Abstract
We report on optical properties of reduced $\text{BaTiO}_{3-\delta}$ at different doping levels including insulating and metallic samples. In all the samples, including metallic one, we observe structural phase transitions from the changes in the infrared-active phonon modes. Metallic ground state is confirmed by the Drude-type low-frequency optical reflectance. Similar to $\text{SrTiO}_{3-\delta}$ we find that the midinfrared-absorption band in $\text{BaTiO}_{3-\delta}$ appears and grows with an increase in the oxygen-vacancy concentration. Upon decrease in temperature from 300 K, the midinfrared band shifts slightly to higher frequency and evolves into two bands: the existing band and a new and smaller band at lower frequency. The appearance of the new and smaller band seems to be correlated with the structural phase transitions.
Introduction
Barium titanate ($\text{BaTiO}_3$) is a prototypical perovskite oxide and a cornerstone material in condensed matter physics, primarily known for its ferroelectric properties. While its cousin, strontium titanate ($\text{SrTiO}_3$), has been extensively studied in its metallic state, $\text{BaTiO}_3$ presents a more complex challenge. When oxygen atoms are removed to create "reduced" $\text{BaTiO}_{3-\delta}$, the material transforms from a classic insulator into a metal. This study explores the interplay between conducting charge carriers and the crystal lattice, specifically investigating whether the structural phase transitions characteristic of the insulating phase persist in the metallic regime.
A central question in this field is why $\text{BaTiO}_3$ and $\text{SrTiO}_3$ diverge so sharply in their electronic properties upon doping. In $\text{SrTiO}_3$, electrons behave as "large polarons" and move relatively freely. In contrast, electrons in $\text{BaTiO}_3$ appear to be "small polarons," tightly bound to lattice distortions. Understanding this difference requires a detailed look at how electrons interact with the "soft" phonon modes that drive structural transitions.
Methods
The researchers examined four single-crystal samples of $\text{BaTiO}_{3-\delta}$, labeled B1 through B4, representing a range from undoped insulators to heavily reduced metals with carrier concentrations up to $n \approx 2.0 \times 10^{20} \text{ cm}^{-3}$.
Characterization involved several complementary techniques:
- Transport Measurements: DC resistivity, Hall effect, and Seebeck coefficients were measured to define the metallic versus insulating behavior across temperatures.
- Optical Spectroscopy: The primary analytical tool was optical reflectance, measured over a broad frequency range from the far-infrared ($50 \text{ cm}^{-1}$) to the ultraviolet ($40,000 \text{ cm}^{-1}$).
- Data Analysis: Electron Energy Loss Spectroscopy (EELS) data was integrated to extend the energy range to 175 eV. Kramers-Kronig analysis was then applied to transform reflectance data into optical conductivity, allowing for the separation of contributions from lattice vibrations (phonons) and free electrons.
Results and Discussion
The study provides compelling evidence for the existence of a "ferroelectric metal." Even in the most metallic sample (B4), clear signatures of structural phase transitions—from cubic (C) to tetragonal (T), orthorhombic (O), and finally rhombohedral (R) phases—were observed in the transport and optical data.
As shown in Figure 4, the resistivity $\rho(T)$ exhibits distinct anomalies at temperatures corresponding to the structural phase transitions. These transitions are also reflected in the optical reflectance spectra (Figure 6), where phonon mode splitting occurs across all four phases, even in samples with high carrier concentrations.
A key discovery is the "colossal softening" of the lattice vibrations with doping. Unlike $\text{SrTiO}_3$, where adding electrons stiffens the soft phonon mode, $\text{BaTiO}_3$ shows a significant drop in the soft mode frequency (approximately $110 \text{ cm}^{-1}$) upon doping. This renormalization suggests exceptionally strong electron-phonon coupling, consistent with the small polaron model where electrons distort the surrounding Ti-O bonds.
Furthermore, a broad mid-infrared (MIR) absorption band was observed to grow with oxygen vacancy concentration. At low temperatures, this band splits into two features (M1 and M2). The emergence of the sharper M2 band, centered around $840\text{--}890 \text{ cm}^{-1}$, appears directly correlated with the structural phase transitions and may be linked to positional disorder of the Ti ions.
Conclusion
This research highlights the unique electronic environment of $\text{BaTiO}_{3-\delta}$, where structural transitions and metallic conductivity coexist. The observed phonon softening and MIR band behavior provide strong support for the small polaron nature of charge carriers in this material.
However, a remaining ambiguity is whether these effects are driven solely by the added electrons or by the oxygen vacancies themselves. Future studies utilizing alternative dopants like Lanthanum or Niobium, which introduce electrons without creating vacancies, will be essential to isolate the intrinsic effects of charge carriers. Additionally, refined theoretical models are needed to fully explain the relationship between Ti-ion disorder and the complex mid-infrared spectra.
Summary updated