Fabia Athena “Adaptive Engineering of Devices for Energy-Efficient Electronics”

Abstract:
Recent advances in artificial intelligence have been driven by the co-evolution of algorithms, data, and specialized accelerators that deliver the required parallelism and throughput. At the same time, the energy required to sustain this progress has been rising rapidly. In modern computing systems, data movement can consume more energy than the computation itself. Accordingly, architectures that bring memory closer to logic, enabled by materials and devices compatible with three-dimensional integration, offer a promising solution to this challenge. 

In this seminar, I will present an adaptive engineering framework that tailors physical properties (e.g., transport, interfacial effects) of electronic devices to specific figures of merit, thereby enabling specialized devices for three-dimensional integration. I will illustrate this approach through two distinct back-end-of-line-compatible amorphous oxide electronics platforms. First, I will discuss amorphous oxide semiconductor-based gate-all-around transistors, a promising platform for near-memory computing, where preserving channel integrity during integration is critical for low-leakage operation. I will show that a channel-last interface integration approach suppresses channel degradation, thereby yielding steep turn-on and low-leakage operation. Second, I will discuss metal oxide resistive memories, a compelling platform for analog in-memory computing, for which a key challenge is achieving low reset current density without exacerbating resistance-state variability. I will show that the unique atomic configuration of MAX phases enables modulation of local thermal transport and defect dynamics, leading to ultra-low reset current density with reduced stochasticity and reliable operation at elevated temperatures. I will conclude by discussing how adaptive engineering could catalyze a paradigm of domain-specific devices for vertically integrated energy-efficient adaptive microelectronics.

Bio:
Fabia Farlin Athena is an Energy Postdoctoral Fellow in Electrical Engineering at Stanford University, where she studies oxide semiconductor gate-all-around transistors and on-chip memories for energy-efficient microelectronics. Athena obtained her PhD and MS in Electrical and Computer Engineering from the Georgia Institute of Technology, where her research focused primarily on metal oxide-based resistive memory devices for analog in-memory computing, and she received the Sigma Xi Best PhD Thesis Award 2025. During her PhD, Athena also interned at IBM, where she studied deep-learning applications of a related class of resistive memory on IBM’s Analog AI Hardware platform. Her contributions have been recognized with accolades, including the IBM PhD Fellowship 2022, Cadence Technology Scholarship 2023, MRS Graduate Student Award 2023, VLSI Technology Highlight Paper 2025, the Stanford Energy Postdoctoral Fellowship 2024, and Forbes 30 Under 30 North America in Energy 2025.

Details

• Date: Thursday, April 2
• Time: 10:00 am - 11:00 am
• Category: Special Seminar
• Location: Grier A (34-401A)

Host

• Tayo Akinwande
• Email: chadcoll@mit.edu