Spectrochimica Acta Part A 57 (2001) 2001– 2008
www.elsevier.com/locate/saa
Spectroscopic properties of Er3 + , Yb3 + and Er3 + /Yb3 +
doped metaphosphate glasses
A. Speghini a,*, R. Francini b, A. Martinez b,1, M. Tavernese b, M. Bettinelli a
a
Dipartimento Scientifico e Tecnologico, Uni6ersità di Verona, Ca’ Vignal, Strada Le Grazie 15, I-37134 Verona, Italy
b
Dipartimento di Fisica and Istituto Nazionale di Fisica della Materia, Uni6ersità di Roma Tor Vergata,
Via della Ricerca Scientifica 1, I-00133 Rome, Italy
Received 14 September 2000; accepted 11 December 2000
Abstract
The absorption and emission spectroscopies of Er3 + doped and Er3 + /Yb3 + codoped Ca(PO3)2, Sr(PO3)2 and
Ba(PO3)2 glasses have been studied. From the Judd– Ofelt intensity parameters, the spontaneous emission probabilities of some relevant transitions and the radiative lifetimes of several excited states of Er3 + have been calculated. The
decay curves of the Er3 + emission at 1.5 mm have been measured at different temperatures. The data have been fitted
using a stretched exponential function and the obtained experimental lifetimes have been compared with the
calculated radiative lifetimes. The difference between the experimental and calculated lifetimes is attributed to the
presence of traces of OH groups in the host glasses. The absolute OH content in some glasses has been determined
from the infrared spectra. The emission spectra at 1.5 mm of the Er3 + ion in the codoped glasses have been measured
at different temperatures. The integrated emission intensities decrease significantly on passing from room temperature
to 13 K, suggesting a temperature dependence of the rate of the energy transfer process between Yb3 + and Er3 + .
© 2001 Elsevier Science B.V. All rights reserved.
Keywords: Lanthanide ions; Phosphate glasses; Luminescence; Energy transfer; Decay curves
1. Introduction
Erbium doped glasses have been widely employed for several optical applications, especially
Dedicated to Professor Harald P. Fritzer on the occasion
of his retirement.
* Corresponding author. Tel.: + 39-045-8027900; fax: +39045-8027929.
E-mail address: [email protected] (A. Speghini).
1
Present address: Centro de Investigaciones en Optica, A.
C., Apartado Postal 1-948, 37150 León, Guanajuato, Mexico.
in the field of optical amplifiers for fibre communications [1] and eye-safe lasers [2,3]. Application
of these glasses requires the evaluation of spectroscopic properties, such as the absorption and
luminescence spectra and the decay times of the
excited states [4]. Among the oxide glasses that
have been investigated for these purposes, particular attention has been devoted to the metaphosphate hosts, which are characterised by
favourable chemical and physical properties [5].
The optical spectra and transition probabilities of
several Er3 + doped metaphosphate glasses have
1386-1425/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S 1 3 8 6 - 1 4 2 5 ( 0 1 ) 0 0 4 7 3 - 5
2002
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
been reported by some of us in previous papers
[6–8], and these materials have shown interesting
spectroscopic properties. Particularly, stimulated
emission cross sections for the 4I13/2 “ 4I15/2 transition higher than for some commercially available
laser glasses have been evaluated for some compositions [8]. The performances of these materials
can be enhanced by codoping with Yb3 + which
can be pumped by a diode laser around 980 nm in
the 2F5/2 state which can subsequently transfer the
excitation to the 4I11/2 state of Er3 + , giving rise to
4
I13/2 “ 4I15/2 sensitised emission [9– 12]. In this
paper, we extend the previous investigations carried out for metaphosphate hosts by considering
Er3 + doped and Er3 + /Yb3 + codoped Ca(PO3)2,
Sr(PO3)2 and Ba(PO3)2 glasses. We present the
absorption and luminescence spectra, and decay
curves of emission at 1.5 mm from the 4I13/2 state
of Er3 + , obtained at temperatures ranging from
293 to 10 K. The Yb3 + “Er3 + energy transfer is
evidenced and discussed.
2. Experimental
Lanthanide doped metaphosphate glasses
with the molar composition 0.99M(PO3)2 –
0.01Ln(PO3)3 (M =Ca, Sr, Ba, Ln=Er, Yb) and
(0.99−x)M(PO3)2 – 0.01Er(PO3)3 – xYb(PO3)3
(x =0.01, 0.05, 0.10) were prepared by mixing
appropriate quantities of reagent grade metal carbonate, ammonium dihydrogen phosphate and
lanthanide oxide. The samples were melted at
1250°C for 2 h in alumina crucibles. Each melt
was then cast in a brass mould and annealed for
12 h at 450°C. The samples were cut to a thickness of about 0.5 cm and carefully polished for
optical measurements.
Absorption and emission spectra and decay
curves were measured at various temperatures
using the equipment described in Ref. [10].
Additional room temperature absorption spectra in the infrared (IR) region were measured
using a Nicolet Magna 760 FTIR spectrometer.
3. Results and discussion
The room temperature absorption spectra of
Er3 + and Yb3 + doped and codoped metaphos-
phate glasses show inhomogeneously broadened
f“ f transitions. As an example, the absorption spectra at room temperature of the
0.99Ba(PO3)2 –0.01Er(PO3)3 and 0.99Ba(PO3)2 –
0.01Yb(PO3)3 glasses are shown in Figs. 1 and 2,
respectively.
The experimental oscillator strengths for the
Er3 + doped glasses were fitted on the basis of the
Judd–Ofelt parametrization scheme [7,8]. The results have been previously reported in Ref. [8] and
are again tabulated in Table 1. The values of the
dk parameters are in the range usually observed
for oxide glasses [13] and are very similar to the
values found for alkaline and alkaline-earth aluminium phosphate glasses [14]. The dk parameters allow the calculation of the total spontaneous
emission probabilities A, branching ratios i and
radiative lifetimes ~rad of the most important excited states of the Er3 + ion [7,8]. The results of
these calculations for the 0.99Ba(PO3)2 –
0.01Er(PO3)3 glass are given in Table 2. The calculated radiative lifetimes of the excited states
emitting in the IR are in the range obtained for
Er3 + in phosphate glasses containing alkalineearth ions [15].
From the absorption spectra of the Yb3 +
doped glasses, the experimental oscillator
strengths P of the 2F5/2 ’ 2F7/2 transition were
obtained for the three alkaline-earth host glasses.
The experimental oscillator strengths are higher
than for a silicate glass [4]. The spontaneous
emission probabilities A for the 2F5/2 “ 2F7/2 transition, calculated following the procedure outlined
in Ref. [16], are given in Table 3. The contribution
of the magnetic dipole transition to the oscillator
strength in the 2F7/2 “ 2F5/2 transition of Yb3 + is
Table 1
Judd–Ofelt parameters for the Er3+ ion in different hosts
Host
d2 (pm2)
V4 (pm2)
V6 (pm2)
r.m.s.
(10−7)
Ca(PO3)2
Sr(PO3)2
Ba(PO3)2
5.48
5.04
5.02
1.29
1.29
1.27
0.76
0.93
0.92
2.75
1.12
2.33
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
2003
Fig. 1. Room temperature absorption spectrum of the 0.99Ba(PO3)2 – 0.01Er(PO3)3 glass.
less than 10% and it was neglected in the calculation of the spontaneous emission probabilities A,
in agreement with Ref. [14]. The A values are in
the range of the values found for other phosphate
glasses and are higher than for alkaline silicate
glasses [14].
Emission spectra of the Er3 + doped metaphosphate glasses were measured at different tempera-
Fig. 2. Room temperature absorption spectrum of the 0.99Ba(PO3)2 – 0.01Yb(PO3) glass.
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
2004
Table 2
Calculated total spontaneous emission probabilities A, branching ratios i and radiative lifetimes ~rad for the main emitting
states of Er3+ in the 0.99Ba(PO3)2–0.01Er(PO3)3 glass
Initial state
Final state
A (s−1)
i
(2H11/2,4S3/2)a
4
F9/2
I9/2
4
I11/2
4
I13/2
4
I15/2
1.7
48.6
34.7
332.2
1114.3
0
0.03
0.02
0.22
0.73
4
6.5
59.2
56.5
1032.6
0.01
0.05
0.05
0.89
0.866
I11/2
I13/2
4
I15/2
1.9
35.2
102.5
0.01
0.25
0.74
7.164
4
23.0
112.7
0.17
0.83
7.374
117.8
1
8.492
4
4
F9/2
I9/2
I11/2
4
I13/2
4
I15/2
4
4
4
I9/2
4
4
I11/2
I13/2
I15/2
4
4
4
I13/2
I15/2
~rad (ms)
0.653
ature and at 12 K are shown in Fig. 4. All the
measured decay curves are slightly non-exponential. This behaviour was already found in the case
of Eu3 + doped germanate glasses [17] and is
attributed to the random distribution of the impurity ions in the glass [18]. The emission decay
curves were fitted employing a stretched exponential function of the form [19,20]:
n
I(t)= I(0) exp −
t
~
h
(1)
where I(t) is the emission intensity after the
pulsed excitation, ~ the lifetime of the excited
state, I(0) a constant, and 0B hB1. The results
of the fit are given in Table 4. The values obtained
for h range from 0.77 to 0.90. The experimental
lifetimes obtained with this procedure are similar
to the values found for Er3 + in silicate glasses
(5–6 ms) [21] and smaller than those found in
a
Thermalisation of these two states was taken into account,
assuming an energy difference of 826 cm−1 [7].
Table 3
Refractive index n, barycenter |, oscillator strength P and
spontaneous emission probability A for the 2F5/2 “ 2F7/2 transition of Yb3+ for the 0.99M(PO3)2–0.01Yb(PO3)3 glasses
M
n
| (cm−1)
P (10−6)
A (s−1)
Ca
Sr
Ba
1.548
1.558
1.585
10460
10477
10442
4.08
4.36
3.99
713
774
729
tures ranging from room temperature to 15 K.
The 980 nm excitation line of the diode laser
populates the 4I11/2 state, which in turn non-radiatively relaxes to 4I13/2. From this state only the
emission transition to the 4I13/2 ground state is
possible. The room temperature emission spectra
for the metaphosphate glasses are shown in Fig.
3.
Decay curves of the Er3 + doped metaphosphate glasses at 1.5 mm were measured at a pulsed
excitation of 980 nm at room and low temperatures. As an example, the decay curves for the
0.99Ba(PO3)2 – 0.01Er(PO3)3 glass at room temper-
Fig. 3. Room temperature emission spectra in the near IR of:
(a) the 0.99Ca(PO3)2 – 0.01Er(PO3)3; (b) 0.99Sr(PO3)2 –
0.01Er(PO3)3; and (c) 0.99Ba(PO3)2 – 0.01Er(PO3)3 glasses.
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
LiLa(PO3)4 glasses (9.5– 9.3 ms) [9]. For all the
Er3 + doped phosphate glasses, the experimental
lifetimes increase on decreasing the temperature.
2005
Table 5
Calculated absorption coefficient hOH at 3000 cm−1 and absolute content of the hydroxyl groups for the 0.99M(PO3)2–
0.01Yb(PO3)3 metaphosphate glasses
M
hOH (cm−1)
[OH] (ppm)
Ca
Sr
Ba
3.37
3.19
3.34
101
96
100
Moreover, the calculated radiative lifetimes are
considerably longer than the experimental ones
(see Table 3). This behaviour cannot be simply
explained on the basis of a multiphonon relaxation by taking into account the highest phonon
energy of the host glasses, relative to the phosphate group stretching vibrations. In fact, using a
simple energy gap model [22] and the parameters
suitable for phosphate glasses [23], the multiphonon relaxation rate is predicted to be negligible for the 4I13/2 “ 4I15/2 relaxation. It is well
known that the trace impurities of OH groups are
effective quenchers of IR emission [9,22,24]. In
phosphate glasses, the OH groups have a broad
and intense band at about 3000 cm − 1 [25,26]. The
absorption coefficient hOH, defined as
log
hOH =
Fig. 4. Decay curves of the 4I13/2 “ 4I15/2 transition of Er3 + for
the 0.99Ba(PO3)2 –0.01Er(PO3)3 glass at: (a) room temperature; and (b) at 12 K after pulsed excitation at 980 nm. Solid
line: stretched exponential model.
Table 4
Calculated radiative lifetimes ~rad and experimental time constants ~ of the 4I13/2 state obtained by ‘stretched exponential’
[18–20] fitting of the fluorescence decay curves at 1.5 mm on
980 nm pulsed excitation of the 0.99M(PO3)2–0.01Er(PO3)3
(M=Ca, Sr and Ba) metaphosphate glasses
M
~rad (ms)
T (K)
~ (ms)
Ca
10.1
300
50
3.9
4.2
Sr
8.7 [7]
300
12
3.2
3.4
Ba
8.5
300
11
4.5
6.2
T0
TD
d
where T0 is the transmittance at 1400 nm, TD the
transmittance at 3000 cm − 1 and d the thickness of
the glass sample, can be used to calculate the
absolute OH content according to the following
equation [27]:
OH content (ppm) = 30 hOH
The OH contents in some phosphate glasses under
discussion are given in Table 5. The obtained
values are in the range reported for other alkalineearth phosphate glass systems prepared in air
(hOH 3–6 cm − 1) [26,28]. The differences between the calculated and the experimental lifetimes could be explained by the presence of OH
groups, causing a decrease of the lifetimes with
respect to the radiative ones. This explanation is
compatible with the increase of the lifetimes on
2006
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
passing from room temperature to low temperature (see Table 3). A similar behaviour was also
observed for silicate glasses [21].
The emission spectra of the Er3 + ion in the
codoped Er3 + /Yb3 + barium metaphosphate
glasses were measured in the near IR at different
temperatures ranging from room temperature to
15 K, pulsed excitation at 980 nm, where both
Yb3 + and Er3 + absorb. As an example, we show
the emission spectra for the 0.89Ba(PO3)2 –
0.01Er(PO3)3 – 0.10Yb(PO3)3 glass at different
temperatures in Fig. 5. Inspection of the figure
shows that the integrated emission intensity decreases on passing from room temperature to 15
K. A very similar behaviour is also observed for
the other codoped barium metaphosphate glasses
under investigation. The temperature dependence
of the integrated emission intensity of the 4I13/2 “
4
I15/2
transition
for
the
0.89Ba(PO3)2 –
0.01Er(PO3)3 – 0.10Yb(PO3)3 glass is shown in Fig.
6. Although in the 0.99Ba(PO3)2 – 0.01Er(PO3)3
glass the integrated intensity of the 4I13/2 “ 4I15/2
transition decreases by about 22% from 295 to 15
K, for the codoped 0.89Ba(PO3)2 – 0.01Er(PO3)3 –
0.10Yb(PO3)3 glass the decrease in the same tem-
perature range is more pronounced (about 43%).
This behaviour is observed for all the three
codoped barium metaphosphate glasses and is
most probably because of the temperature dependence of the probability of the energy transfer
process
2
F5/2(Yb3 + )+ 4I15/2(Er3 + )
“ 2F7/2(Yb3 + )+ 4I11/2(Er3 + )
This is presumably because of a mismatch between the energy levels of the Yb3 + donor and
the Er3 + acceptor. An accurate description on the
behaviour of energy transfer needs the knowledge
of the spectral overlap between the 2F5/2 “ 2F7/2
transition of Yb3 + and the 4I11/2 ’ 4I15/2 transition
of Er3 + and will be the subject of a future investigation.
The luminescence decay curves of the 4I13/2 state
in the 0.98Ba(PO3)2 –0.01Er(PO3)3 –0.01Yb(PO3)3
glass at 1.5 mm were measured at room temperature and at 12 K after pulsed excitation at 980
nm. The curves were fitted reasonably well with a
function of the form rise-decay. The initial rise of
the luminescence denotes again the presence of an
energy transfer from the Yb3 + to the Er3 + ions.
Fig. 5. Emission spectra in the near IR for the 0.89Ba(PO3)2 – 0.01Er(PO3)3 – 0.10Yb(PO3)3 glass measured at 295 K (solid), 200 K
(dashed), 100 K (dotted), 15 K (dash –dotted).
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
2007
Fig. 6. Temperature dependence of the integrated emission of the 4I13/2 “ 4I15/2 band for the 0.89Ba(PO3)2 – 0.01Er(PO3)3 –
0.10Yb(PO3)3 glass. The maximum value of the integrated intensity has been arbitrarily set to 100.
The temperature dependence of the integrated
emission of the 4I13/2 “ 4I15/2band of Er3 + in the
0.89Ba(PO3)2 – 0.01Er(PO3)3 – 0.10Yb(PO3)3 glass
(see Fig. 6) strongly suggests that the rate of the
energy transfer from Yb3 + to Er3 + ions should
decrease and therefore the risetime should increase with the decreasing temperature. The values of the risetimes obtained from the fit at room
temperature and at 12 K are 0.27 and 0.30 ms,
respectively. The behaviour of these values as a
function of the temperature is qualitatively compatible with the above-mentioned temperature dependence of the integrated emission band at 1.5
mm of Er3 + . We note that the values of the
obtained risetimes are slightly longer than those
found for some Er3 + /Yb3 + codoped LiLa(PO3)4
glasses (0.040– 0.175 ms) [9], suggesting that presumably the energy transfer is less efficient in the
present glass. The values of the lifetimes of the
4
I13/2 state are 5.63 and 6.84 ms at room temperature and at 12 K, respectively. They are similar to
the lifetimes obtained for the glasses doped only
with the Er3 + ion (see Table 3) suggesting that
the presence of OH groups is not negligible in the
codoped glass also (see above).
4. Conclusions
In this paper we have reported a preliminary
analysis of the absorption and emission spectroscopy of Er3 + or Yb3 + doped and Er3 + /Yb3 +
codoped Ca(PO3)2, Sr(PO3)2 and Ba(PO3)2 host
glasses. The Judd–Ofelt intensity parameters for
Er3 + in the different hosts, calculated from the
absorption spectra, are in the range observed for
other oxide glasses. The values of the spontaneous
emission probabilities for the transition 2F5/2 “
2
F7/2of Yb3 + are smaller than for borate glasses,
similar to alkaline phosphate glasses and higher
than alkaline silicate glasses. The decay curves of
the Er3 + emission at 1.5 mm of the Er3 + doped
glasses could be fitted reasonably well with a
stretched exponential function, indicating the
presence of a significant disorder at dopant sites.
The experimental lifetimes are considerably lower
with respect to the calculated radiative lifetimes.
This behaviour is probably not because of the
intrinsic multiphonon relaxation and can be accounted for by the presence of traces of OH
groups in the glass hosts, in agreement with the
observation made for other phosphate and silicate
2008
A. Speghini et al. / Spectrochimica Acta Part A 57 (2001) 2001–2008
glasses. Following excitation in the Yb3 + absorption profile, a significant decrease in the emission
intensity of the 4I13/2 “ 4I15/2 transition of Er3 + on
decreasing the temperature is observed for Er3 + /
Yb3 + codoped glasses. This behaviour suggests
that the energy transfer from Yb3 + to Er3 + is
temperature dependent, probably because of the
mismatch between the energy levels of Er3 + and
Yb3 + . The temperature dependence of the energy
transfer probability could be due to a possible
mismatch between the emission transitions of the
donor and the absorption transitions of the acceptor when upper crystal fields of 2F5/2 and 4I15/2 are
thermally populated, or to assistance of phonons.
Further work to obtain a better understanding on
the energy transfer between the Yb3 + and Er3 +
ions is in progress.
Acknowledgements
The authors gratefully thank Erica Viviani
(Università di Verona, Italy) for her expert technical assistance.
References
[1] W.J. Miniscalco, J. Lightwave Technol. 9 (1991) 234.
[2] D. Vivien, Ann. Chim-Sci. Mater. 20 (1995) 211.
[3] A.-F. Obaton, C. Labbé, P. Le Boulanger, B. Elouadi, G.
Boulon, Spectrochim. Acta A 55 (1999) 263.
[4] G. Fuxi, Optical and Spectroscopic Properties of Glass,
Springer/Scientific and Technical Publishers, Berlin/
Shanghai, 1991.
[5] J.E. Marion, M.J. Weber, Eur. J. Solid State Inorg. Chem.
28 (1991) 271.
[6] G. Ingletto, M. Bettinelli, L. Di Sipio, F. Negrisolo, C.
Aschieri, Inorg. Chim. Acta 188 (1991) 201.
[7] S. Zaccaria, M. Casarin, A. Speghini, D. Ajò, M. Bettinelli, Spectrochim. Acta A 55 (1999) 171.
[8] A. Speghini, M. Bettinelli, D. Dal Dosso, S. Zaccaria, D.
Ajò, in: P. Vincenzini, G.C. Righini (Eds.), Ninth
CIMTEC — World Forum on New Materials, Symposium X — Innovative Light Emitting Materials, Techna
Srl, Faenza, Italy, 1999, p. 181.
[9] A.-F. Obaton, J. Bernard, C. Parent, G. Le Flem, C.
Labbé, P. Le Boulanger, et al., J. Appl. Phys. 4 (1998) 315.
[10] R. Francini, F. Giovenale, U.M. Grassano, P. Laporta, S.
Taccheo, Opt. Mater. 13 (2000) 417.
[11] R. Francini, F. Giovenale, U.M. Grassano, Radiat. Eff.
Defects Solids 149 (1999) 137.
[12] G.C. Righini, M. Bettinelli, M. Brenci, C. Duverger, M.
Ferrari, M. Fossi, et al., in: A.J. Glass, J.W. Goodman, M.
Chang, A.H. Guenther, T. Asakura (Eds.), SPIE Proceedings, vol. 3749, SPIE, Bellingham, Washington, USA,
1999, p. 755.
[13] C. Görller-Walrand, K. Binnemans, in: K.A. Gschneider
Jr., L. Eyring (Eds.), Handbook on the Physics and
Chemistry of Rare Earths, vol. 25, Elsevier, Amsterdam,
1998, p. 101.
[14] H. Takebe, T. Murata, K. Morinaga, J. Am. Ceram. Soc.
79 (1996) 681.
[15] H. Takebe, Y. Nageno, K. Morinaga, J. Am. Ceram. Soc.
77 (1994) 2132.
[16] W.T. Carnall, H. Crosswhite, H.M. Crosswhite, Energy
level structure and transition probabilities of the trivalent
lanthanides in LaF3, Argonne National Laboratory Report, 1977.
[17] R. Rolli, G. Samoggia, A. Speghini, M. Bettinelli, M.
Montagna, M. Ferrari, in: J. McKittrick, A. Kitai, C.
Ronda, B. DiBartolo (Eds.), Luminescent Materials. In:
Material Research Society Symposium Proceedings, vol.
560, MRS, Pittsburgh, 1999, p. 169.
[18] D.L. Huber, Mol. Cryst. Liq. Cryst. 291 (1996) 17.
[19] D.L. Huber, Phys. Rev. B 31 (1985) 6070.
[20] D.L. Huber, Phys. Rev. E 53 (1996) 6544.
[21] M. Bettinelli, M. Brenci, R. Dall’Igna, M. Ferrari, G.
Mazzoleni, G.C. Righini, et al., in: P. Vincenzini, G.C.
Righini (Eds.), Ninth CIMTEC — World Forum on New
Materials, Symposium X — Innovative Light Emitting
Materials, Techna Srl, Paenza, Italy, 1999, p. 249.
[22] H. Ebendorff-Heidepriem, D. Ehrt, M. Bettinelli, A.
Speghini, in: S. Jiang, S. Hokkanen (Eds.), SPIE Proceedings. In: Rare-Earth-Doped Materials and Devices III, vol.
3622, SPIE, Bellingham, Washington, USA, 1999, p. 19.
[23] J.M.F. van Dijk, M.F.H. Schuurmans, J. Chem. Phys. 78
(1983) 5317.
[24] H. Ebendorff-Heidepriem, D. Ehrt, J. Philipps, T. Töpfer,
A. Speghini, M. Bettinelli, R. Wong Sun Fat, in: S. Jiang
(Ed.), SPIE Proceedings. In: Rare-Earth-Doped Materials
and Devices IV, vol. 3942, SPIE, Bellingham, Washington,
USA, 2000, p. 29.
[25] G.A.C.M. Spierings, Phys. Chem. Glasses 23 (1982)
101.
[26] H. Ebendorff-Heidepriem, W. Seeber, D. Ehrt, J. NonCryst. Solids 163 (1993) 74.
[27] H. Toratani, H.E. Meissner, T. Izumitani, S.E. Stokowski,
J. Non-Cryst. Solids 95 – 96 (1987) 701.
[28] H. Ebendorff-Heidepriem, W. Seeber, D. Ehrt, J. NonCryst. Solids 183 (1995) 191.