Researchers from the University of Sydney, working with IBM, identified and quantified key factors limiting the performance of quantum computers and demonstrated ways to reduce their impact.
The findings improve understanding of how errors emerge during quantum computations and could help advance the reliability of quantum technology. The research focused on quantum error correction, a key requirement for scaling quantum computers into useful machines.
Quantum computers are highly sensitive to noise, instability, and external interference.
Because qubits are fragile, even small disturbances from their environment can cause quantum information to be lost.
Quantum error-correction systems are designed to repeatedly check qubits for mistakes while calculations are taking place.
However, those checks can introduce new errors of their own.
The University of Sydney and IBM researchers analyzed the role of mid-circuit measurements, which occur when certain qubits are measured at intermediate stages of a quantum operation.
These measurements collapse selected qubits into classical states while allowing other qubits to maintain coherence.
The process provides immediate feedback on how the broader quantum operation should be managed.
But mid-circuit measurements take time, requiring the rest of the quantum operation to idle while measurements are completed.
The research team found that this idling noise is a major performance bottleneck for quantum error correction.
Using a 156-qubit IBM Quantum Heron r2 superconducting quantum processor, the researchers tested how different error-correction methods preserved quantum information and supported quantum logic operations.
The team redesigned the error-correction circuitry to reduce idling time caused by mid-circuit measurements.
This substantially improved performance, increasing logical qubit survival rates from below 90% to more than 96% for each error-correction cycle.
The researchers also found that measurement noise is one of the dominant limitations affecting the reliability of quantum logic operations on current quantum devices.
The work stems from the University of Sydney’s collaboration with IBM, announced in 2024, to advance quantum error correction and benchmark different approaches to fault-tolerant quantum computing.
That collaboration is funded by the Intelligence Advanced Research Projects Activity, a U.S. government research funding agency.
The research also builds on an international collaboration and talent exchange program between the University of Sydney and University College London focused on next-generation quantum technologies.
The paper is titled “Characterising the failure mechanism of error-corrected quantum logic gates.”
KEY QUOTES:
“Quantum computers will become even more useful if we can reliably detect and correct errors while calculations are taking place. This joint project with IBM helps us understand which parts of today’s quantum hardware are introducing the most problems and where engineering improvements will have the greatest impact.”
“This occurs many, many times during each step of the quantum computation. Each such mid-circuit measurement takes time and everything else in the operation has to ‘idle’ while the measurement is completed. This is a major stumbling block. But we can’t get around this step – it is an essential element of quantum error correction. What we have done in this study is pin down quantitatively what kind of performance we need out of these error checks. This is vital to design systems that can scale up and work.”
“Testing these ideas on advanced quantum hardware allows us to better understand the practical challenges involved in scaling up quantum computing systems. This kind of collaboration is essential if we want to develop quantum technologies that are useful outside the laboratory.”
Professor Stephen Bartlett, Project Lead, Director of Sydney Nano, University of Sydney
“Quantum error correction is essential for building large-scale quantum computers, but it introduces a very complex set of engineering challenges. We wanted to identify which physical processes were limiting performance on modern quantum devices. What we found is that the act of measuring qubits during a calculation can itself create instability. By redesigning how those measurements are performed, we were able to significantly improve the reliability of the logical qubits.”
Dr Robin Harper, Lead Author, Sydney Nano and the School of Physics
