研究目的
To investigate the total ionizing dose radiation responses of 55-nm SONOS memory cells in pulse and dc modes, focusing on threshold voltage shifts, memory window changes, and off-state current variations, and to analyze the underlying charge loss mechanisms.
研究成果
The memory cells operated in pulse mode exhibit better radiation-hard capability compared to those in dc mode, with normalized memory window retention of 76% versus 60% after 300 krad(Si) radiation. This is attributed to fewer tunnel oxide traps generated in pulse mode, which reduces charge loss during radiation. The findings suggest that TID evaluations using dc mode may overestimate degradation, and improvements in tunnel oxide quality or nitride trap distribution could enhance radiation hardness.
研究不足
The specific thicknesses of the ONO stack layers are not disclosed due to commercial confidentiality. The study is limited to 55-nm technology nodes and may not generalize to other nodes or memory types. Radiation effects are evaluated only up to 300 krad(Si), and long-term or higher dose effects are not explored. The mechanisms are analyzed based on existing physical models, but direct measurement of trap densities or interface states is not performed.
1：Experimental Design and Method Selection:
The study compares the radiation effects on SONOS memory cells operated in pulse mode (using pulse voltage for programming/erasing) and dc mode (using direct voltage sweeping). Electrical characterizations are performed using a semiconductor device analyzer, and radiation experiments are conducted with a 60Co gamma source.
2：Sample Selection and Data Sources:
The devices under test are 2x2-bit NOR-type flash memory mini-arrays fabricated using a 55-nm CMOS logic technology node. Each cell consists of a memory transistor (SONOS device) and a select transistor.
3：List of Experimental Equipment and Materials:
Agilent B1500A semiconductor device analyzer, 60Co gamma radiation source, transmission electron microscope (TEM) for cross-sectional imaging.
4：Experimental Procedures and Operational Workflow:
Memory cells are programmed/erased using Fowler-Nordheim tunneling with specific voltage conditions (7 V/-7 V for pulse mode with 2 ms/6 ms durations, and 7 V/-7 V for dc mode). Id-Vg curves are measured before and after radiation doses of 30, 100, and 300 krad(Si) at a dose rate of 50 rad(Si)/s. No bias is applied during radiation.
5：Data Analysis Methods:
Threshold voltage (Vth) is defined at a drain current of 1 μA, memory window (MW) as the Vth difference between programmed and erased states, and off-state current (Ioff) at zero gate bias. Data are analyzed to compare degradation between modes.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
