Atomistic Simulations of Heat Removal in Integrated Circuits by Alexander Gabourie

Alexander Gabourie


Alex’s research focuses on understanding the thermal properties of ultrathin, two-dimensional (2D) materials using nanoscale simulations. He specializes in characterizing 2D materials that are interacting with common insulators found in integrated circuits, a configuration difficult to simulate or measure. He also dedicates time to developing open-source, GPU-accelerated algorithms to advance his own research as well as those in the 2D material community.


Technologies from cell phones to self-driving cars perpetually demand more computational performance, with an increasing emphasis on efficiency. To meet these demands, major performance-reducing system bottlenecks, like high circuit temperatures and slow data transfers, must be eliminated. By stacking computing and memory units, and by leveraging nanomaterials like two-dimensional (2D) layered semiconductors, three-dimensional (3D) integrated circuits (ICs) promise to address these issues with energy-efficient, high-performance computation. Unfortunately, this solution may exacerbate some thermal problems, requiring careful consideration of materials’ thermal properties.

To understand the thermal properties of 2D materials, we use atomistic molecular dynamics (MD) simulations. These simulations can calculate necessary thermal properties, like thermal conductivity and thermal boundary conductance, in 3D-IC-analagous structures where experiments struggle. Our calculations show that, in general, interactions with insulators found in ICs drastically degrade 2D materials’ thermal properties, potentially limiting their application in 3D ICs. However, atom-level insights from detailed MD methods developed in-house suggest the possibility to engineer the optimal thermal interfaces for 2D materials. This work, enabled by the ARCS fellowship, may generate key guidelines needed for thermal management of circuits utilizing 2D materials.


6 + 13 =