Graphene accelerates spin-charge transfer to boost valley polarization in 2D magnetic heterostructures

Tsinghua University Press

As conventional electronics approach their scaling limits, researchers are exploring new ways to process information beyond the charge of electrons alone. One promising option is valleytronics, which uses the valley degree of freedom in two-dimensional semiconductors as an additional information carrier. Monolayer transition metal dichalcogenides such as WS₂ are especially attractive for this purpose because their electronic valleys are coupled to electron spin and can be selectively addressed with circularly polarized light. However, intervalley scattering can quickly weaken valley polarization, limiting the performance of valleytronic devices.

A research team from Xiamen University has now developed a strategy to overcome this limitation by controlling how fast spin-polarized charges move across an interface. In a new study reported in Nano Research, the team introduced graphene into a WS₂/CrBr₃ van der Waals heterostructure and found that graphene proximity coupling sharply accelerated interlayer spin-charge transfer. The transfer time between WS₂ and CrBr₃ decreased from 17.3 picoseconds to 1.3 picoseconds, while the maximum valley polarization and valley polarization splitting of trion excitons in WS₂ increased by 32% and 261%, respectively.

The team published their research in Nano Research on July 23, 2026.

“Most previous efforts have focused on suppressing intervalley scattering to preserve valley polarization,” said Zhiming Wu, a corresponding author of the study from Xiamen University. “Our work shows that actively accelerating spin-selective charge transfer is another effective route. By using graphene as a nearby charge extraction layer, we were able to reshape the exciton dynamics and greatly amplify the valley response.”

The researchers fabricated three types of samples: monolayer WS₂, WS₂/CrBr₃, and WS₂/CrBr₃/graphene heterostructures. Optical measurements showed that the WS₂/CrBr₃ structure already exhibited chirality-dependent valley polarization at low temperature due to spin-selective charge transfer into the magnetic CrBr₃ layer. After graphene was added, the effect became much stronger. The maximum degree of valley polarization of trion excitons reached 59.8%, and the valley polarization splitting increased to 37.9%.

To explain the enhancement, the team combined first-principles calculations with time-resolved photoluminescence measurements. The calculations showed that under photoexcitation, the conduction-band offset in WS₂/CrBr₃ decreases substantially, which weakens the driving force for interlayer transfer. By contrast, when graphene is present, photogenerated charges can be rapidly extracted into the graphene layer, allowing the heterostructure to retain a relatively large conduction-band offset. This helps sustain much faster spin-charge transfer from WS₂ to CrBr₃. The accelerated transfer occurs on a timescale much shorter than the exciton lifetime, creating a larger imbalance between spin populations in the two valleys and thereby producing the giant valley polarization splitting.

The team also demonstrated the practical potential of the structure by fabricating circularly polarized photodetectors. Compared with devices based on WS₂/CrBr₃, the WS₂/CrBr₃/graphene photodetectors showed a higher circular anisotropy ratio, reaching 88%, and a 370% improvement in photoresponsivity. These results indicate that proximity coupling can be used not only to tune fundamental exciton behavior, but also to improve the performance of polarization-sensitive optoelectronic devices.

“This study provides a new design principle for valleytronic materials and devices,” Wu said. “In the future, similar proximity-coupling strategies may be extended to other two-dimensional magnetic heterostructures to realize faster, more sensitive, and more energy-efficient spintronic and optoelectronic technologies.”

Other contributors to the study include Ying Ye, Mengyu Liu, Jian Huang, Jingwen Guo, Yuxiang Zhang, Wei Wu, Jiangpeng Zhou, Boyu Zhou, Chenhao Zhang, Xuanli Zheng, Xu Li, Yaping Wu, and Junyong Kang from Xiamen University.

The work is supported by the National Natural Science Foundation of China (Grant Nos. 62374144, 62374143, 62274139 and 62304188), the Natural Science Foundation of Fujian Province (No. 2025J011001), the Natural Science Foundation of Xiamen (No. 3502Z20227009), and the Basic Research Funds for Central Universities (Grant Nos. 20720220025 and 20720230018).