Quantum technologies from ultrasensitive sensors to next-generation information processors depend on the ability of quantum bits, or qubits, to maintain their delicate quantum states for a sufficiently long time to be useful.
One of the most important measures of this stability is the spin coherence time. Unfortunately, qubits may lose coherence because their environment is "noisy," for example, due to the presence of nuclear isotopes or other interference that disturbs the qubit.
Two-dimensional (2D) materials—or atomically thin sheets—can offer quiet environments for qubits, as their reduced thickness naturally lowers the number of isotopes that interact with the qubit.
In a paper recently published by npj 2D Materials and Applications , researchers from the University of Chicago Pritzker School of Molecular Engineering (UChicago PME) have created a high-throughput computational strategy, creating a new, data-driven approach to finding ideal 2D materials and substrates.
"With only a few 2D materials explored so far as qubit hosts, the field has lacked a comprehensive roadmap to identify new candidates, especially since 2D materials must be placed on a supporting substrate in realistic devices," said first author Michael Toriyama, a postdoctoral researcher in UChicago PME's Galli Group .
The new paper outlines a novel computational strategy to predict qubit coherence times across thousands of 2D materials interfaced with substrates. Using an automated framework built on the "cluster correlation expansion" method—a powerful way to simulate how isotopes interact with a qubit—the team calculated spin coherence times for more than one thousand monolayers, discovering 189 that could potentially support coherence times longer than those of diamond, a popular host of spin qubits.
"We found that materials such as WS2 and several Au-oxyselenides appear to be particularly promising, showing predicted coherence times in the tens of milliseconds—exceptional values for solid-state systems," said UChicago PME Prof. Giulia Galli , senior author of the study.
These compounds share two common features: They contain very few nuclei with strong magnetic moments and many of their atoms naturally occur in spin-free isotopes.
"Their structural motifs, such as square-planar transition-metal–oxygen units, may also lend themselves to hosting qubits with desirable electronic properties," said co-author Jiawei Zhan, a UChicago PME PhD candidate who performed several electronic structure calculations of the promising materials.
But qubits do not live in free-floating monolayers. They sit on substrates. The team therefore evaluated more than 1,500 2D material-substrate combinations, revealing that substrates can significantly degrade coherence unless they are chosen carefully. Materials like certain oxides, for example ceria and calcium oxide, which have intrinsically low nuclear-spin noise, help preserve the long spin coherence time of the 2D host. This finding provides a clear guideline on how to design high-performance 2D spin-qubit devices by selecting both a quiet host material and a quiet substrate.
To make such large-scale screening possible—and to accelerate future discovery—the authors also developed analytical models that capture the essential physics behind decoherence in 2D materials and heterostructures, inspired by the previous work of Tohoku University Assoc. Prof. Shun Kanai on 3D materials.
"These simple, structure-based formulas allow fast estimates of coherence times without running expensive simulations," said Kanai, a co-author of the study.
With analytical models, the authors expanded their search to nearly 5,000 additional 2D materials from public databases, identifying over 500 new candidates with long predicted coherence times.
The broader message of the work is clear: the space of potentially useful 2D quantum materials is far richer than previously known. By combining high-throughput simulations, data-driven modeling, and physical insight, the study provides the community with a blueprint for systematically discovering next-generation qubit hosts in 2D systems. It also hints at an exciting direction: using artificial intelligence-inspired generative models similar to ChatGPT to design entirely new 2D materials optimized for quantum coherence.
"As quantum technologies move from the laboratory to practical devices, this kind of data-driven strategy will be essential," said Galli. "It transforms what was once trial-and-error exploration into a rational search across a vast design space—bringing the goal of robust, scalable, quantum-enabled devices closer to reality."
