19 August 2024 in Nanoscale
We successfully demonstrated a 2T0C-FeDRAM with a record-long multibit retention time exceeding 2000 seconds and 4-bit (19 states) capability, attributed to a VSN of 0 V leveraging the non-volatile properties of ferroelectrics.
12 December 2022 in Nanoscale Advances
We demonstrate the programmable light intensity of a micro-LED by compensating threshold voltage variability of thin-film transistors by introducing a non-volatile programmable ferroelectric material, HfZrO2 into the gate stack of the TFT.
19 November 2025 in Optics Express
We present silicon nitride photonics on glass (SING), demonstrated at the wafer-level, establishing its feasibility for high-density optical redistribution layers (RDLs) in panel-level packaging. Demonstrated passive components support 106 Gbps transmission, enabling scalable wafer-level optical interconnects with grating couplers, ring resonators, MMI couplers, MZIs, and stitched waveguides. With a compact 100 µm bend radius, SING enables a 10× higher density than LDW or IOX glass platforms, while remaining compatible with large-panel processing. This positions SING as a promising and scalable solution for next-generation optical RDLs.
25 November 2025 in IEEE Photonics Technology Letters
Realizing high-efficiency grating couplers (GCs) on glass substrates for high-bandwidth Co-packaged Optics (CPO) systems is challenging due to optical leakage into the substrate. In this work, we address this challenge by designing, fabricating, and characterizing a uniform silicon nitride (SiN) GC with an aluminum bottom reflector on a glass substrate. The device achieves a peak coupling efficiency of −5.1 dB and a 1-dB bandwidth of 47 nm, a 4.3 dB improvement over the non-reflector device. Furthermore, we demonstrate 106 Gbps Pulse Amplitude Modulation 4-level (PAM-4) transmission, confirming practical applicability. These results provide a valuable foundation for high-performance SiN GCs on a glass substrate platform, enabling their integration into advanced CPO for glass substrate-based optical interconnects in System-on-Wafer (SoW).
25 March 2026 in Optics Letters
In this Letter, we propose a stitching-loss-tolerant silicon nitride (SiN) waveguide structure that alleviates optical loss caused by reticle stitching, which inevitably arises in i-line stepper lithography suitable for low-cost, large-area fabrication. At a wavelength of 1310 nm, the measured stitching loss of straight waveguides is 0.105 dB/interface. In contrast, tapered waveguides with stitching widths (
W
s
) of 2 µm and 3.5 µm demonstrate significantly reduced losses of 0.0678 and 0.0213 dB/interface. These results pave the way for glass-based panel-level optical interconnect enabled by cost-effective fabrication, addressing a critical bottleneck in next-generation optical packaging.
21 May 2026 in IEEE Journal of the Electron Devices Society
Metal–ferroelectric–metal–insulator–semiconductor (MFMIS) ferroelectric field-effect transistors (FeFETs) using an amorphous indium–tin–zinc-oxide (a-ITZO) channel and Hf0.5Zr0.5O2 (HZO) ferroelectric layer are fabricated. By engineering the area ratio (AR = AI/AF) between the MIS transistor region and the MFM ferroelectric-capacitor region, the memory window (MW) can be continuously tuned while maintaining robust erase operation. To probe time-dependent behavior, single-sweep readouts are performed after program/erase pulses with a controlled delay time (Tdelay). We found that, as Tdelay increases, the threshold voltage (Vth) shift of the program becomes larger than that of the erase under pulsed conditions, producing a progressive reduction—and at low AR, a possible inversion—of the MW. This asymmetric delay dependence has not been reported previously for oxide-semiconductor MFMIS FeFETs. Guided by the experimental trends, we propose a band-diagram-based mechanism that explicitly incorporates oxygen-vacancy-related traps at the HZO/floating-gate interfaces. The model attributes the observed evolution to the interplay among ferroelectric polarization relaxation, charge trapping/de-trapping in the floating gate. These findings establish AR engineering and delay-aware operation as effective levers to optimize MW, retention, and erase behavior in AOS-based MFMIS FeFETs, paving the way for compensation-free display drivers and embedded non-volatile memory.