Advances in donor-based spin qubits in silicon: spin relaxation and flip-flop qubits

Abstract

Donor spins in silicon allow for extremely long storage of quantum information and provide accurate single spin control. While these properties make them attractive for quantum computation, to build a large-scale quantum computer some major challenges still need to be addressed. One key issue is coupling two donor quantum bits (qubits) with high fidelity, in a scalable manner, without requiring extremely accurate donor placement. In this thesis, we propose a new type of qubit, the flip-flop qubit, a combination of the electron-nuclear spin states of the phosphorus donor, that can be controlled by microwave electric fields. A dipole is created when separating the donor electron from the nucleus, allowing two-qubit gates mediated by electric dipole-dipole interaction at donor distances of several hundred nanometres. Gate fidelities are predicted to be within fault-tolerance thresholds for quantum error correction codes, using realistic charge noise values. Strong coupling of the qubit to superconducting resonators can also be achieved. This idea can be extended to couple nuclear spins to electric fields by adding a magnetic drive, applied simultaneously with the electric drive in a Raman-like configuration. Both qubits can be incorporated in a large scale quantum processor. Two types of flip-flop qubit prototypes, one suited for direct dipole-dipole coupling and one for coupling to a resonator, have been designed, fabricated and measured. Fundamental functionalities have been established and the coupling of a charge qubit to a resonator has been observed. When building a large scale quantum computer, precise knowledge of the fundamental physics of the donor system is of key importance. To this end, we analyse the electron spin relaxation. We find that the spin relaxation is caused by phonon emission at high magnetic fields (>3 T), but becomes dominated by evanescent-wave Johnson noise at lower fields. We also find evidence of spurious spin relaxation caused by electron tunnelling to a charge reservoir, preventable by appropriate tuning of the donor electrochemical potential. Overall, the achievements made in this thesis bring us a step closer to achieving a scalable spin-based quantum computer in silicon.