Scientists use neural network to engineer atomic-scale quantum emitter in 2D material

Scientists have engineered a promising new quantum defect using computational modeling. Published in Nature Communications, the research highlights how cobalt, a common metal, could be key to building future quantum computers.

The team began by simulating more than 700 potential defects in tungsten disulfide (WS2), a material with desirable electronic properties. To sift through this massive dataset, the researchers used a deep learning approach. In particular, a neural network trained on simulated and experimental data enabled the researchers to pinpoint cobalt substitution for sulfur (CoS) as an especially promising defect for creating quantum bits (qubits). This neural-network-powered analysis relied on Intel Xeon CPUs and Nvidia Tesla GPUs. In particular, the paper describes an “Intel Xeon E5-2623 v3 CPU with 8 cores and 64 GB of memory combined with a Tesla K80 with 4992 CUDA cores” for training the neural network.

In the computational screening, the researchers used the density functional theory and other first-principles methods, which evaluated the quantum properties of hundreds of potential defects and identified CoS as a top candidate.

In addition to publishing the research in Nature Communications, the scientists have publicly shared the dataset on the website Quantum Defect Genome.

Building quantum bits atom by atom

There were several core finding from the researchers from Lawrence Berkeley National Laboratory and other institutions, including:

1. The cobalt substitution created localized energy levels within the bandgap of WS2.
2. Computational analysis showed that the CoS defect had a computed zero-phonon line (ZPL) of 0.96 eV, falling within the desirable telecom wavelength range. This means this defect has the potential to emit light used in fiber-optic communication.
3. The defect exhibited a spin-doublet ground state, making it potentially useful as a spin-photon interface for quantum applications.
4. Researchers successfully fabricated the CoS defect using site-selective scanning tunneling microscopy (STM) manipulation, demonstrating a pathway from theoretical prediction to experimental realization.
5. Experimental measurements using STM and scanning tunneling spectroscopy (STS) confirmed the theoretical predictions, showing defect states experimentally 0.64 eV below the conduction band minimum (CBM).

Advancing quantum computing

What does all of this mean for the nascent field of quantum computing, which seems to be perpetually about a decade away from widespread commercialization? Namely, the research could help pave the way toward the development of practical quantum bits (qubits) that operate at room temperature and telecom wavelengths.

“In our approach, theoretical screening guides the targeted use of atomic-scale fabrication,” said Alex Weber-Bargioni, one of the study’s principal investigators and a scientist at Berkeley Lab’s Molecular Foundry, in a press release. “Together, these methods open the door for researchers to accelerate the discovery of quantum materials with specific functionalities that can revolutionize computing, telecommunications, and sensors.”

A press release describing the research also notes that the approach used rapid computing techniques to predict the characteristics of hundreds of materials, pinpointing a select few that show the most potential. The scientists then relied on precise fabrication techniques to further evaluate the properties of the most promising candidates.

A potential roadmap for quantum defect discovery

In the Nature Communications paper, the researchers describe the advance as a powerful combination of high-throughput computational screening and precise atomic-scale fabrication. The identification of the CoS defect in WS2 could open up new possibilities for quantum information applications, especially in the telecom range. This approach demonstrates how computational methods can guide experimental efforts while also potentially speeding the discovery and development of quantum defects across an array of 2D materials.