Auger decay of the molecular field split S 2p core excited states in HS radical

Article abstract of DOI:10.1063/1.473035

High-resolution synchrotron-radiation-excited electron spectroscopy is used to show for the first time that there exists a state-dependent anomalous Auger decay ratio also in the case of Auger decay in molecular fragments, created after resonant excitation. The molecular field splitting of a free radical is observed for the first time. Moreover, no evidence of the Auger resonant Raman effect is observed when dissociation occurs.

Full text of DOI:10.1063/1.473035

Auger decay of the molecular field split S 2p core excited states in HS radical
A. Naves de Brito, S. Svensson, S. J. Osborne, A. Ausmees, A. Kivimäki, O.-P. Sairanen, E. Nõmmiste, H.
Aksela, S. Aksela, and L. J. Saethre
Citation: The Journal of Chemical Physics 106, 18 (1997); doi: 10.1063/1.473035
View online: http://dx.doi.org/10.1063/1.473035
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/106/1?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Heavy atom nitroxyl radicals. VI. The electronic spectrum of jet-cooled H2PO, the prototypical phosphoryl free
radical
J. Chem. Phys. 135, 214307 (2011); 10.1063/1.3664903
Heavy atom nitroxyl radicals. V. An experimental and ab initio study of the previously unknown H2PS free radical
J. Chem. Phys. 135, 214306 (2011); 10.1063/1.3662416
Dissociation of core-valence doubly excited states in NO followed by atomic Auger decay
J. Chem. Phys. 133, 154315 (2010); 10.1063/1.3505297
Auger decay calculations with core-hole excited-state molecular-dynamics simulations of water
J. Chem. Phys. 124, 064307 (2006); 10.1063/1.2166234
Phosphorus hyperfine structure in the electronic spectrum of the HPCl free radical
J. Chem. Phys. 121, 1405 (2004); 10.1063/1.1758698
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

Auger decay of the molecular field split S 2p core excited states
in HS radical
A. Naves de Britoa)
The Finnish Synchrotron Radiation Facility at MAX-lab, University of Lund, S-22100 Lund, Sweden
S. Svensson, S. J. Osborne, and A. Ausmeesb)
Department of Physics, Uppsala University, Box 530, S-75121 Uppsala, Sweden
b)
¨
˜
A. Kivimaki, O.-P. Sairanen, E. Nommiste, H. Aksela, and
S. Aksela
The Finnish Synchrotron Radiation Facility at MAX-lab, University of Lund, S-22100 Lund, Sweden, and
Department of Physical Sciences, University of Oulu, FIN-90570 Oulu, Finland
L. J. Saethre
Department of Chemistry, University of Bergen, N-5007 Bergen, Norway
͑Received 5 June 1996; accepted 24 September 1996͒
Using synchrotron radiation, very high resolution electron spectra at the 2p→6a1 and 2p→3b2
resonances of H2S have been recorded. Auger transitions from molecular field split S 2p hole states
*
of HS fragment, created by fast dissociation of the core excited H2S, are found to reveal different
partial rates depending on the symmetry of the final state. The experiments have made it possible to
determine, for the first time, the molecular field splitting of the core levels in a free radical. The
*
molecular field splitting between the 3␴ and 1␲ (‘‘2p3/2’’) levels in the HS fragment is
1/2
3/2
found to be 90͑5͒ meV, which significantly differs from the molecular field splitting of the ‘‘S
2p3/2’’ levels in the H2S molecule. Resonant Auger decay in the molecule before dissociation is also
observed. © 1997 American Institute of Physics. ͓S0021-9606͑97͒00901-X͔
I. INTRODUCTION
The spin-orbit splitting between the S 2p3/2 and S
2p1/2 core levels in H2S has earlier been measured by using
Auger electron and x-ray photoelectron spectroscopies.7 It
was found that there is an apparent difference in spin–orbit
splitting when comparing the Auger and photoelectron re-
sults. Using high-resolution synchrotron radiation excited
Auger and photoelectron spectra this apparent ambiguity was
explained. The difference in splitting is due to a new propen-
sity rule for Auger decay from the molecular field split core
hole states.4 The observed anomalous decay ratio in H2S
pointed out the general necessity for reinterpretation of a
number of high-resolution normal Auger spectra of mol-
ecules.
Recent experimental progress in synchrotron radiation
instrumentation and high-resolution electron spectrometers
has provided new tools for the analysis of electronic struc-
ture of atoms and molecules. As an example the Auger reso-
nant Raman effect has recently been demonstrated in the vuv
region.1 This implies that the width of the electron lines is
governed by the combined broadening of the incident radia-
tion and the electron spectrometer. The method provides an
excellent tool to determine fine details of intensity distribu-
tion, offering a sensitive probe to electron correlation effects.
High-resolution synchrotron-radiation-excited photoelectron
spectroscopy also gives the possibility to determine the split-
ting of core hole states due to the molecular field.2–4 Further-
more, the comparison with Auger spectra provides a power-
ful tool to study the molecular geometry through its impact
in determining the propensity rules of Auger transitions.
For molecules, the resonance Auger process, which can
be studied with tuneable synchrotron radiation, may also in-
clude a complex dynamical process. The time scales of the
dissociation of the core excited molecule and the decay of
the core hole may be in the same order of magnitude; Auger
lines can in some cases be described as originating from
decay processes in dissociated fragments. However, the full
description of this phenomenon is more complicated, a de-
tailed discussion can be found in the recent studies.5,6
In the one-electron approximation, the ground state elec-
tron configuration of H2S is (C2v symmetry͒,
2p64a122b225a122b1 6a103b2
.
2
0
Upon 2p core excitation to either the 6a1 or 3b2 empty
orbitals the H2S molecule dissociates very rapidly into
HS ϩH fragments before the Auger decay.8–10 It is essential
*
*
to notice that the same core excited states in HS are reached
in both cases, independent of whether the 6a1 or 3b2 orbital
is initially occupied in H2S. After the fast neutral dissocia-
a͒
Present address: Department of Physics, University of Brasilia, 70910-900
*
tion of H2S, the core excited HS fragment can decay via
Brasilia-DF, Brazil.
b͒
Auger transition to six possible states in the ⌳⌺ coupling
Permanent address: Institute of Physics, Estonian Academy of Sciences,
Riia 142, EE-2400 Tartu, Estonia.
scheme
18
J. Chem. Phys. 106 (1), 1 January 1997
0021-9606/97/106(1)/18/6/$10.00
© 1997 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

Naves de Brito et al.: 2p core excited states in HS
19
2
2
2
5␴22␲ ͒ 3⌺Ϫ, 5␴22␲ ͒ 1⌬, 5␴22␲ ͒ 1⌺ϩ,
͑A͒ An anomalous Auger decay ratio is present and de-
pends on the symmetry of the final state.
͑B͒ One does not observe Raman behavior of the lines
when fast dissociation takes place before the Auger decay;
this was recently also shown in the case of HCl.5
͑
͑
͑
͑
3
3
4
5␴22␲ ͒ 3⌸, 5␴12␲ ͒ 1⌸, 5␴02␲ ͒ 1⌺ϩ.
͑1͒
͑
͑
In this numeration of the orbitals the core orbitals are in-
cluded. Consequently, the S 1s core orbital is for example
denoted 1␴.
In H2S, considering the molecular field interaction and
spin–orbit splitting the S 2p1/2 and S 2p3/2 orbitals are split
into three orbitals7 with the denotation
͑C͒ Higher vibrational states are populated in the inter-
mediate core excited state leading to the appearance of
prominent ‘‘hot bands’’ in the decay spectrum associated
with the 2pϪ13b2 core excited configuration. No difference
1
was observed earlier when comparing the decay spectra from
the 2pϪ13b2 and 2pϪ16a1 core excited configurations.
Resonant Auger decay in the molecule before dissocia-
tion is also observed. One of the valence photoelectron lines
associated to the 4a1 primary ionization, is found to increase
by as much as 75% in intensity at the energy of the
2p1/2→3b2 excitation. In all resonant Auger spectra a broad
background appears in the region where the structures con-
1
1
3e1/2 ‘‘2p1/2’’͒,4e1/2 and 5e1/2 ‘‘2p3/2’’͒.
͑2͒
͑
͑
*
However, in the HS fragment the Cϱv molecular symmetry
group applies and the following nomenclature ͑including
spin orbit interaction͒ will be used:
1␲ ‘‘2p1/2’’͒,3␴ and 1␲ ‘‘2p3/2’’͒.
͑3͒
͑
1/2
͑
3/2
1/2
*
Therefore, in both H2S and HS the ‘‘2p3/2’’ state is mo-
lecular field split. Obviously, the energy splitting does not
need to be the same. Furthermore, after dissociation there are
*
nected to the decay in the HS fragment are located. These
structures can be assigned to resonant Auger decay in the
H2S molecule. The presence of such molecular background
has recently been studied in great detail in the case of HCl.6
*
three possible core excited states in the HS species, namely,
1
1
4 2
1␲
͒
2␲͒4 2⌸, 3␴
͒
͑
2␲͒
1/2
⌺
͑
͑
͑
1/2
II. EXPERIMENT
1
and 1␲
͒
2␲͒4 2⌸.
͑
͑4͒
͑
3/2
The experiment was performed at the beamline 51, the
Finnish beamline, at the MAX I storage ring in Lund, Swe-
den. This beamline consists of a short undulator14 giving
useful radiation in the photon energy range of 60–600 eV. A
modified SX-700 plane grating monochromator ͑PGM͒ is in-
stalled at the beamline.15 A differential pumping section be-
tween the monochromator and the experimental station re-
duces the pressure by 5 orders of magnitude, thus reducing
the risk for contamination of the monochromator optical el-
ements. For the kinetic energy analysis of the electrons we
used a hemispherical-sector-type electron spectrometer,16
mounted in the magic angle. For this type of spectrometer
with the retarding lens the resolution does not depend sig-
nificantly upon the initial kinetic energy of the electrons. The
experiments were performed using 20 eV pass energy. At
this pass energy the spectrometer resolution is measured to
be around 65 meV.
The H2S gas was made through the reaction of iron͑II͒
sulphide ͑Aldrich, 99.9%͒ with 20 wt % hydrochloric acid:
FeS(s)ϩ2HCl(aq)ϭH2S(g)ϩFeCl2(aq). The HCl solu-
tion was placed in a pressure-equalizing dropping funnel on
top of a round-bottom flask with excess amounts of FeS, and
the flask connected to the gas vessel. The system was evacu-
ated and the HCl solution added dropwise. In order to avoid
impurity from H2O(g) and HCl(g), the gas was passed
through a trap cooled with liquid nitrogen.
In the present report we will analyze, in detail, the reso-
nant Auger decay from each of these states to each one of the
six final states given by Eq. ͑1͒. We will focus on two inter-
esting questions:
͑1͒ What consequences will radiationless resonant Ra-
man scattering have in molecules when dissociation is in-
volved?
͑2͒ When neutral fragmentation of the molecule occurs
before the decay of the core excited state,8–11 does the
anomalous decay ratio also occur in the resonant Auger de-
cay of the fragment?
To answer these questions we have studied the resonant
Auger decay from the 2pϪ16a1 and 2pϪ13b2 core excited
states in H2S. The 6a1 and 3b2 orbitals correspond to the
symmetric and antisymmetric combination of the antibond-
ings S–H orbitals, respectively. The related Auger spectrum
*
has been shown to result from Auger decay in the HS frag-
8–10
*
ment after fast dissociation of the core excited H2S
molecule. This Auger decay after fast molecular dissociation
was first observed for the HBr at the 3d→␴ excitation.12
*
Recently an experiment was reported showing the presence
of resonant Raman effect in the HBr molecule when disso-
ciation is not involved.13 An anomalous decay ratio was also
observed, i.e., when molecular field split core hole states
appear, decay from only one state was observed. The Raman
and non-Raman behavior of the resonance Auger process in
HCl has been recently investigated and discussed.5 In the
present experiment we have repeated the measurements of
Ref. 8, but with the much higher resolution that can be
achieved with the equipment at the MAX laboratory. Vibra-
tionally resolved spectra have been obtained and it has been
possible to study small energy shifts and line broadening
effects.
III. RESULTS AND DISCUSSION
In Fig. 1 the electron spectra taken at three different
photon energies: ͑a͒ 164.35, ͑b͒ 165.50, and ͑c͒ 166.95 eV
are shown. The photon energies were chosen so that at
164.35 eV only the ‘‘2p3/2’’→6a1 excitation is reached, at
165.50 eV both the 2p→6a1 ;3b2 excitations are possible,
and at 166.95 eV only the ‘‘2p1/2’’→3b2 excitation is ac-
From the measured spectra we have found that:
J. Chem. Phys., Vol. 106, No. 1, 1 January 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

20
Naves de Brito et al.: 2p core excited states in HS
The following assignment can be made for the lines in
the upper spectrum (h␯ϭ164.35 eV) in Fig. 1: The lines
3
around 148 eV correspond to transitions to the (2␲2) ⌺Ϫ
final state in the HSϩ ion. The lines at around 147, 146, and
144
eV
correspond
to
transitions
to
the
(2␲2) ⌬,(2␲2) ⌺ϩ and (5␴12␲1) ⌸ final states, re-
spectively. The same bands appear in the lower spectrum
(h␯ϭ166.95 eV) shifted by about 1.2 eV towards higher
kinetic energy. The spectrum in the middle contains an over-
lap of both spectra taken at 164.35 and 166.95 eV. Extensive
vibrational progressions can be observed for different states
in Fig. 1. In this paper we will not make the extensive analy-
sis of the details of the vibrational progressions, this will be
presented elsewhere.
1
1
3
The splitting between the ⌺Ϫ peaks in the 2p1/2 and
3
2p3/2 spectra is 1209͑5͒ meV, while the splitting between the
3⌸ peaks in the 2p1/2 and 2p3/2 spectra is only 1158͑8͒ meV,
the splitting was obtained from least-squares fits to each
spectrum in Fig. 1. The fits were done in the following steps:
First, the valence photoelectron lines were fitted using a va-
lence photoelectron spectrum taken just below the S 2p
threshold. A broad background Gaussian was added. The
spectra at the 2p3/2→6a1 resonance ͑a͒ in Fig. 1, and at the
2p1/2→3b2 resonance ͑c͒ in Fig. 1, were then fitted using
Voigt line profiles for the resonant peaks. The results were
used to fit the spectrum ͑b͒ in Fig. 1. The four lower vibra-
tional states in the 3⌸ state were used to align the spectra ͑a͒
*
FIG. 1. The Auger decay spectra from the HS core excited fragment taken
at different excitation energies: ͑a͒ 164.35 eV, ͑b͒ 165.50 eV, and ͑c͒ 166.95
eV. The photon energies were chosen so that at 164.35 eV only the
‘‘2p3/2’’→6a1 excitation is reached, at 165.50 eV both the
2p→6a1 ;3b2 excitations are reached and at 166.95 eV only the
‘‘2p1/2’’→3b2 excitation is reached in H2S. Note that the small line at
146.1 eV in ͑c͒ is due to a so-called ‘‘hot band.’’ See the text.
cessible in H2S. The monochromator exit slit was set to
600 ␮m ͑excitation bandwidth of ca. 1 eV͒ to increase count
rate but still allow selective excitation to the empty orbitals
described above. The excitation profile for such a large slit
can be approximated by a square function. In Fig. 2 we show
the total ion yield spectrum of H2S and we have indicated the
excitation bandwidths for the cases ͑a͒–͑c͒ in Fig. 1. Using
such a broad excitation band we smear out the valence pho-
toelectron lines that can readily be assigned in Fig. 1. In this
paper we will not focus on the direct and resonant molecular
photoemission contribution but will concentrate the discus-
3
and ͑b͒ in Fig. 1. The (2p1/2Ϫ1) ⌸ vibrational states and the
3
(2p1/2Ϫ1) ⌺Ϫ state were used to align the spectra ͑b͒ and
͑c͒. The average splitting between each vibrational line in the
3
3
(2p1/2Ϫ1) ⌸ and (2p3/2Ϫ1) ⌸ states, respectively, was
found to be 1158͑8͒ meV, as reported above. The line at
146.1 eV in Fig. 1͑c͒ as well as the lines in Fig. 1͑b͒ 148.9
eV, and lines at 148.6 and 149.8 eV in Figs. 1͑b͒ and 1͑c͒ are
assigned to the transitions from the second vibrational state
in the core excited state to final HSϩ states.17 In the literature
these states have been referred to as ‘‘hot bands.’’ The reader
should note that these lines are not present at the lower ex-
citation photon energy spectrum shown in Fig. 1͑a͒, where
only the 6a1 state is populated. Therefore, the hot bands are
connected with the possibility to populate the second core
excited state 3b2. The existence of these states does not in-
terfere with the calculation of the spin–orbit splitting be-
*
sion on the Auger decay in the HS core excited fragment.
3
tween the ⌸ peaks, since either they lie outside the vibra-
tional progression used to obtain the shift, Figs. 1͑b͒ and
1͑c͒, or are absent, Fig. 1͑a͒.
To illustrate the differences we have aligned the upper
and lower spectra in the following way. In Fig. 3 we show
the comparison between decay spectrum at the 2p3/2→6a1
(h␯ϭ164.35 eV, upper͒ and 2p1/2→3b, (h␯ϭ166.95 eV,
lower͒ resonances, respectively. The upper spectrum was
3
shifted so that the ⌸ bands match each other.
In Fig. 4 we show two Auger spectra taken at the
2p1/2 ;2p3/2→6a1 ;3b2 resonance energy of 165.5 eV. The
valence photoelectron bands are also indicated. As men-
tioned above, the most intense 4a1 peak resonates at the
2p1/2→3b2 excitation energy. In Fig. 4 ͑lower spectrum͒
this broad peak is indicated and its presence hides the vibra-
FIG. 2. The H2S total ion yield spectrum in the region 163.8 to 167.5 eV
photon energy, as obtained by using a monochromator slit size of 10 ␮m.
The photon band width for 600 ␮m slit is indicated in the figure, corre-
sponding to the spectra ͑a͒–͑c͒ in Fig. 1. The Gaussian shown in order of
increasing
photon
energy
correspond
to
the
excitations
‘‘2p3/2’’→6a1 ;‘‘2p3/2’’→3b2 ;‘‘2p1/2’’→6a1 ;‘‘2p1/2’’→3b2, respec-
tively.
J. Chem. Phys., Vol. 106, No. 1, 1 January 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

Naves de Brito et al.: 2p core excited states in HS
21
3
Ϫ
FIG. 5. The result of a curve fit to the structures of the 1⌬ and
states of the spectrum ͑a͒ in Fig. 1.
⌺
final
FIG. 3. The spectra corresponding to ‘‘clean’’ transitions from the ͑a͒
3␴ and 1␲ (‘‘2p3/2’’), and ͑b͒ 1␲ (‘‘2p1/2’’) initial states, respec-
1/2
3/2
1/2
tively. The spectra are aligned as discussed in the text.
bonding state, the decay of this manifold to a single final
state shows a similar Raman behavior: Due to energy con-
servation the shape of an Auger peak ͑due to an electronic
transition͒ and its energy follow the photon bandwidth, even
when the lifetime width is narrower than the photon band-
width. The shape of the peak is further affected by vibra-
tional interference effects. In the spectrum taken using
80 ␮m slit, we are well below the upper limit to observe a
narrowing and distortion in the molecular Auger peaks. In-
deed, a clear effect is seen in the 4a1 peak, which has a
strong contribution of participator resonant Auger transi-
tions. No clear slit size effect is observed on the narrow
lines. Since the photon bandwidth at 80 ␮m slit is substan-
tially smaller than that of the resonance width one can con-
clude that the Auger resonant Raman effect is not present,
when dissociation occurs before the Auger decay. This indi-
cates that the decay of the resonance in this case can be
thought of as a two-step process. The first is the rapid ejec-
tion of a ground state H atom and the formation of core
excited radical. The kinetic energy released in the dissocia-
tion is entirely transferred to the H atom. The second step is
the resonant Auger decay in the core excited fragment. This
separation of the dissociation and decay in two steps also
implies that no dispersion of the Auger electron energy with
the photon energy should be observed. The dispersion of the
photon energy would be entirely seen on the kinetic energy
release of the fragmentation. An analogy can be drawn to the
normal Auger decay following core ionization, where a part
of available energy is carried away as the kinetic energy of
photoelectron. From the viewpoint of resonant scattering
theory the situation is analogous to a normal Auger decay in
an atom. Now the molecular dissociation has produced a
manifold of nuclear continuum states both in the intermedi-
ate excited state and in the final state. This leads to the van-
ishing of resonant Raman narrowing in the Auger spectrum.
In Fig. 5 we show the result of a fitting procedure. In this
particular case two lines, corresponding to the transitions
tional band associated to the Auger transition to the 1⌸ final
state. The monochromator resolution using 600 ␮m exit slit
was measured from the valence 2bϪ1 1 state to be about 1.0
eV in total width. For 80 ␮m exit slit the resolution is esti-
mated to be 120 meV.18 A full width at half maximum
͑FWHM͒ of 0.9 eV was measured for the 2pϪ16a11 and
2pϪ13b12 core excited state in the H2S absorption spectrum
in Fig. 2. Our experiment shows a clear difference in the
Raman behavior between the molecular transitions and the
transitions taking place in the fragment. For free atoms, the
Auger resonant Raman effect becomes observable if the
width of the exciting radiation is comparable with the life-
time width of an excited state. Under such conditions the
Raman features, line shape narrowing and linear dispersion,
are observed. If the broadness of a core excited molecular
state is due to a manifold of close-lying vibronic levels of a
*
FIG. 4. The Auger spectra from the HS fragment at 165.5 eV photon
energy obtained using different slit sizes of 600 and 80 ␮m of the mono-
chromator. In the high-resolution spectrum ͑lower, 80 ␮m slit͒ it is seen that
the resonance Auger transitions to the 4a1 molecular final state are dis-
cerned due to their smaller width.
2
from the (3␴1/22␲4) 2⌺ and (1␲3/22␲4) ⌸ levels, are
taken into account in the fitting procedure. We have consid-
J. Chem. Phys., Vol. 106, No. 1, 1 January 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

22
Naves de Brito et al.: 2p core excited states in HS
ered the transitions to the ϭ0 and ϭ1 vibrational sub-
(3␴1/22␲4) and (1␲3/22␲4) core excited states decay to the
v
v
Ј
Ј
3
Ϫ
levels in the 1⌬ and 3⌺Ϫ final states. From the fitting proce-
dure a splitting of 90 ͑5͒ meV was obtained between the
⌺
and 1⌬ final states. However, only the (3␴1/22␲4) core
3
excited state decays prominently only to the ⌸. While the
decay to the ⌸ behaves as the anomalous decay process
3
2
2
(3␴1/22␲4) ⌺ and (1␲3/22␲4) ⌸ levels. This value devi-
ates significantly from the molecular field splitting of the
2p3/2 core hole states in H2S, which is 110 ͑5͒ meV.4 In
order to get reasonable fits it was necessary to include a very
broad Gaussian centered around 146 eV. This Gaussian is
attributed to the decay in the molecule. The decay occurs
from a part of the molecular potential energy surface of the
core excited molecule having a steep gradient to a final state
with a different slope. The positions of the two fitted molecu-
lar field split lines in the 3⌺Ϫ final state ͑Fig. 5͒ are indicated
by two vertical lines separated by 90 meV in Fig. 3. As one
can see from Fig. 3, the fitted line at high kinetic energy of
observed for H2S normal Auger, no strong propensity rule is
3
1
observed in the decay to the ⌺Ϫ and ⌬ states.
The anomalous decay ratio of the core excited state in
*
the HS fragment can qualitatively be understood by the
same arguments as were used in a study of the normal Auger
spectrum of HCl and DCl.21 The decay takes place in an
*
isoelectronic system when compared to the HS radical. If
we, as usual, orient the fragment along the z axis the
px ,py ,pz atomic orbital contributions to the three HS S 2p
molecular field split core orbitals are
␲
␴
␲
:⌿ϭ0.577 p ϩip ͒␤Ϫp ␣ ,
͓͑
͑5͒
͑6͒
͑7͒
͔
1/2
1/2
3/2
x
y
z
3
the ⌺Ϫ peak in the ‘‘2p3/2’’ spectrum coincides with the
3
Ϫ
:⌿ϭ0.408 Ϫp ϩip ͒␣ϩ2p ␤ ,
͓͑
͔
⌺
line in the ‘‘2p1/2’’ spectrum.
x
y
z
3
Ϫ
The
⌺
and 1⌬ lines in the ‘‘2p3/2’’ spectrum are
:⌿ϭ0.707 p ϩip ͒␣ .
͓͑
͔
x
y
considerably broadened when compared to the ‘‘2p1/2’’
spectrum. This is in contrast to what is observed for the
decay to the ⌸ final state. In Fig. 3 one can notice that the
Here, ␣ and ␤ refer to the spin. One should note that only
3
the ␲ core orbital lacks a contribution from pz . One then
3/2
FWHM is 175 meV for the transition from the ‘‘2p3/2’’ core
has to analyze all the six possible transitions to the HSϩ final
excited state to the 3⌺Ϫ final state, and only 127 meV for the
states of ⌺Ϫ and ⌸ symmetry.
3
3
3
transition from the ‘‘2p1/2’’ core excited state to the ⌺Ϫ
One finds that from the orientation of the orbitals one
cannot expect large depressions of the intensity for the tran-
sitions from the ␲1/2 and ␴1/2 core orbitals. There is always a
possible combination that will give rise to a substantial Au-
ger intensity.
final state. As pointed out above the discrepancy comes from
the fact that the ‘‘2p3/2’’ is molecular field split whereas
3
‘‘2p1/2’’ is not. The ⌸ peak in the ‘‘2p3/2’’ spectrum is
just slightly broadened as compared to the same peak in the
‘‘2p1/2’’ spectrum. The intensity ratios of the two molecular
For the ␲ core vacancy the situation is different. For
3/2
3
1
3
field split components for the ⌺Ϫ and ⌬ final states have
transitions to the ⌺Ϫ final state the decay involves a va-
3
been determined to be 1.4 ͑1͒ and 1.1 ͑1͒, respectively ͑Fig.
cancy in the ␲ core orbital and a ⌺Ϫ final state. Since
3/2
5͒. These branching ratios hold for both the
ϭ0 and
both have a pxϩpy atomic orbital contribution no depression
is expected. However, for the transition to the ⌸ final state
v
Ј
3
1
3
ϭ1 vibrational sublevels in the ⌺Ϫ and ⌬ final states.
v
Ј
As can be seen from Fig. 5, the line corresponding to tran-
sitions to the 1⌺ϩ final state is substantially broadened, mak-
ing it impossible to draw firm conclusions on the molecular
field splitting of the initial core excited state.
the final state configuration has one orbital (␲) with a
pxϩpy contribution and another (␴) with a pz contribution.
We have two possibilities for the Auger decay. The ␲ va-
lence electron ͑with pxϩpy contribution͒ can fill the ␲
3/2
Interference effects may also be present in the decay
core hole and the pz (␴Ϫ1) electron is promoted to the con-
tinuum. Since the core orbital in this case has pxϩpy contri-
bution no depression is expected. The second possibility is
that the pz (␴) would fill the core hole and the ␲ electron is
promoted to the continuum. This second path is, however,
strongly depressed because the ␲3/2 molecular field split core
orbital has no pz contribution.
*
spectrum of HS . These effects have been studied for the S
2p normal Auger decay of H2S.19 However, such interfer-
ence effects should hardly be discernible if we consider the
lifetime of the core ionized state, which was found to be
around 70 meV, and the relatively large vibrational energy
splitting of the states involved. One might still have a life-
time interference effect originating from the two molecular
field split components of the ‘‘S 2p3/2’’ core ionized state.
An analysis shows that such an effect might be stronger in
the case of close lying states, but a very small effect was
predicted in the branching ratio.20 In conclusion, although
interference effects may be present they play a minor role as
compared to the strong propensity rule we observe in the
molecular fragment.
As a summary, the decay involving a ␲ initial core
3/2
3
hole and the ⌸ final state is depressed. This depression is
not expected to be as strong as it was in the case of normal
Auger decay in case of a vacancy in the 4e1/2 core orbital in
H2S.4 Experimentally this can be confirmed since we ob-
3
serve a weak decay involving ␲ and ⌸. This is based on
3/2
the fact that the FWHM of the 3⌸ vibrationally resolved
states is 148 meV at 167 eV photon energy while at 164.3 eV
the FWHM is 133 eV. So at 167 eV photon energy, where
decay of both ␲ and ␴ could be observed the peaks are
Also the apparent 2p spin–orbit splitting is different
when comparing the lines that correspond to the transitions
to the 3⌺Ϫ and 3⌸ final states ͑1209 and 1158 meV, respec-
tively͒. From these observations and using information
learned from the normal Auger spectrum of H2S one can
explain the discrepancy. Both the molecular split
3/2
1/2
broader indicating a small decay involving the ␲ core
3/2
hole.
For diatomic molecules one could probably expect a de-
pression of the transitions from the states involving a va-
J. Chem. Phys., Vol. 106, No. 1, 1 January 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47

Naves de Brito et al.: 2p core excited states in HS
23
cancy in the ␲ core orbital as a general rule. A more de-
4 S. Svensson, A. Ausmees, S. J. Osborne, F. Gel mukhanov, H. Ågren, A.
Ј
3/2
¨
˜
Naves de Brito, O.-P. Sairanen, A. Kivimaki, E. Nommiste, H. Aksela,
tailed explanation of analogous calculations made for the
H2S molecule in the case of the normal Auger decay can also
and S. Aksela, Phys. Rev. Lett. 72, 3021 ͑1994͒.
5 E. Kukk, H. Aksela, S. Aksela, F. Gel mukhanov, H. Ågren, and S.
Ј
be found in a recent study by Gel mukhanov et al.22
Ј
Svensson, Phys. Rev. Lett. 76, 3100 ͑1996͒.
6
¨
˜
E. Kukk, H. Aksela, O.-P. Sairanen, S. Aksela, A. Kivimaki, E. Nom-
miste, A. Ausmees, A. Kikas, S. J. Osborne, and S. Svensson, J. Chem.
Phys. 104, 4475 ͑1996͒.
IV. CONCLUSIONS
7 S. Svensson, A. Naves de Brito, M. P. Keane, N. Correia, and L. Karlsson,
Phys. Rev. A 43, 6441 ͑1991͒.
In conclusion, using high-resolution synchrotron-
radiation-excited electron spectroscopy, we have shown for
the first time that there exists a state-dependent anomalous
Auger decay ratio also in the case of Auger decay in molecu-
lar fragments, created after resonant excitation. We have ob-
served, for the first time, the molecular field splitting of a
8 S. Aksela, H. Aksela, A. Naves de Brito, G. M. Bancroft, and K. H. Tan,
Phys. Rev. A 45, 7948 ͑1992͒.
9 A. Naves de Brito and H. Ågren, Phys. Rev. A 45, 7953 ͑1992͒.
10 A. Naves de Brito, N. Correia, B. Wannberg, P. Baltzer, L. Karlsson, S.
Svensson, M. Y. Adam, H. Aksela, and S. Aksela, Phys. Rev. A 46, 6067
͑1992͒.
*
free radical, namely the ‘‘S 2p3/2’’ level in the HS frag-
11 H. Aksela, S. Aksela, M. Ala-Korpela, O.-P. Sairanen, M. Hotokka, G. M.
Bancroft, K. H. Tan, and J. Tulkki, Phys. Rev. A 41, 6000 ͑1990͒.
12 P. Morin and I. Nenner, Phys. Rev. Lett. 56, 1913 ͑1986͒.
13 Z. F. Liu, G. M. Bancroft, K. H. Tan, and M. Schachter, Phys. Rev. Lett.
72, 621 ͑1994͒.
ment, and found this to be 90 meV, i.e., significantly smaller
than the corresponding splitting in H2S molecule. Moreover,
no evidence of the Auger resonant Raman effect is observed
when dissociation occurs.
14 Å. Andersson, S. Werin, T. Meinander, A. Naves de Brito, and S. Aksela,
Nucl. Instrum. Meth. A 362, 586 ͑1995͒.
15
¨
ACKNOWLEDGMENTS
S. Aksela, A. Kivimaki, A. Naves de Brito, O.-P. Sairanen, and S. Svens-
son, Rev. Sci. Instrum. 65, 831 ͑1994͒.
16 S. J. Osborne, A. Ausmees, J.-O. Forsell, B. Wannberg, G. Bray, L. B.
The authors would like to thank Ing. J.-O. Forsell for
assistance. This work has been supported by the Swedish
Natural Science Research Council ͑NFR͒, by the Royal
Swedish Academy of Sciences, and by the Research Council
for the Natural Sciences of the Academy of Finland. A. N. B.
wants to thank the Brazilian Natural Science and Technology
Development Council ͑CNPq͒ for financial support during
part of this project.
¨
Dantas, S. Svensson, A. Naves de Brito, A. Kivimaki, and S. Aksela,
Synchr. Rad. News 7, 25 ͑1994͒.
A. Naves de Brito, Al. Naves de Brito, O. Bjorneholm, J. Soares Neto, A.
Bueno Machado, S. Svensson, A. Ausmees, S. J. Osborne, L. J. Saethre,
17
¨
¨
˜
H. Aksela, O.-P. Sairanen, A. Kivimaki, E. Nommiste, and S. Aksela,
Theo. Chem. Phys. In press 1996.
18
¨
˜
S. Aksela, A. Kivimaki, O.-P. Sairanen, A. Naves de Brito, E. Nommiste,
and S. Svensson, Rev. Sci. Instrum. 66, 1 ͑1995͒.
19 A. Cesar, H. Ågren, A. Naves de Brito, S. Svensson, L. Karlsson, M. P.
Keane, B. Wannberg, P. Baltzer, P. G. Fournier, and J. Fournier, J. Chem.
Phys. 93, 918 ͑1990͒.
1
¨
A. Kivimaki, A. Naves de Brito, S. Aksela, H. Aksela, O.-P. Sairanen, A.
20 A. Cesar and H. Ågren, Phys. Rev. A 45, 2833 ͑1992͒.
Ausmees, S. J. Osborne, L. B. Dantas, and S. Svensson, Phys. Rev. Lett.
21
¨
˜
H. Aksela, E. Kukk, S. Aksela, O.-P. Sairanen, A. Kivimaki, E. Nom-
miste, A. Ausmees, S. J. Osborne, and S. Svensson, J. Phys. B 28, 4259
͑1995͒.
54, 1142 ͑1993͒.
2 J. N. Cutler, G. M. Bancroft, D. G. Sutherland, and K. H. Tan, Phys. Rev.
Lett. 67, 1531 ͑1991͒.
3 J. N. Cutler, G. M. Bancroft, and K. H. Tan, J. Chem. Phys. 97, 7932
͑1992͒.
22 F. Gel mukhanov, H. Ågren, S. Svensson, H. Aksela, and S. Aksela, Phys.
Ј
Rev. A 53, 1379 ͑1996͒.
J. Chem. Phys., Vol. 106, No. 1, 1 January 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.239.20.174 On: Sun, 23 Nov 2014 11:18:47