A new approach to reduce decoherence in superconducting qudit-based quantum processors

Quantum computers, which operate leveraging quantum mechanics effects, could soon outperform traditional computers in some advanced optimization and simulation tasks. Most quantum computing systems developed so far store and process information using qubits (quantum units of information that can exist in a superposition of two states).

In recent years, however, some physicists and engineers have been trying to develop quantum computers based on qudits, multi-level units of quantum information that can hold more than two states.

Qudit-based quantum systems could store more information and perform computations more efficiently than qubit-based systems, yet they are also more prone to decoherence.

Decoherence is a loss of quantum information resulting from interactions between quantum units of information and their surrounding environment. Devising reliable strategies to reduce decoherence is a long-standing goal within the fields of quantum physics and quantum computing.

Researchers at University of Southern California (USC) and University of California-Berkeley (UC Berkeley) have developed new protocols for the dynamical decoupling (DD) of qudit-based systems, which could reduce decoherence in these systems. These protocols, outlined in a paper published in Physical Review Letters, were implemented and experimentally verified on a superconducting transmon processor.

“Multi-level quantum systems, or qudits, hold untapped potential for improving quantum information processing and quantum simulation,” Daniel Lidar and Irfan Siddiqi, co-senior authors of the paper, told Phys.org.

“Such qudit systems have more complicated system-bath and quantum crosstalk interactions than their qubit counterparts and operating these systems and scaling them up is challenging, particularly in light of their increased noise sensitivity and complexity.”

DD is a promising technique to protect quantum states from decoherence, which was found to effectively suppress noise and unwanted coherent interactions across a variety of qubit-based quantum computing platforms.

Lidar, Siddiqi, and members of their research teams set out to also explore the potential of this technique for reducing decoherence in qudit systems, specifically those that rely on superconducting transmon-based qudits.

“Initially, our two teams were tackling this challenge separately, but a chance encounter at the 2024 March Meeting between members of our teams led to a close theoretical-experimental collaboration that enabled us to devise and demonstrate effective strategies for suppressing qudit decoherence and crosstalk,” said Lidar.

DD was first developed several decades ago, building on the results of an experiment carried out by physicist Erwin Hahn in 1950, known as the spin echo experiment. As part of his research, Hahn was able to reverse dephasing effects by applying radiofrequency (RF) pulses to a system made of spinning particles.

The principles underpinning this experimental approach are the basis behind DD techniques that have proved to be effective in reducing decoherence in quantum systems. Since Hahn first conducted his experiment, however, his methods have been significantly improved.

“The quantum information community has greatly developed these early ideas in recent years, showing that DD can be viewed as a type of symmetrization operation that cancels out unwanted interactions to any desired order,” explained Lidar.

“In the simplest instance of a qubit coupled to a bath that causes dephasing, DD can be understood as operations that constantly time-evolve the system forward and backward, and this effective time-reversal negates the system-bath interaction.”

As part of the study, Siddiqi and his team developed new experimental strategies, which account for generic system-bath interactions that are known to affect the entire multi-level transmon Hilbert space.

Subsequently, they took advantage of the high degree of local control they could attain over superconducting transmon-based quantum systems to devise DD protocols that effectively suppress the more complex sources of noise and undesired intra-system interactions in these systems.

“To achieve this, we used insights from the device physics in combination with the group-theoretic foundations of DD,” said Lidar.

“It is remarkable how effective fairly simple DD sequences proved to be at combating the noise and unwanted interactions that afflict these complex multi-level systems. These results were achieved at minimal extra resource cost: we did not need any additional qubits or qudits to demonstrate error suppression.”

So far, the new DD protocols introduced by Lidar and his team were found to be highly effective in suppressing decoherence in a superconducting qudit-based quantum processor. In the future, they could be applied to other qudit-based systems, potentially contributing to their advancement and real-world deployment.

“We believe that integrating multi-level DD techniques of the type we developed here will be an essential component in allowing qudit-based processors to become more competitive with qubit approaches in areas such as quantum error correction and quantum simulation,” said Lidar and Siddiqi.

The researchers are now conducting further studies aimed at assessing qudit-based quantum processors and improving their performance. For instance, they are trying to identify simulations of physical systems or phenomena that could be more naturally emulated by these multi-level quantum architectures.

“Many of the systems or models of interest for our field, from neutrinos to lattice gauge theories, are more naturally emulated on quantum hardware with a similar Hilbert space structure,” added Lidar and Siddiqi.

“For example, neutrinos have three flavor states; it is, therefore, quite natural to study their dynamics on a qutrit (or quantum three-level) system. In the area of improving quantum processor performance, we are actively working on how to improve fault tolerance and quantum error correction by exploiting the larger Hilbert space available in these systems while integrating DD-based noise suppression.

“We believe that this approach will lead to reliably lower resource requirements and improved thresholds for fault-tolerant quantum computing.”

© 2025 Science X Network