Near-Lossless 2D Semiconductor Phase Modulators Using Hybrid Tungsten Oxyselenide/Graphene Electrodes

Why It Matters

The breakthrough removes the long‑standing trade‑off between modulation efficiency and optical loss, enabling scalable, low‑power integrated photonics for data centers and emerging quantum technologies.

Key Takeaways

• Hybrid TOS/graphene electrodes suppress graphene absorption at 1550 nm
• Modulation efficiency achieved 0.202 V·cm with 0.08 dB loss
• Device integrates WS₂, hBN, and SiN microring platform
• Approach decouples phase shift from intensity attenuation
• Enables low‑power photonic interconnects and quantum circuits

Pulse Analysis

The drive toward ever‑denser optical interconnects has been hampered by a persistent efficiency‑loss dilemma: improving electro‑optic modulation typically incurs higher absorption, especially when graphene serves as the top electrode at telecom wavelengths. Conventional graphene or indium‑tin‑oxide contacts introduce noticeable insertion loss, limiting the scalability of silicon‑nitride or silicon‑photonic platforms. Researchers at NTU and partner institutions have now demonstrated a hybrid electrode that sidesteps this trade‑off, preserving the high carrier mobility of graphene while rendering it virtually transparent at 1550 nm.

The key innovation lies in converting a monolayer of WSe₂ into tungsten oxyselenide (TOS) through UV‑ozone treatment, which acts as a strong p‑type dopant for the overlying graphene. This heavy doping pushes graphene’s Fermi level deep into the valence band, suppressing interband transitions near 1550 nm and reducing optical absorption to near‑zero while simultaneously lowering sheet resistance. Stacked with a hexagonal boron nitride dielectric and a monolayer WS₂ electro‑optic layer on a silicon‑nitride microring, the heterostructure delivers a modulation efficiency of 0.202 V·cm with only 0.08 dB extinction‑ratio variation.

The near‑lossless performance opens a practical pathway for energy‑efficient photonic links, on‑chip optical phased arrays, and quantum photonic circuits that demand precise phase control without sacrificing signal strength. Because the TOS/graphene stack is compatible with standard silicon‑nitride foundry processes, it can be integrated into existing manufacturing lines with minimal redesign. Future work may explore scaling the approach to larger arrays, extending the concept to other 2D semiconductors, and optimizing the dielectric thickness for even lower drive voltages, further tightening the link between photonic integration and electronic power budgets.