Research Projects
1) Mechanical, electronic, optical, magnetic and transport properties of chiral nanomaterials
Chiral matter, i.e., structures with non-superimposable mirror images, can often display fascinating electronic, optical, transport and magnetic properties. These materials also usually feature strong coupling between mechanical strains and quantum mechanical properties, allowing the possibility of tuning such properties using deformations (i.e., strain engineering). Thus, chiral nanomaterials offer a great many opportunities for impacting the design of novel electromagnetic, photonic and quantum hardware devices. Specific examples of chiral nanomaterials include quasi-one-dimensional materials such as nanotubes, nanowires, nanoribbons and nanocoils, as well as many common proteins and biomolecules.
In order to have predictive tools for studying chiral nanomaterials, we have been developing a suite of rigorously formulated computational techniques for simulating such materials ab initio. The resulting tools allow for the ground state and excited state properties of chiral structures to be calculated accurately and efficiently at the level of Kohn-Sham Density Functional Theory. We have been employing these tools for:
-
Studying mechanical and electronic properties of various chiral nanomaterials, as well as their responses to imposed strains. (See here, here and here)
-
Investigating and harnessing anomalous transport phenomena (the Chiral Induced Spin Selectivity Effect) in such materials for quantum device applications. (See here)
-
Investigating the possibility of flat-band physics, strong electronic correlations and phenomena such as non-reciprocal superconductivity in such materials. (See here)
See a video lecture (recorded at the Institute of Pure and Applied Mathematics, UCLA, 2022) on this topic here .
2) Interplay of mechanics and quantum physics effects in materials with dispersionless electronic states
In recent years, materials exhibiting dispersionless electronic states (or flat bands) have enjoyed attention due to their association with collective electronic properties such as superconductivity, as well as their connections to other exotic states of electronic matter. Examples of such materials include moiré heterostructures (e.g. twisted bilayer graphene) and various bulk materials based on specialized lattices (e.g. Kagome metals). The mechanics underlying the fascinating properties of many of these materials --- particularly, the interplay of atomic scale mechanical effects (e.g. structural relaxation or strain response) and their electronic scale quantum mechanical features --- is often poorly understood, making it challenging to adopt these materials for practical applications.
In collaboration with a number of research groups, we have been using a variety of computational techniques to gain a better understanding of such effects. Our recent efforts in this direction include:
-
Exploration of mechanical effects on flat bands in low dimensional materials (See here).
-
Development of atomistic and multiscale models for studying atomic relaxation effects and defects (e.g. strain solitons) in van-der Waals heterostructures (See here). [Collaboration with Admal Research Lab at UIUC]
3) Machine Learning based Electronic Structure Prediction: From Nanomaterials to Bulk Solids
Electronic structure calculations, based on Kohn-Sham Density Functional Theory (KS-DFT), serve as the workhorse of computational materials simulations and are a crucial tool in guided materials discovery. The ground state electron density, the principal output of KS-DFT, contains a wealth of material information and conventional computations of this quantity scale cubically with system size. This makes predictions of the electron density via machine learning (ML) models attractive for large and complex systems. However, the large computational cost of KS-DFT also tends to make it difficult to develop quantifiably accurate ML models that are applicable across many scales and system configurations.
Along with collaborators, we have been making use of a number data science tools and specialized first principles data generation processes, to create a new generation of machine learning models that are systematic, reliable and highly efficient (both in terms of model training and model prediction costs). Our efforts along these directions include:
-
Development of an interpretable machine learning model for predicting the electronic fields in one-dimensional nananomaterials undergoing strain. (See here)
-
Development of uncertainty aware machine learning models that can predict the electronic fields in bulk systems, including systems featuring disorder and defects, and applying these models to the study of compositionally complex alloys. (See here) [Collaborations with Computational Science and Machine Learning Lab at Michigan Tech]
4) Understanding the Quantum Properties of Biomolecules
In recent years, there has been an emerging consensus that quantum effects (involving spin) in certain biomolecules may be critical to a variety of life processes. This includes magnetic field detection for animal migration, olfaction, metabolic and growth regulation in cells, and optimization of charge transport in proteins. Additionally, quantum simulations of the molecular processes responsible for these phenomena appear to suggest that the biomolecules at the heart of these processes are highly optimized to work within environmental parameters that closely match the surroundings in which they evolved (e.g., high sensitivity to low magnetic fields such as those associated with the Earth's). In particular, these molecules appear to function with utmost accuracy in wet and warm physiological environments, suggesting that they have evolved to resist or profit from the noise in their particular environment. These observations provide the motivation for investigating the properties of these molecules ab initio, so as to understand and harness their exceptional properties for technological applications.
Following these lines of thought, we have been looking into the properties of certain nanoparticles and molecules of biological origin so as to understand their properties. Our efforts along these directions include:
-
Investigating the structural and decoherence properties of the so called Posner molecule, which has been suggested to play a key role in cognitive processes. (See here and here)
-
Use of specialized first principles and machine learning tools (described above) to gain insights into the electronic properties of certain biomolecules and structures.
