Yubo Qi
University of Alabama at Birmingham, AL·Physics
How electrons move through a material decides whether it conducts, glows, or carries current without resistance. At the atomic scale electrons behave in ways that have no everyday analogue, and these quantum effects underlie superconductors, solar cells and light-emitting diodes.
- Fields
Physics
Materials Science
Computer Science
- Topics
density functional theory
Wannier functions
deep learning
- Grants
- 1 known grant, through Jan 2027
What they work on
- Develop a deep-learning-based, multifunctional computational method that reproduces density functional theory predictions of electron behavior while scaling to systems far larger than direct calculations allow.
- Combine Wannier functions, a localized description of electrons that keeps large calculations tractable, with deep learning models to represent electronic structure and dynamics efficiently.
- Apply the method to twisted two-dimensional systems, where slightly rotated atom-thin sheets form very large repeating patterns, to simulate quantum effects that are inaccessible to conventional calculations.
- Apply the method to two-dimensional heterostructures built from stacked dissimilar atom-thin materials, and release the resulting tools as open-source packages for the condensed matter physics community.
More about the research
Predicting them from first principles is possible but expensive: density functional theory, the standard computational workhorse for calculating electron behavior in materials, grows rapidly costlier as the number of atoms increases, which puts large systems out of reach. This project asks whether machine learning can break that bottleneck.
Why it matters
Simulating electron behavior in large material systems is currently limited by computational cost, which slows the search for materials behind superconductors, solar cells and light-emitting devices. Faster, machine-learning-accelerated quantum simulations could make twisted and layered two-dimensional materials practical to study computationally, and the planned open-source packages would put those capabilities in the hands of other condensed matter theory groups.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Yubo Yang
Hofstra University, NY·Physics
Discover and characterize strongly correlated electronic phases in two-dimensional moiré materials, such as generalized Wigner crystals, using many-body wavefunction methods that go beyond mean-field theory.
Carmen Lilley
University of Illinois Chicago, IL·Materials Science
Develop a machine-learning workflow that predicts elastic properties, stress-strain behavior and fracture strength of 2D lateral heterostructures with interface defects, at accuracy comparable to density functional theory (DFT).
Aldo Romero
West Virginia University, WV·Materials Science
Extend the DMFTwDFT code into a unified open-source framework combining DFT, DFT+U and dynamical mean field theory on the VASP and SIESTA codes, supporting both collinear and noncollinear magnetic configurations.
The school
University of Alabama at Birmingham: 45 labs on PI Pages, $866M a year on research, 166 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.