Evolution on K α in Fe and Lβ
satellites in Au in SPring-8
Nobuyuki Shigeoka, Hirofumi Oohashi, and Yoshiaki Ito
Laboratory of Atomic and Molecular Physics,
Institute for Chemical Research, Kyoto University,
Gokasho, Uji, Kyoto 611-0011 Japan
Aurel M. Vlaicu, Atsushi Nisawa, Hideki Yoshikawa,
Sei Fukushima, and Mamoru Watanabe
Advanced Materials Laboratory, in SPring-8,
Mikazuki, Hyogo 679-5198 Japan
(Dated: April 2, 2003)
With the advent of the third generation synchrotron radiation, we can perform experiments on
the threshold behavior of the satellites including the excitation dynamics in atoms. Especially, xray emission spectroscopy is a suitable tool to study the satellites on the electron correlation. The
contributions of direct ionization, i.e, the shake-off process in Fe and indirect ionization of CosterKronig transition in Au have not been investigated to the x-ray emission spectra yet. We present
the contribution of the spectator holes to the processes around the threshold.
PACS numbers: 32.30.Rj, 32.80.Fb, 32.80.Hd
I.
complex, by Sternemann et al . [6] measuring the valence
fluorescence satellites KM - N 2,3M transition of a solid
in Ge target, and more recently, by Raboud et al . [7]
measuring the KL x-ray emission of Ar induced by impact with monoenergetic photons to investigate the K +L
double excitation from threshold to saturation.
The contributions from 2s and 2p spectator transitions
and 3d spectator transitions (hidden satellites) have not
been investigated in Fe. We report the effect on the spectator transitions in Fe.
The measurements were carried out at BL15XU,
Spring-8, Ako, with a curved crystal x-ray spectrometer [8, 9]. A double crystal Si(311) monochromator with
a bandpass of ∼ 3 eV and a flux of > 1012 photons/sec
was used. The sample was a polycrystalline high-purity
Fe foil. The fluorescence spectrometer employed the Johann geometry with a 1.5 m diameter Rowland circle on
Si (400) crystal providing < 1 eV resolution. BL15XU is
a helical type undulator. Therefore it is easy to reduce
the harmonic components with the slit. The coherent
radiation out of the monochromator is onto the sample
in the sample chamber of the spectrometer. The light
then goes into the crystal housing in which three kinds
of crystals are mounted. The optical focusing condition
can be met by moving the sample, crystal and detector
to satisfy the Rowland geometry.
The measurement was carried out changing the excitation energy from 7,850 eV to 10,000 eV [9] in order to investigate the energy-dependency on K α3,4 satellites’ intensity. Spectrometer angle was fixed at which we can get
K α3,4 satellites and then excitation energy were tuned in
the region between 7,850 eV and 10,000 eV and scanned
in order to obtain the spectral profile. An onset of K α3,4
satellite emission for threshold energy was estimated with
the results. The K α1 contribution was removed by subtracting a Lorentzian tail fitted to the spectrum outside
ORIGIN OF THE SHAKE-OFF PROCESS IN
FE K α3,4 SATELLITES
Most of studies on the contributions from the effects
of shake processes in solids and vapors etc to x-ray absorption have mainly been carried out by Italian group,
Slovenian group, and Japanese group, respectively, in order to elucidate the electron-electron correlation in atom,
vapor, liquids, and solids. The sharp multielectron photoexcitation features due to resonant and shake-up, and
the extended shake-off saturation profiles are of special
interest in x-ray absorption spectroscopy.
The x-ray absorption of 3d transition metals was examined in the energy region of K +L double photoabsorption. However, no significant features attributable
to multiple photoexcitation are found in the spectra, so
that K +L edges for 3d elements were not confirmed [1, 2].
The result could be explained by theoretical predictions due to lower shake-up probabilities for K +L transitions [2]: the transition edges observed in x-ray absorption spectra are due to the shake-up process only, i.e. the
probabilities of resonance and shake-up for K +L transition for 3d elements are lower than the accuracy of the
detection and it is difficult to obtain the pure long-range
shake-off profile in XAFS oscillation.
With the advent of the third generation synchrotron
radiation, we can do experiments on the threshold behavior of the satellites including the excitation dynamics
in atoms, molecules, and solids. Especially, x-ray emission spectroscopy is a suitable tool to study the satellites on the electron-electron correlation. First detailed
photoexcitation measurements were performed by Deslattes et al . [3], where both emission and absorption spectroscopy are combined to examine multielectron vacancies on atomic Ar, by Deutsch and co-workers [4, 5] finding a pure shake-off behavior of the Cu K α x-ray satellite
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
20
the energy range of the satellites. The observed intensity
of the satellite with excitation energy is shown in Fig. 1,
and the ratio of the satellite intensity to the diagram intensity, together with Thomas model fitting in Fig. 2.
counts / I0
1.5x10
nying K ionization in Fe [14]. The saturation intensity,
0.70 % of the K α1 line, is in excellent agreement with
the high-energy, x-ray-tube-measured intensity of 0.73 %,
and the sudden-approximation theoretical intensities of
0.56 % [14]. It is interesting that a sum of [1s2s] and
[1s2p] transition probabilities is good consistent with the
observed one. Threshold energy of satellites obtained by
Thomas model is 7950 eV. This value is corresponding
to the ionization energy of Fe [1s]+Co [2p]. The growth
of satellite intensity with the excitation energy indicates
that the 2p spectator holes, which are considered as the
origin of the K α3,4 satellite emissions, are mainly created
by shake-off in Fe.
Deutsch et al . [15] reported that 3d spectator holes
broaden the widths of the corresponding x-ray emission
lines by up to a few tenths of a eV, that is, the presence of
a spectator hole during the emission process introduces
additional splitting in the initial and final energy levels
and consequently increases considerably the number of
the distinct transition lines, and it is possible to separate
out the contribution of these transitions from those of
the diagram ones by fitting the measured line shape by
an ab initio calculated transition array. The width of
the FWHM in Fe K α1 was invetigated with excitation
energy and is shown in Fig. 3. The feature of the data
-3
K + L*3L*2 L*1
1.0
0.5
7.5
8.0
8.5
9.0
excitation energy [eV]
9.5
10.0x10
3
FIG. 1: Intensity at K α3,4 satellite position with excitation
energy [10]
Int Kα3,4 / Int Kα1,2,3,4[%]
0.6
0.4
Present work
Direct beam from Fe x-ray tube 30 kV
[1s2p] shake-off (Mukoyama)
Thomas model fitting
0.2
4.2
K + L2,3*
threshold
K + M4,5*
threshold
0.0
Kα1
Kα2
4.0
8000
8500
9000
9500
10000
FWHM of peak [eV]
excitation energy [eV]
FIG. 2: Ration of intensity of K α3,4 to intensity of K α1,2,3,4 ,
together with Thomas model fitting [11]
The experimental intensity data were obtained only for
high-energy excitation in conventional x-ray tubes presented by lozenge in Fig. 2. The feature of the data
is the smooth increase of the satellite intensity over a
wide energy range above threshold. An abrupt intensity
jump is predicted for shake-up at threshold while shakeoff should increase smoothly from zero [1, 2, 12, 13]. The
shape of the curve in Fig. 2 marks the behavior as a
pure shake-off process similar to that in Cu [4, 5]. Nonrelativistic Hartree-Fock-Slater calculations yield shake
probability of 0.11, 0.56, 0.41, 3.10, and 9.72 % for the
respective 2s,2p, 3s, 3p, and 3d shake electrons accompa-
3.8
3.6
3.4
3.2
3.0
7000
7500
8000
8500
9000
excitation energy [eV]
9500
10000
FIG. 3: FWHM of the K α1 and K α2 with excitation energy
shows the abrupt increase of the width till a few tens eV
above the [1s3d] threshold. This tendency is in contrast
21
with that in [1s2p] transitions. The matured width is
about 0.5 eV broader than that at [1s3d] threshold. This
value is significantly corresponding to that reported by
Deutsch et al . [15]. It is considerable that the width is
broadened by the presence of an additional 3d spectator
hole in the atom by as much as a few tenths of an eV.
Although it is difficult to elucidate the significance and
causes of these differences, the investigation on the influence of additional holes on level widths in atoms is very
important in atomic physics. Further measurements are
planned to elucidate the details of the [KM ] double photoexcitation in the energy region.
II.
with a thickness of 50 µm. The fluorescence spectrometer employed the Johann geometry with a 1.5 m diameter
Rowland circle on Si(444) crystal providing < 1 eV energy resolution. BL15XU adopts an insertion device of a
planer type undulator. This system can remove the harmonic components out with the slit only [19]. The coherent radiation out of the monochromator is onto the sample in the sample chamber of the spectrometer. The light
then goes into the crystal housing in which three kinds
of crystals are mounted. The optical focusing condition
can be met by moving the sample, crystal and detector to satisfy the Rowland geometry. The measurements
were carried out changing the excitation energy around
L1 edge in order to investigate the energy-dependency on
the visible satellites’ intensity. Spectrometer angle was
fixed at which we can get Lβ20 and Lβ200 satellites and
Scintillation Counter (SC) or CCD detectors scanned in
order to obtain the spectral profile, and then excitation
energies were tuned around L1 edge. The observed absorption spectra in L edges were used to determine the
values of the excitation energy.
It is generally considerable that the double-hole state
L3 M 4 which is the initial state of Lβ200 , is caused by two
processes: one is L3 - M 4 shake-off process. This process
is the direct ionization, so that it’s transition probability depends on the excitation energy. The onset of this
process can be estimated to be L3 + M ∗4 (* means Z +
1 element’s binding energies.). Another process is L1 L3 M 4 C-K transition. This is indirect ionization process
and therefore independent of the excitation energy. The
result suggests that the L3 M 4 double-hole state is mainly
due to C-K transition. The evolutions of the Lβ2,3,15
emission spectra around the L1 edge are shown in Fig. 4.
BEHAVIOR AU Lβ2 VISIBLE SATELLITES
AROUND L1 THRESHOLD
It is difficult to analyze L x-ray emission spectra by
excitation with electrons and high energy photons such
as fluorescence x-ray because the three L subshells can
be ionized and the subsequent redistribution of initial vacancies occurs by L-LM Coster-Kronig transition. It is
usually believed that the L x-ray satellite lines usually appear on a slightly higher-energy side than their diagram
line. The satellites corresponding to M spectator holes
lead to lines that can be resolved well from the parent
lines, whereas those corresponding to N spectator holes
almost coincide with the diagram lines [16, 17]. It is well
known that the Coster-Kronig transition reappears heavy
elements of Z > 74. Therefore, it is very interesting
to investigate theoretically and experimentally satellites
caused by such a transition involved the thresholds using
tunable photon energies. The development over recent
years of tunable high-brilliance hard x-ray beams from
dedicated synchrotron source such as the third generation has given impetus to the atomic physics, especially,
the evolution experiments as mentioned above. The precision and power of these sources permit the exploration
of the electron excitation dynamics with unprecedented
detail and resolution. It is generally known that Au Lβ2
diagram line has two satellites, Lβ20 and Lβ200 on its higher
energy side. Their energy shifts from the diagram line are
enough large to confirm their existences in the data. The
Lβ20 and Lβ200 satellites have previously been assigned to
the L3 M 5 - N 5 M 5 and L3 M 4 - N 5 M 4 transitions, respectively. However, the mechanism of the creation of M 4 or
M 5 spectator hole has been not clarified yet. M i spectator hole can be created by either or both L3 M i shake-off
process or/and L1 - L3 M i C-K transition. According to
the report of Chen et al . [18], C-K transition is allowed
for i = 4,5 in the case of 79 Au. In the present study, the
behavior of the Lβ2 visible satellites are investigated by
the evolution of the photo-excited Lβ2 emission spectra
in SPring-8, in order to elucidate the mechanism of the
origin in the satellites. The measurements were carried
out using a curved crystal x-ray spectrometer at BL15X
in Spring-8, Hyogo [8, 9]. A Si double-crystal monochromator was used. The sample was a high-purity Au foil
0.20
Lβ2'
0.15
Lβ2''
0.10
Lβ2
1.4
0.05
0.00
1.2
11.62
11.64
11.66
Energy [eV]
11.68
11.70x10
3
Intensity (arb. unit)
1.0
Lβ3
Lβ15
0.8
0.6
14370 eV
0.4
14362 eV
14354 eV
0.2
14300 eV
14200 eV
0.0
11.50
11.55
11.60
11.65
11.70
14500 eV
14366 eV
14358 eV
14350 eV
14250 eV
14000 eV
11.75x10
3
Energy [eV]
0
FIG. 4: Dependence of Au Lβ2,3,15 diagram lines and, Lβ2
00
and Lβ2 satellite lines on excitation energy [20]
The spectra Lβ2,15 have hitherto been studied only for
high-energy excitations in conventional x-ray tubes [21–
23]. The most outstanding feature of the data is the
abrupt increase of the satellite intensity over a considerable energy range around the threshold as seen in
22
Fig. 4. The observed spectra were analyzed by fitting the
Lβ2,3,15 diagrams and Lβ2 satellites by single Lorentzian
profile. We obtained relative intensities of satellites to
the diagram line. Lβ3 diagram line and Lβ2 satellites appear just below the L1 edge with the excitation energy.
Lβ3 disappears with the excitation energy below the L1
threshold due to the transition to L1 subshell. It is found
that the relative intensities of both satellites to the diagram line Lβ2 increase along with L1 absorption spectra.
The behavior is similar to the abrupt edgelike behavior
observed in single-electron-correlated spectra. Moreover,
a fine structure was first confirmed around the L1 edge.
In near future, we try to investigate the fine structure
with a high-resolution spectrometer. We presented here
a study of the origin of the satellites ascribed to C-K
transition in Au. The contribution of C-K transition to
Au Lβ satellites were clearly confirmed, that is, Au Lβ2
visible satellites are mainly caused by L1 - L3 M i (i = 4,5)
C-K transition. Further investigations will be executed
for in order to elucidate the details of the dependence of
the intensity in the each satellite around the thresholds
in heavy elements.
[1] A. Kodre, M. Hribar, I. Arcon, D. Glavic-Cindro, M.
Stuhec, R. Frahm, and W. Drube, Phys. Rev. A 45,
4682(1992).
[2] Y. Ito, H. Nakamatsu, T. Mukoyama, K. Omote, S.
Yoshikado, M. Takahashi, and S. Emura, Jpn. J. Appl.
Phys. 32, Suppl.32-2, 58 (1993).
[3] R. D. Deslattes, R. E. LaVilla, P. L. Cowan, and A.
Henins, Phys. Rev. A27, 923 (1983).
[4] M. Deutsch, O. Gang, K. Hamalainen, and C. C. Kao,
Phys. Rev. Lett. 76, 2424(1996).
[5] M. Fritsch, C. C. Kao, K. Hamalainen, O. Gang, E.
Forster, and M. Deutsch, Phys. Rev. A57, 1686(1998).
[6] C. Sternemann, A. Kaprolat, M. H. Krisch, and W.
Schulke, Phys. Rev. A61, 020501(R) (2000).
[7] P.-A. Raboud, M. Berset, J.-Cl. Dousse, Y.-P. Maillard,
O. Mauron, J. Hoszowska, M. Polasik, J. Rzadkiewicz,
Phys. Rev. A65, 062503 (2002).
[8] Y. Ito, A. M. Vlaicu, T. Tochio, K. Mutaguchi, H.
OoHashi, N. Shigeoka, T. Shoji, J. Harada, S. Emura,
and T. Uruga, in Proceegings of International Workshop
on Physical Properties of Matter by x-ray absorption and
emission spectroscopy, 24-26 July 2000, Osaka EllectroCommunication University, Osaka, Japan, p32.
[9] Y. Ito, A. M. Vlaicu and T. Tochio, CP576, Application
of Accelerators in Research and Industry edited by J. L.
Duggan and I. L. Morgan (AIP)(2001).
[10] N. Shigeoka et al ., to be submitted.
[11] N. Shigeoka et al ., to be submitted.
[12] S. J. Schaphorst, A. Kodre, J. Ruscheinski, B. Crasemann, T. Aberg, J. Tulkki, M. H. Chen, Y. Azuma, and
G. S. Brown, Phys. Rev. A47, 1953 (1993).
[13] Y. Ito, H. Nakamatsu, T. Mukoyama, K. Omote, S.
Yoshikado, M. Takahashi, and S. Emura, Phys. Rev.
A46, 6083 (1992).
[14] T. Mukoyama and K. Taniguchi, Phys. Rev. A36, 693
(1987).
[15] M. Deutsch, O. Gang, G. Holzer, J. Hartwig, J. Wolf, M.
Fritsch, and E. Forster, Phys. Rev. A52, 3661 (1995).
[16] L. Salgueiro, J.Phys. (Paris), Colloq 48, C9-609 (1987).
[17] A.-M. Vlaicu, T. Tochio, T. Ishizuka, D. Osawa, Y. Ito,
T. Mukoyama, A. Nisawa, T. Shoji, and S. Yoshikado,
Phys. Rev.A58, 3544 (1998).
[18] M. H. Chen et al ., Atom. Data Nucl. Data Tables 19, 97
(1977).
[19] A. Nisawa et al ., to be published.
[20] H. Oohashi et al ., to be submitted.
[21] J. H. William, Phys. Rev. 45, 71 (1934).
[22] F. K. Richtmyer, S. W. Barnes, and E. Ramberg, Phys.
Rev.46, 843 (1934).
[23] P. Amorim, L. Salgueiro, F. Parente, and J. G. Ferreira,
J. Phys. B21, 3851(1988).
23