Researchers have experimentally demonstrated topological valley quasi-bound states in the continuum within terahertz valley photonic crystals, enabling direct free-space excitation while maintaining low-loss on-chip transport. The peer-reviewed findings were published in Nature Communications on September 1, 2026.
Topological valley photonics supports on-chip light routing protected against backscattering, but standard valley edge states sit outside the light cone. Because of that placement, external free-space radiation cannot excite them directly. Previous radiative devices relied on on-chip-to-free-space emission through strongly leaky modes, an approach where elevated radiative coupling degrades transport efficiency. Achieving direct free-space-to-on-chip excitation without sacrificing topological protection had remained an open technical hurdle.
To overcome the boundary, the research team implemented geometric modulation and Brillouin-zone folding. This process folds the valley edge states into the light cone, converting them into topological valley quasi-bound states in the continuum. The resulting configuration yields controllable radiative coupling while preserving robust valley transport characteristics.
The team directly visualized the signal transport using time-domain terahertz near-field imaging. Experimental measurements confirmed tunable propagation lengths and demonstrated that waves traveled through sharp bends without backscattering losses. The researchers also fabricated and tested splitter and junction architectures, validating both free-space-to-on-chip excitation and subsequent on-chip signal routing.
The study was authored by researchers at Nanyang Technological University, Tongji University, and Westlake University, including Minggui Wei, Jianwei Liu, Manjima Sudheer, Elbert E. M. Chia, Baile Zhang, Yang Long, and Gui-Geng Liu. Funding support included grants from the Singapore National Research Foundation, the Singapore Ministry of Education, the National Natural Science Foundation of China, Westlake University, the Westlake Education Foundation, and the Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship program.
