New Qubit Measurements for Scalability
Achieving high qubit counts in near-term quantum computers is challenging. Top problems include refining how qubits are measured and error correction. Parametric amplifier devices are typically used for measuring qubits; however, the device amplifies weak signals it picks up from the qubits to conduct the readout, resulting in unwanted noise and qubit decoherence. Also, the bulky size of the amplification chain is challenging to work around as qubit counts increase in size-limited refrigerators.
Now, however, the Aalto University research group Quantum Computing and Devices (QCD) is using thermal bolometers as ultrasensitive detectors, as demonstrated in an April 10 Nature Electronics paper. The paper shows that bolometer measurements can be accurate enough for single-shot qubit readout.
Bolometric energy sensing is a fundamentally different kind of measurement. Since a bolometer measures power, or photon number, it is not bound to add quantum noise stemming in the way that parametric amplifiers are. Instead, it subtly senses microwave photons emitted from the qubit via a minimally invasive detection interface, which is roughly 100 times smaller than its amplifier counterpart, making it extremely attractive as a measurement device.
The researchers demonstrated that their nanobolometers could be an alternative to conventional amplifiers. From the onset, the bolometers were accurate enough for single-shot readout, free of added quantum noise, and they consume 10,000 times less power than the typical amplifiers—all in a tiny bolometer.
Single-shot fidelity is used to determine how accurately a device can detect a qubit’s state in just one measurement vs. an average of multiple measurements. The team was able to obtain a single-shot fidelity of 61.8% with a readout duration of roughly 14 microseconds. When correcting for the qubit’s energy relaxation time, the fidelity jumps up to 92.7%.
The team says it can expect to see bolometers approaching the desired 99.9% single-shot fidelity in 200 nanoseconds. Swapping the bolometer material from metal to graphene, which has a lower heat capacity and can detect very small changes in its energy quickly, is a possibility. By removing unnecessary components between the bolometer and the chip itself, even greater improvements in the readout fidelity can be made, achieving a smaller and simpler measurement device that makes scaling up to higher qubit counts more feasible.