State injection, lattice surgery and dense packing of the deformation-based surface code

Abstract: Resource consumption of the conventional surface code is expensive, in part due to the need to separate the defects that create the logical qubit far apart on the physical qubit lattice. We propose that instantiating the deformation-based surface code using superstabilizers makes it possible to detect short error chains connecting the superstabilizers, allowing us to place logical qubits close together. Additionally, we demonstrate the process of conversion from the defect-based surface code, which works as arbitrary state injection, and a lattice surgery-like CNOT gate implementation that requires fewer physical qubits than the braiding CNOT gate. Finally we propose a placement design for the deformation-based surface code and analyze its resource consumption; large scale quantum computation requires $\frac{25}{4}d^2 +5d + 1$ physical qubits per logical qubit where $d$ is the code distance, whereas the planar code requires $16d^2 -16d + 4$ physical qubits per logical qubit, for a reduction of about 55%.