Direct simulation and machine learning structure identification unravel soft martensitic transformation and twinning...

Area of Science:

• Materials Science
• Statistical Physics
• Soft Matter Physics

Background:

• Martensitic transformation involves diffusionless phase transitions, crucial in materials science and statistical physics.
• This phenomenon, typically seen in crystalline solids, is increasingly observed in soft materials, posing challenges due to complex molecular structures.

Purpose of the Study:

• To directly simulate and investigate the mesoscale structural transition in liquid crystalline blue phases (BP).
• To unravel the dynamical processes of martensitic transformation in soft matter using advanced computational methods.

Main Methods:

• Simulated the mesoscale structural transition of blue phase II (BP II) using Langevin-type equations for the orientational order parameter with thermal fluctuations.
• Employed machine-learning-aided analysis to identify and track local structural changes during the phase transition.

Main Results:

• Successfully simulated the transformation from a perfect lattice of BP II to a twinned lattice of blue phase I (BP I).
• Identified nucleation of BP I initiated by the breakup of disclination line defect junctions, followed by network deformation.
• Demonstrated reversible transformation between twinned BP I and BP II via temperature variation.

Conclusions:

• Machine learning-driven simulations provide unprecedented insights into martensitic transformation dynamics in soft materials.
• The study elucidates the role of topological defects in driving phase transitions in complex ordered systems.