much larger than the
potential well depth. As a consequence helium stays
liquid at any temperature and the solid phase can be
produced only under high pressure which forces the
localization of the helium atoms. Due to their very
low mass, He atoms have a large thermal de Broglie
wavelength (�DB � 7 � at T= 1.6 K) and are thus
strongly delocalized. When the solid is created under
pressure the atomic wave functions exhibit a large mu-
tual overlap giving the crystal a macroscopic quantum
nature. The substances whose zero-point energy E0 is
comparable to their potential energyUare called quan-
tum solids (liquids) and are characterized by the pa-
rameter � = E /U0 [7]. The value of � for solid He is
larger than 1, while for all other solid rare gases it is
smaller than 1.
Quantum solids are very soft and extremely com-
pressible. These properties distinguish helium crystals
from all other rare gas matrices and play an important
role in the formation of the so-called atomic bubbles
which are so central to the research reported below.
3. The atomic bubble model
The structure of the trapping site of an alkali atom
in liquid or solid He is described by the so-called bub-
ble model. In this model helium is treated as a contin-
uous medium characterized by its elastic properties.
This is justified by the quantum nature of the matrix,
i.e., the strong overlap of the helium atoms wave
functions arising from the zero point oscillations. The
interaction between any alkali impurity atom and the
surrounding He is strongly repulsive due to the Pauli
repulsion between the valence electron of the atom
and the closed S-shell of the He atoms. Because of its
quantum nature the helium crystal is very compress-
ible, both by external and by internal pressure. The
repulsion of helium by an embedded alkali atom exerts
an internal pressure by which helium atoms are ex-
pelled from the volume occupied by the valence elec-
tron of the alkali. In this way a small cavity is formed
which has been called an «atomic bubble». The shape
of the bubble reflects that of the trapped atom’s elec-
tron orbital, and its size is determined by the balance
between the impurity—helium repulsion and the ex-
ternal pressure on the bubble surface. The equilibrium
configuration can be found by minimizing the total en-
ergy of the impurity plus bubble system here modelled
via
E U d pV
S
M
d
tot bubble
bubble
� � �
� �
�
�
( ) ( )
( )
R R R
R
�
3
2
3
8
�
. (1)
The first term in Eq. (1) represents the interaction
between the alkali atom and the He atoms, where
( )R describes the He density as a function of the dis-
tance R to the impurity atom. The pair potential
U( )R depends on the electronic state of the dopant
and has an angular dependence for states with a
nonvanishing orbital momentum L. For the sphericaly
symmetric ground states of alkali elements, it depends
only on R = | |R and the defect structure can be para-
meterized using two parameters: the bubble radius and
the thickness of the bubble interface. Theoretical pair
potentials U( )R for different electronic states of all
alkali elements interacting with a He atom in its gro-
und state are given in the literature [8].
The energy of the bubble in condensed helium is
composed of the pressure-volume work, pVbubble , the
energy from surface tension, �Sbubble , and the volume
kinetic energy, i.e., the excess zero-point energy due
to the localization of the He atoms at the bubble inter-
face, given by the fourth term of Eq. (1).
For a spherically symmetric defect, the helium den-
sity
( )R is described by the radial trial function
� �
� � �
( , , )
,
{ [ ( )]exp [ ( )]},
R R
R R
R R R R R
b
b
b b
�
�
�
� � � � � �
0
1 10 Rb
�
�
�
(2)
where
0 is the bulk helium density and the parame-
ters Rb and � are measures of the bubble radius and
bubble interface thickness, respectively, and which
are adjusted to minimize the total defect energy Etot .
This approach was first introduced by Jortner et al.
[9] to model a free electron in liquid He, for which the
expression «electron bubble» was coined, and was
later extended to atomic impurities in liquid [10,11]
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1299
and in solid [12] He. It successfully models optical ab-
sorption and emission spectra using the standard adia-
batic line-broadening theory [13], in which lineshapes
are calculated from the Fourier transform of the opti-
cal dipole autocorrelation functionC( )� represented by
� �C i R R d R( ) exp [ ( ( ) )] ( )�
�
� � � �� 1 3exp � , (3)
where �
( )R is the shift of the dopant’s transition en-
ergy due to a single He atom located at position R.
According to the Franck—Condon principle, the ab-
sorption (emission) line shape may then be calculated
assuming a bubble configuration corresponding to the
ground (excited) state.
4. Experimental setup
All our experiments reported in this paper were
performed using the same basic cryogenic setup, one
which allows us to create a He crystal, to dope it with
alkali (Cs, Rb) atoms and molecules, and to perform
spectroscopic studies in the optical, microwave, or in
the radio frequency domain, either with or without
static magnetic and electric fields. In this section we
will first describe the cryostat and the doping proce-
dure, and then present details of the experimental ar-
rangements for the different studies.
4.1. Sample preparation
The He crystal is grown in a cubic (inner dimen-
sions 6�6�6 cm, wall thickness 1.7 cm) pressure cell
made of copper. The cell has five optical windows, one
on each side and one on the top of the cell. It is im-
mersed in a liquid helium bath contained in a specially
designed cryostat with optical windows that are coax-
ial with the cell windows (Fig. 2). Before the experi-
ment, the cryostat is filled with 80 liters of liquid He
at 4.2 K and then cooled to 1.5 K by pumping on the
liquid He. The holding time of the cryostat depends
on the heat load (absorption of laser radiation, high
voltage leakage currents, heat conduction via electri-
cal connections to the cell volume, etc.). It typically
allows us to perform continuous measurements during
2–4 days. After filling the bath, helium gas is admit-
ted into the pressure cell via a liquid nitrogen cold
trap to remove condensable gases. This transfer goes
via a thin (2 mm diameter) capillary that minimizes
heat exchange. The helium used to form the matrix co-
mes from a 200 bar storage bottle (purity 99.9999%)
connected to the capillary via a buffer volume, which
allows the control of the helium pressure in the cell by
a needle valve. The temperature is measured by two
germanium resistors located in the pressure cell and in
the He bath, respectively.
After reaching the working temperature of 1.5 K, a
He crystal is produced by increasing the pressure in
the cell to a value above the solidification pressure
(26.4 bar), typically to about 29.5 bar.
Implantation of alkali atoms in the He crystal is
performed by laser ablation from the target (cut glass
ampule containing 0.5 g of the metal under investiga-
tion) installed under an inert gas atmosphere at the
bottom of the pressure cell in the preparation process.
The successive stages of the implantation process are
shown schematically in the lower panels of Fig. 3. The
top row of the same figure shows photographs taken
through the side-windows of the cryostat and pressure
cell during the implantation. First, the beam of a fre-
quency doubled Nd:YAG laser (� = 532 nm, pulse en-
ergy 20 mJ, repetition rate 1 Hz) is focused, by a lens
mounted just above the top window of the pressure
cell, onto the alkali metal target. The heat deposited
into the sample by each pulse melts a portion of the
crystal just above the target (Fig. 3,a). Simulta-
neously with the melting of the crystal, individual at-
oms, molecules, and clusters are ejected from the tar-
get and are distributed throughout the molten helium
region by convection (Fig. 3,b). By moving the lens
upwards during this phase we displace the focal point
and hence — the position of the molten part in the
crystal until the column of liquefied helium has risen
to the upper end of the window. After the desired col-
umn height is reached, the Nd:YAG laser is switched
off and the molten He resolidifies, thereby trapping
the implanted species in the previously molten col-
umn-like region (Fig. 3,c). This doping procedure was
initially reported [14] for Ba and Cs and was later ap-
plied in all our studies including those with Rb and
Rb/Cs mixtures. The implanted species are trapped
in the solid matrix and do not escape from the trap-
ping volume for many hours. This represents the main
1300 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
Pressure
cell
Fluorescence
Spectroscopy
laser
Nd:YAG-laser
Helmholtz
coils
Cryostat
LensCapillary
Isolation
vacuum
LN2LHe
Fig. 2. Vertical cut trough the cryostat, showing the pres-
sure cell and the triaxial system of Helmholtz coils inside
the helium bath.
advantage of using solid He as compared to liquid He,
where the implanted atoms are quickly washed out of
the observation volume by convection.
Depending on implantation conditions (He pres-
sure, laser pulse energy, repetition rate and the total
number of pulses) the doped part of the crystal has ei-
ther a greyish or bluish color. The color is due to
strong absorption and scattering of light by metallic
clusters produced by the ablation. The peak of the op-
tical absorption spectrum of this background is in the
near infrared (for Cs) or in the red (for Rb) part of
the spectrum, and depends on the cluster size distribu-
tion. The comparison of measured extinction spectra
with calculations based on the Mie theory of light
scattering by small particles [15] reveals a typical
cluster size on the order of 50 nm and a number den-
sity of about 1010 cm �3, comparable to or even larger
than the density of isolated atoms (typically about
109 cm �3).
The signal of interest for our spectroscopic studies
is produced either by atoms or by molecules trapped in
the crystal. Because of the softness of the He crystal
and of the large number of defects due to the brute
force preparation process, the implanted atoms do dif-
fuse within the doped region. When an atom comes
near another atom or cluster, it will be attracted by
van der Waals forces to form a molecule or a larger
cluster, thereby reducing the number of atoms avail-
able for the experiments. The number of atoms can be
increased again by destroying the clusters and mole-
cules by applying pulses from the same Nd:YAG laser,
focused in the center of the doped region (Fig. 3,d)
with a reduced pulse energy and a lower repetition
rate than used in the implantation process (to avoid
melting of the crystal).
4.2. Setup for optical spectroscopy
A top view of the typical setup for purely optical
studies is shown in Fig. 4. The dopants in the helium
crystal are excited by radiation from an optical para-
metric oscillator (OPO) pumped by the third har-
monic (� = 355 nm) of a pulsed Nd:YAG laser. The
tuning range of the OPO covers the large spectral in-
tervals of 780–2400 nm (idler beam) and 450–700 nm
(signal beam). The repetition rate of the laser pulses is
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1301
a b c d
Fig. 3. Different steps of the implantation process shown schematically (lower panels) and as frames of a video recording
taken through one of the lateral window of the cryostat and the pressure cell (upper row). In (d) a frame was selected in
which fluorescence induced by the Nd:YAG laser can be seen. Details are given in the text.
Spectrograph
Cryostat
Solid He
Implanted
atoms
CCD camera
InGaAs photodiode
Optical
parametric
oscillator
� = 450–700 nm
and 780–2400 nm
� = 355 nm
Nd:YAG
laser
Fig. 4. Horizontal cross section of the cryostat and the
pressure cell (not to scale) and top view of the optical
spectroscopic setup.
10 Hz and the average light power is controlled by
means of a �/2 plate and a polarizer. During the ex-
periments, the OPO power is kept at a level of 1 mW
to prevent melting of the He crystal. For spectroscopy
in the range between 720 and 780 nm, which cannot be
accessed by the OPO, we use a tunable cw Ti:Sa laser
pumped by a frequency doubled Nd:YVO3 laser.
The laser induced fluorescence from the sample vol-
ume ( � 3 mm3) is collimated by a lens in the cryostat
and focused into a grating spectrograph which has a
resolution of 0.2 nm. To access different parts of the
spectral domain, two different photodetectors are
mounted to the spectrometer. Visible and near-infra-
red fluorescence is detected by means of a CCD cam-
era which allows the recording of complete spectra,
while for wavelengths above 1 �m an InGaAs photo-
diode is used. In the latter case the grating is rotated
by a stepper motor, and the spectra are recorded point
by point.
This setup was used in our studies of laser-induced
fluorescence of atomic Cs, Cs*HeN exciplexes [16,17],
atomic Rb, Rb*HeN exciplexes [18], and Rb2 mole-
cules [19].
4.3. Magnetic resonance spectroscopy
Optically detected magnetic resonance spectros-
copy relies on driving magnetic dipole transitions be-
tween split Zeeman levels in combination with optical
preparation and detection and will be discussed in de-
tail in Sec. 6. Some of the main elements of the corre-
sponding experimental setup are shown in Fig. 2. The
static magnetic field is produced by three pairs of
superconducting Helmholtz coils mounted orthogo-
nally to each other around the pressure cell, inside the
He bath. A highly stable field is produced by the coils
operated in a self-sustained (persistent) current mode
[20]. The recording of ultra-narrow magnetic reso-
nance lines requires special care in the choice of
low-temperature (and in some cases high-voltage)
compatible nonmagnetic components and in the design
of the pressure cell. For example, indium — the most
convenient material for vacuum/pressure seals at low
temperatures becomes superconductive and may main-
tain a permanent electric current, whose magnetic
field will perturb the measurements. In order to avoid
this problem, all sealings in the pressure cell are made
of pure aluminum. Laboratory fields and gradients are
suppressed by a cylindrical three-layer magnetic
shield surrounding the whole cryostat.
The sample is irradiated by a circularly polarized cw
laser beam resonant with the D1 transition of the im-
planted atoms. For Cs a near infrared diode laser
(1 mW) is used, whereas for Rb we use 1 mW of radia-
tion from the Ti:Sa laser. Laser-induced fluorescence
from the implanted atoms is collected in the same way
as in the optical spectroscopic experiments and de-
tected either by a photomultiplier or an avalanche
photodiode. In these experiments we do not use the
spectrograph but rather an interference filter to sup-
press the scattered laser light. As discussed in Sec. 5,
the fluorescence of Rb atoms in solid He is so weak that
it escaped observation for a long time. For this reason
the magnetic resonances in Rb were detected by moni-
toring the transmitted intensity of the pumping laser.
The magnetic transitions are induced by an oscillat-
ing magnetic field produced by a second system of
Helmholtz coils mounted inside the pressure cell
(shown in Fig. 5, not visible in Fig. 2). The current to
these coils is supplied by a waveform generator
through an electric feedthrough at the bottom of the
1302 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
Polycarbonate
body
rf coil
a b
Glass
electrode
HV
Feedthroughs
WindowCs atoms
Cooper cell
Circ. pol.
laser beam
rf-coils
Fluorescence
to detection
Mirror
Field plates
Mx
Mz
Cs target
Fig. 5. (a) the bottom flange of the pressure cell of Fig. 2 with one of the rf-coils and one of the HV electrodes shown;
(b) top view of the set-up for magnetic resonance experiments with electric fields.
cell. The ODMR technique was also used to study the
microwave transition between the two hyperfine lev-
els of the ground state of Cs in solid He (
= 9.2
GHz). In that case, the oscillating magnetic field was
produced by a microwave frequency synthesizer and
irradiated through one of the fused silica optical win-
dows of the pressure cell.
A separate subsection of this paper is devoted to
our very recent studies of the Stark effect on the mag-
netic resonance transitions in Cs. For that study, the
ODMR setup described above was modified in order
to apply a large static electric field to the Cs-doped
He crystal. A bipolar high voltage produced by two
identical power supplies of opposite polarity was de-
livered to the sample via HV cables traversing the top
flange of the cryostat, then the helium bath, inside of
which they were connected to specially designed
feedthroughs mounted on the bottom of the pressure
cell (Fig. 5). The feedthroughs are connected to a pair
of electrodes (spacing 6.0 mm), cut from float glass
and coated on one side with a transparent conductive
layer of tin oxide. Pure liquid and solid He are excel-
lent electric insulators. In practice, the magnitude of
the applied electric field strength is limited by the
leakage current through the doped He crystal. In-
creasing the field strength above 50 kV/cm results in
the melting of the sample and produces an electric
breakdown in the cell.
5. Optical spectroscopy
5.1. Atomic fluorescence
First systematic spectroscopic studies of atoms in
solid He were carried out in experiments with Ba, for
which a good agreement with the predictions of the
bubble model was obtained [10,12,21]. Later those
studies were extended to alkali elements. Here we dis-
cuss the absorption and fluorescence of cesium atoms
in solid He studied in great detail in [12,17,20]. Ce-
sium is the only alkali element which emits a strong
atomic fluorescence in a solid He environment. The
light alkalis Li, Na, and K do not emit any fluores-
cence neither in solid nor in liquid He. Rb represents
an intermediate case, as it emits fluorescence in liquid
He, but its fluorescence is strongly quenched with in-
creased He pressure [22]. Only recently, a very faint
fluorescence of atomic Rb was observed in solid (hcp)
He following excitation of the D1 and D2 transitions
[18]. The quenching of the atomic fluorescence is due
to the formation of molecular complexes by excited nP
states of alkali atoms and the surrounding helium at-
oms, so-called exciplexes [12,23]. As discussed in
Sec. 5.2, the formation of the exciplex becomes possi-
ble when the alkali—helium interaction is stronger
than the spin—orbit coupling in the alkali atom.
Therefore the probability of the exciplex formation
is much larger for lighter alkalis, in which the
spin—orbit coupling is weak, and it increases with the
density of He, i.e., with rising He pressure, or when
going from liquid to solid phase.
Typical fluorescence spectra of Cs in solid He are
shown in Fig. 6,a for the bcc and hcp crystalline
phases. Only fluorescence on the D1 transition can be
observed, as atoms excited to the 6P3 2/ state are
quenched by the interaction with the matrix to form
either atoms in the excited 6P1 2/ state or to form
Cs*HeN exciplexes. An absorption spectrum recorded
by scanning the excitation laser and monitoring the
intensity of the D1 fluorescence, is shown in Fig. 6,b,
where both the D1 and the D2 absorption lines are de-
tected. The position of the D1 line center measured in
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1303
875 880 885 890
Emission wavelength, nm
hcp
�1
bcc
30
780 800 820 840 860 880
D1
D1
D2
a b
Excitation wavelength, nm
1100
Fl
u
o
re
sc
e
n
ce
,a
rb
. u
n
its
Fig. 6. Experimental emission spectrum of the D1 line of Cs following D1 excitation in the bcc and hcp phases of solid
4He (a) and excitation spectrum of the D1 and D2 transitions detected via fluorescence on the D1 line of Cs in hcp (b).
The resonance wavelengths of the same transitions in the free Cs atom are 894 and 852 nm, respectively.
absorption and emission is plotted in Fig. 7 as a func-
tion of He pressure.
As one can see in Figs. 6 and 7, both the absorption
and the emission lines in solid He are strongly broad-
ened and blueshifted with respect to the free Cs atom.
The shift and the broadening are more pronounced in
the absorption spectra than in the emission spectra.
Very similar spectra were obtained in a pressurized
liquid He environment, where a quantitative agree-
ment with the predictions of the bubble model was de-
monstrated [11]. The blueshift of the spectral lines
can be understood from the changes of the bubble size
that occur during an optical absorption—emission
cycle (Fig. 8). The blueshift is due to the fact that the
electronic density distribution of the excited 6P state
of Cs is spread over a larger volume than that of the
ground state. The interaction with the bubble shifts
both levels towards higher energies, but the excited
state, due to its larger interaction energy, shifts more
than the ground state and hence the transition wave-
length shifts to the blue. The absorption takes place in
a smaller bubble, whose size is determined by the Cs
ground state and the blue shift is particularly large.
After the excitation of the impurity atom, the bubble
expands and a new equilibrium configuration minimiz-
ing the total energy is realized on a time scale much
shorter than the radiative lifetime of the 6P1 2/ state.
The emission then occurs in a larger bubble that re-
flects the size of the excited state and therefore the
broadening and the shift of the emission lines are
smaller.
The bubble model also explains the observed in-
crease of the spectral width and the rate at which the
lines shift with He pressure, as shown in Fig. 7, how-
ever the abrupt changes of the absorption/emission
wavelengths at the liquid—bcc and bcc—hcp phase
transitions have not yet been explained quantitatively
(see Sec. 7).
The only atoms of which optical absorption and
emission spectra in solid helium were studied are
85Rb, 87Rb, 133Cs, 137Ba, and 169Tm. While the
(optical) valence electron of alkali atoms interacts
with the helium matrix with an energy comparable to
the spin—orbit interaction, the optical electron of the
lanthanide thulium behaves quite differently. Its opti-
cal transition occurs between two unfilled shells (4f
and 5d) which are shielded from the matrix by outer
electrons which do not participate in the optical tran-
sition. Ishikawa et al. [24] have measured absorption
and emission lines from Tm in superfluid and solid he-
lium as well as excited state lifetimes. Because of the
shielding from matrix effects the authors of [24] could
observe widths of optical resonance lines on the order
of 0.1 nm (limited by the resolution of the spectrome-
ter). To our knowledge this is the only experiment be-
sides our own which has reported an optical study of
atoms in solid helium. We also mention here the re-
lated optical spectroscopic studies of electron bubbles
in solid He, pioneered by Mezhov-Deglin and Golov
[25–27].
1304 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
hcpbcc
28 30 32 34 36 38 40
11780
11800
11820
11840
11860
848
847
846
845
844
843
hcpbcc
26 28 30 32 34
11300
11320
11340
11360
11380
11400
11420
884
882
880
878
876
E
xc
ita
tio
n
w
av
e
n
u
m
b
e
r,
cm
–
1
E
xc
ita
tio
n
w
av
e
le
n
g
th
,n
m
E
m
is
si
o
n
w
av
e
le
n
g
th
, n
m
E
m
is
si
o
n
w
av
e
n
u
m
b
e
r,
cm
–
1
a
b
He pressure, bar
Fig. 7. Dependence of the D1 absorption (a) and emission
(b) lines of Cs in solid He on He pressure. The experimen-
tal data are shown as dots and the straight lines are reso-
nance positions calculated [20] using the spherical bubble
model. The D1 transition in the free Cs atom is at 894 nm
(11186 cm�1).
6P1/2 6P1/2
6S1/2 6S1/2
Absorption
(850 nm)
Fluorescence
(880 nm)
Bubble shape
relaxation
Bubble shape
relaxation
ps
ps
30 ns
Fig. 8. Expansion and shrinking of the atomic bubble dur-
ing an optical absorption—emission cycle.
5.2. Alkali—helium exciplexes
The term «exciplex» stands for excited state com-
plex and refers to molecular complexes which form
bound or quasibound states only when one of their
constituents is in an excited state. Exciplexes com-
posed of an alkali atom in the nP state and one or sev-
eral ground state He atoms were considered for the
first time in [12,23] to explain the quenching of the
laser-induced fluorescence from light alkalis in liquid
and solid helium. A first experimental confirmation of
that proposal was obtained by a group in Kyoto in a
series of experiments that covered all alkali elements
in liquid He and cold He gas [28–30]. Such complexes
were also observed and extensively studied in experi-
ments with alkali doped superfluid helium
nanodroplets [31–36]. We also mention related exper-
imental studies of Ag*He2 [37] and Mg*HeN [38,39]
exciplexes in liquid He and that of Ba He�* [40]
in cold He gas. In solid helium, the formation of
Cs*HeN exciplexes was reported by our group for the
first time in [16] and described with more experimen-
tal and theoretical details in [17,41]. More recently
we have extended these studies to Rb*HeN in solid
He [18].
The exciplexes are typically detected via their la-
ser-induced fluorescence which can be excited at the
wavelengths of the resonant atomic transitions. The
characteristic feature of exciplex emission, and the
reason why it has escaped observation for many years,
is its strong red shift with respect to the corresponding
absorption lines. For instance the Cs*HeN can be
formed after excitation at 800 nm, while it fluoresces
at 1400 nm. Typical spectra of Cs*HeN and Rb*HeN
in the hcp phase of solid He excited at the D2 atomic
transitions of Cs and Rb, respectively are shown in
Figs. 9,a and 9,b.
The theoretical model for describing the optical
properties of exciplexes developed in [17,41] is based
on adiabatic alkali—helium pair potentials [8], which
are strongly anisotropic for the nP states of the alkalis.
The alkali—helium interaction at intermediate inter-
atomic distances is dominated by the Pauli repulsion
between their valence electrons. When a He atom ap-
proaches from a direction along which the electronic
density of the alkali orbital is high, it experiences a
strong repulsion. However, the nP orbitals possess
nodal planes, or nodal axes, along which the elec-
tronic density is zero so that the helium atom can come
close enough to experience a van der Waals attraction
by the alkali’s core. In this case a short lived weakly
bound or quasibound complex can be formed. For the
light alkalis, in which the spin—orbit interaction is
much weaker than the interaction between the impu-
rity and the He atoms, the excited state orbital is well
represented by a dumbbell-shaped Pz orbital which al-
lows several He atoms to be bound around its waist.
For the heavier alkalis, Rb, and in particular for Cs,
the spin—orbit interaction which generates the fine-
structure splitting of the nP state into nP1 2/ and
nP3 2/ states becomes comparable to the impurity—
helium interaction and the symmetry of the state is
given by its total angular momentum J.
The nP1 2/ state is spherically symmetric and hence
repulsive for He. The nP3 2/ state, on the other hand,
has two distinct orbitals depending on the projection
| |MJ of J on the internuclear axis. The approaching
He atom sees either a repulsive dumbbell-shaped or-
bital in the M /J � � 1 2 configuration oriented along
the direction of the approach, or an attractive ap-
ple-shaped orbital (M /J � � 3 2 configuration) with
two dimples on opposite sides, again along the direc-
tion of approach [23].
In agreement with the discussion above, we
observed the diatomic and triatomic exciplexes
Cs( /6 3 2P )He2, Rb(5P3 2/ )He1, and Rb(5P3 2/ )He2,
whose emission lines at 10500, 12400, and 11800 cm �1 ,
respectively, are shown in Fig. 9. However, the two
strongest and most redshifted emission bands originate
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1305
Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
6000 8000 10000 12000
Emission wavenumber, cm–1
a
b
Cs He* 6Cs He* 7
Cs He* 2
Rb He* 1
Rb He* 2
Rb He* 6Rb He* 7
Cs D1
Rb
D /D1 2
Fig. 9. Experimental emission spectra of Cs*HeN (a) and
Rb*HeN (b) exciplexes in hcp solid He excited at the D2
transitions of Cs and Rb, respectively. The atomic D1 emis-
sion line of Cs and the D /D1 2 doublet of Rb are visible on
the blue side of the exciplex spectra. The vertical dotted
line in (b) separates two spectral regions with different
vertical scales. The emission of atomic Rb and Rb*He12,
exciplexes on the right part of the figure is several orders
of magnitude weaker than any other spectral feature.
from different complexes. Due to the (spin—orbit un-
coupling) mixing of the two fine-structure compo-
nents by the alkali—helium interaction the nP1 2/
state acquires the character of the nP3 2/ state and its
orbital becomes dumbbell-shaped. This mixed state al-
lows several He atoms to be bound around its waist, as
in the case of the lighter alkalis. These complexes
fluoresce at 7200 cm �1 (Cs*HeN) and 7400 cm �1
(Rb*HeN).
The model presented in [17,18,41] has allowed us
to calculate the emission spectra of the exciplexes as
well as their vibrational and total energies. However,
the interaction of the exciplex with the surrounding
He bulk has so far not been taken into account. Based
on the model, the number of bound He atoms can be
estimated to be N = 6 or 7 (see Fig. 9). Alternative
theoretical studies of Cs*HeN [42], Rb*HeN [29],
and the closely related K*HeN [43] suggest N = 6 as
the most probable number of bound He atoms.
Owing to the very high He density and the compact
bubble structure in solid helium, sHeMIS has proven,
in comparison to other experimental techniques (he-
lium nanodroplets, cold helium gas), to be particu-
larly well suited for the formation and investigation of
exciplexes with a maximally allowed number of bound
He atoms. Due to the large rate of collisions with sur-
rounding helium atoms, the attachment of He atoms
in solid He proceeds at a much faster rate than in any
other environment and stops only when all vacancies
are occupied, i.e., by the filling of the two dimples of
the nP3 2/ orbital or of the ring-shaped belt around the
waist of the perturbed nP1 2/ orbital. All intermediate
complexes occur only as transients which have no time
to fluoresce and who thus do not contribute to the
emission spectrum. A very similar behavior was ob-
served in liquid He [28–30], with the difference that
in that case the perturbation of the 6P1 2/ state of Cs
was not strong enough to allow the formation of the
ring-shaped complex. The largest complex seen in su-
perfluid helium is the triatomic Cs( )/6 3 2P He2 excip-
lex [28].
5.3. Alkali dimers
Absorption and emission spectra of alkali molecules
(dimers) were extensively studied in experiments with
alkali-doped helium nanodroplets [36,44–49], where
all homonuclear and some heteronuclear dimers were
formed and investigated. However, until recently,
there had only been very few investigations of alkali
dimers in bulk condensed helium. In superfluid He
only some (unassigned) bands of Na2 and Li2 have
been reported [50].
In our recent experiments [19] we observed for the
first time alkali molecules in solid He. In those experi-
ments we studied the fluorescence spectrum of
Rb-doped solid He under laser-excitation in the broad
spectral range from 450 to 1000 nm and found a re-
markable result. Besides atomic and exciplex emis-
sion, a single additional spectral feature was observed
at 1042 nm (Fig. 10,a). We have assigned this emis-
sion band to the forbidden X g
1� � (1)3� u transi-
tion of Rb2. We base this assignment on the spectral
structure of the free dimer and on the observation of a
long lifetime of the fluorescing state which points to
its metastable character (Fig. 10,b). Our most strik-
ing observation is the fact that this (single) emission
band can be excited on 9 distinct absorption bands
of Rb2 in the range of 450 to 900 nm (Fig. 11),
including transitions originating from the X g
1�
ground state and from the lowest triplet state (1)3� u .
The measured and calculated spectral positions of
these bands are given in Table 1. The calculation of
1306 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
1030 1040 1050 1060 1070 0 50 100 150 200
Time, s
ba
Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
Emission wavenumber, nm
2
1
Fig. 10. (a) Experimental (points) and calculated (solid curve) emission spectra of (1)3�u � X g
1� transition in Rb2.
The theoretical curve was obtained by shifting the calculated spectrum by 20 nm to the red. (b) Measured pulse shapes of
the molecular fluorescence at 1042 nm (curve 1) and scattering of the excitation laser light at 842 nm (curve 2), which
represents the time resolution of the photodetector as the actual width of the laser pulse (5 ns) is not resolved.
the theoretical band positions assumes that only the
lowest vibrational state of the emitting electronic
level is populated, but does not take the interaction
with the He matrix into account. We attribute the
(small) discrepancies between the calculated and mea-
sured band positions to this fact.
Table 1. Calculated and measured wavelengths of Rb2 ab-
sorption bands (in nm) in solid He. The labels refer to the
corresponding peaks in Fig. 11.
Band Label �theor �exper
X g A u
1 1� �� a 878 842
( )1 3�u � ( )1 3�g b 735 742
X g B u
1 1� �� c 664 653
( )1 3�u � ( )3 1�g d 623 622
( )1 3�u � ( )2 3�g e 590 580
( )1 3�u � ( )2 3�g e 586 580
( )1 3�u � ( )3 3�g f 507 524
X g
1� � ( )2 1�u g 477 445
X g
1� � ( )2 1�u g 464 445
We have also studied the time dependence of this
fluorescence in an experiment using pulsed excitation
and found a width of the fluorescence pulse of 60 �s
(Fig. 10,b), much longer than the characteristic life-
time of any allowed electronic transition. The fluores-
cence pulse has finite rise and decay times, which
point to the formation of a state that does not exist in
the free dimer. We assign [19] this behavior to the for-
mation of a molecular exciplex state, from which the
observed fluorescence emanates.
The experimental observations show that in solid
He, due to the interaction with the matrix, all la-
ser-excited molecular states are quenched. The
quenching results in the population of the metastable
(1)3� u state which is the lowest excited state of this
molecule. The perturbation of the molecule by the sur-
rounding helium partly lifts the selection rule that
forbids the radiative transition from that (triplet)
state to the singlet ground state so that the transition
X g
1� � (1)3� u can be observed.
We have also observed photodissociation of the
Rb2 molecule into two Rb atoms, one in the ground
state and the other in the excited 5P1 2/ or 5P3 2/
state. The latter emits fluorescence at the same wave-
length as do individual Rb atoms excited either at D1,
or D2 transition. The photodissociation spectrum re-
corded by scanning the excitation wavelength and de-
tecting the atomic fluorescence contains the forbidden
molecular bands (1)3� u � (2)3� u (650 nm) and
X g
1� � (3)1� g (490 nm) in addition to the bands b,
d, and e shown in Fig. 11 and Table 1.
6. Optical pumping and magnetic resonance
Our initial motivation for entering the field of
sHeMIS was the expectation that the helium quantum
matrices would be ideally suited for performing high
resolution experiments involving spin polarized de-
fects, such as the search for a permanent electric di-
pole moment of the electron [51]. This expectation
was based on two facts: on one hand, 4He has neither
an electronic nor a nuclear magnetic moment, so that
it has no first order coupling to the spin of the para-
magnetic defect. On the other hand, it had been known
from the spectroscopy of atoms and ions in superfluid
helium as well as from studies of isolated electron
bubbles in both liquid and solid helium that the im-
planted defects locally impose, to a large extent, their
own symmetry, i.e., the shape of their electron distri-
bution on the spatial distribution of helium atoms sur-
rounding the defect. Unlike in any other solid state
matrix where the symmetry of the dopant is in general
lowered due to local field effects, which themselves
couple to the dopant’s spin via the spin—orbit inter-
action, one therefore did not expect any coupling
(again in lowest order) of the dopant’s spin to the lo-
cal host structure, provided that the dopant had a
spherically symmetric electronic wave function, as is
the case, e.g., for atomic S1 2/ , P1 2/ , 1
0S states. The
spherically symmetric S1 2/ ground state of alkali met-
als seemed particularly well suited for testing this hy-
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1307
a
b
c
d
e
f
g
Excitation wavelength, nm
500 600 700 800 900
Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
Fig. 11. Measured excitation spectrum of the Rb2 fluores-
cence at 1042 nm. Experimental data are shown as points,
and the solid lines are fitted Gaussians. The assignment of
all peaks is given in Table 1. The dotted vertical lines
mark the spectral regions covered by the three excitation
sources described in the text. The vertical scales differ for
the different intervals. Note that the excitation at the
bands b, d, and e also branches to photodissociation
channels.
pothesis, and it was expected that a spin polarized
sample of such atoms would have an exceptionally
long electronic spin relaxation time.
6.1. Optical pumping
The first experimental demonstration of optical
pumping of alkali atoms in condensed helium was per-
formed with 85Rb, 87Rb, and 133Cs in superfluid he-
lium [52]. In those experiments, spin polarization was
created by optical pumping of rubidium or cesium at-
oms implanted by laser sputtering into liquid helium.
The efficient process of optical pumping was dis-
covered in the 1950’s in vapor phase atomic samples
and has since found numerous applications. The reso-
nant scattering of a circularly polarized light beam by
the atoms transfers angular momentum from the light
to the atoms. In this way the irradiation of the sample
with circularly polarized light leads to the creation of
spin orientation (dipole polarization) [53] by the re-
distribution of populations among the ground state
Zeeman sublevels. In all optical-pumping sHeMIS ex-
periments carried out to date the optical pumping used
excitations to the lowest lying P1 2/ state, i.e., the
so-called D1(nS nP/ /1 2 1 2� ) transition, whose wave-
lengths (for Cs and Rb) lie in the near infrared. The
hyperfine interaction splits the ground and first
excited states into two levels with total angular
momenta of F = I � 1/2, where I denotes a nuclear
spin. Due to the large homogeneous linewidth of the
absorption line in condensed helium (cf. Sec. 5) the
hyperfine structure of the transition cannot be re-
solved, so that the selection rules are those of a
J J� � �1
2
1
2 transition. This situation is equiva-
lent to exciting the transition in a vapor or an atomic
beam using a spectrally broad light source, as is en-
countered, e.g., in the optical pumping by resonance
radiation from a discharge lamp. One can show that in
this case excitation with linearly polarized light does
not lead to a redistribution of populations, so that op-
tical pumping can only be achieved with circularly po-
larized light. It is thus not possible to create an align-
ment (quadrupole polarization) in the ground state by
optical pumping.
After a number of absorption—emission cycles on
the D1 transition nearly all atoms become pumped to
the state | ,F I / M F� � � �1 2 which does not absorb
circularly polarized light and which therefore is re-
ferred to as «dark state». The polarized sample does
not fluoresce and any polarization destroying effect,
such as a magnetic resonance transition, leads to a re-
vival of fluorescence. This forms the basis of the opti-
cal detection of magnetic resonance.
In 1995 our group showed that efficient optical
pumping is also possible with alkali atoms embedded
in the cubic phase of solid 4He [54] and, with a re-
duced efficiency, in the anisotropic hexagonal phase of
the matrix [55,56]. In those experiments we demon-
strated our original assumption that the spin polariza-
tion of alkali atoms in solid helium may indeed be very
long lived. In a subsequent study we investigated the
optical pumping process in the bcc and in the hcp
phase in detail [57]. We found that optical pumping is
of the repopulation type, i.e., that the spin polariza-
tion created during the optical pumping process in the
excited P1 2/ state, which itself lives for a few 10 ns, is
not destroyed by the interaction with the helium ma-
trix. This is surprising at first sight as it is known that
this process, whose origin is spin—orbit coupling in-
duced by collisions, occurs in helium gas and becomes
more efficient with growing helium pressure. How-
ever, in solid helium the collision rate with helium at-
oms on the bubble interface is so high that the Cs spin
has no time to couple efficiently to the short-lived col-
lision induced orbital momentum.
Optical pumping of cesium in the anisotropic hcp
phase of helium is much less efficient than in the iso-
tropic bcc phase. We showed that the degree of spin
polarization achievable in hcp 4He depends strongly
on the value of the magnetic holding field B0 (see
Fig. 19,b), while in bcc phase it was found to be
independent of B0 [57]. This effect is one of the multi-
ple manifestions of bubble deformations discussed in
Sec. 7.
6.2. Magnetic resonance
A high degree of spin polarization is a prerequisite
for sensitive magnetic resonance experiments. Mag-
netic resonance on electron bubbles had been previ-
ously studied in condensed helium (for a partial re-
view of those studies see e.g. [2]) and in those
experiments the electrons were polarized using the
Boltzmann factor in a strong external magnetic field.
To our knowledge this polarization technique has
never been applied to atomic defects in condensed he-
lium, for which the creation of polarization by optical
pumping is orders of magnitude more efficient (cf
Sec. 6.1). The optical properties of a polarized atomic
medium depend on the orientation of the spin polar-
ization with respect to the light polarization. In this
way the spin depolarization by transverse fields
(Hanle effect), by resonant oscillating fields (mag-
netic resonance) or by crystalline fields and gradients
(zero field resonance) can be detected by optical
means. In the simplest arrangement, the laser beam
used for the creation of the polarization by optical
pumping can be used to detect the alterations of that
polarization arising from interactions with external
fields. This is the very essence of the powerful method
1308 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
of ODMR, which not only allows the creation of a
large degree of spin polarization but also its highly ef-
ficient detection in dilute atomic samples. Most spin
related experiments on alkali atoms in condensed he-
lium were carried out using the ODMR technique.
The optical pumping process produces population
imbalances, i.e., spin polarization � � �Fz F 3 4, in both
hyperfine levels of the ground state. Due to the nu-
clear magnetic moment, the gyromagnetic ratios of
both states differ slightly in magnitude, in addition to
having opposite signs, and the corresponding Zeeman
splittings will be slightly different. As a consequence,
the magnetic resonance transitions between Zeeman
sublevels in the F = 3 and F = 4 states will occur at
slightly different frequencies. Since the probing laser
interacts simultaneously with both hyperfine states
one can observe both states in a single scan of the rf
frequency (Fig. 12,a).
As mentioned above, Kinoshita et al. [52] were the
first to observe magnetic resonance on alkalis im-
planted in superfluid helium using the ODMR tech-
nique. The longer relaxation time obtained in the cu-
bic phase of solid helium allowed us to observe mag-
netic resonance lines in such matrices with a width
(HWHM) of only 10 Hz [58]. We took advantage of
these narrow lines to build an optically pumped mag-
netometer in the Mx geometry [58] and were able to
demonstrate a magnetometric sensitivity of 2.6 pT for
an integration time of 1 second.
As discussed in Sec. 5, rubidium emits fluorescence
in liquid helium, but the fluorescence intensity is
strongly quenched when the helium pressure is in-
creased and it was long believed that Rb would not
fluoresce at all in solid helium. It is only recently that
we could observe a faint fluorescence from Rb in solid
helium [18]. The relatively strong fluorescence of Rb
in superfluid matrices allowed Kinoshita et al. [52] to
observe ODMR signals in that phase. To measure
magnetic resonance from Rb in solid helium we used
the fact that Rb still has resonant optical absorption,
despite the fact that its deexcitation is mainly
radiationless in the atomic channel. Because of the
low optical thickness (10 �5–10 �4) the absorption is
difficult to detect. Using the Mx variant (Fig. 13) of
ODMR, in which the magnetic resonance process in-
duces a modulation of the transmitted intensity at the
rf frequency, we were able to detect magnetic reso-
nance signals from rubidium atoms embedded in the
bcc phase of solid helium [59]. In that experiment,
low-frequency intensity noise was rejected by phase
sensitive lock-in detection, which allowed us to ex-
tract the weak signal modulation due to the reso-
nantly driven spin precession. In this way we have
shown that the helium matrix does not affect the
Land� gJ -factors of 85Rb, 87Rb, and 133Cs at a level
of 2·10 �4, and that the gI factors coincide at least at
the level of 10% with those of free atoms [59]. The
study also revealed that the optical pumping process
in Rb is of the depopulation type, in which spin pola-
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1309
a
b
Microwave frequency, GHz
Radio frequency, kHz
Fl
u
o
re
sc
e
n
ce
si
g
n
al
, a
rb
. u
n
its
Fl
u
o
re
sc
e
n
ce
si
g
n
al
, a
rb
. u
n
its
Fig. 12. Intramultiplet (a) and intermultiplet (b) mag-
netic resonance spectra of Cs in bcc 4He. a and b refer to
the transitions A and B shown in Fig. 15. Note the dif-
ference of the linewidths in the two cases.
a b
Fluorescence Absorption
Fig. 13. Geometries for optically detected magnetic reso-
nance spectroscopy with static (B0) and oscillating (B1, rf
or microwave) magnetic fields: (a) Mz in which the fluo-
rescence shows a resonant DC change when the Larmor fre-
quency coincides with the oscillating field frequency, and
(b) Mx geometry, in which the transmitted light intensity
is modulated at the frequency of the oscillating field.
rization in the excited state is destroyed by the matrix
interaction, whereas in cesium, repopulation is the
dominant mechanism. This is a direct consequence of
the fact that the excited Rb state decays via an inter-
mediate exciplex state to the ground state, a process in
which spin polarization is destroyed. In principle
it should therefore also be possible to monitor the
ground state spin polarization via the far red detuned
emission of exciplex fluorescence, although this has
not yet been attempted.
6.3. Relaxation times
Longitudinal relaxation. In our first experiments
on optically pumped cesium in solid helium we have
determined T1, i.e., the longitudinal relaxation time of
the spin polarization, using the method of «relaxation
in the dark» [54]. We found T1 to vary in the range of
1–2 seconds when the external holding field was var-
ied in the range of 10 nT to 100 �T [57]. In the origi-
nal optical pumping study [52] in He II, the T1 time
was not measured. Only recently did Furukawa et al.
[60] determine the T1 time of spin polarized cesium
in superfluid helium. In that experiment great care
was taken to suppress contributions from the loss of
atoms due to convection and a lower bound of
T1 = 2.24(19) s was found for the longitudinal relax-
ation time. The T1 times of 1 or 2 seconds observed
with cesium atoms in solid [54] and in superfluid [60]
helium, respectively, are longer than the ones ob-
served in vapor cells in which inert buffer gases or spe-
cial surface coatings are used to prevent depolarizing
collisions with the cell walls.
The question as to why the T1 times in condensed
helium are only 1–2 seconds and not orders of magni-
tude longer has not yet found a quantitative answer. It
is known that static quadrupolar deformations of the
atomic bubbles in hcp matrices depolarize the atoms
on a time scale of a few 10 �s (cf. Sec. 7). Because of
its quantum nature, the spherical bubble interface in
the bcc phase undergoes large zero point oscillations
which can be decomposed into oscillations of different
multipole orders, and the atomic spins will couple to
the quadrupolar deformations. Bubble surface oscilla-
tion frequencies can be estimated to lie in the gigahertz
range and are much faster than the depolarization rate
due to a deformed bubble. In that case the depolarizing
effect of a given instantaneous deformed bubble confi-
guration will not be very effective, and one has to ap-
ply the theory of motional narrowing [61] to infer the
effective lifetime of the spin polarization.
Transverse relaxation. Transverse relaxation times
are more relevant for practical applications as they de-
termine the width of magnetic resonance lines. They
can either be inferred from the widths of these lines,
from free induction decay signals, or from spin echo
experiments. In the first experiments on optically
pumped rubidium and cesium atoms in superfluid he-
lium, Kinoshita et al. [52] observed magnetic reso-
nance linewidths on the order of 50 �T which corre-
spond to a spin coherence relaxation time (T2 time) on
the order of 1 �s. They found that strong convection
currents in He II carried the atoms out of the observa-
tion volume. The large rf intensity required for ob-
serving the spin flip transitions under those conditions
is at the origin of the observed relatively short coher-
ence times.
In solid helium matrices the atomic diffusion time
is orders of magnitude longer than the observation
times realized in liquid helium and the magnetic reso-
nance lines were expected to be substantially narrower
in such matrices. In 1995 we reported the first obser-
vation of magnetic resonance signals from cesium at-
oms trapped in the isotropic bcc phase of solid 4He
[54]. In those experiments we recorded level crossing
signals (ground state Hanle effect) in longitudinal
and transverse fields as well as magnetic resonance
lines [54]. The widths of both the Hanle resonances
and the magnetic resonances (extrapolated to low rf
power) gave consistent values of 300 nT, which corre-
sponds to an effective T2 time of 150 �s. Although
more than two orders of magnitude larger than the
corresponding T2 times in superfluid helium, those
values were almost four orders of magnitude lower
than the T1 time in solid helium. After improving our
apparatus we have redetermined the T2 times in 1996
[58] using the technique of «FID (free induction
decay) in the dark» and found a lower limit
T2 108 3� ( ) ms for the transverse spin relaxation time,
which is only one order of magnitude below the T1
time. There is no a priori reason why the T1 and T2
times of dilute Cs samples in condensed helium should
be significantly different. The T2 value of 108 ms can
be explained by a magnetic field inhomogeneity on the
order of 10 �7. Although a spin echo experiment would
be the technique of choice for measuring the intrinsic
T2 time, such an experiment was not yet carried out.
6.4. Multiphoton transitions in the ground state
of Cs in bcc 4He
The normal magnetic resonance transition is a
�M � � 1 magnetic dipole transition, in which one rf
photon of a given helicity is absorbed between adja-
cent Zeeman sublevels . The selection rules imply that
resonances between states with magnetic quantum
numbers differing by �M N� � can also be driven by
the simultaneous absorption of N photons of the same
helicity. In the case of a linear Zeeman splitting of the
different M levels all these processes are resonant at
1310 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
the same frequency, i.e., when the rf frequency�rf co-
incides with the Larmor frequency �0. When one in-
creases the magnetic field, the quadratic Zeeman ef-
fect splits the magnetic levels of a state with angular
momentum F into 2 1F � components between which
2F individual one-photon resonances can be driven
(Fig. 14,a) which form a spectrum of 2F equally
spaced lines (Fig. 14,b, bottom). This forms the basis
for observing the higher order transitions. Increasing
the rf power the 2 1F � individual two-photon reso-
nances appear next as narrow peaks located at the
midpoints between the power broadened and satu-
rated one-photon resonances (Fig. 14,a). This behav-
ior continues when the rf power is further increased,
and broadens/saturates the two-photon resonances,
after which 2 2F � individual three-photon resonances
appear, this time again at the positions of the one pho-
ton resonances. The procedure repeats until the high-
est order process, i.e., a transition between the levels
M F� � and M F� involving the absorption of 2F
photons is reached. The long spin relaxation times and
the narrow magnetic resonance lines of cesium in bcc
solid helium make this sample well suited for the
study of those processes. In an experimental study
[62] we have indeed observed (Fig. 14,b) all
multiphoton processes in the F � 4 hyperfine ground
state up to the process of simultaneous absorption of
8 rf photons (narrowest line in top spectrum of
Fig. 14,b). In a subsequent detailed theoretical analy-
sis we have investigated the influence of different
relaxation mechanisms on the shape of these multipho-
ton spectra [63].
6.5. Hyperfine transitions
Magnetic resonance transitions can not only be
driven within a given F multiplet (process A in
Fig. 15), but also between the two hyperfine multi-
plets (process B in Fig. 15). In the free cesium atom
the corresponding resonance frequency is on the order
of 9.2 GHz (clock transition). We have studied this
intermultiplet transition with cesium in the bcc and in
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1311
a b
Fl
u
o
re
sc
e
n
ce
,
ar
b
. u
n
its
MF
N = 1
N = 1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
F = 4
2.0
1.5
1.0
0.5
0
� �rf ref– , kHz
–2 0 2 4 6 8 10 12
4
3
2
1
0
–1
–2
–3
–4
Fig. 14. (a) Possible transitions involving absorption of N rf photons between the Zeeman sublevels of the F = 4 ground
state of the Cs atom, in a magnetic field which produces a linear and quadratic Zeeman shift. (b) Magnetic resonance
spectra showing the multiphoton transitions of (a). The rf power increases from the bottom to the top.
– – – –
Fig. 15. Hyperfine-Zeeman structure of the ground state
of cesium in a static magnetic field of 4.3 �T with
intramultiplet (A) and intermultiplet (B) magnetic reso-
nance transitions observed in solid helium. The solid dots
represent relative sublevel populations achieved by optical
pumping with circularly polarized light.
the hcp phase of 4He [64]. In bcc phase we found that
the transition is blue shifted by approximately 200
MHz with respect to the transition in the free atom
(see Fig. 18,b) and that this shift increases with pres-
sure at a rate of � 1.4 MHz/bar. The blue shift can be
explained by the fact that the He matrix compresses
the wave function of the valence electron, so that
| ( )|
/
�6
2
1 2
0S R � and thus the hyperfine coupling con-
stant of the Fermi contact interaction increases. The
quantitative value of the shift and its pressure depend-
ence can be predicted in a satisfactory way by the bub-
ble model [20,65]. The linewidth of the hyperfine
transition in bcc 4He was found to be on the order of
100 kHz, i.e., 10 4 times larger than the width of the
intermultiplet magnetic resonance transitions dis-
cussed above.
6.6. The Stark effect of Cs in bcc 4He
Recently we have investigated the effect of a static
electric field on the properties of the cesium ground
state [66]. The motivation for this study was twofold:
firstly, the quadratic Stark effect constitutes a back-
ground signal which may induce systematic errors in
experiments searching for electric dipole moments,
and secondly, there has been a 40-year-old discrepancy
between theoretical and experimental values of the
tensor polarizability of the cesium ground state. The
energy of the 6S1 2/ ground state magnetic sublevel
| ,F M� is shifted by a static external electric fieldE ac-
cording to
�E S F M S F M( , , ) ( , , ) ,/ /6
1
2
61 2 1 2
2� � � E (4)
where the polarizability �( , , )/6 1 2S F M is given by
�
� � �
( , , )
( ) ( )
( )
(
/
( ) ( ) ( )
6
3 1
2
1 2
0
2
0
3
2
3
2
S F M
F F
M F F
I
�
� � �
� �
I � 1)
. (5)
In Eq. (5) one distinguishes three contributions,
viz., the scalar polarizability, �0
2( ), which arises in se-
cond order perturbation theory and which leads to a
common (i.e., F and M independent) shift of all
levels, an F-dependent scalar polarizability, �0
3( ),
which arises in third order perturbation theory (in-
volving the hyperfine interaction), and a third order
tensor polarizability, �2
3( ), which provides F and M
dependent level shifts. For cesium, the values
of the three contributions are in the ratio
� � �0
2
0
3
2
3( ) ( ) ( ): : � 1:10 �5:10 �7. It was the tiny tensor
polarizability �2
3( ) which was at the center of our in-
terest. A few years ago we had remeasured �2
3( ) in a
thermal atomic beam and had confirmed earlier exper-
imental results [67]. Because of the narrow magnetic
resonance lines obtained with Cs in bcc 4He, an
ODMR measurement in an external field was the
method of choice for this study.
For the Stark shift experiments we equipped the
pressure cell as described in Sec. 4.3. By using a set of
mirrors, we could easily switch between the Mx and
the Mz configurations (Fig. 13) by a simple beam
translation (Fig. 5,b).
In the strongly polarized ground state, the
magnetic resonance is dominated by the
| , | ,F M� � � � �4 4 4 3 transition and the differential
Stark shift of the two involved levels appears as a shift
of the magnetic resonance frequency proportional to
E 2 due to the �2
3 4( )( ) contribution of Eq. (5). The in-
set in Fig. 16 illustrates the magnetic resonance shift
in an electric field as measured in the Mz geometry
(Fig. 13,a). Because of an unexplained drift of the
magnetic resonance frequency, the measurement of the
Stark shift from a series of individual scans in differ-
ent electric fields yielded a large scattering of the data
points. We have therefore used a faster recording
method which consisted of locking the rf frequency to
the Larmor frequency [58] using feedback in the Mx
configuration (Fig. 13,b). This has allowed us the di-
rect, and therefore faster, recording of the Larmor fre-
quency changes upon application of the electric field.
The resulting (quadratic) electric field dependence of
the Larmor frequency is shown in Fig. 16, from which
one can extract the tensor polarizability � �
2
3 . The re-
sult � �
2
3 24 3 31 30 10( ) . ( )F � � � � � Hz/(kV/cm)2 is
in good agreement with the earlier atomic beam exper-
iments (see [67] and references therein). In parallel
we reanalyzed the theoretical calculation of the Stark
1312 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
0
2
4
6
8
10
12
14
R
e
so
n
an
ce
sh
ift
,H
z
–40 –20 0 20 40
E, kV/cm
20.45 20.55 20.65Fl
u
o
re
sc
e
n
ce
,a
rb
.u
n
its
17 Hz
E = 40 kV/cmE = 0
rf , kHz
Fig. 16. Quadratic Stark shift of the magnetic resonance
frequency in a magnetic field of 6 �T recorded with a
feedback lock of the rf frequency to the center of the mag-
netic resonance line in the Mx geometry. The inset shows
a particular value of the Stark shift recorded in the rf
scanning mode.
shift, thereby identifying terms omitted in earlier
treatments as well as a sign error. The numerical eval-
uation of those terms has allowed us to reconcile ex-
perimental and theoretical values of � �
2
3 after 40 years
of discrepancy [66,68].
7. Deformed bubbles
As discussed in Sec. 5, the spherical bubble model
described in Sec. 3 has proven to be very effective for
describing atomic absorption and emission spectra in
an isotropic environment such as the liquid or bcc
phase of solid He. Here we want to address the exten-
sion of the model to bubbles with static or dynamic
deformations. The uniaxial hcp phase of solid helium
has anisotropic elastic constants, so that the repulsive
force exerted by the dopant is counteracted by aniso-
tropic restoring forces, leading to deformed bubbles.
As the anisotropies are small, it is reasonable to des-
cribe the deformation in lowest order by a quadru-
polar shape which can be parametrized as
R
R
P Pb( )
( cos ) (cos ) (cos ) ,
�
� � � � � �
0
2
0 21 3 1 2� � � �
(6)
where the Pl (cos )� are Legendre polynomials. The
bubble radius parameter Rb of Eq. (2) depends on the
azimuthal angle � measured with respect to the bub-
ble axis, and � characterizes the bubble deformation.
In a similar way as described in Sec. 3, the three pa-
rameters R0, �, and � can be found by minimizing the
defect energy. Besides these static deformations in the
hcp phase one also has to consider dynamical defor-
mations (in both phases). The latter are due to the
zero-point fluctuations of the helium atoms on the
bubble interface which may be expressed as the sum
over (uncorrelated) shape oscillations of different
multipole orders.
Different effects have been identified which can be
traced back to dynamic and static deformations. Dy-
namic quadrupole deformations (oscillations) were
first considered in [69] for explaining the characteris-
tic doubly-shaped contour of the D2 excitation line of
Rb and Cs in superfluid He (see also Fig. 6,b). Such
anisotropic oscillations split the absorption line due to
a dynamic Jahn—Teller effect. A more general ap-
proach including monopole (breathing), dipole, and
quadrupole modes of the bubble oscillations was ap-
plied in [70,71] for the analysis of the excitation spec-
tra of Yb � ions and Ca atoms in liquid helium. As
mentioned above in Sec. 6.3 quadrupolar oscillations
may also be responsible for the finite, but rather long
spin relaxation time of alkali atoms in superfluid and
bcc matrices, and finally the symmetric monopole
(breathing mode) oscillations were suggested to ex-
plain the relatively large width of the hyperfine mag-
netic resonance transition [20,64] (Sec. 6.5).
Figure 17 gives an overview of bubble effects when
going, first from vacuum to spherical bubbles and
then to bubbles with a static deformation. The effect
of a spherical bubble on the optical transitions was
discussed in detail above. Here we address the addi-
tional effects induced by a (small) static quadrupolar
bubble deformation superimposed on the spherical
bubble as it arises when going from bcc to hcp 4He.
As shown in Fig. 17 the deformation affects the opti-
cal transitions, the hyperfine structure and the
Zeeman structure of the individual hyperfine levels.
When going from the bcc to the hcp phase by increas-
ing the helium pressure at constant temperature, the
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1313
a b c
d
–
–
–
–
–
Fig. 17. Wavelength of the D1 transition and hyperfine frequency of the free Cs atom (a); effects of the spherical bubble
in bcc 4He on the optical and microwave transitions (b); additional perturbations of the optical, the hyperfine and the
Zeeman transitions due to deformed bubbles (c); the deformed bubble seen as a spherical bubble with a quadrupolar
change of He density at its interface and the corresponding perturbed ground state of Cs interpreted as an S orbital with
a D orbital admixture (d).
atomic density rises and one would naively expect that
any line shifts induced by the symmetric bubble in bcc
phase would further increase at the phase transition.
In this sense the optical absorption frequency and the
hyperfine transition frequency, which are both shifted
to the blue by the spherical bubble, would experience
a further blue shift. However, the experiments have
revealed that both frequencies jump to the red at the
bcc � hcp phase transition, after which they move
again towards the blue upon a further pressure in-
crease (Figs. 7 and 18,b).
The theoretical treatment of deformation induced
effects starts from the interpretation of a bubble de-
formation in terms of a prolate (oblate) quadrupolar
change of the surrounding helium density: He atoms
are added (removed) along one direction and removed
(added) along the orthogonal directions. For symme-
try reasons the corresponding perturbation operator
will have the form
V f r Pdef bub � ( ) ( )2 � , (7)
which has the symmetry of a second rank tensor that
perturbs the 6S state by admixtures of nD states ac-
cording to
|
| |
6
6
1 2
S
S nDn
n
n
n
�
� �
� � �
�
�
�
�
�
, (8)
where the S–D mixing coefficients �n are given by
�n
S nD
nDV S
E E
�
� �
�
| |
.
def bub 6
6
(9)
The energy minimization procedure applied to the
(deformed) bubble model then establishes a relation
between the bubble deformation parameter � and the
S–D mixing coefficients �n . In this way we have inter-
preted the hyperfine frequency red shift at the
bcc � hcp transition as originating from a loss of the
S character of the ground state wave function accord-
ing to Eq. (8). This has allowed us [56] to infer a
6S–5D mixing coefficient of 3% and from that a bub-
ble deformation parameter of � = 6–7%.
For the explanation of the observed red shift of the
D1 absorption line [20] we propose the following
qualitative explanation. The spherical 6S ground state
does not fit into the deformed bubble and the interac-
tionVdef bub increases its energy. At the same time the
spherically symmetric excited 6P1 2/ state is mixed by
Vdef bub with the nearby 6P3 2/ state to produce a de-
formed state, which has a better fit to the bubble
shape, so that the energy increase of the exited state is
smaller than the one of the ground state. As a result
the transition wavelength becomes red-shifted. A fur-
ther increase in pressure tends to render the bubble
more spherical, hence the ensuing correlated blue shift
of the transition frequencies. We are currently work-
ing on the quantitative investigation of this hypo-
thesis.
Finally, we discuss the most remarkable effect that
deformed bubbles have on the intramultiplet Zeeman
transitions (transitions A in Fig. 15). In Fig. 18,a we
compare the ultranarrow magnetic resonance lines ob-
tained with Cs in the bcc phase to the extremely broad
magnetic resonance spectrum observed in the hcp
phase under otherwise identical conditions. The spec-
trum is broadened by more than three orders of magni-
tude and shows a rich substructure which indicates a
lifting of the Zeeman degeneracy induced by the bub-
ble deformation. The validity of this assumption has
been demonstrated in a convincing way by the obser-
1314 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
Radio frequency, kHz
Fl
u
o
re
sc
e
n
ce
si
g
n
al
, a
rb
. u
n
its
Fl
u
o
re
sc
e
n
ce
si
g
n
al
, a
rb
. u
n
its
Microwave frequency, GHz
a b
F
Fig. 18. Comparison of intramultiplet (Zeeman) magnetic resonance spectrum of cesium in bcc and hcp 4He (a). Compa-
rison of intermultiplet (hyperfine) magnetic resonance spectrum of cesium in bcc and hcp 4He. Both spectra were re-
corded in a single scan in which the phase transition was induced by a pressure change during the scan between the two
resonances (b).
vation of a zero magnetic field magnetic resonance
spectrum (Fig. 19,a) recorded with the arrangement
of Fig. 13,a, with B0 = 0. As shown in the inset, the
main features of the spectrum can be understood as re-
sulting from a quadrupolar shift of the energies of the
levels | , ,/6 1 2S F M� according to
�
F M Q M F F, [ ( )] .� � �3 12 (10)
In the Fig. 19,b we show the dependence of the de-
gree of spin polarization P Jz z� � � and of the longitu-
dinal relaxation time T1 on the magnetic holding
field. Both quantities are much smaller than in bcc
4He due to the faster spin relaxation experienced in
hcp 4He. In a low magnetic field, the cesium spin
couples to the bubble axis and only its projection
onto this axis is conserved. The maximal energy of
the coupling may be estimated from Fig. 19,a to be on
the order of 10 kHz. Since the orientations of the
bubble axes in space are most probably distributed
randomly (polycristalline sample) an effective spin
depolarization is generated. In very large magnetic
fields the spin couples to the magnetic field and its
projection along the field becomes the conserved
quantity. In this sense the magnetic field dependen-
cies shown in the Fig. 19,b can be interpreted as a
magnetic field induced decoupling of the spins from
their interaction with the deformed bubbles. The
kinks in these dependencies mark the intermediate re-
gion between the discussed extreme cases, and the
fact that the kinks occur at a Larmor frequency of
� 10 kHz (in accordance with the highest energy fea-
ture of the zero-field magnetic resonance spectrum
Fig. 19,a) gives further support to this interpretation.
Note that this view is analogous to the well-known
Breit—Rabi problem in which the magnetic field de-
couples pairs of angular momenta from their mutual
(fine or hyperfine) interaction in low fields.
Naively one might expect that Eq. (10) represents just
the eigenvalues of the interaction Vdef bub (Eq. (7)).
However the matrix elements � �6 61 2 1 2S V S/ /| |def bub
vanish asVdef bub acts only in the space of orbital vari-
ables. This is similar to the quadratic Stark effect due
to the perturbation V ez rPSt � � E 1(cos )� which
does not lift the Zeeman degeneracies in the first and
second order. However, it is known [72] that the sec-
ond rank tensor part of the third order interaction
T0
2( )( )Stark � [ ]( )V H VSt hf St! ! 0
2 produces a quad-
rupolar lifting of the Zeeman degeneracies in S /1 2
state. In a similar way one can show that the eigen-
energies of Eq. (10) correspond to the expectation
values of the second rank operator
T V H V0
2
0
2( ) ( )( ) [ ]def bub def bub hf def bub� ! ! . (11)
The coupling of the atomic spin, or rather the
atomic magnetic moment, to the axis of a nonspherical
bubble is very weak with an interaction energy on the
order of 10 kHz, i.e., some 3·10 �7 cm �1. The weakness
may be appreciated by comparing it to the coupling of
nuclear quadrupole moments to electric field gradients
in uniaxial crystals, which are typically measured in
megahertz.
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1315
Radio frequency, kHz
Fl
u
o
re
sc
e
n
ce
si
g
n
al
, a
rb
. u
n
its
,
,
,
a b
Fig. 19. Zero field magnetic resonance spectrum in the F = 4 ground state of cesium trapped in hcp 4He. The inset shows
the level structure expected from a third order perturbation by quadrupolar bubble deformations. The lines are slightly rf
power broadened and the weaker lines at the midpoints between the main lines correspond to �M � 2 transitions induced
by the absorption of 2 rf photons. The magnetic resonance spectrum shown in Fig. 18,a can be understood as the Zeeman
splitting of the spectrum shown here (a). Spin polarization (top) and longitudinal relaxation rate (bottom) as a function
of the magnetic field. Decoupling occurs at a Larmor frequency of 10 kHz, which coincide with the rightmost resonance
of the zero field spectrum (b).
8. Search for an electric dipole moment
of the electron
Our original motivation for investigating alkali at-
oms in condensed helium matrices was the hope that
such samples might be used for the search of a perma-
nent electric dipole moment (EDM) of the electron
[51,73–75]. An EDM violates the discrete symmetries
of parity and time reversal invariance and the exis-
tence of such a moment would advance modern parti-
cle physics theories beyond their presently best de-
scription in terms of the standard model. EDM
experiments are carried out with neutral species in-
volving an unpaired electron and aim at detecting a
shift of a magnetic resonance frequency proportional
to the strength of an externally applied electric field.
Sensitive experiments therefore call for narrow mag-
netic resonance lines, i.e., long spin relaxation times
and environments which can sustain large electric
fields. It was speculated [51] that alkali atoms in solid
helium may fulfill the former condition, while it was
known that pure liquid or solid helium has a very high
dielectric strength. As discussed above it was later
found that long spin coherence times are impossible to
realize in He II, but it was subsequently shown that
one can indeed observe extremely narrow magnetic
resonance lines with a good signal/noise ratio with Cs
and Rb in solid helium.
After a decade of efforts towards an EDM experi-
ment with cesium in bcc 4He we have now reached the
conclusion that it is not possible to realize an experi-
ment which is competitive to other past and ongoing
alternative approaches. This conclusion is based on
multiple reasons. Although the combination of mag-
netic resonance linewidth and signal/noise ratio are
comparable to what can be reached in thermal atomic
beams, we found recently that the maximum electric
field that a doped sample can hold is on the order of
50 kV/cm, which is more than a factor of two less
than the field used in the Tl beam experiment which
has produced the currently lowest upper limit on the
electron EDM [76]. We have also found that the mag-
netic resonance frequency shows a slow drift (about
2 mHz/s) in time [20,66]. The origin of the drift is
not clear, although we have shown that it is not corre-
lated with magnetic field drifts in our 3-fold shield
nor with pressure or temperature changes of the crys-
tal. Such a drift is of course detrimental for EDM
expriments in which one wants to compare magnetic
resonance frequencies in a geometry in which E is par-
allel to B0 with one where E is antiparallel to B0. In
Sec. 7 we have shown that a bubble deformation on
the order of 6–7% causes an interaction energy of
10 kHz. The present upper limit of the electron EDM
of 10 �27 e·cm corresponds to an EDM of the cesium
atom of approximately 10 �25 e·cm, which, in an elec-
tric field of 50 kV/cm, induces an electric Larmor
frequency dE/h of 1 �Hz. In order not to be perturbed
by bubble deformations the latter have to be stable at
a level of � " 10 �6, when assuming a quadratic rela-
tion between the bubble deformation and the corre-
sponding shift of the Zeeman levels (see Eq. (11)). In
order to avoid systematic effects related to bubble de-
formations, one has to ascertain that any deformation
of the 1 nm sized bubbles correlated with the direc-
tion of the electric field is less than 10 �13 cm. At
present we have no idea how to achieve such precision.
Besides these fundamental and technical limita-
tions there are also practical limitations which have
forced us to abandon the EDM related branch of our
research. Successful EDM experiments require a long
data integration time in order to reach the required
statistical sensitivity and for performing control mea-
surements of systematic effects. Efficient experiments
thus need a highly reliable apparatus, in which most
of the data acquisition can be automatized, a goal
which is hard to realize with the experiments on doped
helium crystals. The helium holding time of the
cryostat equipped for Stark spectroscopy is on the or-
der of 2 days, out of which typically 10 hours can be
used for actual data taking. The present system does
not allow a refilling of the helium bath without bring-
ing the bath up to 4 K and destroying the crystal. De-
spite many efforts in the development of low tempera-
ture and magnetic resonance compatible sealing
techniques for the windows and the different
feedthroughs of the cell [20] superleaks which mani-
fest only at the cryogenic temperatures appear fre-
quently following thermal cycling of the cell. More-
over, the reproducibility of the signal quality in each
newly grown and doped crystal is another unsolved is-
sue. All of these features make the experiments on
doped crystals very demanding and far from allowing
their automatization. Last, but not least we mention
the recent trend to move from paramagnetic atoms to
paramagnetic molecules for searching an electron
EDM.
In polar paramagnetic molecules with one heavy
atom (such as YbF or PbO) whose axes are aligned by
an external field E ext the electron experiences an in-
ternal electric field E intwhich is up to five orders of
magnitude larger than E ext [77]. This large enhance-
ment factor, compared to atoms makes such samples
the most promising candidates to search for the elec-
tron EDM.
9. Concluding remarks
Starting from the initial proposal by Sergei
Kanorsky from the Lebedev Physical Institute in Mos-
1316 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
cow to use heavy paramagnetic atoms embedded in
superfluid helium in an alternative experiment to
search for a permanent electric dipole moment of the
electron we have initiated in the early 1990’s a related
low temperature experiment in the laboratory of T.W.
H�nsch at the Max-Planck-Institute for Quantum Op-
tics in Garching (Germany). Today, 15 years later, we
have reached the conclusion that such EDM experi-
ments are not feasible with the present day technology
and knowledge. However, we may claim that our ef-
forts regarding the many aspects of such a difficult ex-
periment have been quite fruitful in the sense that
they have provided a wealth of information in the for-
merly unexplored field of solid helium matrix isola-
tion spectroscopy. The present review has given an ac-
count of these different facets. Many unexpected
features were discovered and extensions of the basic
atomic bubble model were able to explain most of
them. Because of lack of manpower, we were not able
to pursue in depth many of the interesting and promis-
ing side roads which have opened up, such as the study
of the exciplex lifetimes and the recently observed
quenching of molecular excited states. Nonetheless,
many questions could be answered such as the mystery
about the quenching of fluorescence from the lighter
alkalis by the discovery of exciplexes, the long stand-
ing discrepancy between experiment and theory of
tensor polarizabilities, and insights into bubble defor-
mation induced effects.
It is also interesting to note the intimate relation of
the present studies with the field of doped helium
nanoclusters which has risen, in parallel to our activi-
ties, over the past decade. In particular the recent dis-
covery of exciplex spectra as well as of homo- and
heteronuclear dimers in solid helium has further con-
nected the two fields.
The current difficult funding situation makes our
own future in the field of sHeMIS uncertain and the
time may have come to pass the sHeMIS torch to other
groups. We are aware at least of one group (Ph.
Jacquier and J. Dupont-Roc at LKB in Paris) who
have taken up this research and are preparing a chal-
lenging experiment to measure the anapole moment of
the cesium nucleus. The anapole moment is the lowest
order nuclear magnetization multipole moment which
violates parity but conserves time reversal invariance.
It was shown by M.A. Bouchiat and C. Bouchiat [78]
that an atom with an anapole moment will exhibit a
linear Stark effect when placed in an environment
which has a quadrupolar symmetry. Based on our dis-
covery of quadrupolar bubble deformations it was sug-
gested in [78] to use Cs in an uniaxial hcp He crystal
for realizing such an anapole moment experiment.
Acknowledgements
The research work presented above is the fruit of
more than a decade of experimental and theoretical
efforts carried out by a number of undergraduate stu-
dents, Ph. D. students, and postdocs, and with the in-
valuable support of mechanical and electronics techni-
cians and engineers. All their individual contributions
are acknowledged. The doctoral thesis works that
have emerged from this line of research were presented
by Markus Arndt [79] and Stephen Lang [80] at the
Ludwig-Maximilians Universit�t in Munich, by Taro
Eichler [81] at the Friedrich-Wilhelms-Universit�t in
Bonn and more recently by Daniel Nettels [82],
Reinhard M�ller-Siebert [20] and Simone Ulzega [66]
at the University of Fribourg. One of us (A.W.) ac-
knowledges his colleague and friend Sergei Kanorsky
who submitted the original EDM proposal to the
Max-Planck-Institute for Quantum Optics and with
whom he initiated the first heroic steps of this re-
search. We thank Paul Knowles for his critical read-
ing of the manuscript.
The research was funded partly by the individual
host institutions and by the national funding agencies
in Germany (Deutsche Forschungsgemeinschaft,
DFG) and in Switzerland (Swiss National Science
Foundation, SNF). We particularly acknowledge the
support of SNF by the grants 21-59451.99,
20-67008.01, and 200020-103864.
1. J.P. Toennies and A.F. Vilesov, Annu. Rev. Phys.
Chem. 49, 1 (1998).
2. S.I. Kanorsky and A. Weis, Optical and Magneto-op-
tical Spectroscopy of Point Defects in Condensed He-
lium, Academic Press (1998); B. Bederson and H.
Walther, Adv. in Atomic, Molecular and Optical Phy-
sics 38, 87.
3. R.E. Boltnev, G. Frossati, E.B. Gordon, I.N.
Krushinskaya, E.A. Popov, and A. Usenko, J. Low
Temp. Phys. 127, 245 (2002).
4. E.P. Bernard, R.E. Boltnev, V.V. Khmelenko, V.
Kiryukhin, S.I. Kiselev, and D.M. Lee, J. Low Temp.
Phys. 134, 133 (2004).
5. E.B. Gordon, A. Usenko, and G. Frossati, J. Low
Temp. Phys. 130, 15 (2003).
6. E.B. Gordon, Fiz. Nizk. Temp. 30, 1009 (2004) [Low
Temp. Phys. 30, 756 (2004)].
7. J. Wilks, The Properties of Liquid and Solid Helium,
Clarendon Press, Oxford (1967).
8. J. Pascale, Phys. Rev. A28, 632 (1983).
9. J. Jortner, W.R. Kestner, S.A. Rice, and M.H. Co-
hen, J. Chem. Phys. 43, 2614 (1965).
10. S.I. Kanorsky, M. Arndt, R. Dziewior, A. Weis, and
T.W. H�nsch, Phys. Rev. Â50, 6296 (1994).
11. Ò. Kinoshita, Ê. Fukuda, Y. Takahashi, and T.
Yabuzaki, Phys. Rev. A52, 2707 (1995).
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1317
12. S. Kanorsky, A. Weis, M. Arndt, R. Dziewior, and T.
H�nsch, Z. Phys. Â98, 371 (1995).
13. A.P. Hickman, W. Steets, and N.F. Lane, Phys. Rev.
Â12, 3705 (1975).
14. M. Arndt, R. Dziewior, S. Kanorsky, A. Weis, and T.
H�nsch, Z. Phys. Â98, 377 (1995).
15. H.C. van de Hulst, Light Scattering by Small Parti-
cles, Dover Publications, New York (1981).
16. D. Nettels, A. Hofer, P. Moroshkin, R. M�ller-Siebert,
S. Ulzega, and A. Weis, Phys. Rev. Lett. 94, 063001
(2005).
17. P. Moroshkin, A. Hofer, D. Nettels, S. Ulzega, and A.
Weis, J. Chem. Phys. 124, 024511 (2006).
18. A. Hofer, P. Moroshkin, S. Ulzega, and A. Weis, Phys.
Rev. A74, No 4 (2006); arXiv:physics/0605181 (2006).
19. P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis, Phys.
Rev. A74, 032504 (2006).
20. R. M�ller-Siebert, Ph. D. thesis, University of Fri-
bourg, Switzerland (2002), unpublished.
21. S.I. Kanorsky, M. Arndt, R. Dziewior, A. Weis, and
T.W. H�nsch, Phys. Rev. Â49, 3645 (1994).
22. Ò. Kinoshita, Ê. Fukuda, Ò. Matsuura, and T.
Yabuzaki, Phys. Rev. A53, 4054 (1996).
23. J. Dupont-Roc, Z. Phys. Â98, 383 (1995).
24. Ê. Ishikawa, A. Hatakeyama, K. Gosyono-o, S. Wada,
Y. Takahashi, and T. Yabuzaki, Phys. Rev. Â56, 780
(1997).
25. A.I. Golov and L.P. Mezhov-Deglin, JETP Lett. 56,
514 (1992).
26. A.I. Golov and L.P. Mezhov-Deglin, Physica
Â194–196, 951 (1994).
27. A.I. Golov, Z. Phys. Â98, 363 (1995).
28. Ê. Enomoto, Ê. Hirano, M. Kumakura, Y. Takahashi,
and T. Yabuzaki, Phys. Rev. A66, 042505 (2002).
29. K. Hirano, K. Enomoto, M. Kumakura, Y. Takahashi,
and T. Yabuzaki, Phys. Rev. A68, 012722 (2003).
30. K. Enomoto, K. Hirano, M. Kumakura, Y. Takahashi,
and T. Yabuzaki, Phys. Rev. A69, 012501 (2004).
31. J. Reho, J. Higgins, Ñ. Callegari, K.K. Lehmann,
and G. Scoles, J. Chem. Phys. 113, 9686 (2000).
32. J. Reho, J. Higgins, K.K. Lehmann, and G. Scoles,
J. Chem. Phys. 113, 9694 (2000).
33. F.R. Bruhl, R.A. Trasca, and W.E. Ernst, J. Chem.
Phys. 115, 10220 (2001).
34. Ñ.P. Schulz, P. Claas, and F. Stienkemeier, Phys.
Rev. Lett. 87, 153401 (2001).
35. G. Droppelmann, O. B�nermann, Ñ.Ð. Schulz, and F.
Stienkemeier, Phys. Rev. Lett. 94, 023402 (2004).
36. O. B�nermann, M. Mudrich, M. Weidemuller, and F.
Stienkemeier, J. Chem. Phys. 121, 8880 (2004).
37. J.L. Persson, Q. Hui, Z.J. Jakubek, M. Nakamura,
and M. Takami, Phys. Rev. Lett. 76, 1501 (1996).
38. Y. Moriwaki and N. Morita, Eur. Phys. J. D5, 53
(1999).
39. Y. Moriwaki, K. Inui, K. Kobayashi, F. Matsushima,
and N. Morita, J. Mol. Structure 786, 112 (2006).
40. Y. Fukuyama, Y. Moriwaki, and Y. Matsuo, Phys.
Rev. A69, 042505 (2004).
41. P. Moroshkin, D. Nettels, A. Hofer, S. Ulzega, and A.
Weis, Proc. of SPIE 6257, 625706 (2006).
42. M. Zbiri and Ñ. Daul, Phys. Lett. A341, 170 (2005).
43. T. Takayanagi and M. Shiga, Phys. Chem. Chem.
Phys. 6, 3241 (2004).
44. F. Stienkemeier, J. Higgins, W.E. Ernst, and G.
Scoles, Phys. Rev. Lett. 74, 3592 (1995).
45. J. Higgins, Ñ Callegari, J. Reho, F. Stienkemeier,
W.E. Ernst, M. Gutowski, and G. Scoles, J. Phys.
Chem. 102, 4952 (1998).
46. J. Reho, J. Higgins, and K. Lehmann, Faraday Dis-
cuss. 118, 33 (2001).
47. F.R. Bruhl, R.A. Miron, and W.E. Ernst, J. Chem.
Phys. 115, 10275 (2001).
48. W.E. Ernst, R. Huber, S. Jiang, R. Beuc, M. Movre,
and G. Pichler, J. Chem. Phys. 124, 024313 (2006).
49. M. Mudrich, O. B�nermann, F. Stienkemeier, O. Du-
lieu, and M. Weidemuller, Eur. Phys. J. D31, 291
(2004).
50. Y. Takahashi, K. Sano, T. Kinoshita, and T. Yabu-
zaki, Phys. Rev. Lett. 71, 1035 (1993).
51. M. Arndt, S.I. Kanorsky, A. Weis, and T.W.
H�nsch, Phys. Lett. A174, 298 (1993).
52. T. Kinoshita, Y. Takahashi, and T. Yabuzaki, Phys.
Rev. Â49, 3648 (1994).
53. D. Budker, W. Gawlik, D.F. Kimball, S.M. Roches-
ter, V.V. Yashchuk, and A. Weis, Rev. Mod. Phys.
74, 1153 (2002).
54. M. Arndt, S.I. Kanorsky, A. Weis, and T.W.
H�nsch, Phys. Rev. Lett. 74, 1359 (1995).
55. S. Lang, M. Arndt, T.W. H�nsch, S.I. Kanorsky, S.
Lucke, S.B. Ross, and A. Weis, Fiz. Nizk. Temp. 22,
171 (1996) [Low Temp. Phys. 22, 129 (1996)].
56. S. Kanorsky, S. Lang, T. Eichler, K. Winkler, and A.
Weis, Phys. Rev. Lett. 81, 401 (1998).
57. S. Lang, S.I. Kanorsky, T. Eichler, R. M�ller-Siebert,
T.W. H�nsch, and A. Weis, Phys. Rev. A60, 3867
(1999).
58. S.I. Kanorsky, S. Lang, S. L�cke, S.B. Ross, T.W.
H�nsch, and A. Weis, Phys. Rev. A54, R1010 (1996).
59. T. Eichler, R. M�ller-Siebert, D. Nettels, S. Kanor-
sky, and A. Weis, Phys. Rev. Lett. 88, 123002 (2002).
60. T. Furukawa, Y. Matsuo, A. Hatekeyama, Y. Fuku-
yama, T. Kobayashi, H. Izumi, and T. Shimoda,
Phys. Rev. Lett. 96, 095301 (2006).
61. D. Pines and Ñ.Ð. Slichter, Phys. Rev. 100, 1014
(1955).
62. D. Nettels, R. M�ller-Siebert, S. Ulzega, and A. Weis,
Appl. Phys. Â77, 563 (2003).
63. D. Nettels, R. M�ller-Siebert, and A. Weis, Appl.
Phys. Â77, 753 (2003).
64. S. Lang, S. Kanorsky, M. Arndt, S.B. Ross, T.W.
H�nsch, and A. Weis, Europhys. Lett. 30, 233 (1995).
65. S. Kanorsky and A. Weis, in: NATO Advanced Study
Institute «Quantum Optics of Confined Systems», Klu-
wer Academic Publishers (1995).
66. S. Ulzega, Ph. D. thesis, University of Fribourg,
Switzerland (2006), unpublished.
1318 Fizika Nizkikh Temperatur, 2006, v. 32, No. 11
P. Moroshkin, A. Hofer, S. Ulzega, and A. Weis
67. Ñ Ospelkaus, U. Rasbach, and A. Weis, Phys. Rev.
A67, 011402 (2003).
68. S. Ulzega, A. Hofer, P. Moroshkin, and A. Weis,
arXiv:physics/0609107 (2006).
69. T. Kinoshita, K. Fukuda, and T. Yabuzaki, Phys.
Rev. Â54, 6600 (1996).
70. Y. Moriwaki and N. Morita, Eur. Phys. J. D13, 11
(2001).
71. Y. Moriwaki and N. Morita, Eur. Phys. J. D33, 323
(2005).
72. J.R.P. Angel and P.G.H. Sandars, Proc. R. Soc.
London A305, 125 (1968).
73. A. Weis, S. Kanorsky, M. Arndt, and T.W. H�nsch,
Z. Phys. Â98, 359 (1995).
74. A. Weis, Hunting the Electron Electric Dipole Mo-
ment, Plenum Press (1997), Electron Theory and
Quantum Electrodinamics — 100 Years Later, J.P.
Dowing (ed.), Proc. of ASI B358, 149.
75. A. Weis, S. Kanorsky, S. Lang, and T.W. H�nsch,
Can Atoms Trapped in Solid Helium be Used to
Search for Physics Beyond the Standard Model?
Springer (1997); K. Jungmann, J. Kowalski, I. Rein-
hard, and F. Tr�ger (eds.), Atomic Physics Methods in
Modern Research, Lecture Notes in Physics, 499, 57.
76. B.C. Regan, E.D. Commins, Ñ.J. Schmidt, and D.
DeMille, Phys. Rev. Lett. 88, 071805 (2002).
77. J.J. Hudson, B.E. Sauer, M.R. Tarbutt, and E.A.
Hinds, Phys. Rev. Lett. 89, 023003 (2002).
78. M.A. Bouchiat and Ñ. Bouchiat, Eur. Phys. J. D15,
5 (2001).
79. M. Arndt, Ph. D. thesis, Ludwig Maximilians Univer-
sit�t M�nchen (1995), Max-Planck-Institut f�r Quan-
tenoptik, Garching, MPQ-Report 197.
80. S. Lang, Ph. D. thesis, Ludwig Maximilians Uni-
versit�t M�nchen (1997), unpublished.
81. T. Eichler, Ph. D. thesis, University of Bonn, Ger-
many (2000), unpublished.
82. D. Nettels, Ph. D. thesis, University of Fribourg,
Switzerland (2003), unpublished.
Spectroscopy of atomic and molecular defects in solid 4He
Fizika Nizkikh Temperatur, 2006, v. 32, No. 11 1319
|
| id | nasplib_isofts_kiev_ua-123456789-120885 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 0132-6414 |
| language | English |
| last_indexed | 2025-12-07T16:50:44Z |
| publishDate | 2006 |
| publisher | Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України |
| record_format | dspace |
| spelling | Moroshkin, P. Hofer, A. Ulzega, S. Weis, A. 2017-06-13T08:59:48Z 2017-06-13T08:59:48Z 2006 Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) / P. Moroshkin, A. Hofer, S. Ulzega, A. Weis // Физика низких температур. — 2006. — Т. 32, № 11. — С. 1297–1319. — Бібліогр.: 82 назв. — англ. 0132-6414 PACS: 32.30.–r, 32.60.+i, 33.35.+r, 33.50.–j, 67.80.–s, 76.70.Hb https://nasplib.isofts.kiev.ua/handle/123456789/120885 A little more than a decade ago our team extended the field of defect spectroscopy in
 cryocrystals to solid ⁴He matrices, in both their body-centered cubic (bcc) and hexagonally
 close-packed (hcp) configurations. In this review paper we survey our pioneering activities in the
 field and compare our results to those obtained in the related fields of doped superfluid helium and
 doped helium nanodroplets, domains developed in parallel to our own efforts. We present experimental
 details of the sample preparation and the different spectroscopic techniques. Experimental
 results of purely optical spectroscopic studies in atoms, exciplexes, and dimers and their interpretation
 in terms of the so-called bubble model will be discussed. A large part of the paper is devoted
 to optically detected magnetic resonance, ODMR, processes in alkali atoms. The quantum nature
 of the helium matrix and the highly isotropic shape of the local trapping sites in the bcc phase
 make solid helium crystals ideal matrices for high resolution spin physics experiments. We have investigated
 the matrix effects on both Zeeman and hyperfine magnetic resonance transitions and
 used ODMR to measure the forbidden electric tensor polarizability in the ground state of cesium.
 Several unexpected changes of the optical and spin properties during the bcc—hcp phase transition
 can be explained in terms of small bubble deformations. The research work presented above is the fruit of
 more than a decade of experimental and theoretical
 efforts carried out by a number of undergraduate students,
 Ph. D. students, and postdocs, and with the invaluable
 support of mechanical and electronics technicians
 and engineers. All their individual contributions
 are acknowledged. The doctoral thesis works that
 have emerged from this line of research were presented
 by Markus Arndt [79] and Stephen Lang [80] at the
 Ludwig-Maximilians Universitt in Munich, by Taro
 Eichler [81] at the Friedrich-Wilhelms-Universitt in
 Bonn and more recently by Daniel Nettels [82],
 Reinhard M ller-Siebert [20] and Simone Ulzega [66]
 at the University of Fribourg. One of us (A.W.) acknowledges
 his colleague and friend Sergei Kanorsky
 who submitted the original EDM proposal to the
 Max-Planck-Institute for Quantum Optics and with
 whom he initiated the first heroic steps of this research.
 We thank Paul Knowles for his critical reading
 of the manuscript.
 The research was funded partly by the individual
 host institutions and by the national funding agencies
 in Germany (Deutsche Forschungsgemeinschaft,
 DFG) and in Switzerland (Swiss National Science
 Foundation, SNF). We particularly acknowledge the
 support of SNF by the grants 21-59451.99,
 20-67008.01, and 200020-103864. en Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України Физика низких температур Quantum Liquids and Solids Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) Article published earlier |
| spellingShingle | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) Moroshkin, P. Hofer, A. Ulzega, S. Weis, A. Quantum Liquids and Solids |
| title | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) |
| title_full | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) |
| title_fullStr | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) |
| title_full_unstemmed | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) |
| title_short | Spectroscopy of atomic and molecular defects in solid ⁴He using optical, microwave, radio frequency, magnetic and electric fields (Review Article) |
| title_sort | spectroscopy of atomic and molecular defects in solid ⁴he using optical, microwave, radio frequency, magnetic and electric fields (review article) |
| topic | Quantum Liquids and Solids |
| topic_facet | Quantum Liquids and Solids |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/120885 |
| work_keys_str_mv | AT moroshkinp spectroscopyofatomicandmoleculardefectsinsolid4heusingopticalmicrowaveradiofrequencymagneticandelectricfieldsreviewarticle AT hofera spectroscopyofatomicandmoleculardefectsinsolid4heusingopticalmicrowaveradiofrequencymagneticandelectricfieldsreviewarticle AT ulzegas spectroscopyofatomicandmoleculardefectsinsolid4heusingopticalmicrowaveradiofrequencymagneticandelectricfieldsreviewarticle AT weisa spectroscopyofatomicandmoleculardefectsinsolid4heusingopticalmicrowaveradiofrequencymagneticandelectricfieldsreviewarticle |