H. Kobayashi et al. / Physica B 263—264 (1999) 346—349
347
and try to explain both phenomena by the double-
well potential model which has succeeded in the
The change of the Young’s modulus, dE, due to
the relaxation is expressed as
analysis of Ge-doped SiO glasses [3].
ꢀ
ꢀ
D
G
A 1
1
D
!AdE
E
B"+ ¹DG ꢀ 1#(uq)
A
2k ¹
B
G
ꢀ
sech
dD.
ꢀ
ꢀ
ꢅ
ꢅ
G
2. Experiment and specimens
(2)
The experimental method applied to the meas-
urements of the internal friction and the Young’s
modulus is a composite oscillator type with
Eqs. (1) and (2) are fitted to the internal friction
and the Young’s modulus data by the Monte Carlo
ꢀ
method, and 50 sets of parameters (º , A , D ) are
G
G
G
ꢂ
a quartz of 50;3.5;3.5mm , which produces lon-
determined. A series of º values are chosen for the
G
gitudinal waves [5]. This method has only one
oscillational frequency of 51 kHz but the capability
of high resolution suitable for measurements of
small changes of internal friction. Glass specimens
fitted curves to be smooth.
To determine the potential distribution uniquely,
any functional form for the distribution was not
assumed. Firstly the fitting of internal friction was
made under the condition of a symmetric potential
were obtained as follows. SiO soot rods were ob-
ꢀ
tained from SiCl by the vapor-phase axial depos-
ꢄ
ꢀ
(D "0 for all i’s), and the modulus change is
G
ition (VAD) method, and then the rods were heat
calculated by using the determined A values. Sec-
G
treated at a high temperature in an atmosphere of
ꢀ
ondly, both the asymmetry parameter D ’s and the
G
relaxation strength A ’s are used as the fitting para-
G
SF and He gases. These rods were sintered at
ꢄ
a temperature of 1500°C, and SiO —4 mol% F and
meters, and the internal friction and the elastic
ꢀ
—8 mol% F glasses were obtained. Pure SiO glass
ꢀ
modulus change are fitted simultaneously. When
which was obtained by the direct method was pre-
the parameters A ’s are determined, the density
G
pared and used as the standard of SiO glasses.
of relaxation strength, P(º )"(A #A )/
ꢀ
G
G>ꢃ
G
2(º !º ), is calculated.
G>ꢃ
G
3. Analysis
4. Discussion
The observed relaxation peak around 35 K is con-
sidered to be a superposition of widely distributed
Debye-type mechanical relaxation peaks. For the
analyses the internal friction is expressed as [3,6]
The internal frictions of three types of SiO glass-
es were measured in the temperature range of
ꢀ
1.6—250 K. The results are shown in Fig. 1, in which
the symbols O, # and X show those of pure
SiO , SiO —4 mol% F and —8 mol% F glasses. At
ꢀ
D
G
A 1
uq
D
G
\ꢃ
pQ "+
ꢀ
sech
dD.
B
ꢀ
ꢀ
ꢀ
ꢀ
¹ DG ꢀ 1#(uq)
A
2k ¹
ꢅ
G
low temperatures below 5 K, the internal friction is
almost independent of temperature, which is
thought to be due to the TLS relaxation [7]. In
order to compare the values of the internal friction
of the glasses, they are normalized in Fig. 1 by those
\ꢂ
(1)
where Q is the quality factor, A is the magnitude of
G
contribution from ith Debye relaxation, ¹ is the
temperature, u is the angular frequency of sound,
D is the asymmetric energy of the double-well po-
at 2 K, which are 1.49;10
(pure SiO ),
ꢀ
\ꢂ
1.41;10
\ꢂ
(—4 mol% F), and 1.63;10
(—8
ꢀ
tential, and D is the maximum value of the distri-
mol% F). These values of three glasses are not very
different except for their peaks, which show a few
% point differences at 35 K. In Fig. 2, the Young’s
moduli of the three kinds of glasses are shown, in
which the symbols O, # and X show the values of
pure SiO , SiO —4 mol% F and —8 mol% F glasses
G
bution of the asymmetry energy for the i’s potential.
q is the relaxation time in the asymmetric potential
and given by, q"q exp(º /k ¹)sech(D/k ¹),
ꢅ
G
where º is the barrier height of the ith potential
G
and q is a constant.
ꢅ
ꢀ
ꢀ