Content uploaded by Tetsuya Ishikawa
Author content
All content in this area was uploaded by Tetsuya Ishikawa
Content may be subject to copyright.
This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address: 94.154.24.1
This content was downloaded on 31/10/2013 at 21:27
Please note that terms and conditions apply.
Unusual under-threshold ionization of neon clusters studied by ion spectroscopy
View the table of contents for this issue, or go to the journal homepage for more
2013 J. Phys. B: At. Mol. Opt. Phys. 46 164023
(http://iopscience.iop.org/0953-4075/46/16/164023)
Home Search Collections Journals About Contact us My IOPscience
IOP PUBLISHING JOURNAL OF PHYSICS B: ATOMIC, MOLECULAR AND OPTICAL PHYSICS
J. Phys. B: At. Mol. Opt. Phys. 46 (2013) 164023 (5pp) doi:10.1088/0953-4075/46/16/164023
Unusual under-threshold ionization of
neon clusters studied by ion spectroscopy
K Nagaya1,2, A Sugishima1,2, H Iwayama1,2,HMurakami
1,2,MYao
1,2,
H Fukuzawa2,3, X-J Liu2,3,KMotomura
2,3,KUeda
2,3,NSaito
2,4,
LFoucar
2,3,5, A Rudenko2,6,MKurka
2,3,7,K-UK
¨
uhnel2,7, J Ullrich2,7,8,
A Czasch9,RD
¨
orner9,RFeifel
10, M Nagasono2, A Higashiya2,
M Yabashi2, T Ishikawa2, T Togashi2,11,HKimura
2,11 and H Ohashi2,11
1Department of Physics, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan
2RIKEN, XFEL Project Head Office, Hyogo 679-5148, Japan
3Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577,
Japan
4National Metrology Institute of Japan, AIST, Tsukuba 305-8568, Japan
5Max-Planck-Institut fuer Medizinische Forschung, Jahnstraße 29, D-69120 Heidelberg, Germany
6J R Macdonald Laboratory, Department of Physics, Kansas State University, Manhattan, KS 66506,
USA
7Max-Planck-Insitut f¨
ur Kernphysik, D-69117 Heidelberg, Germany
8Physikalisch-Technische Bundesanstalt, D-38116 Braunschweig, Germany
9Institut f¨
ur Kernphysik, Universit¨
at Frankfurt, D-60486 Frankfurt, Germany
10 Department of Physics and Astronomy, Uppsala University, SE-751 20 Uppsala, Sweden
11 Japan Synchrotron Radiation Research Institute, Hyogo 679-5198, Japan
E-mail: nagaya@scphys.kyoto-u.ac.jp
Received 27 February 2013, in final form 25 June 2013
Published 13 August 2013
Online at stacks.iop.org/JPhysB/46/164023
Abstract
We carried out time-of-flight mass spectrometry for neon clusters that were exposed to intense
free electron laser pulses with the wavelength of 62 nm, which induce optical transition from
the ground state (2s22p6) to an excited state (2s22p5nl ) in the Ne atoms. In contrast to Ne+
ions produced by two-photon absorption from isolated Ne atoms, the Ne+ion yield from Ne
clusters shows a linear dependence on the laser intensity (I). We discuss the ionization
mechanisms which give the linear behaviour with respect to Iand expected features in the
electron emission spectrum.
(Some figures may appear in colour only in the online journal)
1. Introduction
‘More is different’ is a famous phrase given by Anderson
[1], who proposed that, as the number of constituent atoms
increases, new phenomena could emerge. In this work, we
investigate ionization processes of clusters irradiated by free
electron laser (FEL) pulses. Here, ‘More is different’ has two
aspects: we can control not only the number of constituent
atoms but also the number of photons over a wide range.
Multi-photon absorption by a single atom, which leads to
the production of a highly charged ion, is a well-known
example for varying the number of photons, while the shift
of the ionization potential (IP) with cluster size is an example
for varying the number of atoms e.g. in the case of the
single-photon absorption. A particularly intriguing situation
is encountered when a cluster is exposed to photons whose
energy is lower than the IP. In this case, if any ionization
event (i.e. under-threshold ionization) takes place, it must be
an effect of ‘More is different’, either due to the number of
photons or to the number of atoms. Furthermore, if the photon
energy is tuned to the Rydberg states or exciton levels, various
auto-ionization processes such as an interatomic Coulombic
decay (ICD) [2,3] and other phenomena like the exciton–Mott
transition (EMT) [4] are expected to occur.
The ICD in general is induced by a two-centre energy
transfer, and conventionally it is triggered by ionizing an
0953-4075/13/164023+05$33.00 1© 2013 IOP Publishing Ltd Printed in the UK & the USA
J. Phys. B: At. Mol. Opt. Phys. 46 (2013) 164023 K Nagaya et al
inner-valence electron. The ICD was proposed by Cederbaum
and has been verified experimentally in rare-gas clusters
[5–8] and molecular clusters [9,10]. Recently, Kuleff et al
[11] proposed a novel ICD mechanism, in which two electrons
are photo-excited from outer valence orbitals to the Rydberg
states and one of them is then emitted by using the relaxation
energy from the other. EMT is defined as an insulator-to-
metal transition, in which the electron–hole correlation plays
a crucial role: an exciton gas that was created by an optical
means can be transformed to an electron–hole plasma when
the exciton density is high enough to induce screening effects
by the overlap of the wavefunctions. In the rare-gas cluster,
the evolution of the Rydberg excited states to the excitons has
been investigated as a function of cluster size [12,13], but the
EMT has not been reported.
In this study, we adopted neon clusters with an average
size Nof 1000 atoms and exposed them to extreme ultraviolet
free electron laser (EUV-FEL) pulses with a wavelength of
62 nm, which corresponds to the optical transition from the
ground state (2s22p6) to an excited state (2s22p5nl )inthe
Ne atom [14]. We found that, whereas the Ne+ion yield from
the uncondensed Ne gas shows a quadratic dependence on
the laser intensity (I), indicating two-photon absorption, the
Ne+ion yield from Ne clusters shows a linear dependence on
the laser intensity. We discuss possible ionization mechanisms
which give rise to such a linear behaviour and predict expected
features in the electron emission spectrum.
2. Experiment
The experiments were performed at the SPring-8 Compact Self
Amplified Spontaneous Emission (SASE) Source (SCSS) test
accelerator in Japan [15]. Our experimental setup was almost
the same as the one reported in [16,17]. Briefly, the cluster
beam crossed the FEL beam at 45◦in the horizontal plane. The
photon energy was tuned to 20 eV (62 nm). The FEL beam was
partially blocked by a 1.5 mm wide horizontal beam stopper
before the ionization region, so that the unfocused beam did not
irradiate the cluster beam directly. The FEL beam was focused
back onto the cluster beam by a multi-layer focusing mirror
fabricated at the Lawrence Berkeley National Laboratory,
whichwasthesameasusedin[16]. Taking all the optical
elements (deflecting and focusing mirrors, etc) between the
radiation source point and the ionization volume into account,
we estimated the power density in the focus spot to be at
most ∼3×1014 Wcm
−2at full power of the FEL, assuming a
diffraction limited focus size of 3 μm in diameter and a pulse
length of 30 fs [18]. The measured spectral fluctuation of the
FEL was 0.3 eV (FWHM) in this experiment.
The cluster beam was prepared by the adiabatic expansion
of a Ne gas through a pulsed 250 μm nozzle. The stagnation
pressure was 4.6 bar and the nozzle temperature was 80 K
[19]. The average cluster size Nwas estimated to be 1000
atoms according to scaling laws [20,21]. To avoid space–
charge effects due to the ionization of a background gas, the
pulsed gas jet was cut to 0.6 mm width and 0.4 mm height
with knife-edge slits and travelled to the focus spot located at
1.7 m downstream from the nozzle.
Time−of−flight [ s]
Intensity [counts/shot]
<GMD>
(a) 1.1
(b) 2.5
(c) 3.2
(d) 3.6
20Ne
22Ne
GMD peak height [a.u.]
counts
1.1 2.5 3.2
3.6
µ
mass [amu]
5.6 5.8 6 6.2
0
0.1
0.2
20 22
024
0
500
Figure 1. Ion TOF spectra of a Ne1000 cluster beam irradiated by
62 nm FEL pulses. The numbers in the figure denote the peak height
of the gas monitor detector (GMD). The inset shows the power
distribution in the GMD for the full and for one-third of the laser
power. Colour areas correspond to the laser power distributions for
each TOF spectrum.
We measured time-of-flight (TOF) spectra with our
momentum imaging spectrometer [16,17]. Fragment ions
were vertically extracted by a uniform electrostatic field. They
travelled through an extraction region (75 Vcm−1electric field
strength, 40 mm in length), an acceleration region (110 V cm−1
electric field strength, 52 mm in length) and a field-free region
(308 mm), and were finally detected by a microchannel-plate
(MCP) detector equipped with a three-layer delay-line anode
(RoentDek HEX120) [22].
The measurements were carried out with the full laser
power and with one-third of it. In the latter case, the intensity
was attenuated by transmitting the laser pulses through a gas
chamber filled with Ar. Since the pulses from the SASE
source fluctuate shot to shot, we measured the intensity
of each pulse by a gas monitor detector (GMD). At the
same time, the TOF spectrum was also measured for each
laser shot. Both signals were recorded by an eight-channel
digitizer (Acqiris DC282×2), and the timing signals were
extracted by a software constant fraction discriminator [23].
This procedure enabled us to deduce precise laser-power
dependences of the TOF spectra. In this study, we discuss
the FEL intensity dependence based on the GMD pulse height
which is proportional to the photon number in each FEL pulse
because of the uncertainty in estimating the focal size of FEL.
3. Results
In figure 1, spectrum (a) displays the TOF spectrum
accumulated during the experiment with one-third of the full
laser power. The corresponding power distribution is shown
as a black area in the inset of figure 1as a function of the
GMD peak height. For the one-third power, the weighted
average of the GMD peak height is 1.1. Since the power
2
J. Phys. B: At. Mol. Opt. Phys. 46 (2013) 164023 K Nagaya et al
0
0.1
0
0.05
0
0.02
0.04
0.06
20 22 24
0
0.02
0.04
<GMD>
3.2
1.1
exp
fit (total)
fit (component)
mass [amu]
2.5
3.6
<GMD>
<GMD>
<GMD>
Intensity [counts/shot]
(d)
(c)
(b)
(a)
Figure 2. Decomposition of the TOF spectrum into four Lorentzian
curves: 20Ne+from the uncondensed gas and from the cluster, and
22Ne+from the uncondensed gas and from the cluster.
distribution is much wider for the full power experiment, we
have divided the data into three parts: (b) below 3.0, (c) from
3.0 to 3.4 and (d) above 3.4 (green, blue and red areas). The
average peak height is 2.5 for (b), 3.2 for (c) and 3.6 for (d).
In each corresponding TOF spectrum, a prominent peak is
observed at the mass-to-charge ratio m/qof 20, and a smaller
one at m/q=22, corresponding to 20Ne+and its isotope
22Ne+, respectively. In addition, broad distributions centred at
m/q=20 and m/q=22 are also seen. The broad features
are due to the fragments from clusters, whereas the sharp
peaks correspond to ions from the uncondensed gas, because
large fragment energies resulting in a broad TOF distribution
can only be produced by Coulomb repulsion between ions
from the cluster. No multiply charged ions such as Ne2+were
observed.
We now examine the laser intensity dependence of the
ion yield to clarify the multiphoton ionization mechanism. For
this purpose, we decompose the TOF spectra into a sum of
two narrow and two wide Lorentzian curves, as displayed in
figure 2. The height and width parameters were determined by
a least-squares fitting. In the fitting procedure, the m/qregion
below 20 was eliminated, because the MCP signal in this region
was contaminated by H2O. We analysed the apparent ratio of
20Ne+to 22 Ne+for uncondensed atoms as a function of FEL
intensity and found a progressive deviation from the natural
abundance (about 9:1) with an increase of the laser power.
Thus, we use the data for 22Ne+hereafter. The benefit of using
the 22Ne+isotope for a quantitative analysis was demonstrated
in a previous paper [24].
In figure 3,the22 Ne+yield is plotted versus the laser
power. The directly measured GMD signal is shown on the
abscissa. The integrated intensity of the sharp peak exhibits a
quadratic dependence on the laser intensity, indicating that the
ionization of isolated atoms was due to two-photon absorption
as expected. In contrast, the ion yields from the clusters show a
linear dependence on the laser intensity. To check the influence
of the choice of how the data are divided, results of the other
such choices, in which the full power data are divided into two
parts, are displayed by diamonds in figure 3(a).
4. Discussion
These experimental results suggest that the ionization process
is qualitatively different between free neon atoms and neon
atoms embedded in clusters. Two-photon ionization was
observed for Ne atoms, while an unexpected linear dependence
of the ion yield on the FEL intensity was found for Ne clusters.
Although it is difficult to clarify the ionization process based
only on these results of ion spectrometry alone, we try here to
discuss the possible cluster ionization mechanisms and predict
the expected features of the electron emission spectrum in each
case.
024
0
1
GMD peak height [a.u.]
22Ne yield [a.u.]
atom
cluster
(a)
51
10−1
100
GMD peak height [a.u.]
22Ne yield [a.u.]
atom
cluster
(b)
Figure 3. (a) 22Ne+yield from uncondensed atoms (closed symbols) and from clusters (open symbols) as a function of the GMD peak
height. The intensities of these two components are re-normalized at the highest point of the GMD scale. The horizontal error bars denote
the standard deviation, and the vertical error bars denote the statistical errors, which are defined by the square root of the integrated number
of 22Ne+. The line guides a quadratic dependence on the laser power for atoms and a linear dependence for clusters. (b) Log–log plot of the
22Ne+yields. The lines are guides for the eyes the same as (a).
3
J. Phys. B: At. Mol. Opt. Phys. 46 (2013) 164023 K Nagaya et al
In the infrared spectral region, tunelling ionization
followed by plasma heating is known to give efficient cluster
ionization even if the photon energy is much smaller than the
IP [25,26]. In such a case, the ponderomotive energy should be
several electron volts or more, and the intensity dependence of
the ion yield would be nonlinear, contrary to what we observe
here. Therefore, we can safely ignore the ionization of the
cluster induced by the electric field of the laser.
The first candidate for the ionization mechanism of neon
clusters is the direct two-photon ionization of the constituent
atoms. In this case, the intensity dependence of the ion
yield should be quadratic, which is however inconsistent
with our experimental results. It should be noted that the
linear intensity dependence could be observed even in such
a two-photon process when the saturation of ionization is
reached for each constituent atom [16,27]. Such saturation
effects can however be excluded in this experiment, since
the simultaneously measured atomic ionization shows a
quadratic intensity dependence. Since the cross-section of
the two-photon ionization of the cluster constituent atoms is
expected to be of the same order of magnitude as that of an
isolated atom, the two-photon route could be excluded as an
ionization process of neon clusters based on the FEL intensity
dependence of the ion yield.
Thus, we have to search for alternative ionization
mechanisms in which the number of emitted ions is
proportional to the number of photon-absorbing atoms.
As a candidate of such an ionization process, a novel
ICD mechanism via resonant excited states is proposed by
Kuleff [11]. When many Rydberg excited atoms are formed
within a cluster, energy transport by the exchange of a virtual
photon becomes possible between the excited states of two
neighbouring atoms, and hence one atom can ionize another
atom utilizing this transition energy. In this case, the number of
generated ions is expected to be approximately one-half of the
number of excited atoms. Since the number of excited atoms
would be proportional to the FEL intensity, this type of ICD is
consistent with our experimental results. If this is the case, the
ICD electron should appear as a peak in the electron emission
spectrum.
Let us consider a situation where one single ion is created
in the cluster. The creation of this single ion can be either
due to the direct two-photon ionization of a single atom in the
cluster or due to the ICD following the sequential excitation
of two atoms by two photons. Since the binding potentials of
atoms in the cluster are distorted by the Coulombic potential
of a single neighbouring ion, the sequential excitation of other
atoms leads to the delocalization of excited electrons and to
the production of nano-plasma. This is the same effect as the
ionization barrier suppression found in the inner ionization of
the cluster [28]. In this case, the ion yield will linearly depend
on the FEL intensity too, since the number of ions is expected
to be proportional to the number of excited atoms. The electron
emission spectrum should be dominated by thermal electrons
from nano-plasma with an additional peak of ‘the first ionizing
atom’ at about 18 eV ( =2hν−IP), independent of whether the
first electron emission is due to direct two-photon ionization
or double excitation ICD.
Now, if the multiple excitation takes place before the first
electron ejection as discussed above, the excited electrons will
be delocalized also because of the lowering of the inner IPs
[28]. As briefly noted in the introduction, the delocalization
of the excited electrons can be viewed as the EMT, which
has been intensively studied for bulk silicon [29]. When a
sufficient number of excitons are created in bulk silicon, it
is well known that the bound electrons are spontaneously
delocalized and transformed to a plasma state by the overlap
of the exciton wavefunctions. Also in this case of a neon
cluster, if the FEL irradiation achieves a high enough exciton
density, electrons will be liberated by the thermoelectronic
emission from the nano-plasma produced by EMT. Then, as in
the previous case, the electron emission spectrum is expected
to show an exponential component corresponding to thermal
electrons. The only difference from the previous case would be
the lack of a sharp peak corresponding to the direct two-photon
ionization of a single atom or a double excitation ICD.
5. Summary
We carried out time-of-flight mass spectrometry on neon
clusters that were exposed to intense free electron laser
(FEL) pulses with the wavelength of 62 nm, which induce
transitions from the ground state (2s22p6) to an excited state
(2s22p5(2P1/2,3/2) 3d) in a Ne atom. In contrast to isolated
Ne atoms, in which a Ne+ion is produced by two-photon
absorption, the Ne+ion yield from Ne clusters shows a linear
dependence on the laser intensity. We have considered possible
ionization mechanisms consistent with the experimentally
observed FEL intensity dependence of the ion yields and
discussed the expected electron emission spectrum for each
case. We believe that the additional information obtainable
from electron emission spectra will be helpful to decide the
underlying mechanism.
Acknowledgments
We thank Dr A I Kuleff, Dr Ph V Demekhin, Professor L S
Cederbaum and Professor U Saalmann for fruitful discussions.
We are grateful to the SCSS Test Accelerator Operation
Group at RIKEN for continuous support in the course of
the studies. We are also grateful to A Belkacem and the
optics group at the LBNL for fabricating the home-made
focusing mirror. This work was supported by the X-ray Free
Electron Laser Utilization Research Project of the Ministry of
Education, Culture, Sports, Science and Technology of Japan
(MEXT), by the grant-in-aid for the Global COE Program
‘The Next Generation of Physics, Spun from Universality and
Emergence’ from the MEXT, by the grants-in-aid (20310055,
21244062) from the Japan Society for the promotion of
Science (JSPS), by the IMRAM project and by the MPG
Advanced Study Group within CFEL. RD acknowledges
support by the DFG FOR 1789. RF thanks the Swedish
Research Council (VR) for financial support.
4
J. Phys. B: At. Mol. Opt. Phys. 46 (2013) 164023 K Nagaya et al
References
[1] Anderson P W 1972 Science 177 393
[2] For a recent review, see Averbukh V et al 2011 J. Electron
Spectrosc. Relat. Phenom. 183 36
[3] Cederbaum L S, Zobeley J and Tarantelli F 1997 Phys. Rev.
Lett. 79 4778
[4] Mott N F 1968 Rev. Mod. Phys. 40 677
[5] Marburger S, Kugeler O, Hergenhahn U and M¨
oller T 2003
Phys. Rev. Lett. 90 203401
[6] ¨
Ohrwall G et al 2004 Phys. Rev. Lett. 93 173401
[7] Jahnke T et al 2004 Phys. Rev. Lett. 93 163401
[8] Ouchi T et al 2011 Phys. Rev. A83 053415
[9] Jahnke T et al 2010 Nature Phys. 6139
[10] Mucke M et al 2010 Nature Phys. 6143
[11] Kuleff A I, Gokhberg K, Kopelke S and Cederbaum L S 2010
Phys. Rev. Lett. 105 043004
[12] Stapelfeldt J, W¨
ormer J and M¨
oller T 1989 Phys. Rev. Lett.
62 98
[13] W¨
ormer J, Karnbach R, Joppien M and M¨
oller T 1996
J. Chem. Phys. 104 8269
[14] Chan W F, Cooper G, Guo X and Brion C E 1992 Phys. Rev. A
45 1420
[15] Shintake T et al 2008 Nature Photon. 2555
[16] Motomura K et al 2009 J. Phys. B: At. Mol. Opt. Phys.
42 221003
[17] Fukuzawa H et al 2009 J. Phys. B: At. Mol. Opt. Phys.
42 181001
[18] Moshammer R et al 2011 Opt. Express 19 21698
[19] Nagaya K et al 2010 J. Electron Spectrosc. Relat. Phenom.
181 125
[20] Hagena O F and Obert W 1972 J. Chem. Phys. 56 1793
[21] Karnbach R, Joppien M, Stapelfeldt J, W¨
ormer J and
M¨
oller T 1993 Rev. Sci. Instrum. 64 2838
[22] Jagutzki O et al 2002 IEEE Trans. Nucl. Sci. 49 2477
[23] Motomura K et al 2009 Nucl. Instrum. Methods
606 770
[24] Sugishima A et al 2012 Phys. Rev. A86 033203
[25] Ditmire T, Donnelly T, Rubenchik A M, Falcone R W
and Perry M D 1996 Phys. Rev. A53 3379
[26] Krainov V P and Smirnov M B 2002 Phys. Rep. 370 237
[27] Sorokin A A, Wellh¨
ofer M, Bobashev S V, Tiedtke K
and Richter M 2007 Phys. Rev. A75 051402
[28] Saalmann U and Rost J M 2003 Phys. Lett. 91 223401
[29] Shah J, Combescot M and Dayem A H 1977 Phys. Rev. Lett.
38 1497
5
A preview of this full-text is provided by IOP Publishing.
Content available from Journal of Physics B Atomic Molecular and Optical Physics
This content is subject to copyright. Terms and conditions apply.
