Development of high-coherence superconducting devices for quantum computing: Fabrication process development, materials analysis, and device characterization
Overview
- Date:Starts 26 September 2024, 10:00Ends 26 September 2024, 13:00
- Location:Kollektorn, MC2, Kemivägen 9, Chalmers
- Language:English
Opponent: Dr. Mollie Schwartz, MIT Lincoln Laboratory, USA
Superconducting quantum circuits are a promising platform for the experimental realization of quantum computers.
One of the main challenges in quantum computing hardware is the limited time over which we can sustain the information encoded in a quantum state: the state is easily perturbed by its environment in a process known as decoherence, leading to a loss of the information stored within.
A major source of decoherence in superconducting circuits are parasitic two-level systems (TLSs), which can couple to the device and act as a source of dielectric loss.
In this thesis, I study the impact of device design, fabrication procedure, and materials properties on the TLS loss of our devices through materials analysis, cryogenic microwave measurements, and simulations.
The devices I study include coplanar waveguide (CPW) resonators, 3D cavity resonators, and aluminium-on-silicon transmon qubits.
In our quantum processor architecture, the CPW resonators used for qubit readout face similar loss sources as the qubits; however, resonators have a faster fabrication and characterization turnaround. Therefore, we use resonators as proxies to investigate the decoherence mechanisms in quantum circuits, and we show how the extracted information can be applied to improve coherence in transmon qubits.
Having identified a dominant source of TLS loss at the substrate-metal interface of our devices, we find that by increasing the grain size of the superconducting films through increasing the film thickness, we can decrease the contribution of intergranular oxide to the TLS loss at this interface. We show that this approach can yield time-averaged energy relaxation times T1 > 200 μs, with the best qubit reaching an average T1 = 270 μs and a highest observed T1 = 501 μs, improving on our previously standard T1 ~ 100 μs.
Studying the performance of our 3D cavity resonators as a function of design, material quality, and fabrication procedure, our improved cavities can reproducibly yield resonance quality factors above 80 million.
We also integrate our quantum processors into a flip-chip architecture to improve the scalability of our devices. We find that our approach does not measurably degrade the performance of the integrated qubits.
Additionally, we study the nonlinearity of the dielectric susceptibility of TLSs through the intermodulation products generated in a resonator driven by two detuned tones. Our analysis method can reconstruct the standard TLS model parameters from a single spectrum measured at relatively high drive powers corresponding to ~1000 photons.