Research & Publications

An archive of my Ph.D. research at UMass Amherst focused on predictive modeling, high-performance computing, semiconductor dynamics, and thermoelectric transport.

Demystifying My Research

At its core, my Ph.D. research focused on understanding how heat and electricity move through microscopic materials—specifically semiconductors—and how we can optimize them for better energy efficiency (a field known as thermoelectricity).

Because these materials are measured at the nanoscale, their behavior isn't governed by classical physics; it's governed by complex quantum effects. To study and predict these interactions, I couldn't just use a standard computer. I had to build custom predictive models and leverage high-performance computing (HPC) and massive cloud/data center computing clusters to run highly parallelized simulations.

The mathematical foundations required to solve these equations, and the need to process vast amounts of simulated data across distributed computing networks, seamlessly bridged my transition into AI and Machine Learning. The same principles of optimizing massive parallel workloads, handling complex predictive modeling, and analyzing high-dimensional data directly apply to architecting and training today's large-scale enterprise AI systems. Specifically, I built physics-aware models for high-throughput discovery of 2D materials and simulated full phonon distributions to evaluate performance bottlenecks—projects that required transforming massive theoretical datasets into actionable insights.

Research Work

Google Scholar
  • Very high thermoelectric power factor near magic angle in twisted bilayer graphene
    Demonstrated that flat bands emerging at low twist angles in twisted bilayer graphene result in sharp features in the electronic density of states, leading to superior and tuneable thermoelectric performance.
    A. Kommini, Z. Aksamija • 2D Materials, 8(2), 025026, 2021.
  • High throughput 2D material discovery for thermoelectric applications / Materials selection rules for optimum power factor
    Calculated the power factor of 2D materials using existing databases by accurately modeling carrier transport, determining optimal material parameters using a novel controlled-scattering approach.
    A. Kommini, Z. Aksamija • Journal of Physics: Materials, 3, 015005, 2020. View Journal
  • Anisotropy in power factor of 2D materials with periodic potential barriers using Wigner-Rode formalism
    Evaluated the anisotropy in power factor that arises from carriers being transported by different processes (energy filtering and confinement) in 2D materials like MoS2.
    A. Kommini, Z. Aksamija • Physical Review Applied, 14, 034037, 2020. View Journal
  • Electron Transport and Thermopower in 2D and 3D Heterostructures – A Theory Perspective
    A comprehensive review discussing the impact of interfaces and heterojunctions on the electronic and thermoelectric transport properties of materials, ranging from graphene to transition metal dichalcogenides.
    A. K. Majee, A. Kommini, Z. Aksamija • Annalen der Physik, 531, 1800510, 2019. View Journal
  • Fowler-Nordheim emission in surface modified free-standing diamond nanomembranes
    Studied field emission from bulk emitters covered with a thin wide-gap semiconductor layer for mass spectrometry applications, exploring resonant tunneling induced enhancement of electron emission.
    C. Henkel, R. Zierold, A. Kommini, C. Thomason, Z. Aksamija, R. H. Blick • Scientific Reports, 9(1), 6840, 2019. View Journal
  • Thermoelectric properties of periodic quantum structures in the Wigner-Rode formalism
    Implemented a comprehensive transport model using the Wigner–Rode formalism to study tunneling, energy relaxation, and barrier impacts in semiconducting nanostructures.
    A. Kommini, Z. Aksamija • Journal of Physics: Condensed Matter, 30(4), 044004, 2018. View Journal
  • Low-temperature enhancement of the thermoelectric Seebeck coefficient in gated 2D semiconductor nanomembranes
    Formulated a novel controlled scattering approach by simulating a confined 3D material to mimic a 2D material with a gate, boosting device thermoelectric properties.
    A. Kommini, Z. Aksamija • Journal of Computational Electronics, 15(1), 27-33, 2015. View Journal
  • An AES-Based Intellectual-Property Identification in System-on-a-Chip (SOC) Design
    Embedded different Advanced Encryption Standard (AES) encoders into a System-on-a-Chip (SOC) based watermarking scheme at the behavior design level to securely identify intellectual property.
    P. G. Gopinath, K. Adithya, B. Ajay • International Journal of Computer Applications, 72(9), 28-33, 2013. View Journal
  • Phonon drag contribution towards TE performance in MoS2: Using full phonon distribution
    Studied the contribution of different phonon modes towards phonon drag by calculating the full phonon distribution including normal processes to inform the selection of semiconducting materials for future thermoelectric devices.
    Research Project

Conference Appearances

  • APS March Meeting Boston, MA (2019) • Materials selection rules for optimum power factor
  • MRS Fall Meeting Boston, MA (2018) • Improving Thermoelectric Power Factor in 2D Materials
  • HPC Day Dartmouth, MA (2017) • Wigner-Rode Formalism simulations
  • ICT Pasadena, CA (2017) • Two presentations on Thermoelectric Properties & Inelastic scattering
  • 20th SISPAD Washington, D.C. (2015) • Enhancement of Seebeck coefficient

News & Features