A silicon-based surface code quantum computer and reducing the overhead of magic state distillation

Since the Kane proposal our understanding of the need for and methods of quantum error correction has developed significantly – motivating improved architectures for quantum computing, in particular based on the surface code. I’ll present an analysis of a novel scheme for implementing a surface code with donor spins in silicon using their dipolar interaction and a repeating mechanical motion.

We performed error correcting threshold simulations that show significant tolerance to range of errors, including a systematic fabrication error in qubit implantation. Architectures for fault-tolerant computing such as these are expected to require potentially significant overhead in time and qubits, particularly due to the need to distill ‘magic states’. We discuss these distillation protocols and see how their overhead can be improved through both ‘balanced investment’ of qubits and tracking correlated errors through these protocols. Using these techniques we make estimates of the run-time and qubit overhead of some post-classical factoring tasks.