From Logic to Atoms: Design Automation for Atomic-Scale Computing
Date:
As CMOS scaling approaches its physical limits and AI workloads keep raising compute demand, Field-coupled Nanocomputing (FCN) has emerged as one of the most promising post-CMOS paradigms: information moves through electric or magnetic field interactions rather than current flow, enabling extremely dense circuits with potentially terahertz-scale operating frequencies. Its most advanced realization, Silicon Dangling Bond (SiDB) logic, pushes this idea to the atomic scale and has already yielded fully functional gates, bringing FCN circuits within reach of practical fabrication.
Closing the design gap to conventional CMOS, however, means rethinking every layer of the design-automation stack. Because FCN confines gates and wires to a single physical plane, logic synthesis must target planarity directly to avoid the wire crossings that later degrade signal reliability. Physical design must then jointly solve placement, routing, and clocking under that same constraint, while the atomic substrate additionally demands defect- and temperature-aware layout and simulation to guarantee that a design still works once fabricated.
This talk follows that flow end-to-end, from Boolean specification to dot-accurate atomic layout, showing how synthesis, physical design, and physical simulation fit together into one coherent methodology. All of it is implemented in the open-source Munich Nanotech Toolkit (MNT), which provides an extensible logic-to-atoms pipeline.