PhD Defense: Akram Ibrahim
Location
Physics : 401
PhD Defense: Akram Ibrahim – Online Event
Date & Time
December 2, 2025, 10:00 am – 12:00 pm
Description
ADVISOR: Dr. Can Ataca
TITLE: Multiscale Atomistic Modeling and Machine Learning for Nanomaterials: Structural Evolution and Device Applications
ABSTRACT: Understanding and controlling the synthesis–structure–property relationships of materials, and faithfully simulating their behavior in physical processes and devices, remain pivotal challenges for developing the next generation of functional materials that underpin emerging nano- and quantum-technologies—including energy storage systems, sensors, catalysts, logic circuits, and memory devices. Progress is impeded by the complexity of engineered materials: their crystal structures are often unknown or, when characterized, exhibit disorder, extended defects, amorphous domains, and complex interfaces. Such features demand the simulation of system sizes and timescales that far exceed the capabilities of conventional electronic-structure methods. Overcoming these challenges requires computational modeling approaches that balance two competing objectives: achieving the accuracy of electronic-structure theory while maintaining the computational efficiency necessary to access the length and time scales relevant to experimental synthesis and device operation.
Multiscale atomistic modeling, augmented by machine learning—particularly machine-learned interatomic potentials capable of accurately and efficiently representing complex potential energy surfaces—bridges this divide by extending the reach of first-principles accuracy from nanometers and picoseconds to microns and seconds. This dissertation develops and applies integrated multiscale atomistic frameworks that combine density functional theory, molecular dynamics, Monte Carlo methods, non-equilibrium Green’s function techniques, and advanced machine-learned interatomic potentials. We demonstrate how these approaches enable the prediction of complex crystal structures and their chemical, electronic, and magnetic properties, as well as the simulation of large-scale phenomena such as crystal growth, defect migration, heterogeneous chemical kinetics, electronic transport in devices, and magnetic phase transitions.
TITLE: Multiscale Atomistic Modeling and Machine Learning for Nanomaterials: Structural Evolution and Device Applications
ABSTRACT: Understanding and controlling the synthesis–structure–property relationships of materials, and faithfully simulating their behavior in physical processes and devices, remain pivotal challenges for developing the next generation of functional materials that underpin emerging nano- and quantum-technologies—including energy storage systems, sensors, catalysts, logic circuits, and memory devices. Progress is impeded by the complexity of engineered materials: their crystal structures are often unknown or, when characterized, exhibit disorder, extended defects, amorphous domains, and complex interfaces. Such features demand the simulation of system sizes and timescales that far exceed the capabilities of conventional electronic-structure methods. Overcoming these challenges requires computational modeling approaches that balance two competing objectives: achieving the accuracy of electronic-structure theory while maintaining the computational efficiency necessary to access the length and time scales relevant to experimental synthesis and device operation.
Multiscale atomistic modeling, augmented by machine learning—particularly machine-learned interatomic potentials capable of accurately and efficiently representing complex potential energy surfaces—bridges this divide by extending the reach of first-principles accuracy from nanometers and picoseconds to microns and seconds. This dissertation develops and applies integrated multiscale atomistic frameworks that combine density functional theory, molecular dynamics, Monte Carlo methods, non-equilibrium Green’s function techniques, and advanced machine-learned interatomic potentials. We demonstrate how these approaches enable the prediction of complex crystal structures and their chemical, electronic, and magnetic properties, as well as the simulation of large-scale phenomena such as crystal growth, defect migration, heterogeneous chemical kinetics, electronic transport in devices, and magnetic phase transitions.
