Zhongli Liu
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
Research Subjects
Zhongli Liu has a total of 142 co-authors. Here are the top co-authors by number of shared publications:
- Xingji Li
Harbin Institute of Technology
27 shared publications Jianqun Yang
Harbin Institute of Technology
17 shared publications- Fengkai Liu
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin, China
12 shared publications - Yadong Wei
Harbin Institute of Technology
11 shared publications - Lei Wu
Harbin Institute of Technology
10 shared publications Xiu-Lu Zhang
10 shared publications
Latest Publications
- Open Access
- 19 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
Constraint-aware HLS optimization via LLM-based multi-role agent with iterative reflexion
High-Level Synthesis (HLS) design productivity is often constrained by manual, expert-driven pragma tuning and the difficulty of balancing performance optimization with strict hardware resource limits. While automated Design Space Exploration (DSE) has been extensively studied, many existing approaches rely on open-loop estimation models or substantial training data, which can lead to physically unrealizable designs under tight resource budgets. In this work, we propose Reflex-HLS , a training-free agentic framework that formulates HLS optimization as a closed-loop, feedback-driven process grounded in physical synthesis outcomes. Reflex-HLS integrates Large Language Models (LLMs) within a Planner–Actor–Critic architecture, where optimization decisions are iteratively refined based on hardware-in-the-loop feedback from the vendor toolchain, ensuring physical realizability throughout the optimization process. Experimental results on PolyBench and MachSuite benchmarks, validated through full Vivado synthesis and implementation, demonstrate that Reflex-HLS achieves speedups of up to 48.41 × (e.g., on jacobi-2d ) while strictly adhering to device resource constraints. Further evaluations across regular, irregular, and structured workloads show that the proposed framework generalizes beyond conventional compute kernels. Comparative studies against representative open-loop DSE methods, large foundation models operating without hardware feedback, and closed-loop heuristic search highlight the trade-offs among optimization paradigms, with Reflex-HLS producing more deployable and resource-efficient designs under tight resource budgets in the evaluated cases.
- 19 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
Open AccessConstraint-aware HLS optimization via LLM-based multi-role agent with iterative reflexion
High-Level Synthesis (HLS) design productivity is often constrained by manual, expert-driven pragma tuning and the difficulty of balancing performance optimization with strict hardware resource limits. While automated Design Space Exploration (DSE) has been extensively studied, many existing approaches rely on open-loop estimation models or substantial training data, which can lead to physically unrealizable designs under tight resource budgets. In this work, we propose Reflex-HLS , a training-free agentic framework that formulates HLS optimization as a closed-loop, feedback-driven process grounded in physical synthesis outcomes. Reflex-HLS integrates Large Language Models (LLMs) within a Planner–Actor–Critic architecture, where optimization decisions are iteratively refined based on hardware-in-the-loop feedback from the vendor toolchain, ensuring physical realizability throughout the optimization process. Experimental results on PolyBench and MachSuite benchmarks, validated through full Vivado synthesis and implementation, demonstrate that Reflex-HLS achieves speedups of up to 48.41 × (e.g., on jacobi-2d ) while strictly adhering to device resource constraints. Further evaluations across regular, irregular, and structured workloads show that the proposed framework generalizes beyond conventional compute kernels. Comparative studies against representative open-loop DSE methods, large foundation models operating without hardware feedback, and closed-loop heuristic search highlight the trade-offs among optimization paradigms, with Reflex-HLS producing more deployable and resource-efficient designs under tight resource budgets in the evaluated cases.
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (10) , 4014-4022
Different Single-Event Burnout Behavior in p-GaN HEMTs Under Zero and Negative Gate Bias
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (10) , 4014-4022
Different Single-Event Burnout Behavior in p-GaN HEMTs Under Zero and Negative Gate Bias
- 27 August 2026
- Published by American Physical Society (APS) in Physical Review B
- Vol. 114 (8) , 084109
Ab initio framework for dopant-resolved multimechanism scattering and charge transport in semiconductors
The conventional empirical treatment of defect scattering fails to accurately capture carrier transport behavior under realistic doping conditions. Here, we demonstrate a precise first-principle computational approach for modeling doping behavior in semiconductors under various doping elements and concentrations. The methodology is applied to cubic silicon carbide (3C-SiC) to evaluate dopant performance. Based on first-principles calculations and self-consistent Fermi level determination, we obtained realistic, temperature-dependent defect concentrations consistent with experiments. Unlike empirical models, defect scattering strongly depends on defect types. In nitrogen- and phosphorus-doped systems, ionized defect scattering dominates at low temperatures, while phonon scattering becomes dominant at higher temperatures. In aluminum-doped systems, an unconventional temperature-dependent trend was found. Phonon and ionized defect scattering collectively contribute to the transport properties at low temperatures, while ionized defect scattering gradually becomes the dominant scattering mechanism at higher temperatures. The calculated electron and hole mobilities both show good agreement with experimental data. The study provides crucial insights into defect-limited carrier transport and offers guidance for optimizing doping strategies in 3C-SiC-based electronic devices.
- 27 August 2026
- Published by American Physical Society (APS) in Physical Review B
- Vol. 114 (8) , 084109
Ab initio framework for dopant-resolved multimechanism scattering and charge transport in semiconductors
The conventional empirical treatment of defect scattering fails to accurately capture carrier transport behavior under realistic doping conditions. Here, we demonstrate a precise first-principle computational approach for modeling doping behavior in semiconductors under various doping elements and concentrations. The methodology is applied to cubic silicon carbide (3C-SiC) to evaluate dopant performance. Based on first-principles calculations and self-consistent Fermi level determination, we obtained realistic, temperature-dependent defect concentrations consistent with experiments. Unlike empirical models, defect scattering strongly depends on defect types. In nitrogen- and phosphorus-doped systems, ionized defect scattering dominates at low temperatures, while phonon scattering becomes dominant at higher temperatures. In aluminum-doped systems, an unconventional temperature-dependent trend was found. Phonon and ionized defect scattering collectively contribute to the transport properties at low temperatures, while ionized defect scattering gradually becomes the dominant scattering mechanism at higher temperatures. The calculated electron and hole mobilities both show good agreement with experimental data. The study provides crucial insights into defect-limited carrier transport and offers guidance for optimizing doping strategies in 3C-SiC-based electronic devices.
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (7) , 3155-3168
RollPlace: Improving Macro Placement via Monte Carlo Rollout Search
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (7) , 3155-3168
RollPlace: Improving Macro Placement via Monte Carlo Rollout Search
- 3 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (3) , 531-538
Single-Event Burnout Effect of Split-Gate-Trench MOSFET in Different Ion Angles
For high-frequency and low-signal-loss circuits in space environments, the application demand and range of Split-Gate-Trench MOSFET (SGTMOSFET) are increasing. However, these devices are susceptible to heavy ion irradiation at different angles through experiments, resulting in catastrophic Single-Event Burnout (SEB) failure in the heavy ion experiment on the 130 V N-channel SGTMOSFET. This study utilized TCAD software to simulate the SEB failure phenomenon in the experiment, investigating its mechanism and predicting failure behaviors at various angles. The results showed that the unique split-gate structure of SGTMOSFET changed the electric field distribution within the device. When heavy ions were incident at specific angles, the changes in carrier concentration caused by the electric field accelerate the occurrence of SEB failure. The influence on Single-Event Gate Rupture (SEGR) was also discussed. Moreover, the simulation results indicated that the trench depth significantly affected the SEB ion incidence angle effect of SGTMOSFET. These findings prompted researchers to pay more attention to the SEB ion angle effect and help formulate angle-related enhanced design strategies, thereby improving the stability of SGTMOSFET in the space environment.
- 3 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (3) , 531-538
Single-Event Burnout Effect of Split-Gate-Trench MOSFET in Different Ion Angles
For high-frequency and low-signal-loss circuits in space environments, the application demand and range of Split-Gate-Trench MOSFET (SGTMOSFET) are increasing. However, these devices are susceptible to heavy ion irradiation at different angles through experiments, resulting in catastrophic Single-Event Burnout (SEB) failure in the heavy ion experiment on the 130 V N-channel SGTMOSFET. This study utilized TCAD software to simulate the SEB failure phenomenon in the experiment, investigating its mechanism and predicting failure behaviors at various angles. The results showed that the unique split-gate structure of SGTMOSFET changed the electric field distribution within the device. When heavy ions were incident at specific angles, the changes in carrier concentration caused by the electric field accelerate the occurrence of SEB failure. The influence on Single-Event Gate Rupture (SEGR) was also discussed. Moreover, the simulation results indicated that the trench depth significantly affected the SEB ion incidence angle effect of SGTMOSFET. These findings prompted researchers to pay more attention to the SEB ion angle effect and help formulate angle-related enhanced design strategies, thereby improving the stability of SGTMOSFET in the space environment.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (1) , 351-357
Simulation of the Single-Event Burnout in Lateral Enhancement-Mode β-Ga 2 O 3 MOSFET Devices
This study investigates the single-event burnout (SEB) mechanisms and hardening strategies of beta-gallium oxide (β-Ga2O3) enhancement-mode (E-mode) metal-oxide-semiconductor field-effect transistor (MOSFET) through technology computer-aided design (TCAD) simulations. Firstly, the accuracy of the simulation results was verified against relevant literature, and the necessity of including the substrate structure in the SEB simulation was demonstrated. Subsequently, transient analysis reveals a four-stage SEB evolution: (a) The large number of electron-hole pairs (EHPs) generated by heavy-ion incidence drift under the influence of the external field is collected by the electrodes. (b) Under the influence of the high electric field at the gate edge and the transient current induced by heavy ions, impact ionization is triggered, generating new EHPs. (c) The uniformly distributed hole pillars and the auxiliary gate depleted the channel, leading to transient current decay. (d) Substrate holes leave the region beneath the gate and are reduced through recombination, while channel holes accumulate at the gate interface, elevating the gate potential and maintaining the conductive channel until thermal burnout occurs. The thermal burnout in the final stage has been identified as the core process of its SEB. Based on this, we propose a source-connected Schottky body contact (SC-SBC) structure to extract excess holes under the gate in order to verify the SEB mechanism and attempt device hardening design. Simulation results demonstrate that this method effectively drains accumulated gate-region holes, suppresses gate potential elevation, and prevents SEB. The results also verified that the rise in gate potential due to hole accumulation beneath the gate is the fundamental mechanism responsible for SEB in lateral E-mode β-Ga2O3 MOSFETs.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Device and Materials Reliability
- Vol. 26 (1) , 351-357
Simulation of the Single-Event Burnout in Lateral Enhancement-Mode β-Ga 2 O 3 MOSFET Devices
This study investigates the single-event burnout (SEB) mechanisms and hardening strategies of beta-gallium oxide (β-Ga2O3) enhancement-mode (E-mode) metal-oxide-semiconductor field-effect transistor (MOSFET) through technology computer-aided design (TCAD) simulations. Firstly, the accuracy of the simulation results was verified against relevant literature, and the necessity of including the substrate structure in the SEB simulation was demonstrated. Subsequently, transient analysis reveals a four-stage SEB evolution: (a) The large number of electron-hole pairs (EHPs) generated by heavy-ion incidence drift under the influence of the external field is collected by the electrodes. (b) Under the influence of the high electric field at the gate edge and the transient current induced by heavy ions, impact ionization is triggered, generating new EHPs. (c) The uniformly distributed hole pillars and the auxiliary gate depleted the channel, leading to transient current decay. (d) Substrate holes leave the region beneath the gate and are reduced through recombination, while channel holes accumulate at the gate interface, elevating the gate potential and maintaining the conductive channel until thermal burnout occurs. The thermal burnout in the final stage has been identified as the core process of its SEB. Based on this, we propose a source-connected Schottky body contact (SC-SBC) structure to extract excess holes under the gate in order to verify the SEB mechanism and attempt device hardening design. Simulation results demonstrate that this method effectively drains accumulated gate-region holes, suppresses gate potential elevation, and prevents SEB. The results also verified that the rise in gate potential due to hole accumulation beneath the gate is the fundamental mechanism responsible for SEB in lateral E-mode β-Ga2O3 MOSFETs.
- 6 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (2) , 348-354
Degradation Patterns of Total Ionizing Dose Effects in SiGe HBTs Based on Forward and Inverse Operations
This paper investigates the performance degradation pattern of total ionizing dose (TID) effect in SiGe HBTs through Co-60 γ ray and 1-MeV electron irradiation experiments. The results demonstrate that forward excess base current (ΔIB) consistently exhibits super-linear degradation, while the degradation pattern of inverse ΔIB shows dependence on the irradiation source. The increase in current stems from two interacting factors: enhanced surface recombination velocity and depletion region expansion. In forward operation, both interface trap and oxide trap charge exhibit the expansion effect on depletion region. However, in inverse operation, interface trap and oxide trap charge exhibit expansion and contraction effects on depletion region, respectively. Consequently, varying generation ratios of ionization defects caused by different experimental conditions may not change the forward degradation pattern, but will change the inverse degradation pattern. Subsequently, the simulation results of various ratios and the experimental results of hydrogen pre-treatment further confirmed the relationship between the ratios of ionization defects and the degradation patterns.
- 6 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 73 (2) , 348-354
Degradation Patterns of Total Ionizing Dose Effects in SiGe HBTs Based on Forward and Inverse Operations
This paper investigates the performance degradation pattern of total ionizing dose (TID) effect in SiGe HBTs through Co-60 γ ray and 1-MeV electron irradiation experiments. The results demonstrate that forward excess base current (ΔIB) consistently exhibits super-linear degradation, while the degradation pattern of inverse ΔIB shows dependence on the irradiation source. The increase in current stems from two interacting factors: enhanced surface recombination velocity and depletion region expansion. In forward operation, both interface trap and oxide trap charge exhibit the expansion effect on depletion region. However, in inverse operation, interface trap and oxide trap charge exhibit expansion and contraction effects on depletion region, respectively. Consequently, varying generation ratios of ionization defects caused by different experimental conditions may not change the forward degradation pattern, but will change the inverse degradation pattern. Subsequently, the simulation results of various ratios and the experimental results of hydrogen pre-treatment further confirmed the relationship between the ratios of ionization defects and the degradation patterns.
- 31 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 72 (12) , 3695-3701
Gate Oxide Breakdown in P-Channel Split-Gate Trench VDMOS Induced by Total Ionizing Dose Effects Under High Gate Bias
This study elucidates the permanent gate oxide breakdown mechanism in P-channel Split-gate Trench Vertical Double-Diffused MOSFETs (P-channel SGT VDMOS) under the synergistic effect of 60Co-γ irradiation and elevated gate bias (VG), representing a critical reliability concern previously less explored in aerospace power device research. Specifically, concurrent exposure to VG of 15 V and 60Co-γ irradiation at total ionizing dose (TID) of 200 krad with a dose rate of 50 rad/s was analyzed to reveal the underlying breakdown mechanism. This failure manifests through abrupt gate current surges exceeding 100 mA, along with a threshold voltage (Vth) shift beyond the normal operating range. Experimental and simulation analyses confirm that elevated VG accelerates both the accumulation of oxide charges (Not) generated by irradiation accumulation and the generation of interface traps (Nit). This process generates electric field distortion surpassing SiO2 critical breakdown strength of 1×107 V/cm. Crucially, the post-irradiation immediate measurement (PIM) demonstrates that the superposition of applied bias fields and radiation-induced electric fields establishes percolation pathways within the gate oxide. Under combined stress conditions of a 15 V VG and 200 krad TID, these pathways trigger permanent gate oxide breakdown. These findings provide essential design guidance for suppressing gate oxide failure in aerospace power systems.
- 31 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 72 (12) , 3695-3701
Gate Oxide Breakdown in P-Channel Split-Gate Trench VDMOS Induced by Total Ionizing Dose Effects Under High Gate Bias
This study elucidates the permanent gate oxide breakdown mechanism in P-channel Split-gate Trench Vertical Double-Diffused MOSFETs (P-channel SGT VDMOS) under the synergistic effect of 60Co-γ irradiation and elevated gate bias (VG), representing a critical reliability concern previously less explored in aerospace power device research. Specifically, concurrent exposure to VG of 15 V and 60Co-γ irradiation at total ionizing dose (TID) of 200 krad with a dose rate of 50 rad/s was analyzed to reveal the underlying breakdown mechanism. This failure manifests through abrupt gate current surges exceeding 100 mA, along with a threshold voltage (Vth) shift beyond the normal operating range. Experimental and simulation analyses confirm that elevated VG accelerates both the accumulation of oxide charges (Not) generated by irradiation accumulation and the generation of interface traps (Nit). This process generates electric field distortion surpassing SiO2 critical breakdown strength of 1×107 V/cm. Crucially, the post-irradiation immediate measurement (PIM) demonstrates that the superposition of applied bias fields and radiation-induced electric fields establishes percolation pathways within the gate oxide. Under combined stress conditions of a 15 V VG and 200 krad TID, these pathways trigger permanent gate oxide breakdown. These findings provide essential design guidance for suppressing gate oxide failure in aerospace power systems.
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 72 (10) , 3364-3374
Physics-Constrained Neural Network-Based Compact Modeling of Displacement Damage for AlGaN/GaN HEMTs
This paper proposes a physics-constrained neural network (PCNN) framework to model displacement damage-induced degradation in AlGaN/GaN HEMTs. The architecture combines feed-forward neural networks (FFNN) with Monotonic Block to enforce the physical correlations between bias voltages, radiation fluence, and drain-source current degradation. A hybrid optimization strategy integrates data-driven learning with differentiable physics regularizations, incorporating analytical constraints based on displacement damage mechanisms to guide training. This ensures both accurate interpolation within the experimental fluence range and physically consistent extrapolation beyond it. Smoothing operators are introduced to enhance the continuity of predicted device characteristics. Experimental validation using 7.6 MeV carbon-ion and 30 MeV fluorine-ion irradiation confirms that the model achieves high accuracy (RMSE < 0.055%) and reliably predicts key physical parameters within the training range (RMSE < 6.5%). Notably, the model maintains reasonable physical trends even at untrained fluence levels. By integrating machine learning with device physics, the proposed framework provides a compact, generalizable approach for modeling radiation effects, offering critical insights for radiation-hardened design and lifetime prediction of GaN-based MMICs in space applications.
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nuclear Science
- Vol. 72 (10) , 3364-3374
Physics-Constrained Neural Network-Based Compact Modeling of Displacement Damage for AlGaN/GaN HEMTs
This paper proposes a physics-constrained neural network (PCNN) framework to model displacement damage-induced degradation in AlGaN/GaN HEMTs. The architecture combines feed-forward neural networks (FFNN) with Monotonic Block to enforce the physical correlations between bias voltages, radiation fluence, and drain-source current degradation. A hybrid optimization strategy integrates data-driven learning with differentiable physics regularizations, incorporating analytical constraints based on displacement damage mechanisms to guide training. This ensures both accurate interpolation within the experimental fluence range and physically consistent extrapolation beyond it. Smoothing operators are introduced to enhance the continuity of predicted device characteristics. Experimental validation using 7.6 MeV carbon-ion and 30 MeV fluorine-ion irradiation confirms that the model achieves high accuracy (RMSE < 0.055%) and reliably predicts key physical parameters within the training range (RMSE < 6.5%). Notably, the model maintains reasonable physical trends even at untrained fluence levels. By integrating machine learning with device physics, the proposed framework provides a compact, generalizable approach for modeling radiation effects, offering critical insights for radiation-hardened design and lifetime prediction of GaN-based MMICs in space applications.