## welcome

Melting was carried out at temperatures as low as possible (maximum 1370 K) for relatively short periods (maximum 0. The melts were poured in copper moulds of room temperature. T h e temperature of the middle part of the curve connected with space charge relaxation is underlined.

Experimental data are given by triangles, circles, crosses, and dots. All curves are guides to the eye. The evaporated aluminium electrodes are blocking for alkali ions and do not diffuse into the specimen. The measuring voltages of reaching nearly the transformation range, the glasses were taken out of the mould and immediately annealed to avoid cracking. Samples were cut and ground u n d e r oil, washed with trichlormethane and stored in a desiccator.

Disks 25 mm diameter and approximately 2 mm thickness were used for electrical measurements. The electrodes were applied by vacuum evaporating a thin film of aluminium. The glasses are not phase separated and, therefore, fulfil the demand for homogeneity. The surface structure of the specimen is not planar due to grinding which gave a roughness in the 2 -3.

T h e frequency scale is reduced to the bulk relaxation time. In the case of the conductivity curves in fig. Discussion The frequency dependence of the effective dielectric constant and the effective electrical conductivity of four alkaliborosilicate glasses are summarized as master curves shown in figs. The bulk relaxation time, z D, is determined by the frequency of the maximum in M " ( w )the imaginary part of the electrical modulus.

We emphasize that the p a r a m e t e r r given in eq. This is in accordance with the behaviour expected for CPA elements (see eq. According to the model for plane electrodes, the conductivity is expected to increase proportionately to co2 in the low-frequency range. The model with the CPA-element considering the roughness predicts a co2n-slope which, due to superposition with discharge processes at the lowest frequencies, is reached in the experiment only immediately below the curvature range.

The discharge discussed in connection with our experimental data begins only under the conditions of nearly complete build-up of the DDSC and high electric field. From this, we conclude that the high electrical field intensity in the neighbourhood of the electrodes alters the charge transport mechanism in this layer resulting in discharge of alkali ions at the cathode.

The discharge process discussed in ref. The relatively low-temperature dependence of this type of frequency dependence justifies the reduction carried out. This is true of cation-conducting alkalioxide glasses. I 10 only on the carrier concentration at the electrodes. A detailed quantitative description of the discharge process requires more very lowfrequency high accuracy investigation.

As discussed above, the effective conductivity decreases by some orders of magnitude at the lowest frequencies (see fig. In comparison with the frequency-independent bulk conductivity, the magnitude of the decrease seems, at first sight, to be proportional to Li20 content. However, it is evident from the non-reduced presentations in fig. Frequency dependence of the effective electrical conductivity, c7, of four alkaliborosilicate glasses at high temperatures.

Experimental data are given by dots. The lower temperature dependence of conductivity and permittivity during the discharge process in comparison with the DDSC-relaxation based on measuring results of a 0.

These statements are supported by admittance measurements of 0. In this case, the quotient of measuring voltage and sample thickness is about three orders of magnitude higher than in the measurements discussed here, the frequency range of the frequency indepen- H.

Conclusions On the basis of the general theory of spacecharge relaxation and experimental results, a model for ion-conducting glasses was developed including especially the influence of drift and diffusion of unipolar ions between plane, blocking electrodes. Particular attention was paid to the detailed investigation of model assumptions as well as their experimental realization, since they are commonly deficient in the literature. The small signal behaviour of six fast ion-conducting borosilicate glasses from 10 -2 to 106 Hz and from 293 to 713 K shows that dynamic space charge processes appear in a clearly distinguished frequency range between bulk relaxation at higher and high-field discharge processes at very low frequencies.

This is mainly due to the sample thickness specimen and small signal application. This procedure is very useful as the master curves show. The relation to the roughness is concluded by analogy from the data of Bates et al. The model presented for space charge relaxation is applicable to any isotropic unipolar solid ionic conductor. The authors are indebted to Dr W. Mueller from the Academy of Sciences, Berlin, for providing the investigated samples.

The authors wish tO thank the referees for their careful reading a n d pertinent suggestions. Appendix 1 Inserting the Poisson equation (8) in eq. If n I d e t e r m i n e d by eq. P(o) (40) from eq. C: Solid State Physics 10 (1977) 1459. Raleigh, in: Electrode Processes in Solid State Ionics, Proc. NATO Advanced Study Institute, ed. Dupuy (Reidel, Dordrecht, 1975) p. Acta 29 (1984) 1381.