Molecular Quantum-dot Cellular Automata (QCA) is an end-of-roadmap alternative to silicon-based computation. Logical operations and data movement are accomplished via Coulomb interactions between QCA cells that have bistable charge configurations. This basic device-device interaction can allow for the computation of any Boolean logic function. Molecular QCA systems are expected to operate at room temperature, could potentially offer densities and speeds that are at least two orders of magnitude beyond what end-of-the-curve CMOS can provide, and are expected to dissipate very little power. Tools exist which allow circuit designs to be directly translated into QCA cell layouts. However, there is currently no manufacturing process that can position QCA molecules to form QCA circuits with the necessary sub-nm precision.
This proposal attacks the positioning problem from both an experimental and a design perspective. The work focuses on the design of computationally interesting QCA systems (i.e. logic that would facilitate tasks like image processing) that might actually be built using a process of self-assembly and guided assembly. The work will develop processes for self-assembly in solution of mesoscale (1-100 nm) circuitboards (DNA structures), to which molecular QCA cells or other components would attach, and use a new process for guided self-assembly of DNA circuitboards on lithographic features on silicon. The systems target, data convolution, can be accomplished with systolic architectures that map well to QCAs device architecture, and the resulting molecular circuitry could eventually provide enhanced data processing capabilities for CMOS chips.
There will be a unique interplay between physical science and computer science with work in design influencing what experiments are actually conducted. Closing the feedback loop, experimental science will refine work in design. The net result should be accelerated progress toward realizable systems.
