| Abstrakt: |
The Scaling of semiconductor devices is the major demand in the world for handy and portable systems. This brings a very big change in the CMOS industry and technology. As we scalethe CMOS beyond its limits, it results in some major problems such as high leakage current, complex circuits, high power dissipation which limits the performance of the device. The purpose of this research work is to divert from operating at high frequency to deliver better and great performance in essence with low power. The CMOS based memories consume more power in modern processor. So, to design the memory for Artificial intelligence and neuromorphic computing the major demand of these applications is efficient device with great thermal stability, better TMR ratio, high density, non-volatility, low leakage and easy to scale down in future. Symica tool is the EDA tool used to design the device with different parameters and helps to calculate various performance parameters. In this paper the device level parameters are discussed and compared with the previously reported data. The proposed design provides better performance than the reported data. The 32 nm node technology is used to design the STT-MRAM.The results show, great thermal stability with less resistances and better TMR ratio. To design high performance STT-MTJ based MRAM, attention is toward the device level parameters. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink