A clever quantum trick brings practical quantum computers closer

Because these changes can happen at random, even a single error can disrupt a calculation. Preventing that disruption is one of the central problems facing quantum engineers.

Protecting Information With Logical Qubits

To reduce these errors, researchers combine many physical qubits into a single logical qubit and apply continuous error correction. This strategy helps preserve quantum information over time, making storage relatively stable. But storing information is only part of the task. To run a quantum algorithm, qubits must be actively manipulated using quantum gates, which are the basic operations that power quantum computation.

Applying those operations without introducing new errors has proven far more difficult than simply keeping qubits stable at rest.

A New Way to Compute While Fixing Errors

A team led by D-PHYS Professor Andreas Wallraff has now demonstrated a method that tackles this problem directly. Working with researchers from the Paul Scherrer Institute (PSI) and theorists led by Professor Markus Müller at RWTH Aachen University and Forschungszentrum Jülich, the group showed how to perform quantum operations between superconducting logical qubits while correcting errors at the same time. Their findings were recently published in Nature Physics.

The work marks an important advance toward fault tolerant quantum computing, where calculations can proceed without being derailed by constant errors.

Why Quantum Error Correction Is Different

Error correction in classical computers relies on copying information. Multiple identical bits can be stored, checked later, and compared. If one flips, a majority vote reveals the correct value. That approach does not work in quantum systems.

"With qubits, things are a lot more complicated," says Dr. Ilya Besedin, a postdoctoral researcher in Wallraff's group and co-leading author of the study alongside PhD student Michael Kerschbaum. Quantum information cannot be copied or cloned. Instead, it must be distributed across entangled qubits. On top of that, quantum systems suffer from phase flip errors, which have no equivalent in classical computing and require their own correction methods.

Error Correction With Surface Codes

One widely used solution involves surface codes. In this approach, the information of a single qubit is spread across several physical data qubits. Error detection relies on repeated measurements of stabilizers, which work alongside the data qubits to form the logical qubit.

These stabilizers are monitored using additional qubits connected to the data qubits. Measuring them reveals whether a bit flip or phase flip has occurred between checks. Z-type stabilizers detect changes in bit value, while X-type stabilizers detect phase changes. Importantly, the data qubits themselves are never directly measured, allowing them to safely store the corrected quantum state.

The Challenge of Performing Logical Operations

The process becomes more complex when researchers want to apply a logical operation such as a controlled-NOT gate between two logical qubits. Errors can occur during the operation itself, and those errors must also be corrected.

"Performing a logical operation in this fault-tolerant way would be relatively easy if we could move our qubits around and connect them arbitrarily to each other," says Kerschbaum. In superconducting quantum processors, however, qubits are fixed in place. Only neighboring qubits can interact, which limits how operations can be carried out.

Splitting the Square With Lattice Surgery

To work within those constraints, the team turned to a method known as lattice surgery. In their experiment, the researchers began with a single logical qubit encoded across seventeen physical qubits. The data qubits and stabilizers were arranged in a roughly square pattern. Over several cycles, stabilizers were measured every 1.66 microseconds to correct both bit flips and phase flips.

At a key moment, three data qubits running through the center of the square were measured. This step effectively divided the surface code into two separate halves. At the same time, measurements of the X-type stabilizers were paused.

"The end result of this operation was that we had two logical qubits entangled with each other," explains Besedin. During the splitting process, bit flip errors continued to be corrected. Afterward, bit flip error correction resumed independently on each half. While this operation does not yet produce a controlled-NOT gate on its own, it can be combined with additional splitting and merging steps to create one.

A First for Superconducting Qubits

"One could say that the lattice surgery operation is the operation, and all the others can be constructed from it," says Besedin.

He adds, "To the best of our knowledge, this is the first time lattice surgery has been performed on superconducting qubits," he adds, "and we still have some way to go. For instance, 41 physical qubits would be required to make the splitting operation on one logical qubit stable against phase flips too. Nonetheless, this demonstration of lattice surgery on superconducting qubits marks an important step towards the ambitious goal of building useful quantum computers with thousands of qubits.