隨著科技的進步，人工智慧及物聯網越來越深入日常生活，此時能夠高速處理資料和執行複雜運算的高效能運算晶片變得日益重要，如何降低耗能也成為另一個重要且迫切的問題。在低溫下操作元件能夠改善亞閾值擺幅，不僅降低功耗還可以提升運算表現。
本篇研究首先建立一個參數萃取平台，以目前業界通用的鰭式電晶體元件模型BSIM-CMG為基礎，由於此模型並沒有加入低溫的效應，因此我們考慮亞閾值擺幅在低溫會接近飽和的趨勢、分析低溫下的閾值電壓和遷移率、統整使用到的參數及其對模型的意義、討論目前BSIM-CMG沒辦法同時吻合低溫下次臨界區和反轉區的閾值電壓並提出兩種想法去解決，一個是調整擴散電流，另一個則是調整平帶電壓的值。最後針對兩個閘極電壓範圍，分別提出可以同時吻合300K、210K、160K、110K、77K、36K的參數組合。

With the advancement of technology, Artificial Intelligence and the Internet of Things are becoming increasingly deeply rooted in our daily lives. High performance computing chips which process data at high speed and perform complex calculations are becoming increasingly crucial. The way of reducing power consumption has become another important and urgent issue. Operating devices at low temperatures improves subthreshold swing which reduce the power consumption and enhance the computing performance.
In this research, we first establish a parameter extraction platform based on the BSIM-CMG, which is a common FinFET model in the industry. Yet this model does not incorporate the effect of low temperature, we consider the trend of subthreshold swing saturation at low temperature. Then we analyze the threshold voltage and mobility at low temperature, and sort the using parameters and their effects to the model. We discuss the reason why the current BSIM-CMG is unable to match the threshold voltage of the sub-threshold region and strong inversion region simultaneously at cryogenic temperatures. In order to solve this problem, we propose two ideas. One is by adjusting diffusion current. The other is by modifying the flat-band voltage. In the end, focusing on the separate ranges of gate voltage, we propose parameter sets that can simultaneously fit at 300K, 210K, 160K, 110K, 77K, and 36K.

[1] Jean-Gabriel Boinot-Tramoni and Marie Gruet, “A Beginner’s Guide to High Performance Computing,” Sep 10, 2020. [Online]. Available: https://medium.com/quantonation/a-beginners-guide-to-high-performance-computing-ae70246a7af
[2] Precedence Research , “High Performance Computing Market Size to Surpass USD 64.65 Bn by 2030,” London, April 04, 2022. [Online]. Available: https://www.globenewswire.com/en/news-release/2022/04/04/2415844/0/en/High-Performance-Computing-Market-Size-to-Surpass-USD-64-65-Bn-by-2030.html
[3] Tryggvi Adalbjornsson, “Iceland’s data centers are booming—here’s why that’s a problem,” June 18, 2019. [Online]. Available: https://www.technologyreview.com/2019/06/18/134902/icelands-data-centers-are-booming-heres-why-thats-a-problem/
[4] Jason Woo, “Develop a low-temperature (77K) device technology to achieve a 25X improvement in performance/power compared to state-of-the-art (SOA) room temperature CPUs,” DARPA, 2021
[5] S. K. Singh, S. Gupta, R. A. Vega and A. Dixit, “Accurate Modeling of Cryogenic Temperature Effects in 10-nm Bulk CMOS FinFETs Using the BSIM-CMG Model,” in IEEE Electron Device Letters, vol. 43, no. 5, pp. 689-692, May 2022, doi: 10.1109/LED.2022.3158495.
[6] P. Kuthe, M. Müller and M. Schröter, "VerilogAE: An Open Source Verilog-A Compiler for Compact Model Parameter Extraction," in IEEE Journal of the Electron Devices Society, vol. 8, pp. 1416-1423, 2020, doi: 10.1109/JEDS.2020.3023165.
[7] The SciPy community. SciPy Reference Guide, 1.8.1. pp. 1329-1332. (2022). [Online]. Available: https://docs.scipy.org/doc/scipy/scipy-ref-1.8.1.pdf
[8] Yogesh Singh Chauhan, Darsen Lu, Sriramkumar Venugopalan, Sourabh Khandelwal, Juan Pablo Duarte, Navid Paydavosi, Ai Niknejad and Chenming Hu, “BSIM-CMG model parameter extraction,” in FinFET Modeling for IC Simulation and Design: Using the BSIM-CMG Standard, first ed. USA: Academic Press, 2015.
[9] Harshit Agarwal, Pragya Kushwaha, Avirup Dasgupta and Ming-Yen Kao. BSIM-CMG 111.0.0 Multi-Gate MOSFET Compact Model Technical Manual. (2019). [Online]. Available: https://bsim.berkeley.edu/models/bsimcmg/
[10] A. Beckers, F. Jazaeri and C. Enz, “Theoretical Limit of Low Temperature Subthreshold Swing in Field-Effect Transistors,” in IEEE Electron Device Letters, vol. 41, no. 2, pp. 276-279, Feb. 2020, doi: 10.1109/LED.2019.2963379.
[11] Electrical4U. “Fermi Dirac Distribution Function,” October 26, 2020. [Online]. Available: https://www.electrical4u.com/fermi-dirac-distribution-function/
[12] A. Beckers, F. Jazaeri, A. Grill, S. Narasimhamoorthy, B. Parvais and C. Enz, “Physical Model of Low-Temperature to Cryogenic Threshold Voltage in MOSFETs,” in IEEE Journal of the Electron Devices Society, vol. 8, pp. 780-788, 2020, doi: 10.1109/JEDS.2020.2989629.
[13] A. Ortiz-Conde, F.J. Garcı́a Sánchez, J.J. Liou, A. Cerdeira, M. Estrada and Y. Yue, “A review of recent MOSFET threshold voltage extraction methods,” in Microelectronics Reliability, vol. 42, pp. 583-596, 2002, doi: 10.1016/S0026-2714(02)00027-6.
[14] S. Takagi, A. Toriumi, M. Iwase and H. Tango, “On the universality of inversion layer mobility in Si MOSFET's: Part I-effects of substrate impurity concentration,” in IEEE Transactions on Electron Devices, vol. 41, no. 12, pp. 2357-2362, Dec. 1994, doi: 10.1109/16.337449.
[15] Harshit Agarwal, Pragya Kushwaha, Avirup Dasgupta and Ming-Yen Kao. (2019) BSIM-CMG code (Version 111.0.0). https://bsim.berkeley.edu/models/bsimcmg/
[16] Arnout Beckers, Farzan Jazaeri, Heorhii Bohuslavskyi, Louis Hutin, Silvano De Franceschi and Christian Enz, “Characterization and modeling of 28-nm FDSOI CMOS technology down to cryogenic temperatures,” in Solid-State Electronics, vol. 159, pp. 106-115, 2019, doi: 0.1016/j.sse.2019.03.033.