Prof. Markus Müller explains the research behind the breakthrough and the obstacles that have yet to be overcome in realizing practical quantum computing.
Jülich, 25 May 2022 – While quantum computers promise to deliver immense computing power for the future, today’s systems are still highly prone to errors. Researchers from Switzerland, Germany, and Canada have now succeeded in correcting errors during operation for the first time. The new technique is regarded as an important milestone and was published in the renowned journal Nature today. Prof. Markus Müller from Forschungszentrum Jülich and RWTH Aachen University was involved in the project.
Prof. Müller, you supported the team at ETH Zürich in carrying out the experiments. Why is this seen as a milestone or breakthrough?
Copyright: Forschungszentrum Jülich / Sascha Kreklau
Quantum computers are highly prone to error, which is perhaps the biggest challenge currently facing us. Correcting these errors is very demanding. They affect the quantum bits, or qubits for short, which are the basic information units of quantum computers.
To build a practical quantum computer, these qubit errors must be detected and corrected quickly enough and repeatedly. If the errors accumulate, they distort the quantum calculations and render them useless. Broadly speaking, there are two fundamental types of errors. Previous error correction techniques were unable to simultaneously detect and correct these errors. The team headed by Andreas Wallraff from ETH Zürich has now presented the first experimental system that can automatically compensate for both error types in a way that enables the practical use of the results of quantum operations.
This technique will also benefit other systems – including the EU project OpenSuperQ at Jülich, on which Forschungszentrum Jülich is collaborating with ETH Zürich and other European partners.
What sort of system achieved this milestone?
Copyright: ETH Zürich / Quantum Device Lab
The error correction technique is run on a chip comprising a total of 17 superconducting qubits. It is operated at a temperature of only 0.01 Kelvin, i.e. just above absolute zero. As information stored in individual qubits is relatively unstable, the key is to avoid using single qubits and instead to use several qubits together – using what is known as a surface code. This method involves distributing the quantum information over several physical qubits. Nine of the seventeen qubits on the chip are arranged in a square 3×3 lattice and together form a logical qubit, the value or unit that a quantum computer uses to perform calculations. This unit is the counterpart of the “bit” in a conventional digital computer, which can usually have the value “0” or “1”.
How exactly does this technique work?
Quantum error correction still faces a fundamental problem. According to the rules of quantum physics, quantum information is lost when it is read out. The remaining eight qubits on the chip are therefore slightly offset from the others and serve as auxiliary qubits. They enable errors to be detected without the information stored in the system being distorted by the readout process.
To detect whether a disturbance occurs in a logical qubit and distorts the information,
the eight additional qubits are quickly and repeatedly measured. This reveals the type of error and where it is likely to have occurred on the chip. The logical qubit – i.e. the quantum information stored in the other nine qubits – is not distorted in this process. The effects of the detected errors can then be eliminated by applying appropriate corrections to the qubits. For most applications, however, it is sufficient to monitor the errors and correct them at the end of the quantum calculation – which was the approach taken in this experiment.
What was your involvement in the development?
Copyright: ETH Zürich / Quantum Device Lab
We were delighted to be able to support our colleagues in this experiment by providing characterization techniques that we had previously investigated in our group. These involve certain protocols that allow the collected measurement information to be processed in such a way that the quality of the logical qubit can be evaluated. This technique enabled the quality of the quantum information stored in the error correction code at the start of the experiment to be determined using a limited number of measurements. It also made it possible to distinguish the error type of residual errors – which can occasionally occur despite the high quality levels used in the experiment – and in particular to differentiate between errors that could and could not be corrected by the code.
A significant hurdle on the path towards practical quantum computing has now been overcome. What’s the next step?
The development of quantum computers is currently a hotly contested field in quantum research, with large corporations such as Google and IBM throwing their hats in the ring. Forschungszentrum Jülich and RWTH Aachen University are involved in a number of research consortia that are building practical quantum computers based on diverse and promising physical platforms. The approaches range from trapped ions in the AQTION and IQuAn projects and neutral atoms in MUNIQC-Atoms to superconducting qubits in the OpenSuperQ and QSolid projects.
The experiments at ETH Zürich are impressive, and show the potential of quantum error correction techniques to protect quantum processors from disturbance. The larger devices that are currently in development can be expected to require more complex technology. However, if they are equipped with error correction protocols, their size means that they will ultimately provide even greater error protection.
Krinner, S et al.
Realizing Repeated Quantum Error Correction in a Distance-Three Surface Code
Nature (published 25 May 2022), DOI: 10.1038/s41586-022-04566-8