Relaxor Ferroelectrics: Back to the Single-Soft-Mode Picture

Abstract

The fluctuations of electric polarization in a disordered ferroelectric substance, relaxor crystal PbMg_{1/3}Nb_{2/3}O_{3} (PMN), were studied using a nonlinear inelastic light-scattering technique, hyper-Raman scattering, within a 5-100 cm^{-1} spectral interval and in a broad temperature range from 20 to 900 K. The split ferroelectric mode reveals a local anisotropy of up to about 400 K. Spectral anomalies observed at higher temperatures are explained as due to avoided crossing of the single primary polar soft mode with a temperature-independent, nonpolar spectral feature near 45 cm^{-1}, known from Raman scattering. The temperature changes of the vibrational modes involved in the measured fluctuation spectra of PMN were captured in a simple model that accounts for the temperature dependence of the dielectric permittivity as well. The observed slowing down of the relaxational dynamics directly correlates with the huge increase of the dielectric permittivity.

Infrared, terahertz, and microwave spectroscopy of the soft and central modes in Pb(Mg1/3Nb2/Nuzhnyy1, Petzelt2, Bovtun3 et al. 2017Phys. Rev. BSelf Cite21160View full textAdd to dashboardCite

“…The authors [17] suggest that the mode lying near 45 cm -1 , which is present at all temperatures and is only weakly temperature dependent, is the F 2g Raman mode activated due to a local doubling of the unit cell (local B-site one-to-one ordering) [10,11,13]. According to this picture, above ~300 K the SM seen below 40 cm -1 is only a singlet which appears to be bilinearly coupled with the F 2g mode due to the B-site disorder, which relaxes the selection rules.…”

“…The inspection of HRS spectra indicates that there might be an important coupling between the F 2g mode and SM [17], which prevents their crossing at high temperatures. Therefore in Fig.…”

“…For fitting the spectra at T < 300 K we had to account for the weak third mode, not seen at 300 K in [25], assigned in Ref. [17] to the F 2g mode. We first added this mode only to the A 1 -response (E  P l ), which is usually more active in the Raman response.…”

“…Since the selection rules concerning SM for the HRS and IR are expected to be the same [15] and the corresponding spectra look qualitatively similar, it is of interest to compare both responses in the whole temperature range more quantitatively. For that in this work we have complemented our previous IR data by new THz measurements with a thin plate of dense PMN ceramics (described in [26]) and carried out the IR reflectivity measurements on single crystals, both up to 900 K. These data were analyzed together with our earlier data in the lower-frequency range [19,20,23] and compared with the HRS data [17].…”

“…Namely, in this case below T B , the complete dielectric response consists only of the contributions of PNDs (their bulk and boundaries) and the response of the paraelectric matrix can be neglected. Below the complete freezing temperature T f ≈ 200-220 K, where the relatively stable ferroelectric phase (still consisting of static PNDs) can be induced by an external electric field, the SM response becomes underdamped, but splits into at least two components [3,17]. In inelastic neutron and light scattering experiments, the overdamped SM appears as a central peak (i.e.…”