J. Phys. B: At. Mol. Opt. Phys. 30 (1997) 403–412. Printed in the UK
PII: S0953-4075(97)76226-6
Observation of electron impact inner-shell excitation of
argon by means of satellite Auger spectra
B Paripás†, Gy Vı́kor‡§ and S Ricz‡
† Department of Physics, University of Miskolc, H-3515 Miskolc-Egyetemváros, Hungary
‡ Institute of Nuclear Research of the Hungarian Academy of Sciences (ATOMKI), H-4001
Debrecen, Pf 51, Hungary
Received 12 July 1996
Abstract. We have studied the inner-shell excited states of argon below the 2p thresholds for
electron-impact excitation in the range of the incident energy E0 = 275–1700 eV. We carried
this out through the measurement of the electrons emitted during the decay of the states. The
corresponding satellite peaks in the spectra of the emitted electrons were separated from the
normal (diagram) LMM Auger peaks by a special evaluating process. We observed the satellite
peaks coming from the 2p−1 (2 P3/2 )4s → 3p−2 (1 D)4s, 2p−1 (2 P3/2 )4p → 3p−2 (1 D)4p and
2p−1 (2 P3/2 )3d → 3p−2 (1 D)4d transitions at all applied incident energies. The intensity of the
satellite peaks increased approaching the ionization threshold (E0 = 275–300 eV). Nevertheless
in the range E0 = 1000–1700 eV we could identify more satellites due to the lower continuous
background and the smaller post-collision interaction effects.
1. Introduction
The inner-shell excited states of argon below the 2p ionization thresholds have been
known for many years. The photoexcitation of these states is widely studied through the
measurement of the energy spectra of the emitted Auger-like electrons (resonant Auger
process) (e.g. de Gouw et al 1995). The electron impact excitation of these states was
studied by Shaw et al (1982) using an electron energy-loss technique. As a second method,
one can also study the energy spectra of the emitted electrons in this case. Unfortunately the
observation of these Auger-like electrons originating from the electron impact inner-shell
excited states is very difficult, because many excitation and ionization channels are open.
In the latter case these lines can be considered as certain satellites on the high-energy tail of
the normal or diagram Auger lines. Though some of these satellites were observed before
by Werme et al (1973), Helenelund et al (1982) and Weber et al (1991), these authors did
not identify them (except for Weber et al 1991). In our paper—besides identification of
these satellites—we try to answer some particular questions which will be specified below.
2. Satellite Auger lines following Ar∗ (2p−1 (2 P3/2,1/2 )nl) inner-shell excitation
During the electron impact inner-shell excitation of Ar the projectile electron, e−
p , having
energy E0 , can excite the Ar atom by the promotion of an inner shell 2p electron to an
empty nl Rydberg orbital. The resulting state can decay dominantly through an LMM
§ On leave from the Institute of Physics, PO Box 57, 11001 Belgrade, Yugoslavia.
c 1997 IOP Publishing Ltd
0953-4075/97/020403+10$19.50 403
404
B Paripás et al
Auger-like manner: one 3p electron fills the inner-shell vacancy and another 3p electron is
ejected (e−
satellite ). During this process the Rydberg electron can either remain in its original
nl orbital (spectator) or can be shaken up to a higher (rarely shaken down to a lower)
excited n0 l state. It is stressed that shake-up process is a very important exit channel n0 > n
(Meyer et al 1991), and during this process the orbital momentum of the Rydberg electron
is not changed (l 0 = l). This inner-shell excitation-decay mechanism is denoted:
ep− + Ar(1 S0 ) → Ar∗ (2p−1 (2 P3/2,1/2 )nl) + e−
sc
−→ Ar+∗ (3p−2 (1 S0 , 3 P0,1,2 , 1 D2 )n0 l) + e−
satellite .
Without Rydberg electrons the above process is called the normal or diagram
L2,3 M2,3 M2,3 Auger process. Then the two initial and five final states form the well known
Auger spectrum consisting of 10 lines.
The existence of the Rydberg electron considerably complicates the measured Augerlike (satellite) spectrum. Not only are the number of lines multiplied by the number of the
possible nn0 l combinations, but the lines are split up according to the coupling of the angular
momenta of the vacancies and the Rydberg electron. The fine structure of the above two
initial and five final configurations result in several hundred possible transitions. Fortunately,
there are some experimental results which allow us to simplify the model satellite spectrum.
First of all, the separation of the levels of the initial 2p−1 (2 P3/2 )nl configurations is less
than 30 meV which is much smaller than their natural widths (approximately 120 meV)
(Shaw et al 1982). Therefore, the fine-structure splitting of this inner-shell excited state
can increase the width of satellite peaks, can distort their shapes but never increases their
number. Secondly, the fine structure of the emitted electron spectra of the photoexcited
2p−1 (2 P3/2 )4s, 2p−1 (2 P3/2 )3d, 2p−1 (2 P3/2 )4d, 2p−1 (2 P3/2 )5d and 2p−1 (2 P1/2 )4s inner-shell
states are known from the experimental results of de Gouw et al (1995). In table 1
we compiled the specifications of the most intense satellites originating from the above
initial states. Since the method of excitation hardly affects the decay of the excited states,
we can also use these data for electron impact excitation. Obviously table 1 does not
contain information about the relative probabilities of the different 2p−1 (2 P3/2 )nl innershell excitations, so the given percentages belonging to different initial configurations are
not comparable.
However, the above list of the possible inner-shell excitation processes is not complete.
In case of electron impact at least two further types of inner-shell excitation process can be
observed.
(i) The 2p−1 (2 P1/2 )nl inner-shell states (except for 2p−1 (2 P1/2 )4s) where the excitation
energy is greater than the ionization threshold (248.63 eV) of the Ar+ 2p−1 (2 P3/2 ). Here
(as a first approximation) the configurations given in table 1, can be used for 2p−1 (2 P1/2 )nl
excitations also, but with energies increased by 2.15 eV (the fine-structure splitting of the
Ar+ 2p−1 (2 P3/2 ) and (2 P1/2 ) states) and with intensities multiplied by 12 (the statistical weight
of (2 P1/2 ) compared to (2 P3/2 )).
(ii) The 2p−1 (2 P3/2 )4p photoexcitation is electric dipole forbidden, but the electron
impact 2p−1 (2 P3/2 )4p excitation becomes dominant just above the ionization threshold
(Shaw et al 1982). The situation must be similar at other 2p−1 (2 P3/2 )np and at the
2p−1 (2 P1/2 )np excitations, too.
Due to the lack of direct experimental data, the energies of the electron ejected from
the above-mentioned inner-shell excited states can be calculated combining the 245.96 eV
excitation energy for the 2p−1 (2 P3/2 )4p initial state (King et al 1977) and the energies of
the final 3p−2 (1 S0 , 3 P0,1,2 , 1 D2 )n0 p states from Cvejanović et al (1994) and from Moore
Excitation of argon by satellite Auger spectra
405
Table 1. The specifications of the dominant lines of the emitted electron spectra originating
from the photoexcited 2p−1 (2 P3/2 )nl inner-shell states. (From the experimental results of de
Gouw et al (1995).)
Initial state
Final state
Energy (eV)
Decay
probability
(%)
2p−1 (2 P
3p−2 (1 S)4s 2 S
3p−2 (1 D)4s 2 D and 3p−2 (3 P)3d 2 D
3p−2 (3 P)4s 2 P
3p−2 (1 P)4s 4 P
207.85
210.08
211.41
211.88
10
46
22
11
2p−1 (2 P3/2 )3d
3p−2 (1 D)4d 2 P, 2 D, 2 F and 3p−2 (3 P)5d 4 D, 4 F
3p−2 (3 P)4d 4 P, 2,4 F, 4 D
3p−2 (1 D)3d 2 P, 2 D,
206.37
208.20
209.62
27
15
25
2p−1 (2 P3/2 )4d
3p−2 (1 D)6d 2 S, 2 P, 2 D, 2 F, 2 G
3p−2 (1 D)5d 2 S, 2 P, 2 D, 2 F, 2 G
3p−2 (3 P)6d 2,4 P, 2,4 F, 2,4 D
3p−2 (3 P)5d 2 P, 2 D and 3p−2 (1 D)4d 2 S
3p−2 (3 P)5d 4 P, 4 D, 2,4 F and 3p−2 (1 D)4d 2 D, 2 F
204.37
205.17
205.98
206.52
206.99
15
20
11
16
11
2p−1 (2 P3/2 )5d
3p−2 (1 D)7d and 3p−2 (1 D)8d
3p−2 (3 P)7d and 3p−2 (3 P)8d
204.00
205.73
36
40
3/2 )4s
Table 2. Calculated energies of the dominant lines of the emitted electron spectra originating
from the electron impact excited 2p−1 (2 P3/2 )np inner-shell states. (Details in text.)
Initial state
Final state
Energy (eV)
2p−1 (2 P3/2 )4p
3p−2 (1 S)4p 2 P
3p−2 (1 D)4p 2 D, 2 P and 2 F
3p−2 (3 P)4p 2 S, 4 S, 2 P and 2 D
3p−2 (3 P)4p 4 D and 4 P
3p−2 (1 D)5p 2 P
206.39
208.70, 208.77, 209.06
210.23, 210.24, 210.37, 210.48
210.65, 210.93
206.62
(1971) (table 2). The decay probabilities (relative intensities) for these transitions are not
known.
Another non-negligible decay path of inner-shell excited states involves the ejection
of two electrons, shown by the photoelectron spectra of Heimann et al (1987). However,
for the 2p−1 (2 P3/2 )nl excitation of Ar it is not an important exit channel (Meyer et al
1991); moreover, this process probably will result in a continuous ejected electron energy
distribution and not a sharp peak-like structure in the measured energy region.
3. Experimental set-up
The Auger electron spectra were obtained by a triple-pass electrostatic electron spectrometer.
The spectrometer (called ESA-21) is primarily used in high-energy ion–atom collisions and
described in detail by Varga et al (1992). Briefly, the spectrometer consists of two stages:
a spherical mirror electron transport and a cylindrical mirror double-pass high-resolution
analyser. The first stage collects the electrons from the collision region of the gas-jet target
406
B Paripás et al
and the primary electron beam and directs them to the object point of the second stage. The
construction of the spectrometer makes it possible to analyse the electrons at 13 different
emission angles simultaneously as the electrons are detected by 13 channeltrons in the socalled detector ring. In our present experiment (because the axis of the built-in electron gun
is perpendicular to the ion beam) we could detect the electrons emitted at a 15◦ , 30◦ , 45◦ ,
60◦ , 75◦ and 90◦ angle with respect to the primary electron beam (at both sides) and with
an acceptance angle of 3.5◦ for all directions. In the present measurement at 0◦ a moveable
Faraday cup was put to collect the projectile electrons. The relative energy resolution of
the spectrometer which can be changed within the interval from 0.01 and 0.35%, was set
to about 0.05% by moveable slits.
The data acquisition was made by a personal computer which controlled and changed
the spectrometer potential, registered events from the counters connected to the channeltrons
and from the current integrator connected to the Faraday cup.
The Ar L2,3 M2,3 M2,3 Auger electron spectrum (and its neighbourhood) was measured at
275, 300, 500, 1000, 1700 eV projectile electron energies. The projectile current varied from
400 to 5000 nA as a function of the projectile energy, due to the smaller beam divergence
at higher energies.
The target pressure was stabilized by a constant (12 mbar) buffer-gas pressure. The
overall pressure in the vacuum chamber was 5.4 × 10−5 mbar with target gas, and the
base pressure was typically 3 × 10−7 mbar. To increase the electron detection efficiency,
a differential pumping system was applied, which provided 10 times better vacuum in the
region of the detectors and the inside of the double cylindrical spectrometer.
The combination of the spherical mirror electron transport and the double-pass
cylindrical mirror electron energy analyser effectively reduces the counts due to the scattered
electrons on the electrodes. This very low instrumental background together with the highenergy resolution and the high and stable transmission of the spectrometer proved to be
ideal for our ‘satellite hunting’ experiment.
4. Data evaluation
The computer fit of the above diagram Auger and satellite lines (and the polynomial
background) to the experimental spectra was made by the EWA least-squares fit program
of Végh (1995).
Considering that the majority of the satellite peaks overlap with the high-intensity
diagram L2,3 M2,3 M2,3 Auger lines, the knowledge of the exact Auger line shapes is essential.
Close to the ionization threshold the Lorentzian Auger line shape, yiL , is condiderably
distorted by the post-collision interaction (PCI). The theoretically distorted line shape, yiK ,
in eikonal approximation is described as follows (Kuchiev and Sheinerman 1987):
yiK = yiL πξ/ sinh(πξ ) exp(2ξ arctan ε);
yiL = Y0 (1 + ε 2 )−1
and
ε = (E0 − Ei )/(0.50L )
where yi is the electron intensity at Ei energy, 0L is the natural line width, E0 is the
energy position of the peak, Y0 is the height of the undistorted peak and ξ is the asymmetry
parameter determined by the collision kinematics. Approaching the ionization threshold, the
effect of PCI is increased (in our case ξ → ∞). Strictly speaking this formula is derived
for triple coincidence experiments where the collision kinematics is entirely fixed. For
Excitation of argon by satellite Auger spectra
407
non-coincidence experiments Sheinerman et al (1994) gave more accurate line shapes in
graphical form. Nevertheless, we used the above equation for diagram Auger lines obtained
in our single measurements and the ξ asymmetry parameter is considered as an adjustable
parameter. In the model spectra we used fixed triplet intensity ratios (Ridder et al 1976)
and the same values of 0L and ξ for all the 10 diagram Auger lines.
In the case of the satellite peaks we disregarded the effect of PCI on the line shapes for
the following reasons.
(i) At a given projectile energy the PCI effect is de facto smaller, because in the case of
inner-shell excitation only the projectile electron leaves the collision area having the total
excess energy (neglecting the very small recoil energy of the target atom). At inner-shell
ionization the excess energy is shared between the two outgoing (ejected and projectile)
electrons, therefore the PCI effects exists even at relatively high electron-projectile energy,
as it was found by Völkel et al (1988).
(ii) The fitting procedure is not very sensitive to the line shape for such small lines.
On the other hand, the Fano resonance can considerably distort the satellite line shapes.
The reason for this is that the inner-shell excited states are degenerate in energy with
the continuum states resulting from the simultaneous direct ionization plus the excitation
process:
−
ep− + Ar(1 S0 ) → Ar+∗ (3p−2 (1 S0 , 3 P0,1,2 , 1 D2 )n0 l) + e−
sc + e
and the amplitudes for the two processes are expected to interfere, since the final states
are indistinguishable. For normal Auger process the amplitude for the direct double outershell ionization is negligible, but here this is not the case. For the line shape of the Fano
resonance the Shore–Balashov parametrization is used (McDonald and Crowe 1992):
yiF = Y0 (1 + aε)/(1 + ε 2 )
and
ε = (E0 − Ei )/(0.50L )
which is added to the direct ionization continuum. The a parameter characterizes the
asymmetry of the resonance profile and depends on the probability of the direct process.
Considering that the fitting procedure is not very sensitive to the line shape of the weak
lines, instead we used the Lorentzian profile (a = 0) unless the satellite line is apparently
asymmetric. When fitting both the above-diagram Auger and satellite line profiles were
convoluted with a Gaussian spectrometer function.
P
The quality of the fit was estimated using the normalized χ 2 , χ 2 = (m − p)−1 ri2 ,
1/2
where ri = (Mi − Si )/σi (residual) and σi = Mi (standard deviation). Here the following
notations are used. The experimental spectrum contains m channels, the number of the
independently adjustable parameter is p, Mi and Si are the counts in the ith channel of the
measured and synthesized spectra, respectively.
5. Results and discussion
In figures 1(a), (b) to 5(a), (b) we show a complete set of ejected electron spectra measured
at 275, 300, 500, 1000, 1700 eV projectile electron energies. Each spectrum is the sum
of the four spectra measured by the four channeltrons positioned as 75◦ and 90◦ emission
angles at both sides. In this paper we do not present the spectra measured at smaller
emission angles. The reason for this is that we could not observe any significant difference
between the satellite structure of the spectra measured at different emission angles, except
that at smaller emission angles the continuous background is higher, so the satellite peak
parameters have generally larger statistical errors. In figures 3 to 5 the counts are displayed
on a square root scale to suppress the high peaks and expand the small satellites.
408
B Paripás et al
Figure 1. The energy spectrum of the ejected electrons measured at 275 eV projectile electron
energy, with the result of the best computer fit when, (a) the synthesized spectrum is built
from diagram lines only, (b) the satellites, marked by arrows and capital letters, are also added.
(Detailed specifications are given in table 3 and table 4 and in the text.)
Figure 2. The same as figure 1, but at 300 eV projectile electron energy.
Figure 3. The same as figure 1, but at 500 eV projectile electron energy.
In the figures the results of the best computer fits are also displayed, but the synthesized
spectra on the left-hand side (a) and the right-hand side (b) of the figures are built in a
different manner: on the left-hand side we built it up from the 10 diagram lines while on
Excitation of argon by satellite Auger spectra
409
Figure 4. The same as figure 1, but at 1000 eV projectile electron energy.
Figure 5. The same as figure 1, but at 1700 eV projectile electron energy.
the right-hand side we added some satellites, the presence of which became obvious on the
basis of the left-hand side residual spectra. Then the best values of the parameters (energy
and height) of the satellites were calculated in the course of a new fitting procedure. This
procedure normally results in more accurate diagram line parameters and a significantly
lower normalized χ 2 value. After this we can repeat the above procedure adding new
satellite(s) to the spectra. The scientific basis of this ‘satellite hunting’ is given by Cumpson
and Seah (1992). When the counting statistics is the only source of the difference between
the measured and the model (synthesized) spectra the χ 2 values are normally distributed
around one. In our case (m−p ≈ 250) it is very probably (> 99%) that above χ 2 ≈ 1.3–1.4
the counting statistics is not the only source of error. On the left-hand side figures χ 2 > 1.4,
so we can state that the systematic (and not the random) differences dominate between the
model and the measured spectra. Hence, we can improve our model via reducing χ 2 .
After adding the satellites to the Auger spectra (right-hand side of the figures) the χ 2 is
significantly reduced proving that we have really overcome an important contribution of the
systematic error. At high projectile energy (in figures 4 and 5) the χ 2 remains far above
the 1.4 value showing that our good counting statistics would allow us to continue the
improvement of the model.
At low-incident electron energy the high continuous background conceals the satellites
except for some very intense ones. Therefore, there are only three satellite lines which are
surely present in all the measured spectra, the parameters of which are shown in table 3.
410
B Paripás et al
Table 3. The parameters of the most-intense satellites which are present in each measured
spectra. The peak areas are given in thousandths of the 2p−1 (2 P3/2 ) → 3p−2 (1 D2 ) diagram line
area, and the energy scale is adjusted to 203.25 eV, the energy of this transition.
Notation
in
Energy
figures
(eV)
Identification
Relative peak area (∗ 10−3 ) at
different projectile energies
275 eV 300 eV 500 eV 1000 eV 1700 eV
A
206.3
2p−1 (2 P3/2 )3d → 3p−2 (1 D)4d, 2 P, 2 D, 2 F 54(9)
(and 2p−1 (2 P3/2 )4p → 3p−2 (1 S)4p 2 P
and 2p−1 (2 P1/2 )4p → 3p−2 (1 D)4p 2 D)
28(4)
5(2)
8(2)
7(4)
B
208.6
2p−1 (2 P3/2 )4p → 3p−2 (1 D)4p 2 P, 2 D, 2 F
60(10)
36(5)
8(2)
3(1)
4(1)
C
210.0
2p−1 (2 P3/2 )4s → 3p−2 (1 D)4s 2 D
41(11)
18(4)
8(2)
5(2)
5(1)
Two of them are identified on the basis of table 1 as the most probable transitions following
the 2p−1 (2 P3/2 )4s and 3d inner-shell excitation (A and C peaks on the figures). When
the Shore–Balashov parametrization is used for line shapes, the values of the asymmetry
parameter a did not differ significantly from zero, so finally these lines are considered
to be symmetrical peaks (Lorentzian convoluted with a Gaussian spectrometer function).
The third line corresponds to the most probable final state following the parity forbidden
2p−1 (2 P3/2 )4p excitation in table 2 (peak B). This peak—in contrast with the above two
peaks—is a definitely asymmetric peak where the asymmetry parameter a = −1.03 ± 0.08
does not depend on the primary energy. Here we cannot give any physical reason for this
behaviour and one cannot even exclude that this asymmetric peak is composed from smaller
symmetric peaks. In spite of the fact that the 2p−1 (2 P3/2 )4p excitation is a dipole forbidden
one, at electron impact the intensity of this B peak is increased (compared to the A and
C peaks, which follow allowed transitions) when approaching the ionization threshold (at
275 and 300 eV primary energy). This behaviour is qualitatively understood, because at
electron impact the dipole forbidden/allowed excitation probability ratio must depend on
the momentum transfer to the atom by the scattered electron. For forward scattering in
the 275–1700 eV primary electron energy range the momentum transfer is decreased by a
factor of 3.7. In accordance with Shaw et al (1982) it is equivalent to approximately one
magnitude decrease of the above ratio. This effect is smaller for other scattering angles and
in our experiment where the scattered electrons are not detected.
As we mentioned before, some of the above satellite peaks have been observed in former
works. For example at 3–5 eV primary electron energy Werme et al (1973) found the above
three lines (peaks 77, 79 and 80), while at the threshold Helenelund et al (1982) found the
first two lines, however neither of them identified these satellite lines. Weber (1990) found
and identified the 2p−1 (2 P3/2 )4p → 3p−2 (1 S, 1 D, 3 P)4p satellites.
We did not observe a peak at 209.6 eV, although in resonant Auger spectra the
2p−1 (2 P3/2 )3d → 3p−2 (1 D)4d 2 P, 2 D, 2 F transition at 206.3 eV and the 2p−1 (2 P3/2 )3d →
3p−2 (1 D)3d 2 P, 2 D transition at 209.6 eV have nearly the same intensity (table 1). A possible
explanation is that the peak at 206.3 eV is mainly the 2p−1 (2 P3/2 )4p → 3p−2 (1 S)4d 2 P and
the 2p−1 (2 P1/2 )4p → 3p−2 (1 D)4p 2 D transitions.
At higher projectile energy—in spite of the fact that the satellite peak intensities are
much smaller relative to the diagram line intensities—the presence of some other satellites
is obvious. This is—in addition to the better statistics—due to the reduced continuous
Excitation of argon by satellite Auger spectra
411
background of the scattered and secondary electrons. The latter feature comes from the
decreased double-differential cross section for ionization in the measured energy region and
from the low instrumental background of our triple-pass spectrometer. Several of the further
lines can be easily identified on the basis of table 1, e.g.:
(i) at 204–204.4 eV: 2p−1 (2 P3/2 )4d, 5d → 3p−2 (1 D)6d, 7d (F on the figures);
(ii) at 207.8 eV: 2p−1 (2 P1/2 )4s → 3p−2 (1 D)4s, 2 D (G) and
(iii) at 211.4–211.9 eV: 2p−1 (2 P3/2 )4s → 3p−2 (3 P)4s 2 P and 3p−2 (3 P)4s 4 P (H).
However, we have found at least two peaks below the 2p−1 (2 P3/2 ) → 3p−2 (1 D2 )
diagram line at 202.2 and 202.6 eV (D and E) which cannot be identified in this manner.
These peaks are obviously present in the spectra measured at any emission angle at 1700
and 1000 eV projectile energy. They are still recognizable in some spectra at 500 eV, but
they were not found in the spectra measured at 300 and 275 eV projectile energy. We would
like to stress that the consideration of the latter two peaks in the model spectra at high (1700
and 1000 eV) projectile energy is very important, their effect is comparable to that of the
three identified satellites in table 3. In coincidence experiment Lohmann et al (1992) found
this peak-like structure below the 2p−1 (2 P3/2 ) → 3p−2 (1 D2 ) diagram line but they tried to
explain it as an angular dependent PCI effect. The parameters of these additional satellites
(identified and unidentified) measured at 500, 1000 and 1700 eV projectile electron energies
are shown in table 4.
In the course of the computer fit we obtained the best values for the other adjustable
parameters. Among them the values of the ξ asymmetry parameter can have significance.
These are 1.35(±0.13), 0.82(±0.06), 0.32(±0.03), 0.24(±0.03) and 0.24(±0.02) at 275,
300, 500, 1000 and 1700 eV projectile electron energies, respectively. Although the ξ
asymmetry parameter is determined by the collision kinematics, we cannot compare these
experimental data as the detailed kinematics is not known (we do not detect the other two
outgoing electrons). Instead for comparison with other results we calculated the shift of the
maximum of the line (1E) using the 1E = (0L ξ )/2 formula (Kuchiev and Sheinerman
1987). These 1E shifts in the above order are: 81(±8), 49(±4), 19(±2), 14(±2) and
14(±1) meV, respectively.
Table 4. The parameters of the additional satellites which were found in the spectra measured at
500, 1000 and 1700 eV projectile electron energies. The peak areas are given in thousandths of
the 2p−1 (2 P3/2 ) → 3p−2 (1 D2 ) diagram line area and the energy scale is adjusted to 203.25 eV,
the energy of this transition.
Relative peak area (∗ 10−3 ) at
different projectile energies
Notation
in
figures
Energy
(eV)
Identification
500 eV
D
202.2
Not identified
—
E
202.6
Not identified
8(2)
F
204.0–204.4.
2p−1 (2 P3/2 )4d, 5d → 3p−2 (1 D)6d, 7d
—
G
207.8
2p−1 (2 P1/2 )4s → 3p−2 (1 D)4s 2 D
—
H
211.4–211.9
2p−1 (2 P3/2 )4s → 3p−2 (3 P)4s 2 P and
→ 3p−2 (3 P)4s 4 P
—
1000 eV
1700 eV
4(2)
9(1)
20(5)
17(3)
7(3)
13(4)
—
5(2)
4(2)
3(2)
412
B Paripás et al
6. Conclusions
We could significantly improve the quality of the computer fit to the measured ejected
electron spectra by means of adding certain satellite lines to the synthesized model spectra.
These satellites correspond to the most probable transitions following the 2p−1 (2 P3/2 )4s, 4p
and 3d inner-shell excitations. At electron impact the 2p−1 (2 P3/2 )4p excitation dominates
when approaching the ionization threshold.
At higher projectile energy, despite the decreasing intensity of the satellite peaks relative
to the diagram lines, we could identify more satellite lines. This is mainly due to the
decreased continuous background of the scattered and secondary electrons which was partly
resulted by the applied low-background triple-pass electron spectrometer.
Acknowledgments
This work has been supported by the Hungarian Scientific Research Foundation (OTKA,
grant no. T016636). The authors are indebted to L Sarkadi for his critical reading of the
manuscript.
References
Cumpson P J and Seah M P 1992 Surf. Int. Anal. 18 345
Cvejanović S, Bagley G W and Reddish T J 1994 J. Phys. B: At. Mol. Opt. Phys. 27 5661
de Gouw J A, van Eck J, Peters A C, van der Weg J and Heideman H G M 1995 J. Phys. B: At. Mol. Opt. Phys.
28 2127
Heimann P A et al 1987 J. Phys. B: At. Mol. Phys. 20 5005
Helenelund K, Hedman S, Asplund L, Gelius U and Siegbahn 1983 Phys. Scr. 27 245
King G C, Tronc M, Read F H and Bradford R C 1977 J. Phys. B: At. Mol. Phys. 10 2479
Kuchiev M Yu and Sheinerman S A 1987 Sov. Phys.–Tech. Phys. 32 879
Lohmann B, Xian-Kun Meng and Keane M 1992 J. Phys. B: At. Mol. Opt. Phys. 25 5223
McDonald D G and Crowe A 1992 J. Phys. B: At. Mol. Opt. Phys. 25 2129
Meyer M, Raven Ev, Sonntag B and Hansen J E 1991 Phys. Rev. A 43 177
Moore C E 1971 Atomic Energy Levels (Washington, DC: National Bureau of Standards)
Ridder D, Dieringer J and Stolterfoht N 1976 J. Phys. B: At. Mol. Phys. 9 L307
Shaw D A, King G C, Read F H and Cvejanović D 1982 J. Phys. B: At. Mol. Phys. 15 1785
Sheinerman S A, Kuhn W and Mehlhorn W 1994 J. Phys. B: At. Mol. Opt. Phys. 27 5681
Varga D, Kádár I, Ricz S, Végh J, Kövér A, Sulik B and Berényi D 1992 Nucl. Instrum. Methods A 313 163
Végh J 1995 Proc. European Conf. on the Applications of the Surface and Interface Analysis (Montreaux)
(ECASIA ’95) ed J Mathieu, B Reihl and D Briggs 689 p QA34
Völkel M, Schnetz M and Sandner W 1988 J. Phys. B: At. Mol. Opt. Phys. 21 4249
Weber W, Huster R, Kamm M and Mehlhorn W 1991 Z. Phys. D 22 419
Werme L O, Bergmark T and Siegbahn K 1973 Phys. Scr. 8 149