Positron-annihilation monitoring of reduction processes in conducting glasses

Article abstract of DOI:10.1016/j.jallcom.2004.06.001

Depth-resolved, Doppler-broadening positron spectroscopy has been used to study structural properties of semiconducting glasses. Bismuth- and lead-doped, silica and germanium glasses with five different compositions were studied. Conducting surface layers in these glasses were obtained by reduction in hydrogen atmosphere. Depth scanning with a slow-positron beam allows to follow the temperature and time evolution of layers modified during reduction processes. However, while reduction processes in inner layers (below a few tens of nm) are clearly seen in changes of the Doppler-broadened positron- annihilation line, the near-to-surface changes are more complex and not fully understood. We prove that the positron spectroscopy can help in optimising technological questions of reduction processes.

Full text of DOI:10.1016/j.jallcom.2004.06.001

Journal of Alloys and Compounds 382 (2004) 257–263
Positron-annihilation monitoring of reduction processes
ଝ
in conducting glasses
a
b
b
b
Damian Pliszka , Bogusław Kusz , Maria Gazda , Konrad Trzebiatowski ,
c,∗
c
c
Grzegorz P. Karwasz , Roberto S. Brusa , Antonio Zecca
a
Institute of Physics, Pedagogical University of Słupsk (IF PAP), Arciszewskiego 22B, Słupsk 76-200, Poland
b
Instytut Fizyki, Politechnika Gdanska, Narutowicza 11/12, Gdansk 80952, Poland
c
Dipartimento di Fisica and INFM, Universita’ di Trento, via Sommarive 14, Trento 38050, Italy
Received 19 September 2003; received in revised form 10 February 2004; accepted 16 February 2004
Abstract
Depth-resolved, Doppler-broadening positron spectroscopy has been used to study structural properties of semiconducting glasses. Bismuth-
and lead-doped, silica and germanium glasses with five different compositions were studied. Conducting surface layers in these glasses
were obtained by reduction in hydrogen atmosphere. Depth scanning with a slow-positron beam allows to follow the temperature and
time evolution of layers modified during reduction processes. However, while reduction processes in inner layers (below a few tens
of nm) are clearly seen in changes of the Doppler-broadened positron-annihilation line, the near-to-surface changes are more complex
and not fully understood. We prove that the positron spectroscopy can help in optimising technological questions of reduction
processes.
©
2004 Elsevier B.V. All rights reserved.
PACS: 73.61.Jc; 61.82.Rx; 68.37.Rt; 61.72.Qq
Keywords: Bismuth–silica glass; Bismuth–germanate glass; Lead–silica glass; Conducting glass; Reduction; Microscopy; Positron spectroscopies
1
. Introduction
The lead-doped glass is the material traditionally used
in electron channel multipliers. We have started exploiting
a new class of bismuth-doped glasses, which necessitate
thermal treatments at lower temperatures than lead-doped
glasses [2]. Bismuth–germanate glasses in comparison with
bismuth–silicate glasses may contain less bismuth so these
materials are complementary to each other.
We have observed, using the atomic force and optical mi-
croscopy, some nanometer precipitations of metal inside the
near-to-surface layer and accumulation of metal nano-drops
on the surface, after long treatment times, 2 h for silica
glasses and about 8 h for germanium glasses [2,3]. The metal
drops are Bi and some Ge in the case of bismuth–germanate
glasses and solely Bi in the case of silica glasses. The pur-
pose of this work is to see the early stages of annealing when
these reduced layers and drops on the surface start to appear,
before these layers can be identified by standard microscopy
techniques.
Bismuth- and lead-doped silicon and germanium-oxide
glasses are used, among other applications, for electron
continuous-dynode multipliers [1]. To make them con-
ducting, reduction procedures in hydrogen atmosphere are
applied. The depth of the reduced layer must be thoroughly
controlled, in order to ensure a high secondary-electrons
yield and a low surface conductivity (see Fig. 1). In
Fig. 1 the surface conductivity versus time of reduction
for samples of series “S” and “G” (see also Table 1) are
reported. Furthermore, not only the secondary-electron
emission coefficient varies with the reduction conditions,
but also the ageing properties of the electron multiplier
devices.
ଝ
We dedicate this article to the memory of Professor A. van Veen, who
has given important contributions to the development of positron physics.
∗
We have applied the positron-annihilation technique,
which is depth-resolved and non-destructive, and permits
Corresponding author. Fax: +39-0461-881696.
E-mail address: karwasz@science.unitn.it (G.P. Karwasz).
0
925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.06.001

2
58
D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
2. Experimental
Conducting lead glasses samples were produced from
SiO2 and PbO2 oxides. Bismuth glasses were prepared
from milled SiO2 or GeO2 powder and bismuth nitrate
4
7
BiNO3(OH)2·BiO(OH). The mixture was decomposed at
◦
00 C and then the obtained mixture was ground again
◦
and melted at 1200 C, see Ref. [2] for details. Five series
of conducting glasses were produced: one lead–silica, two
bismuth–silica and two bismuth–germanium glasses. De-
tails on stoichiometry and thermal treatment are given in
Table 1.
Thermal treatment in hydrogen atmosphere was done to
create surface reduced layers. The annealing temperature
◦
◦
was 300 C for bismuth–silica glasses of series I; 340 C for
bismuth–silica glasses of series II and bismuth–germanate
◦
Fig. 1. Changes in the surface conductivity for silica-based (“S” series)
and germanium-based (“G” series) bismuth-doped glasses (adapted from
Ref. [2]) during heat treatment at 340 C.
glasses of series III and IV, and 400 C for lead–silica glasses
of series V. During reduction both oxygen and water vapour
were desorbed from the samples.
◦
Positron-annihilation spectroscopy is a non-destructive
technique permitting to determine essentially the presence
of defects in solids. Defects, like vacancies and dislocations
constitute trapping centres for positrons injected into sam-
ples and cause the rise of positron lifetimes and changes in
the shape (broadening) of the annihilation gamma line, at
511 keV (see Ref. [5] for details about positron techniques).
In the present measurements the Doppler-broadening of the
annihilation line was studied and a variable-energy positron
beam described elsewhere, was used [6]. Positron implan-
tation energy E was changed from 50 eV to 25 keV. This,
to study defects at sub-nano-scale, i.e. single vacancies,
dislocations, and change in the structure of the materials.
Glass samples, similar to those used in our study [3] were
prepared and characterized at Gdañsk Technical University
while positron measurements were done at Trento Univer-
sity. Partial experimental results of positron investigation
have been already published [3,4] but the complete, com-
parative analysis for different glasses is performed only
now.
Table 1
Stoichiometry, thermal treatments, and positron-annihilation data for surface-conducting glasses
Series
I
Stoichiometry (density)
Symbol
Thermal treatment
Reduced layer
From–to (nm)
S (surface)
From–to (nm)
S (bulk)
Bi0.3Si0.7O1.85 (5.8 g/cm3)
A0
A1
A2
A3
As-received
–
0.516
0.5252
0.504
–
–
0.489
–
0.540
0.542
◦
300 C/3.3 h
0–20
0–25 (?)
0–8
◦
300 C/25 h
25–500
35–8
◦
300 C/91 h
0.54035
II
Bi0.57Si0.43O1.72 (6.46 g/cm3)
Bi0.3Ge0.7O1.85 (5.65 g/cm3)
S0
As-received
–
0–8
–
0.492
0.490505
–
–
–
–
0.469
–
0.535
0.540
◦
S1a
S1b
S1c
340 C/0.3 h
◦
340 C/0.5 h
0–210
0–600
◦
340 C/0.8 h
–
III
B0
B1
B2
B3
B4
As-received
0.495
0.484
0.5084
–
–
–
0.477
–
0.519
0.531
0.532
◦
340 C/0.4 h
0–12
0–6
–
◦
340 C/1.3 h
15–300
5–700
3–1000
◦
340 C/4 h
◦
340 C/5.5 h
–
–
IV
V
Bi0.33Ge0.67O1.84 (5.77 g/cm3)
Pb0.4Si0.6O2.0 (5.46 g/cm3)
G0
G1
G2
G3
As-received
0.493
0.484
–
–
–
0.471
–
0.507
0.525
◦
340 C/0.8 h
4–70
–
–
◦
340 C/2 h
15–120
15–800
◦
340 C/7 h
–
C0
C1
As-received
0.515
–
–
0.497
0.536
◦
400 C/21 h
–
1–700
Thickness of the reduced layers (in nm) and S-parameter are determined using VEPFIT program [11]. Surface and bulk S-parameters are given in sixth
and eighth columns, respectively.

D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
259
1
.6
according to a simplified formula [5]: d = (40/ρ)E (where
3. Results
3
ρ is the sample density in g/cm , d is expressed in nm and
E the energy in keV) gives the positron implantation range
down to 1.0 ␮m for the heaviest glass at 25 keV implanta-
tion energy.
Results of S-parameter measurements and VEPFIT anal-
ysis of the samples reported in Table 1 are presented
in Figs. 2–7. In Fig. 2 we present the S-parameter data
for non-reduced samples (for clarity we show only the
“smoothed”, VEPFIT results) and we compare with the
data from other laboratories for natural quartz [12] and for
A high-purity germanium detector with 1.3 keV resolu-
tion and 24% detection efficiency was used to monitor the
shape of the annihilation line. We have evaluated this energy
133
resolution on the basis of FWHM of the 356 keV Ba line
FWHM was 1.08 keV). The stability of the energy resolu-
(
tion of the detector and the stability of the positron beam
was controlled thoroughly by methods described elsewhere
[6–8].
The S-parameter, defined as the ratio between the num-
ber of counts in the central part of the broadened 511 keV
line and the total number of counts in this line is commonly
used to monitor the presence of defects in solids. In pres-
ence of open volume defects, this parameter rises as com-
pared to a defect-free material, due to the enhanced probabil-
ity of annihilation with valence (i.e. low-momentum) elec-
trons, see [5]. In the present measurements the S-parameter
is defined as the number of counts in the |511 keV − Eγ| ≤
0
.85 keV window compared to the total number of counts
in the |511 keV − Eγ| ≤ 4.25 keV peak. The W-parameter
defined as the ratio of positrons in the “wing” 1.6 keV ≤
|
511 keV − Eγ| ≤ 4.0 keV range to the overall area of the
Fig. 2. The S-parameter dependence for as-obtained glass samples, see
Table 1 for samples description. The fit on the experimental curves
obtained by the VEPFIT [11] programme. Quartz data were taken from
measurements of Fujinami et al. [12]; data for synthetic silica glass are
taken from measurements of Anwand et al. [13]. Brazilian quartz samples,
as measured by Anwand et al. [13] (with energy resolution for the S-curve
of 1.09 keV) show a similar dependence as the data of Ref. [12] but are
much lower (0.42 in bulk) — out of scale on this figure. The quartz and
silica data have been rescaled on abscissa scale, in order to account for
the difference in the sample density.
peak describes annihilation of positrons with high-energy
electrons (i.e. core). The S versus W curves permit to extract
partial information about chemical elements surrounding an
open volume defect.
The experimental uncertainty mainly depends on the
counts accumulated in the annihilation line. Several spec-
5
tra, up to typically 2 × 10 counts have been accumulated
per point. Lower total counts characterise points at positron
implantation energies below 1 keV, see Ref. [8]. As a
conservative number we give a 0.5% statistical error on
S-parameter values above 1 keV implantation energy and
somewhat higher uncertainty at lower energies.
Present spectra have not been deconvoluted for the energy
resolution of the detector: as the reference value we quote
the value of S = 0.523 from our laboratory for the bulk of
defect-free, p-type, floating-zone grown silicon, which we
used previously [9,10] for normalization of silicon-related
spectra. With this normalization, the S-parameter in bulk
quartz, as given in review [5] would amount to 0.465 and
for SiO2 layers thermally grown in dry deposition processes
it would be 0.51.
The S versus E curves have been analysed with the
variable-energy-positron fit (VEPFIT) [11] program that
models the diffusion and trapping of positrons after their
implantation and thermalization. The program permits
to determine the depth of up to six layers with relative
S-parameters and the positron diffusion length in each layer.
Additionally, it determines the S-parameter for the sam-
ple surface and the S-parameter for epithermal positrons
Fig. 3. Positron-annihilation S-parameter vs. positron implantation energy
for bismuth–silica glasses (series I, with Si molar contents higher than
Bi). Modelling of the S-parameter curves, has been performed using the
VEPFIT [11] program. Increase of S-parameter in the formed reduced
layer with rising reduction time is clearly seen.
(injected to the sub-surface layer).

2
60
D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
Fig. 6. The S-parameter for bismuth–germanium glasses (series IV, similar
to series III).
Fig. 4. Positron-annihilation S-parameter vs. positron implantation energy
for bismuth–silica glasses (series II, with majority of Bi). Modelling of the
S-parameter curves has been performed using the VEPFIT [11] program.
Si/Pb glass (sample “C0”) is probably a VEPFIT artefact —
experimental data are much scattered, due to a lower count-
ing rate in those measurements.) This behaviour resembles
the dependence for defect-free quartz [12], see Fig. 2. Also
for the synthetic quartz, the S-parameter falls monotonically
with the positron implantation energy, see Ref. [13]. From
Fig. 2 one notes that the surface S-value are almost the same,
silica glass [13]. Please, note that due to different energy
resolutions between our and other [12,13] experimental
set-ups, only general shapes of curves can be compared and
not their absolute values.
The values of S-parameter show rather big differences
between non-reduced and reduced glasses — reduction pro-
cesses generally rise the S-parameter, compare Fig. 2 with
Figs. 3–7. It seems that for the S-parameter, the changes
in the chemical composition of glass are less important
than reduction processes. For reduced glass samples, the
S-parameter shows non-monotonic dependencies, rising
with the depth for higher reduction times.
0.515 for glasses “A0” and “C0”, in which silicon is the ma-
jority component over bismuth or lead, respectively. On the
other hand, in sample “S0” containing more Bi than Si and
in samples Ge/Bi (“B0”, “G0”) the surface S-value is lower
(about 0.493). Note that both for quartz [12] and for the na-
tive oxide on Si from Trento laboratory [8] the S-values at
the surface are low, about 0.495.
The bulk S-parameter value for non-reduced samples does
not correlate with solely one of the metallic components:
the highest bulk value (0.497) has been measured for Si/Pb
The S-parameter for all as-received glasses falls mono-
tonically with the positron implantation energy. (A bump for
Fig. 5. The S-parameter for bismuth–germanium glasses (series III). Note
a complex change of the S-level with the reduction time for the first
Fig. 7. The S-parameter for lead–silicon glasses (series V). Modelling of
the S-parameter performed using the VEPFIT [11] program.
1
0 nm depth.

D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
261
glass, a lower (0.489) for Si/Bi glass “A0” and the value
of 0.47 for Ge/Bi. The S bulk value in Si and Ge glass
decreases as the content of Bi increases. What seems to be
dependent on the glass matrix type is the diffusion length,
which amounts to 15–20 nm for Si-based glasses (samples
the whole series only one reduced layer is seen. In sam-
ple “S1a” it resembles the thin layer in “A1” and extends
down to about 10 nm, with the S-parameter only slightly
higher (0.49) than in the non-reduced bulk. In samples
“S1b” and “S1c” — a single layer with a high S-parameter
(0.54), extending down to 200 and 600 nm, respectively, is
observed.
“
A0”, “S0” and, less evident, also “C0”) and to 35–40 nm
for Ge-based glasses.
Structural changes in reduced samples (of series II and
IV) were previously studied [2,3] by the atomic force mi-
croscopy, optical microscopy and X-ray diffraction meth-
ods. For example, after 44 h reduction time in Si/Bi sam-
ples (series II) rather big, about 10 ␮m, microcrystals of
rhombohedric Bi on the surface and a 50 ␮m thick layer
of Bi-granules embedded in the glass were observed. For
the same long reduction time, Ge/Bi samples show a thin,
Germanium-based bismuth glasses (Figs. 5 and 6) show a
somewhat more gradual rise of the S-parameter value in the
reduced inner layer than it was seen for “S1b–S1c” samples.
For example, both for “S1b” and “B2” samples the modified
inner layer extends down to 200–300 nm but for “B2” sample
the S-parameter is lower, about 0.52, than the “saturation”
value (0.532 for sample “B4”). Similarly, in samples “G”,
the S-parameter rises gradually in the (inner) reduced later,
from 0.47 for non-reduced bulk (“G0”) to 0.507 in sample
“G2” and 0.525 in sample “G3”.
The changes of S-parameter in the first 10 nm in Bi/Ge
glasses are quite complex. Reduction times shorter than 1/2 h
(samples “B1” and “G1”) cause a decrease of the S-value in
the first 10 nm, below the values for “as-obtained” glasses,
see Figs. 5 and 6. In sample “B2”, treated for 1.2 h we ob-
served a rise of the S-parameter in the first few nm. With
longer reduction times, when the inner modified layer is well
developed, the near-to-surface S-parameter is again low, be-
low the S-received value (samples “G2”, “G3”, “B3”, “B4”).
Somewhat similar changes can bee seen also for “A” Si/Bi
series, but for Ge/Bi glasses they are more evident. We re-
call the conductivity curve, Fig. 1, in which lowering of the
surface conductivity is observed for Ge/Bi glasses for 1–2 h
annealing time.
4
␮m layer of embedded granules and nano-drops of Bi with
some admixture of Ge metallic phase on the surface [2].
In positron measurements we observe, with long reduction
times (see Table 1), a general rise of the S-parameter, to
about 0.53–0.54, almost independently on the kind of glass.
Such high values have been measured for highly defected
SiOx non-stoichiometric layers, see [9] (we reported there
the S-parameter values normalized to silicon bulk — without
normalization the S-parameter would amount to 0.53–0.54,
depending on oxygen deficiency in SiOx). In the reduced
layers nano-precipitates of Bi in GeOx and SiOx matrices are
formed. In such a structure positrons are expected to mainly
annihilate in defects at the interfaces of nano-precipitates
and the open structures of SiOx [9] and GeOx, giving the
observed increase of the S-parameter.
Let us consider first Si/Bi glasses: the two series “A1–A3”
(
Fig. 3) and “S1a–S1c” (Fig. 4) show some essential dif-
VEPFIT analysis hardly gives any quantitative insight on
these near-to-surface changes of the S-parameters. Mod-
elled S-values depend on several input data, like epithermal
positron scattering length and the diffusion length, and
these vary with changes in the glass structure (and den-
sity). Qualitative analysis would indicate different type of
changes in inner and outer layers. In the inner layer a more
defected structure, resembling SiOx [10], while in the outer
— a more compact (S rises but remain lower than in SiOx),
metal-enriched composition is observed. For example, Then
and Pantano [1] observed a silicon rich zone in the outer
20–50 nm layer of the surface layer. Other techniques,
more element-specific than S-parameter measurements,
must be applied for understanding differences between
the two layers observed in our present and previous [2]
measurements.
Changes in lead glasses, series “C” are clear although
experimental errors are somewhat higher than in series
I–IV. The reduced layer in sample “C1” extends beyond the
positron-implantation range of our apparatus.
It is important to note that in spite of the fact that
S-parameters in non-reduced samples differ much between
single glasses (from 0.47 to about 0.50), in all reduced
glasses the S-parameter values are closer (0.53–0.54).
The S-parameters in reduced layers prove to be almost
ferences. Samples of series “A”, with 30% molar contents
of Bi (relative to Si), require very long (up to several days)
reduction times in order to obtain satisfactory high surface
conductivity and secondary-electron emission coefficients
[3]. From S-curves one notes that 3 h reduction time (sample
“
A1”) do change this glass, apart from the very superfi-
cial, 20 nm deep layer — a slight rise of the S-parameter
is observed, see Table 1. In sample “A2”, heated for 25 h,
the near-to-surface layer seems to disappear if compared to
“
A1” — a modified layer with a high value of S-parameter
(0.54) extends from 25 to 500 nm depth. However, in order
to reproduce the S-curve in sample “A2” we had to assume
two layers and a non-modified bulk. The near-to-surface
layer extends down to 25 nm and shows the S-value (0.50)
slightly higher than that in non-reduced bulk (0.49); the
positron diffusion length in this layer is about 5 nm, much
shorter than in the bulk. In sample “A3” (heated for almost
4
days) both, the near-to-surface (first few nm) and the
deeper (below 35 nm) layers have the same, high (0.54)
S-parameter indicating a new structure with more open
volumes.
In samples “S1a–S1c”, silica-based with a higher con-
tents of bismuth than series in “A”, we do not observe
such a complex reduction dynamics as in series “A”. For

2
62
D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
defects, then the W/S plot shows some separated straight
segments [14].
In Fig. 8 we compare two types of glasses, Si/Pb and
Si/Bi. The points for these two glasses form two distinct
triangles (with vertices signed by circles): annihilation sites
are not the same in the two types of glasses. This confirms
the conclusions from “as-received” curves, Fig. 2 — the
annihilation sites are glass-specific and not related to one of
the components, Si or Pb (Bi). For the non-reduced sample
“
S0” only one surface and one bulk annihilation-sites are
present. For samples “S1a–S1c” another type of annihilation
sites appear.
For Ge/Bi glasses, see Fig. 9, the spread of points is
somewhat large but essentially full points (“G” series”) and
open points (“B”) series occupy the same region of the
S–W plane: annihilation sites are the same in both glasses.
However, the univocal identification of the atoms near the
positron-trapping sites would require investigation with
positron-coincidence techniques [15].
Fig. 8. Positron-annihilation parameters W vs. S plot for selected
bismuth–silicon and lead–silicon glasses (series II and V). Experimental
S/W points lay on two different triangles for the two glass series, with
vertices signed by circles. In “as-received” “S0” glasses the (open) points
lay on a segment between the upper (open) circle, corresponding to a
surface-annihilation site and the middle (open) circle corresponding to a
non-modified bulk. The lower vertex of the open-circle triangle corre-
sponds to annihilation sites in the modified glass. Similar dependencies
are observed of Si/Pb glass (full points and heavy-line triangle).
4. Conclusions
Present analysis of positron measurements prove to be
complementary to other methods used in characterisation
of conducting glasses. The positron beam method allows to
follow the growth of the modified layer with the reduction
time, in particular at early stages of the process. For example,
silica glass samples “A3” (Bi) and “C1” (Pb) show that the
independent on the kind of glass (0.54 with experimental
error for silica glasses, 0.53 for germanate glasses).
To get some more insight on the nature of positron-trapping
sites, in Figs. 8 and 9 we show plots of the W-parameter
versus the S-parameter for four series of samples. The W/S
plots allow to determine how many types of annihilation
centres co-exist in the glass. If the annihilation takes place
only in the surface-related and in the bulk defect-free sites,
all W/S points lay on one segment. If several kinds of anni-
hilation centres co-exist in the material, surface, bulk and
◦
◦
used reduction times (91 h at 300 C and 21 h at 400 C,
respectively) caused “over reduction” of the samples: the
reduced layer thickness extends over 1000 nm (see Figs. 3
and 7). Samples of series II–IV, annealed at slightly higher
◦
temperature (340 C) have thinner reduced layers, so it is
possible to observe them for the entire section (Figs. 4–6).
We observe also some changes in near-to-surface layers,
like for samples “A2”,“S1a”, “B1”, “G1”. However, positron
measurements are not able to show the presence of any
metallic layers on the surface and indicate only a complex
dynamics of the reduction processes in the near-to-surface
layers.
The conclusion that the bulk values of S-parameter de-
pends on the whole composition of the glass and does not
correlate just to one metallic component is important. It in-
dicates that positrons annihilate in sites surrounded by sev-
eral different atoms — the glass seems to be homogeneous
at a nano-scale.
Essential difference between silica-based and germanium-
based bismuth glasses is that in Ge/Bi glasses more gradual
changes of the S-parameter in modified layers are observed
than in Si/Bi ones. Also, for similar treatments, like for
samples “S1c” and “G3”, the reduced layer remains some-
what thinner for germanate glasses. Therefore, Ge/Bi types
of glasses seem to be more appropriate for technological
applications than Si/Bi glasses.
Fig. 9. Positron annihilation W vs. S plot for selected bismuth–germanium
glasses (series III and IV). These data show a bigger scatter than the data
from Fig. 8, but all of them lay in the same region of the S/W plane,
indicating a similar nature of trapping centres in both samples.

D. Pliszka et al. / Journal of Alloys and Compounds 382 (2004) 257–263
263
References
[9] R.S. Brusa, G.P. Karwasz, N. Tiengo, A. Zecca, F. Corni, R. Tonini,
G. Ottaviani, Phys. Rev. B 61 (2000) 10154.
[
10] R.S. Brusa, G.P. Karwasz, G. Mariotto, A. Zecca, R. Ferragut, P.
Folegati, A. Dupasquier, G. Ottaviani, R. Tonini, J. Appl. Phys. 94
[
[
1] A.M. Then, C.G. Pantano, J. Non-Cryst. Solids 120 (1990) 178.
2] B. Kusz, K. Trzebiatowski, M. Gazda, L. Murawski, J. Non-Cryst.
Solids 328 (2003) 137.
(2003) 7483.
[
11] A. van Veen, H. Schut, J. de Vries, R.A. Hakvoort, M.R. Ijpma, in:
P.J. Schultz, G.R. Massoumi, P.J. Simpson (Eds.), Positron Beams for
Solids and Surfaces, AIP Conference Proceedings, vol. 218, London,
Ontario, 1990, p. 171.
[
[
[
[
[
[
3] D. Pliszka, M. Gazda, B. Kusz, K. Trzebiatowski, G.P. Karwasz, W.
Deng, R.S. Brusa, A. Zecca, Acta Phys. Polon. 99 (2001) 465.
4] B. Kusz, D. Pliszka, M. Gazda, R.S. Brusa, K. Trzebiatowski, G.P.
Karwasz, A. Zecca, L. Murawski, J. Appl. Phys. 94 (2003) 7270.
5] P. Asoka-Kumar, K.G. Lynn, D.O. Welch, J. Appl. Phys. 76 (1994)
935.
6] R.S. Brusa, G.P. Karwasz, M. Bettonte, A. Zecca, Appl. Surf. Sci.
16 (1997) 59.
7] A. Zecca, R.S. Brusa, M. Duarte-Naia, Measur. Sci. Technol. 5
1994) 61.
[
[
[
[
12] M. Fujinami, N.B. Chilton, K. Ishii, Y. Ohki, J. Appl. Phys. 74
(1993) 4540.
4
13] W. Anwand, G. Brauer, M. Hasegawa, O. Dersch, F. Rauch, Acta
Phys. Polon. A 99 (2001) 321.
14] M. Clement, J.M.M. de Nijs, P. Balk, H. Shut, A. Van Veen, J. Appl.
Phys. 79 (1996) 9029.
15] W. Deng, D. Pliszka, R.S. Brusa, G.P. Karwasz, A. Zecca, Acta
Phys. Polon. A 101 (2002) 875.
1
(
8] A. Zecca, M. Bettonte, J. Paridaens, G.P. Karwasz, R.S. Brusa,
Measur. Sci. Technol. 9 (1998) 1.