Nanoplasma Dynamics of Single Large Xenon Clusters Irradiated with Superintense X-Ray
Pulses from the Linac Coherent Light Source Free-Electron Laser
T. Gorkhover,
1
M. Adolph,
1
D. Rupp,
1
S. Schorb,
1,2
S. W. Epp,
3,4
B. Erk,
3,4
L. Foucar,
3,5
R. Hartmann,
6
N. Kimmel,
7,8
K.-U. Ku
¨
hnel,
4
D. Rolles,
3,5
B. Rudek,
3,4
A. Rudenko,
3,4
R. Andritschke,
7,8
A. Aquila,
9,10
J. D. Bozek,
2
N. Coppola,
9,10
T. Erke,
9
F. Filsinger,
11
H. Gorke,
12
H. Graafsma,
9
L. Gumprecht,
9
G. Hauser,
7,8
S. Herrmann,
7,8
H. Hirsemann,
9
A. Ho
¨
mke,
3,4
P. Holl,
6
C. Kaiser,
4
F. Krasniqi,
3,5
J.-H. Meyer,
9
M. Matysek,
13
M. Messerschmidt,
2
D. Miessner,
7,8
B. Nilsson,
9
D. Pietschner,
7,8
G. Potdevin,
9
C. Reich,
6
G. Schaller,
7,8
C. Schmidt,
3,4
F. Schopper,
7,8
C. D. Schro
¨
ter,
4
J. Schulz,
9,10
H. Soltau,
6
G. Weidenspointner,
7,8
I. Schlichting,
5,3
L. Stru
¨
der,
3,7,8,14
J. Ullrich,
3,4,15
T. Mo
¨
ller,
1
and C. Bostedt
2,
*
1
Institut fu
¨
r Optik und Atomare Physik, Technische Universita
¨
t Berlin, Hardenbergstr. 36, 10623 Berlin, Germany
2
Linac Coherent Light Source, SLAC National Accelerator Laboratory, Stanford, California 94309, USA
3
Max Planck Advanced Study Group, Center for Free-Electron Laser Science, Notkestrasse 85, 22607 Hamburg, Germany
4
Max-Planck-Institut fu
¨
r Kernphysik, Saupfercheckweg 1, 69117 Heidelberg, Germany
5
Max-Planck-Institut fu
¨
r medizinische Forschung, Jahnstr. 29, 69120 Heidelberg, Germany
6
PNSensor GmbH, Otto-Hahn-Ring 6, 81739 Mu
¨
nchen, Germany
7
Max-Planck-Institut Halbleiterlabor, Otto-Hahn-Ring 6, 81739 Mu
¨
nchen, Germany
8
Max-Planck-Institut fu
¨
r extraterrestrische Physik, Giessenbachstrasse, 85741 Garching, Germany
9
Photon Science, DESY, Notkestrasse 85, 22607 Hamburg, Germany
10
European XFEL GmbH, Albert-Einstein-Ring 19, 22761 Hamburg, Germany
11
Fritz-Haber-Institut der Max-Planck Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
12
Forschungszentrum Ju
¨
lich, Institut ZEL, 52425 Ju
¨
lich, Germany
13
Inst. f. Experimentalphysik, Universita
¨
t Hamburg, 22607 Hamburg, Germany
14
Universita
¨
t Siegen, Emmy-Noether Campus, Walter Flex Str. 3, 57068 Siegen, Germany
15
Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, 38116 Braunschweig, Germany
(Received 13 February 2012; published 15 June 2012)
The plasma dynamics of single mesoscopic Xe particles irradiated with intense femtosecond x-ray
pulses exceeding 10
16
W=cm
2
from the Linac Coherent Light Source free-electron laser are investigated.
Simultaneous recording of diffraction patterns and ion spectra allows eliminating the influence of the laser
focal volume intensity and particle size distribution. The data show that for clusters illuminated with
intense x-ray pulses, highly charged ionization fragments in a narrow distribution are created and that the
nanoplasma recombination is efficiently suppressed.
DOI: 10.1103/PhysRevLett.108.245005 PACS numbers: 52.50.Jm, 32.80.Aa, 36.40.Gk, 52.25.Tx
Free-electron lasers (FELs) such as the Linac Coherent
Light Source (LCLS) deliver extremely intense and coher-
ent x-ray flashes with a femtosecond pulse length, opening
the door for imaging single nanometer-sized objects with
atomic resolution in single shots [1,2]. All matter irradiated
by the intense x-ray pulses will be highly ionized within
femtoseconds. So far, investigations about the ionization
dynamics under such conditions have focused on ensem-
bles of atoms and small systems on the one hand [3–6] and
macroscopic solids on the other hand [7]. A detailed under-
standing of the x-ray pulse-matter interaction for inter-
mediate, i.e., mesoscopic, systems is of utmost relevance
for x-ray imaging applications. Most samples to be imaged
with intense x-ray laser pulses will be in the size regime of
a few tens to hundreds of nanometers and first results on
their pulse-length dependent scattering response are con-
troversial [8,9].
For studying fundamental questions of the light-matter
interaction, atomic rare gas clusters have proven to be ideal
targets [10,11]. They are intermediate between atoms and
bulk solids, and energy dissipation into surrounding media
is virtually absent due to their finite size. First results from
short-wavelength FELs have shown that even in the
vacuum-ultraviolet spectral regime, the clusters can still
be very efficiently heated with inverse bremsstrahlung
(IBS) [12]. For shorter wavelength, IBS becomes negli-
gible and the ionization starts being dominated by multi-
step photoionization [13,14]. In the intense x-ray focus of
the LCLS with fluences exceeding 10
5
photons=
A
2
atoms
are sequentially ionized from the inside out starting with
the inner-shell electrons followed by subsequent inner-
shell vacancy decay [3]. In clusters, after a certain number
of ionization events, the particle is charged to a degree that
its Coulomb energy becomes higher than the kinetic energy
of the sequentially ejected electrons and a nanoplasma is
formed. In dense nanoplasma, the cluster ionization can be
further enhanced by energy-exchanging electron collisions
[15]. For photon energies of 100 eV, large clusters are
PRL 108, 245005 (2012)
PHYSICAL REVIEW LETTERS
week ending
15 JUNE 2012
0031-9007=12=108(24)=245005(5) 245005-1 Ó 2012 American Physical Society
transiently highly charged [16] before they disintegrate in a
hydrodynamic expansion accompanied by recombination
of the cluster core [17–19].
In this Letter we present pioneering results on the ion-
ization and recombination dynamics of single nanometer-
sized samples in intense x-ray pulses from LCLS. We have
used a novel coincident imaging and spectroscopy ap-
proach for the investigation of individual clusters in single
shots. Analysis of the scattering patterns allows down-
selection of the data to events with only a single particle
with a defined size in the x-ray focus, therefore providing a
unique insight into the intense x-ray pulse-matter interac-
tion. Being able to determine the exposure intensity and
size for every single particle, we find drastically different
plasma dynamics for different power densities which are
not observable in experiments averaging over the focal
volume and cluster size distribution [5]. At the highest
intensities in the center of the focus, fragments from a
highly charged nanoplasma are detected with an unex-
pected absence of low charge states indicating that plasma
recombination during the cluster expansion is efficiently
suppressed. The results show that for the high peak-power
and photon energies available at LCLS, the ionization and
disintegration dynamics of nanoscale samples differ sub-
stantially from those found in earlier experiments in the
soft x-ray regime.
The experiments were performed in the CAMP end
station [20] at the AMO beam line of LCLS [21]. The
FEL pulses with a photon energy of 800 eV and average
pulse energies of 1.5 mJ were focused to a spot size of
about 6 micron (FWHM) diameter. The electron bunch
length was set to 200 fs and the x-ray pulse length has
been measured to be about 2=3 of that value [22], yielding
an estimated x-ray pulse length of 130 fs for the current
experiment. With a beam line transmission on the order of
20%, the resulting power densities exceed 10
16
W=cm
2
.
Clusters with average radii of 30 nm were formed by
supersonic expansion of xenon at 9 bar backing pressure
through a 200-micron conical nozzle with a half-opening
angle of 4
cooled to 245 K. The pulsed cluster jet was
skimmed twice and a piezo-driven slit skimmer right be-
fore the interaction region was adjusted such that, on
average, less than one particle was in the FEL focus as
evidenced by the single-shot scattering patterns. The clus-
ter beam was crossed with the focussed x-ray beam above
the aperture of an ion time-of-flight spectrometer.
Conically shaped spectrometer electrodes allowed for a
free line of sight from the interaction region to the photon
detector. The scattered photons were detected by single
photon counting pnCCDs covering scattering angles from
0.5
–5
[20]. All data were stored on a shot-to-shot basis
and correlated to the FEL parameters.
The experimental setup as well as three representative
single-cluster scattering patterns are shown in Fig. 1. The
images are used to fit the cluster sizes and filter the data
such that only hits from single clusters with similar radii of
FIG. 1 (color online). Experimental setup (top panel) and
single-shot scattering images from individual clusters (bottom
panel). The cluster beam is crossed with the focused x-ray beam
above the aperture of a time-of-flight spectrometer. The scattered
photons are detected by pnCCDs covering scattering angles from
0.5
to 5
. The scattering signal exhibits a strong intensity
variation due to the different positions (a)–(c) of the clusters
in the focus as depicted in the inset.
Intensity [arb.units]
50004000300020001000
123452654
Xenon charge state q
Averaged data
c
Single power density ion spectra
a
b
10
16
W/cm
2
< 10
14
W/cm
2
10
15
W/cm
2
5500500045004000350030002500200015001000500
time-of-flight [ns]
Focal volume integrated ion spectra
10
16
W/cm
2
atom
10
16
W/cm
2
10
15
W/cm
2
Zoom x 10
0
0.5
1
FIG. 2 (color online). Top panel: Single shot ion spectra re-
corded in coincidence with the scattering images (a)–(c) shown
in Fig. 1. Bottom panel: Focal volume integrated ion yield
spectra for cluster (top) and atomic (bottom) targets. The
single-shot spectra strongly correlate with the position of
the cluster in the FEL focus and are strikingly different from
the integrated data.
PRL 108, 245005 (2012)
PHYSICAL REVIEW LETTERS
week ending
15 JUNE 2012
245005-2
30 2nmare used for further analysis, effectively elim-
inating the influence of the particle size distribution com-
mon to experiments with supersonic expansion sources.
The total number of scattered photons for individual im-
ages is proportional to the FEL power density experienced
by the respective single clusters. The scattered photon
intensity varies strongly from shot to shot [cf. bottom
panels (a)–(c) in Fig. 1] even for clusters of similar size
due to the fluctuations of their position in the FEL focus
waist. The integrated scattering signal, i.e., the impinging
photon intensity experienced by the cluster declines by
about one order of magnitude between cases (a), (b), and
(c), respectively.
The ion yield spectra from single clusters recorded in
coincidence with the scattering patterns from Fig. 1 are
shown in the top panel of Fig. 2. Additionally, in the bottom
panel of Fig. 2, focal-volume integrated spectra are pre-
sented for two different power densities as well as an atom
reference spectrum. The three single-cluster spectra reveal
strong power-density dependencies. For the best-hit
case (a) the ion yield is centered around a charge state of
q ¼ 26
þ
and q<4
þ
are completely absent. At the same
time, 26
þ
is also the high-charge state cutoff in the atomic
reference recorded at a similar peak power density (bottom
panel Fig. 2) which will be discussed further below. The
broadening of the peak in the cluster spectrum is attributed
to the kinetic energy release of the cluster fragments, but
the existence of charge states beyond the atomic ionization
limit cannot be excluded from the current measurement.
For the cluster illuminated with about a factor of ten less
power density [case (b)], the charge distribution shifts to
significantly lower q. Proceeding to a factor 100 less power
density [case (c)], only singly-charged fragments are left,
very similar to experiments in the single x-ray photon limit
[23]. The single-cluster data are in striking contrast to the
focal volume integrated results in the lower panel of Fig. 2.
Here, the spectra are dominated by singly-charged frag-
ments and low charge states in agreement with an earlier
study [5]. Decreasing the power density by an order of
magnitude, similar to the conditions of the single-particle
experiments described above, leads to a significant sup-
pression of the high charge states while the intensity of low
q remains rather similar.
The coincident imaging and spectroscopy data from
single clusters demonstrate the existence of strong
power-density-dependent dynamical processes that are
washed out in the integrated data. This becomes particu-
larly obvious for the highest power densities, where low
charge states are entirely absent, a feature that has never
been observed in nonlinear light-matter interaction experi-
ments in other spectral regimes [10,11,24]. To shed more
light on the transition from the low to the high power
density spectra, the dominant charge state as well as the
average charge state from single clusters with similar radii
(30 2nm) are compared in Fig. 3. The average charge
state (represented by open circles Fig. 3) is determined as
the center of gravity of the ion yield spectra. It rises
monotonically from q 5
þ
to q 15
þ
within the inves-
tigated power-density window, indicating an increase in
overall energy absorption by the cluster with increasing
power densities. The consequences are a higher mean
charge state and higher kinetic energy releases, which
both contribute to the observed slope. The dominant charge
state (represented by black diamonds in Fig. 3) is defined
as the most abundant one in the ion yield spectra. It rises
rapidly and saturates just above 26
þ
. It should be noted that
the charge states are deduced from the spectra without
considering initial kinetic energy releases which push the
ion peaks to shorter flight times. The observed saturation
points toward similar photoionization processes for xenon
atoms in the gas phase and in the cluster. Q ¼ 26
þ
is also
the highest observed charge state in the atomic xenon
spectra (cf. Fig. 2). The ground state ionization potential
for Xe
25þ
is with 890 eV [25] slightly higher than the
incident photon energy of 800 eV, indicating that sequen-
tial multiphoton ionization via resonant excitation and
intermediate excited states can occur [26], similar to ob-
servations in argon atoms and clusters [6].
The absence of low charge states and the rapid increase
of the dominant q with increasing power densities points to
a very efficient suppression of recombination processes
in the x-ray induced nanoplasma. It is noted that this
observation is quite surprising, as for the most highly
charged clusters less than 1% of the photoactivated elec-
trons can leave before the photoemission is frustrated in the
30
25
20
15
10
5
Xe charge state
10
8642
FEL intensity [arb.units]
1,2
1,3
1,4
1,5
1,6
1,7
1,8
1,9
2,2
3,2
time of flight [
µ
dominant charge
average charge
a
b
FIG. 3 (color online). Average (open circles) and the dominant
charge state (filled diamonds) for increasing FEL intensities. The
FEL intensities are deduced from the integrated scattering in-
tensity of the similarly sized clusters and plotted on a relative
scale. The dashed lines are guides to the eye and demonstrate the
different behavior of a constantly growing average charge state
versus the rapid saturation of the dominant charge state. The ion
charge states are estimated without considering the initial ion
kinetic energy release.
PRL 108, 245005 (2012)
PHYSICAL REVIEW LETTERS
week ending
15 JUNE 2012
245005-3
increasing cluster Coulomb potential [13]. Our findings are
also vastly different as compared to results of earlier
experiments with about one order of magnitude lower
photon energies where very efficient nanoplasma recom-
bination mechanisms have been found [17,19]. The initial
photoionization processes are similar for both cases, and
they are guided by the atomic photoabsorption cross sec-
tions and subsequent inner-shell vacancy decay. After frus-
tration of the photoemission, a nanoplasma is formed in
which the electrons thermalize within femtoseconds
through energy-exchanging collisions [7,15]. The present
results show that the high x-ray photon energies, and thus
high electron excess energies and mean charge states, must
play an important role for the nanoplasma evolution.
For a further discussion of the nanoplasma dynamics, we
have implemented a simplified model of the cluster expan-
sion in the time-dependent plasma code
FLYCHK [27].
FLYCHK is a population kinetics model that includes all
the relevant ionization and recombination processes for
arbitrary atomic elements. The disintegration of large clus-
ters can be approximated by a hydrodynamic expansion in
which the hot quasifree electrons force the cold ions to
expand.
FLYCHK is then used to calculate the ionization and
population distribution in the expanding nanoplasma up to
several nanoseconds after the x-ray pulse. As the micro-
scopic ionization dynamics cannot be described with
FLYCHK, two different initial states in qualitative agreement
with the experimental data shown in Fig. 2 are assumed. In
the first scenario representing the high x-ray power density
case, a quasiequilibrium plasma with a mean charge state
around the highest observed q ¼ 26
þ
and solid Xe density
is taken as the starting point. To represent the conditions
from the regions with about an order of magnitude less
power density, a colder nanoplasma with a mean charge
state of 8
þ
is assumed. The results of the FLYCHK calcu-
lations and the corresponding experimental results from
Figs. 2(a) and 2(b) are displayed in Fig. 4. In comparison
with the experiment,
FLYCHK reproduces qualitatively the
measured ion distribution of case (a) including the sup-
pressed recombination to lower charge states but predicts
too much recombination for case (b). The simulations
show that for the hot, fast expanding plasma in case (a)
the high initial electron temperature lowers three-body
recombination rates during the time scale where the plasma
is still dense and that the high charge states freeze out
whereas in (b) the initial lower electron temperatures lead
to a slower expansion and therefore a longer period of
efficient recombination. It should be noted that our model
assumes homogeneous plasma densities thus neglecting
any shell effects [17] and starts at quasiequilibrium con-
ditions. Charge redistribution within the nanoplasma and
shell explosion can lead to a significant broadening of the
ion distribution and higher observed charge states [18].
Further,
FLYCHK does not take into account three-body
recombination into higher Rydberg states above n ¼ 10,
which have been shown to become significant for clusters
explosions induced with intense infrared pulses after sev-
eral ps of expansion [28].
In summary, thanks to the single-shot coincidence
method pioneered in this work, the x-ray induced ioniza-
tion dynamics of Xe clusters could be studied under un-
precedented well-controlled experimental conditions, i.e.,
at well-defined power density and for well-characterized
particle size. This way, the strong limitation of integrating
over very different power densities and cluster sizes in the
focal volume common to almost all studies of nonlinear
light-matter interaction could be overcome. At the highest
power density of (5 10
16
W=cm
2
), predominantly ions
with charge states peaking at 26
þ
are observed while low
charge states are virtually absent. The results give strong
evidence that the narrow charge distribution is due to
inefficient electron-ion recombination. The inefficient re-
combination is attributed to the high initial temperature of
the x-ray produced nanoplasma.
This research was carried out at the Linac Coherent
Light Source (LCLS) at the SLAC National Accelerator
Laboratory. LCLS is an Office of Science User Facility
operated for the U.S. Department of Energy Office of
Science by Stanford University. The authors want to thank
Y. Ralchenko and R. W. Lee for their support with the
FLYCHK simulations as well as T. Fennel, J.-M. Rost, and
U. Saalmann for helpful discussions. We acknowledge the
Max Planck Society for funding the development and
operation of the CAMP instrument within the ASG at
Intensity [arb.units]
3025201510
50
Xe charge state
initial q=8+
FLYCHK
Simulation
initial q=26+
Intensity [arb.unit]
30252015105
Xe charge
Single shot ion spectra
a
b
FIG. 4 (color online). Comparison between FLYCHK simula-
tions and experimental single-shot ion data (inset). Both the data
and simulation show a strong suppression of recombination for
the highly excited x-ray-induced nanoplasma (initial q ¼ 26
þ
)
in the center of the FEL focus and a significant recombination for
lower excitation (initial q ¼ 8
þ
). For the initial q ¼ 26
þ
nano-
plasma, the high charge states freeze out due to rapid expansion
of the cluster and low three-body recombination rates for hot
electron temperatures.
PRL 108, 245005 (2012)
PHYSICAL REVIEW LETTERS
week ending
15 JUNE 2012
245005-4
CFEL and funding from BMBF 05KS4KT1, 05KS7KT2
and 05K10KT2, DFG BO 3169/2-2, as well as HGF
Virtuelles Institut VH-VI-302.
*Corresponding author.
bostedt@slac.stanford.edu
[1] H. N. Chapman et al., Nature (London) 470, 73 (2011).
[2] M. M. Seibert et al., Nature (London) 470, 78 (2011).
[3] L. Young et al., Nature (London) 466, 56 (2010).
[4] M. Hoener et al., Phys. Rev. Lett. 104, 253002
(2010).
[5] H. Thomas et al., Phys. Rev. Lett. 108, 133401 (2012).
[6] S. Schorb, D. Rupp, M. L. Swiggers, R. N. Coffee, M.
Messerschmidt, G. Williams, J. D. Bozek, S.-I. Wada, O.
Kornilov, T. Mo
¨
ller, and C. Bostedt, Phys. Rev. Lett. 108,
233401 (2012).
[7] S. M. Vinko et al., Nature (London) 482, 59 (2012).
[8] A. Barty et al., Nature Photon. 6, 35 (2012).
[9] L. Lomb et al., Phys. Rev. B 84, 214111 (2011).
[10] T. Ditmire, T. Donnelly, A. M. Rubenchik, R. W. Falcone,
and M. D. Perry, Phys. Rev. A 53, 3379 (1996).
[11] T. Fennel, K.-H. Meiwes-Broer, J. Tiggesba
¨
umker, P.-G.
Reinhard, P. M. Dinh, and E. Suraud, Rev. Mod. Phys. 82,
1793 (2010).
[12] H. Wabnitz et al., Nature (London) 420, 482 (2002).
[13] C. Bostedt et al., Phys. Rev. Lett. 100, 133401 (2008).
[14] H. Iwayama et al.,
J. Phys. B 42, 134019 (2009).
[15] C. Bostedt, H. Thomas, M. Hoener, T. Mo
¨
ller, U.
Saalmann, I. Georgescu, C. Gnodtke, and J.-M. Rost,
New J. Phys. 12, 083004 (2010).
[16] C. Bostedt et al., Phys. Rev. Lett. 108, 093401 (2012).
[17] M. Hoener, C. Bostedt, H. Thomas, L. Landt, E. Eremina,
H. Wabnitz, T. Laarmann, R. Treusch, A. de Castro, and T.
Mo
¨
ller, J. Phys. B 41, 181001 (2008).
[18] H. Thomas, C. Bostedt, M. Hoener, E. Eremina, H.
Wabnitz, T. Laarmann, E. Plo
¨
njes, R. Treusch, A. R. B.
De Castro, and T. Mo
¨
ller, J. Phys. B 42, 134018 (2009).
[19] B. Ziaja, H. N. Chapman, R. Santra, T. Laarmann, E.
Weckert, C. Bostedt, and T. Mo
¨
ller, Phys. Rev. A 84,
033201 (2011).
[20] L. Stru
¨
der et al., Nucl. Instrum. Methods Phys. Res., Sect.
A 614, 483 (2010).
[21] J. D. Bozek, Eur. Phys. J. Special Topics 169, 129 (2009).
[22] S. Schorb et al., Appl. Phys. Lett. 100, 121107 (2012).
[23] M. Hoener, C. Bostedt, S. Schorb, H. Thomas, L. Foucar,
O. Jagutzki, H. Schmidt-Boeking, R. Do
¨
rner, and T.
Mo
¨
ller, Phys. Rev. A 78, 021201 (2008).
[24] C. Bostedt, M. Adolph, E. Eremina, M. Hoener, D. Rupp,
S. Schorb, H. Thomas, A. R. B. De Castro, and T. Mo
¨
ller,
J. Phys. B 43, 194011 (2010).
[25] T. A. Carlson, C. W. Nestor, Jr., N. Wasserman, and J. D.
Mcdowell, At. Data Nucl. Data Tables 2, 63 (1970).
[26] B. Rudek et al. (to be published).
[27] H.-K. Chung, M. Chen, W. Morgan, Y. Ralchenko, and R.
Lee, High Energy Density Phys.
1, 3 (2005).
[28] T. Fennel, L. Ramunno, and T. Brabec, Phys. Rev. Lett. 99,
233401 (2007).
PRL 108, 245005 (2012)
PHYSICAL REVIEW LETTERS
week ending
15 JUNE 2012
245005-5
