Surface phase transitions in BiFeO${}_3$ below room temperature

We combine a wide variety of experimental techniques to analyze two heretofore mysterious phase transitions in multiferroic bismuth ferrite at low temperature. Raman spectroscopy, resonant ultrasound spectroscopy, electron paraelectric resonance, x-ray lattice constant measurements, conductivity and dielectric response, and specific heat and pyroelectric data have been collected for two different types of samples: single crystals and, in order to maximize surface/volume ratio to enhance surface phase transition effects, BiFeO${}_3$ nanotubes were also studied. The transition at $T$ $=$ 140.3 K is shown to be a surface phase transition, with an associated sharp change in lattice parameter and charge density at the surface. Meanwhile, the 201 K anomaly appears to signal the onset of glassy behavior.

Figure 1

Specific heat at 0 and 1 T and ac magnetic susceptibility at 0 T (inset) as a function of temperature for BFO single crystal.

Figure 2

Positions of two Raman peaks as a function of temperature, showing shifts beginning at 140 K, with sharp anomalies at 180 K, and further changes of slope at 200 K for a BFO single crystal.

Figure 3

(Top) RUS results for a BFO single crystal (full symbols) and a ceramic sample (open symbols): The resonant frequencies increase between ∼140 and ∼240 K, indicating a hardening of the lattice between these two temperatures. (Bottom) The elastic loss (inverse of the quality factor) shows gradual increase above ∼150 K, with peaks at ∼180 K and at 220–240 K depending on the sample.

Figure 4

Relative change of the reciprocal lattice vector |$q$| as a function of temperature probing the bulk of a single crystal (blue open symbols) and the topmost surface (red solid symbols). The surface data show a rapid expansion of the lattice parameter upon heating above 140 K, and this feature is absent from the bulk.

Figure 5

(Top) $Z$ modulus versus temperature and frequency; (bottom): $Z$ modulus and impedance phase angle as a function of temperature on heating and cooling showing a first-order phase transition at ∼140 K for a BFO single crystal.

Figure 6

Discharge current anomalies in BiFeO${}_3$ single crystals. (Left) pristine samples show two clear anomalies at ∼140 and ∼200 K, though in subsequent runs (right) only the 140 K anomaly is clear, although the 200 K anomaly is still visible for field-cooled samples. The field-cooling dependence of the peak temperature for the 140 K anomaly indicates that this pyroelectriclike current is due to the sudden carrier emission from trap levels triggered by the surface phase transition.

Figure 7

Alpha parameter reflecting the asymmetry of the EPR curves for BFO nanotubes.

Figure 8

(Top panel) Electron paramagnetic susceptibility, showing an increase between ∼125 and 200 K, which is close to the temperature range where elastic softening and anomalous skin expansion has been detected; (middle panel) an exponential fitting to the EPR linewidth yields an extrapolated freezing temperature of ∼33 K; the experimental data depart from the glassy fit for temperatures below 140 K; (lower panel) the gyromagnetic factor also shows a rather abrupt drop below the skin transition temperature. The EPR susceptibility shows a sharp drop at 201 K, whereas the $g$ value (bottom panel) shows a broad maximum near this temperature. These data have been obtained for BFO nanotubes.

Figure 9

Density of states versus energy from ab initio calculations for pure BFO (black) and BFO with $V_{\mathrm{Bi}}$ vacancies in different charge states ($V_{\mathrm{Bi}}^x$, $V_{\mathrm{Bi}}^{\prime }$, $V_{\mathrm{Bi}}^{\prime \prime }$, $V_{\mathrm{Bi}}^{\prime \prime \prime }$). The Fermi energy $E_F$ corresponds to that of the stoichiometric system. In the case of defective systems, this energy lies in the gap, with localized states partially occupied.