High coherence and low cross-talk in a superconducting qubit architecture

In a new report now published in Science Advances, Peter A. Spring and a team of scientists in physics at the Oxford University described qubit coherence and low cross-talk and single-qubit gate errors in superconducting qubit architecture, suited for two-dimensional (2D) lattices of qubits. The experimental setup involved an inductively shunted cavity enclosure with non-galvanic, out-of-plane control wiring, qubits and resonators fabricated on opposing sides of a substrate. The scientists developed a proof-of-concept device featuring four uncoupled transmon qubits, i.e., a superconducting charged qubit with reduced sensitivity to charge noise, to exhibit specific features measured via simultaneous randomized benchmarking. The three-dimensional integrated nature of the control wiring allowed the qubit to remain addressable as the architecture formed larger qubit lattices.

Quantum architect

Efforts to build three dimensional (3D) lattices with multitudes of highly coherent qubits enclosed is an outstanding hardware challenge. Researchers have previously developed superconducting circuits as a promising platform to realize such lattices and form a universal gate set. Typically, two sets of requirements must be met to scale such superconducting lattices, including a method to route control wiring to the circuit allowing all qubits to remain addressable and measurable, while preventing low frequency spurious modes from emerging within the circuit with increasing dimensions. The scaling process should also prevent decoherence channels to qubits and be compatible with gate fidelities beyond the threshold of quantum error correction codes. Physicists had previously overcome wiring limits of edge connected circuits via 3D integrated control wiring as a practical solution. Alternatively, circuits can be enclosed in inductively shunted cavities in two dimensions with a cutoff frequency to cavity modes. Spring et al presented experimental results relative to the latter concept on a four-qubit proof-of-principle circuit, where the circuit architecture featured 3D integrated out-of-plane control wiring, qubits and readout resonators fabricated on opposing sides of a substrate. The team also included a key new feature for compatibility with transmon coherence times, exceeding 100 µs, low cross-talk and single-qubit gate errors.

• 

• 

Device architecture and cross-talk characterization

The researchers obtained images of the cavity enclosure and circuit, where the enclosure base maintained a single central "pillar" and a lid containing a matching cylindrical recess filled with a ball of indium. They arranged the four coaxial transmon qubits in a 2 x 2 lattice with 2 mm spacing and then implemented an out-of-plane wiring design with inductive shunt design, and a circuit layout, where each resonator was coaxially aligned with, and capacitively coupled, to a qubit. The setup allowed the qubit electrodes to be "electrically floating." The team obtained the basic circuit parameters and characterized cross-talk of the device, where the device was a proof-of-principle demonstration of the circuit architecture without intentional couplings, except between qubit-resonator pairs. As a result, Spring et al identified all other couplings as undesired cross-talk. The team then defined the terms of cross-talk and summarized the experimental and simulated parasitic transverse couplings in the device followed by experimental measurements of qubit control line selectivity, and resonator control line selectivity. They also measured the parasitic qubit-resonator coupling to understand the parasitic dispersive shift between qubit and resonator. Followed by single-qubit randomized benchmarking performed on all four qubits separately and simultaneously. The team conducted each of the 31 x 80 experiments, 5,000 times to build statistics and presented the resulting error-per-physical gates, and also performed correlated randomized benchmarking based on simultaneous experimental data. For band structure simulations, Spring et al analyzed the high-frequency structure simulator model of a unit cell that contained ideal dimensions of the central 2 mm x 2 mm, region of the device. They then mapped the band structure during simulations while gathering details on the analytical cutoff frequency, band curvature, and plasma skin and depth predictions within the setup.

• 

• 

Outlook

In this way, Peter A. Spring and colleagues analyzed average qubit coherence times and simultaneous single-qubit gate fidelities in a four-qubit demonstration of a 3D superconducting circuit architecture. Prior to qubit coupling circuitry inclusion, the team highly suppressed the residual crosstalk of the setup. The envisioned optimized device is applicable to study correlated errors generated from high energy radiation in lattices of qubits with high coherence and exponentially suppressed cross-talk. The present architecture contained an inductively shunted cavity enclosure tightly surrounding the circuit, combined with 3D integrated out-of-plane control wiring and reverse-side readout resonators. The outcomes highlighted the low cross-talk of the experimental setup. The enclosure package is reusable by reshaping the indium ball in the lid recess; however, the circuit was not bonded to the enclosure and could not be removed and remounted therefore. The scientists highlighted several shortcoming of the presented device, including the small and variable external resonator-decay rates and dispersive shifts that were non-optimal for qubit readouts. Spring et al. credited the increased coherence in the setup to the fabrication process, which differed from previous implementations of the architecture.