Ultraefficient reconstruction of effectively hyperuniform disordered biphase materials via non-Gaussian random fields

Area of Science:

• Materials Science
• Statistical Physics
• Computational Materials Science

Background:

• Disordered hyperuniform systems exhibit crystal-like suppression of density fluctuations while maintaining liquid-like isotropy.
• These systems possess unique properties beneficial for transport, electronics, and mechanics.
• Hyperuniformity, initially for particle systems, now extends to heterogeneous materials like composites and porous media.

Purpose of the Study:

• To develop an ultraefficient method for reconstructing effectively hyperuniform biphase materials.
• To enable simultaneous generation of microstructure and material property fields.
• To incorporate hierarchical uncertainties in material reconstruction.

Main Methods:

• Utilized second-order non-Gaussian random fields for explicit reconstruction.
• Avoided iterative optimization methods like simulated annealing.
• Developed a single-step reconstruction process for microstructure and properties.

Main Results:

• Demonstrated an ultraefficient, single-step reconstruction of effectively hyperuniform biphase materials.
• Successfully generated both microstructure and latent property fields simultaneously.
• Showcased the method's ability to handle hierarchical uncertainties in heterogeneous materials.
• Validated the method across diverse systems (isotropic, anisotropic, hyperuniform, nonhyperuniform).

Conclusions:

• The proposed method offers a significant advancement over iterative techniques for hyperuniform material reconstruction.
• This approach accelerates the design, analysis, and discovery of novel disordered hyperuniform heterogeneous materials.
• The method's efficiency and versatility make it broadly applicable in materials science and engineering.