We are observing dielectric properties of the glass and glass ceramics. A peak in tan delta is observed and it is shifting w.r.t. frequency as well as temperature.
The tan delta signal or graph is defined as the quotient of the loss modulus (mechanical) or dielectric loss and the storage modulus (mechanical) or permittivity (dielectric), respectively. When a loss peak is clearly evident, I prefer to use its spectrum over that of a tan delta. In mechanical measurements, the highest temperature peak in tan delta is taken to indicate the glass transition temperature (peak temperature). This sometimes carries over to dielectric measurements, but typically, dielectric measurements may indicate more tan delta peaks than one sees mechanically. A peak as a function of frequency is telling you something about the dissipation process that is producing the energy loss. It might correspond to a molecular Debye-type relaxation (dipole resonances or oscillating dipoles). When you say the peak is shifting as a function of temperature, I presume you mean you are measuring at fixed frequency, and at different fixed frequencies the peak temperatures are different. This again comes back to how energy at different frequencies can couple into the "heat bath" of your materials.
Glass and glass ceramics can be considered dielectric material. When dielectric material is placed between capacitor plates it's capacitance increases. The ratio of capacitance with and without dielectric is known as dielectric constant. When a.c. field is applied, as long as polarization follow a.c. field, dielectric constant is independent of frequency and dielectric loss is zero considering no d.c.current flow. However, when polarization does not follow a.c. field, dielectric constant decreases with frequency and dielectric loss and tan delta show a peak at a particular frequency. This peak shifts with the increase in temperature. This is called dielectric relaxation and may be observed below glass transition temperature. In Debye type dielectric relaxation, such peaks are seen easily.
In silicate glasses the dielectric phenomena are mostly determined by the following processes:
- limited mobility of cations in the silicate network (random walk, hopping)
- electrode polarization (which can be large)
- Maxwell-Wagner effect in phase separated glasses where the conductivity of the occluded phase is larger than that of the maxtrix
All these effects may overlap. Frequently the so-called electrical modulus formalism is used for the analysis of dielectric data instead of the complex permittivitiy, as it is easier to sort out electrode effects.
Please fid attached some articles which may be interesting for you