Two-dimensional ferroelectric materials are widely used in advanced technologies and devices due to their unique physical properties. One of the device design strategies is the lateral heterostructures (LHS) which are easy to synthesize and precisely control. To tackle the challenge of large-scale LHS simulations, the Machine Learning Interatomic Potential (MLIP) can be utilized. Here we train the potential model specifically for monolayer GeTe and (Ge,Pb)Te LHS, which can respond to external electric fields. We investigate the thermal hysteresis and the dynamic properties including domain evolution during the ferroelectric phase transition of GeTe. According to the hysteresis loop under different temperatures and electric field frequencies, GeTe shows stable ferroelectric performance and tunability by electric fields. Additionally, we simulate and regulate multiple phases of the (Ge,Pb)Te LHS. With corrections for dipole interactions, the results show the relationship among the different states. Our research demonstrates that GeTe possesses good ferroelectric properties and highlights the application value of MLIP in the design and study of ferroelectric materials.

Close

PrevNext

View figure in article

• Dynamical phase transitions
• Ferroelectric domains
• Ferroelectricity
• Polarization vortices
• 2-dimensional systems
• First-principles calculations
• Machine learning