Transmon-qubit readout using in-situ bifurcation amplification in the mesoscopic regime

TL;DR

This paper introduces a novel transmon qubit readout method using in-situ bifurcation amplification of polaritonic meters in the mesoscopic regime, achieving high fidelity without external amplifiers.

Contribution

The work demonstrates a new in-situ bifurcation-based readout technique utilizing polaritonic meters coupled to a transmon qubit, operating effectively in the mesoscopic regime.

Findings

Achieved 98.6% single-shot fidelity in qubit readout.

Operated successfully in the mesoscopic regime with low photon number.

No external quantum-limited amplifier required.

Abstract

We demonstrate a transmon qubit readout based on the nonlinear response to a drive of polaritonic meters in-situ coupled to the qubit. Inside a 3D readout cavity, we place a transmon molecule consisting of a transmon qubit and an ancilla mode interacting via non-perturbative cross-Kerr coupling. The cavity couples strongly only to the ancilla mode, leading to hybridized lower and upper polaritonic meters. Both polaritons are anharmonic and dissipative, as they inherit a self-Kerr nonlinearity $U$ from the ancilla and effective decay $\kappa$ from the open cavity. Via the ancilla, the polariton meters also inherit the non-perturbative cross-Kerr coupling to the qubit. This results in a high qubit-dependent displacement $2\chi > \kappa, ~U$ that can be read out via the cavity without causing Purcell decay. Moreover, the polariton meters, being nonlinear resonators, present bistability, and bifurcation behavior when the probing power increases. In this work, we focus on the bifurcation at low power in the few-photon regime, called the mesoscopic regime, which is accessible when the self-Kerr and decay rates of the polariton meter are similar $U\sim \kappa$. Capitalizing on a latching mechanism by bifurcation, the readout is sensitive to transmon qubit relaxation error only in the first tens of nanoseconds. We thus report a single-shot fidelity of 98.6 $\%$ while having an integration time of a 500 ns and no requirement for an external quantum-limited amplifier.