Dianyu Qi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Dianyu Qi has a total of 103 co-authors. Here are the top co-authors by number of shared publications:
Yongyu Wu
Zhejiang University
10 shared publications- Dawei Gao
Zhejiang University
9 shared publications - Kun Ren
Zhejiang University
8 shared publications - Wenjing Zhang
Shenzhen University
7 shared publications - Xuan Li
Zhejiang ICsprout Semiconductor, Hangzhou, China
6 shared publications Guangji Li
Zhejiang ICsprout Semiconductor
5 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (11) , 6392-6398
CMOS-Compatible All-Oxide p--n Heterojunction Deep Ultraviolet Photodetector and Imaging
Deep ultraviolet (DUV) photodetection and imaging are essential for applications ranging from secure communications to environmental monitoring, but the integration of high-performance DUV sensors with established Si technology remains challenging. In this work, we demonstrate a CMOS-compatible, self-powered DUV photodetector and imaging based on an all-oxide p–n heterojunction of amorphous Ga2O3 and NiO. The device is fabricated using low-temperature radio frequency (RF) magnetron sputtering, which ensures full compatibility with back-end-of-line CMOS integration. The heterojunction detector exhibits excellent solar-blind performance under zero bias, achieving a responsivity ( ${R}\text {)}$ of 88 mA/W, a detectivity ( ${D} ^{\ast } $ ) of $10^{{11}}$ Jones, an external quantum efficiency (EQE) of 43%, and a fast response time 2O3/NiO stack with CMOS readout integrated circuit (ROIC) for the first time, forming an $8\times 8$ pixel imaging array. Structural and elemental analyses confirm robust integration, and the array demonstrates clear imaging capability under 254 nm illumination. This study provides a scalable solution for achieving high-performance DUV imaging systems by integrating ultrawide bandgap oxide semiconductors with established Si technology, enabling emerging optoelectronic applications.
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (11) , 6392-6398
CMOS-Compatible All-Oxide p--n Heterojunction Deep Ultraviolet Photodetector and Imaging
Deep ultraviolet (DUV) photodetection and imaging are essential for applications ranging from secure communications to environmental monitoring, but the integration of high-performance DUV sensors with established Si technology remains challenging. In this work, we demonstrate a CMOS-compatible, self-powered DUV photodetector and imaging based on an all-oxide p–n heterojunction of amorphous Ga2O3 and NiO. The device is fabricated using low-temperature radio frequency (RF) magnetron sputtering, which ensures full compatibility with back-end-of-line CMOS integration. The heterojunction detector exhibits excellent solar-blind performance under zero bias, achieving a responsivity ( ${R}\text {)}$ of 88 mA/W, a detectivity ( ${D} ^{\ast } $ ) of $10^{{11}}$ Jones, an external quantum efficiency (EQE) of 43%, and a fast response time 2O3/NiO stack with CMOS readout integrated circuit (ROIC) for the first time, forming an $8\times 8$ pixel imaging array. Structural and elemental analyses confirm robust integration, and the array demonstrates clear imaging capability under 254 nm illumination. This study provides a scalable solution for achieving high-performance DUV imaging systems by integrating ultrawide bandgap oxide semiconductors with established Si technology, enabling emerging optoelectronic applications.
- 23 October 2026
- Published by Wiley in Laser & Photonics Reviews
- p. e72117
A Broadband CMOS Image Sensor Enabled by Monolithic 2D/3D Heterogeneous Integration
Si CMOS image sensors (CIS) dominate modern imaging technologies but are fundamentally limited to wavelengths below 1 µm by the Si bandgap. Extending CMOS imagers into the short‐wave infrared (SWIR) typically relies on hybrid integration of non‐Si photodetectors, resulting in increased fabrication complexity and cost. Here we report a broadband CMOS image sensor enabled by a monolithic 2D/3D heterogeneous integration platform. This platform integrates a PtSe 2 /Si heterojunction with an atomically flat van der Waals interface into the CIS pixel through an integration process fully compatible with back‐end‐of‐line (BEOL) process. The platform extends the conventional 4‐transistor (4T) CIS pixel into the SWIR regime. The resulting detector exhibits broadband photoresponse from 400 to 1600 nm, with a peak specific detectivity of 1.4 × 10 12 Jones at 808 nm and representative SWIR detectivities of 3.2 × 10 9 and 1.2 × 10 9 Jones at 1310 and 1550 nm, respectively, with a 3 dB bandwidth of 25 kHz. Pixel‐level measurements and system‐level imaging demonstrations confirm continuous visible‐to‐SWIR operation. By combining monolithic integration, CMOS compatibility, and broadband imaging functionality within a unified architecture, this work provides a scalable pathway for extending mature Si imaging technologies beyond their intrinsic spectral limits.
- 23 October 2026
- Published by Wiley in Laser & Photonics Reviews
- p. e72117
A Broadband CMOS Image Sensor Enabled by Monolithic 2D/3D Heterogeneous Integration
Si CMOS image sensors (CIS) dominate modern imaging technologies but are fundamentally limited to wavelengths below 1 µm by the Si bandgap. Extending CMOS imagers into the short‐wave infrared (SWIR) typically relies on hybrid integration of non‐Si photodetectors, resulting in increased fabrication complexity and cost. Here we report a broadband CMOS image sensor enabled by a monolithic 2D/3D heterogeneous integration platform. This platform integrates a PtSe 2 /Si heterojunction with an atomically flat van der Waals interface into the CIS pixel through an integration process fully compatible with back‐end‐of‐line (BEOL) process. The platform extends the conventional 4‐transistor (4T) CIS pixel into the SWIR regime. The resulting detector exhibits broadband photoresponse from 400 to 1600 nm, with a peak specific detectivity of 1.4 × 10 12 Jones at 808 nm and representative SWIR detectivities of 3.2 × 10 9 and 1.2 × 10 9 Jones at 1310 and 1550 nm, respectively, with a 3 dB bandwidth of 25 kHz. Pixel‐level measurements and system‐level imaging demonstrations confirm continuous visible‐to‐SWIR operation. By combining monolithic integration, CMOS compatibility, and broadband imaging functionality within a unified architecture, this work provides a scalable pathway for extending mature Si imaging technologies beyond their intrinsic spectral limits.
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (9) , 5520-5527
Mechanistic Optimization of Spacer-Type Split-Gate Flash Balancing Yield and Performance
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (9) , 5520-5527
Mechanistic Optimization of Spacer-Type Split-Gate Flash Balancing Yield and Performance
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Electron Device Letters
- Vol. PP (07413106) , 1
A 0.015 μm 2 /bit Split-Gate Quad-Bit-Cell Charge Trapping Flash Memory via CHH Erasure and Self-Convergent CHE Programming
To address narrow threshold voltage margin (VTM) and slow speed issues of quadruple-level-cell (QLC) NAND, and bit-to-bit (B2B) disturbance in traditional multi-bit storage, a 0.015 μm2/bit Quad-Bit-Cell (QBC) charge-trapping flash memory (CTF) was demonstrated via a split-gate dual-bit SONOS architecture, combined with Channel hot-electron induced Hot-Hole (CHH) erasure and gate-controlled self-convergent Channel-Hot-Electron (CHE) programming. It achieved a 4.5 V memory window (MW), endurance up to 105 P/E cycles, and a minimum 0.8V inter-level VTH margin after 10-year extrapolation at 175 ℃, providing a high-density, high-speed, and highly reliable solution for next-generation embedded flash memory.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Electron Device Letters
- Vol. PP (07413106) , 1
A 0.015 μm 2 /bit Split-Gate Quad-Bit-Cell Charge Trapping Flash Memory via CHH Erasure and Self-Convergent CHE Programming
To address narrow threshold voltage margin (VTM) and slow speed issues of quadruple-level-cell (QLC) NAND, and bit-to-bit (B2B) disturbance in traditional multi-bit storage, a 0.015 μm2/bit Quad-Bit-Cell (QBC) charge-trapping flash memory (CTF) was demonstrated via a split-gate dual-bit SONOS architecture, combined with Channel hot-electron induced Hot-Hole (CHH) erasure and gate-controlled self-convergent Channel-Hot-Electron (CHE) programming. It achieved a 4.5 V memory window (MW), endurance up to 105 P/E cycles, and a minimum 0.8V inter-level VTH margin after 10-year extrapolation at 175 ℃, providing a high-density, high-speed, and highly reliable solution for next-generation embedded flash memory.
- 2 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (5) , 2795-2802
Mechanism and Performance Verification of Channel-Hot-Electron Induced Hot-Hole Injection-Based High-Speed Erasure for Split-Gate Charge-Trapping Flash Memory
The mechanism of the novel channel-hot-electron induced hot-hole (CHH) injection for high-speed erasure (ERS) in the split-gate charge trapping flash (CTF) cell is systematically investigated. In this scheme, hot holes generated by impact ionization of channel hot electrons exhibit approximately three times higher density and a significantly wider spatial distribution compared to band-to-band tunneling-induced hot-hole (BBHH) ERS. This enhances the hole injection to the charge trapping layer and the trapped hole-electron neutralization, leading to more efficient ERS. In addition, it is experimentally verified that the CHH mechanism enables approximately$100\times $ faster erasing speed (~0.1 ms) than BBHH, while also achieving better endurance ($10^{{5}}$ ) and retention (ten years at$175~^{\circ }$ C). These advantages make it ideal for high-speed and highly reliable nonvolatile memory (NVM) applications.- 2 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 73 (5) , 2795-2802
Mechanism and Performance Verification of Channel-Hot-Electron Induced Hot-Hole Injection-Based High-Speed Erasure for Split-Gate Charge-Trapping Flash Memory
The mechanism of the novel channel-hot-electron induced hot-hole (CHH) injection for high-speed erasure (ERS) in the split-gate charge trapping flash (CTF) cell is systematically investigated. In this scheme, hot holes generated by impact ionization of channel hot electrons exhibit approximately three times higher density and a significantly wider spatial distribution compared to band-to-band tunneling-induced hot-hole (BBHH) ERS. This enhances the hole injection to the charge trapping layer and the trapped hole-electron neutralization, leading to more efficient ERS. In addition, it is experimentally verified that the CHH mechanism enables approximately$100\times $ faster erasing speed (~0.1 ms) than BBHH, while also achieving better endurance ($10^{{5}}$ ) and retention (ten years at$175~^{\circ }$ C). These advantages make it ideal for high-speed and highly reliable nonvolatile memory (NVM) applications.- 11 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (2) , 524-530
A Process-Robust Split-Gate eFlash Memory With Spacer-Defined Floating Gate in 55-nm Node
To overcome the critical performance-manufacturability-cost trade-off in Internet of Things (IoT) and edge devices, this paper presents a split-gate embedded Flash (eFlash) memory in a 55-nm technology node. The memory utilizes a spacer-defined floating gate (FG) to co-optimize high performance, high manufacturability, and low cost. The proposed cell achieves a low read voltage of 1.5 V, program/erase operations at ±7 V, and a fast program time of 1 μs, enabling high energy efficiency at both the cell and system levels. The self-aligned FG structure inherently eliminates etchinduced variations, demonstrating enhanced process robustness that is confirmed by a high yield in fully functional array testing. This work provides a highly competitive solution for balancing performance, manufacturability, and cost in advanced-node embedded Flash memories.
- 11 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (2) , 524-530
A Process-Robust Split-Gate eFlash Memory With Spacer-Defined Floating Gate in 55-nm Node
To overcome the critical performance-manufacturability-cost trade-off in Internet of Things (IoT) and edge devices, this paper presents a split-gate embedded Flash (eFlash) memory in a 55-nm technology node. The memory utilizes a spacer-defined floating gate (FG) to co-optimize high performance, high manufacturability, and low cost. The proposed cell achieves a low read voltage of 1.5 V, program/erase operations at ±7 V, and a fast program time of 1 μs, enabling high energy efficiency at both the cell and system levels. The self-aligned FG structure inherently eliminates etchinduced variations, demonstrating enhanced process robustness that is confirmed by a high yield in fully functional array testing. This work provides a highly competitive solution for balancing performance, manufacturability, and cost in advanced-node embedded Flash memories.
- 5 March 2026
- Published by Elsevier in Microelectronics Journal
- Vol. 173
High-performance stepped-field-plate LDMOS in BCD technology with robust process window
- 5 March 2026
- Published by Elsevier in Microelectronics Journal
- Vol. 173
High-performance stepped-field-plate LDMOS in BCD technology with robust process window
- 2 March 2026
- Published by IOP Publishing in Japanese Journal of Applied Physics
- Vol. 65 (7)
A 319.3-TOPS/W high-efficient computing-in-memory engine with 55nm high-density multi-level NOR-flash array and event-driven neuromorphic architecture
This paper presents a high-efficiency spiking neural network (SNN) accelerator based on a 1.5T NOR flash device. By leveraging the non-volatile and tunable conductance characteristics of NOR flash, the proposed design enables precise synaptic weight storage and low-power updates in a computing-in-memory architecture. The accelerator adopts a 512×1024 array structure, supporting large-scale SNN parallel inference and event-driven computation. With optimized spike accumulation and neuron firing circuits, the system achieves high computational density with reduced power consumption. Experimental results demonstrate an energy efficiency of 319.3 TOPS/W, highlighting its superiority over SRAM- or RRAM-based implementations and its potential for edge AI applications such as low-power sensing and on-device inference.
- 2 March 2026
- Published by IOP Publishing in Japanese Journal of Applied Physics
- Vol. 65 (7)
A 319.3-TOPS/W high-efficient computing-in-memory engine with 55nm high-density multi-level NOR-flash array and event-driven neuromorphic architecture
This paper presents a high-efficiency spiking neural network (SNN) accelerator based on a 1.5T NOR flash device. By leveraging the non-volatile and tunable conductance characteristics of NOR flash, the proposed design enables precise synaptic weight storage and low-power updates in a computing-in-memory architecture. The accelerator adopts a 512×1024 array structure, supporting large-scale SNN parallel inference and event-driven computation. With optimized spike accumulation and neuron firing circuits, the system achieves high computational density with reduced power consumption. Experimental results demonstrate an energy efficiency of 319.3 TOPS/W, highlighting its superiority over SRAM- or RRAM-based implementations and its potential for edge AI applications such as low-power sensing and on-device inference.
- 1 March 2026
- Published by IOP Publishing in Journal of Semiconductors
- Vol. 47 (3) , 032301
Optimizing 55 nm split-gate memory for compute-in-memory: a focus on floating-gate engineering
The escalating need for high-performance artificial intelligence (AI) computing intensifies the "memory bottleneck" of the von Neumann architecture, prompting extensive exploration of computation-in-memory (CIM) solutions. This study is centered on the optimization of a high-efficiency, low-power "L"-shaped split-gate floating-gate (FG) memory for CIM applications. Fabricated on a 55 nm CMOS platform, the memory devices were systematically investigated through wafer acceptance test (WAT), Sentaurus™ simulations and comprehensive evaluations with the DNN + NeuroSim Framework V2.0. Among devices with diverse FG lengths, the 95-nm FG variant exhibits outstanding performance: it achieves a 5.35 V memory window, reaches a maximum conductance of 16.7 μ S with excellent linearity under the varying voltage and width pulse scheme (VWPS), realizes 32-state multi-level storage, and attains a 92% training accuracy on the CIFAR-10 dataset using the VGG8 neural network.
- 1 March 2026
- Published by IOP Publishing in Journal of Semiconductors
- Vol. 47 (3) , 032301
Optimizing 55 nm split-gate memory for compute-in-memory: a focus on floating-gate engineering
The escalating need for high-performance artificial intelligence (AI) computing intensifies the "memory bottleneck" of the von Neumann architecture, prompting extensive exploration of computation-in-memory (CIM) solutions. This study is centered on the optimization of a high-efficiency, low-power "L"-shaped split-gate floating-gate (FG) memory for CIM applications. Fabricated on a 55 nm CMOS platform, the memory devices were systematically investigated through wafer acceptance test (WAT), Sentaurus™ simulations and comprehensive evaluations with the DNN + NeuroSim Framework V2.0. Among devices with diverse FG lengths, the 95-nm FG variant exhibits outstanding performance: it achieves a 5.35 V memory window, reaches a maximum conductance of 16.7 μ S with excellent linearity under the varying voltage and width pulse scheme (VWPS), realizes 32-state multi-level storage, and attains a 92% training accuracy on the CIFAR-10 dataset using the VGG8 neural network.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (1) , 102-106
Impact of EP-CG Interlayer Dielectric on Performance of L-Shaped Split-Gate Flash Memory
Split-gate L-shaped floating-gate flash cells in 55 nm are reported that reach a 5.6 V window, 10. P/E cycles, and a tight V. distribution for both SLC (Single-Level Cell) and TLC (Tri-Level Cell), while cutting two critical masks. The advance is enabled by a 27 nm triple-oxide (Dry/HTO/Dry) inter-layer dielectric that raises control-gate-to-erase-poly breakdown voltage to 18 V. The sharpening of the floating-gate tip, together with the L-shaped floating gate, suppresses the EP-to-FG coupling ratio (ECR) to 0.05 while keeping the CG-to-FG coupling ratio (GCR) at 0.64, delivering low-voltage erase performance while maintaining well-controlled floating gate control.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (1) , 102-106
Impact of EP-CG Interlayer Dielectric on Performance of L-Shaped Split-Gate Flash Memory
Split-gate L-shaped floating-gate flash cells in 55 nm are reported that reach a 5.6 V window, 10. P/E cycles, and a tight V. distribution for both SLC (Single-Level Cell) and TLC (Tri-Level Cell), while cutting two critical masks. The advance is enabled by a 27 nm triple-oxide (Dry/HTO/Dry) inter-layer dielectric that raises control-gate-to-erase-poly breakdown voltage to 18 V. The sharpening of the floating-gate tip, together with the L-shaped floating gate, suppresses the EP-to-FG coupling ratio (ECR) to 0.05 while keeping the CG-to-FG coupling ratio (GCR) at 0.64, delivering low-voltage erase performance while maintaining well-controlled floating gate control.