Operating Flux-Tunable Superconducting Qubits with High Fidelity
- Author(s): Foxen, Brooks Riley
- Advisor(s): Martinis, John M
- et al.
The experimental challenge of today’s quantum computing engineers is to choose a physical qubit system and whittle away at the multifaceted orders of magnitude improvement needed to build a practical quantum computer. In the first part of this thesis I will provide a general introduction to superconducting qubits and the isolated cryogenic environment in which they operate with an eye towards both the particular design requirements and the challenges we face in accommodating many more qubits in the future. Next, I will present a series of three experiments specifically oriented towards system-level improvements for flux-tunable superconducting qubits. In the first experiment, we develop a new metrology tool to characterize the on-chip settling of magnetic flux waveforms. We then used this technique to develop a new PCB-based packaging solution thereby increasing our package-to-chip wiring limit by at least a factor of ten enabling control of 10x more qubits on a single chip. In the second experiment, we provide a fabrication process for superconducting interconnects which allow for the three dimensional integration of our qubits and a direct factor of 2 improvement in qubit connectivity moving from linear chains to two dimensional grids of qubits enabling n**2 more complex circuits. Finally we implement a hardware-efficient 2-qubit fermionic simulation gateset, proposed to study quantum chemistry, using DC flux control on an adjustable coupling gmon transmon device. This first realization of the complete fSim gateset, of which CZ is a member, yielded nearly a 2x improvement over the best reported Pauli error for a CZ for a solid state system to date, 0.41%, and implements an arbitrary photon conserving and low leakage two qubit unitary operation with a factor of 4 times higher fidelity than a minimally universal gateset using single qubit rotations with only a CZ enabling 4-8x more coherent fSim operations. In total, these improvements have increased the computational complexity of our quantum processor by a factor of 400-800 on the road towards building a practical quantum computer.