Atomistic study of finite temperature properties in ferroelectric BiAlO$_3$

Technical Summary: Atomistic Study of Finite Temperature Properties in Ferroelectric BiAlO3

Problem Statement
Lead-free perovskite ferroelectrics are critical for sustainable technological applications, yet the search for viable alternatives to lead-based materials remains challenging. Bismuth aluminate (BiAlO3) has emerged as a promising candidate, synthesized via high-pressure, high-temperature techniques. While experimental studies have established its rhombohedral ($R3c$) ground state and ferroelectric nature, there is a significant discrepancy between experimental data and theoretical predictions. Specifically, Density Functional Theory (DFT) calculations, which are restricted to zero Kelvin, report spontaneous polarization values ($72.3–82.5 \, \mu\text{C/cm}^2$) that differ substantially from experimental measurements ($9–29 \, \mu\text{C/cm}^2$). Furthermore, existing computational approaches lack the capability to model the finite-temperature behavior of BiAlO3, leaving gaps in understanding the temperature evolution of spontaneous polarization, strain, coercive fields, and the coupling between ferroelectric (FE) and antiferrodistortive (AFD) modes.

Methodology
To address these limitations, the authors developed a first-principles-based atomistic model to simulate BiAlO3 at finite temperatures. The methodology involved the following steps:

1. Phonon Analysis: Using Density Functional Perturbation Theory (DFPT) within the VASP package, the authors identified structural instabilities in the cubic ($Pm\bar{3}m$) phase. This revealed imaginary frequencies corresponding to the $\Gamma$ mode (ferroelectric distortion) and $R$ and $M$ points (antiferrodistortive oxygen octahedra tilts).
2. Effective Hamiltonian Construction: Based on the identified instabilities, an effective Hamiltonian was parameterized. This Hamiltonian includes terms for the ferroelectric local mode ($u$), the antiferrodistortive mode ($\omega$, corresponding to $R_{25}$ distortions), and strain ($\eta$). Crucially, it accounts for the coupling between these modes (FE-strain, AFD-strain, and FE-AFD). The parameters were derived from DFT calculations and are provided in atomic units.
3. Monte Carlo Simulations: The parameterized Hamiltonian was utilized in single-flip Metropolis Monte Carlo (MC) simulations. A supercell of $20 \times 20 \times 20$ unit cells with periodic boundary conditions was employed. Simulations utilized a simulated annealing technique, cooling from 1500 K to 10 K. To accurately capture competing instabilities near the phase transition, the number of MC steps was increased in the vicinity of the Curie temperature ($T_C$).
4. External Perturbations: The model was extended to investigate the effects of external electric fields (along [111], [110], and [100] directions), hydrostatic pressure, and uniaxial/biaxial stresses.

Key Results

• Ground State and Phase Transition: The simulations confirm a structural phase transition from a paraelectric cubic phase to a ferroelectric rhombohedral ($R3c$) phase. The predicted Curie temperature is $1160 \, \text{K}$, and the spontaneous polarization at low temperatures is $81 \, \mu\text{C/cm}^2$ along the [111] direction. The authors note that while the $T_C$ is higher than the experimental value ($>793 \, \text{K}$), likely due to an overestimation of binding energy in the parameterization, the polarization aligns well with DFT ground-state values.
• Mode Coupling: A strong coupling between ferroelectric and antiferrodistortive modes is observed across a broad spectrum of temperatures and electric fields. The AFD order parameter mimics the trend of the FE mode, with both oriented along the [111] pseudocubic direction.
• Electric Field Response: Under an electric field along [111], the material exhibits a single hysteresis loop, confirming the $R3c$ phase. The saturation polarization ($P_{Sat}$) varies slightly with field direction ($57.8–60.2 \, \mu\text{C/cm}^2$), while the coercive field ($E_C$) shows significant anisotropy ($2.4–3.8 \, \text{MV/cm}$). The AFD and strain order parameters exhibit "butterfly" loop behaviors, with strain showing high saturation ($0.021$) and significant remanent strain ($0.018–0.019$).
• Temperature Dependence: As temperature increases, the hysteresis loop area, saturation polarization, remanent polarization, and coercive field decrease. This trend is consistent with other ferroelectric materials.
• Pressure and Stress Effects:
  • Hydrostatic Pressure: Positive hydrostatic pressure suppresses both spontaneous polarization and the Curie temperature (rate of $-45 \, \text{K/GPa}$), while negative pressure has the opposite effect.
  • Uniaxial Stress: Tensile uniaxial stress induces a sequence of two phase transitions (cubic $\to$ tetragonal $\to$ rhombohedral), whereas compressive stress results in a single transition (cubic $\to$ rhombohedral).
  • Biaxial Stress: The trend is reversed; tensile biaxial stress leads to a single transition, while compressive biaxial stress induces two transitions.

Significance and Claims
The paper claims to fill a critical gap in the literature by providing the first computational approach to extend DFT simulations to finite temperatures for BiAlO3. By developing a parameterized effective Hamiltonian, the study offers atomistic insights into the material's behavior under experimentally relevant conditions, including the role of FE-AFD coupling and the response to mechanical loads.

The authors highlight that the high saturation and remanent strains predicted by their simulations suggest BiAlO3 could be a promising candidate for shape memory applications. Additionally, the study demonstrates that the Curie temperature and phase stability of BiAlO3 can be tuned via hydrostatic pressure and mechanical stress, offering potential pathways for optimizing the material for specific device applications. The work underscores the necessity of moving beyond zero-Kelvin DFT calculations to fully understand the functional properties of lead-free ferroelectrics.