-
重离子辐照实验所用的器件为22 nm全耗尽型绝缘体上硅(Fully Depleted Silicon on Insulator, FDSOI)的测试片。芯片和测试板如图1所示,实验被测芯片为22 nm FDSOI工艺的6T SRAM。SRAM器件的存储容量为128 K,具有13个地址位和16个数据位。图1(a)所示为辐照实验前已开帽的SRAM器件,开帽可以保证入射离子能够到达灵敏区。实验前对器件进行了功能测试,以确保待测芯片能够在标准电压以及高低拉偏电压下正常工作。
-
本工作基于中国科学院近代物理研究所HIRFL加速器和北京原子能研究院HI-13串列加速器单粒子效应实验终端,利用不同种类重离子对该款器件进行了单粒子效应实验,实验条件及参数如表1所列。如图2所示, DUT(Device under Test)在HIRFL的TR5单粒子效应测试终端的大气辐照平台进行辐照测试。原子能研究院HI-13串列加速器辐照实验时,DUT被放置于真空罐中进行辐照测试。测试系统每次辐照前都要进行复位和写入数据图形,开始读操作后测试系统不断地循环读取存储阵列的数据并与预设数据图形比对,当发现错误事件后,系统会自动记录相应的错误地址和数据等信息。实验中实时观察器件工作电流,若达到限流,立即断电并记录该实验条件。
表 1 SRAM实验条件及参数
实验条件 实验参数 角度/(°) α = 0,90;β = 0,30,60 核心电压/V 0.54,0.72,0.80,0.90,1.08,1.26 LET/(MeV·cm2·mg−1) 8.7,13.4,22.2,37.4,75.4,78.1 在HIRFL加速器开展辐照实验时,我们利用能量为16 MeV/u的181Ta离子开展重离子辐照实验,研究被测器件饱和区的SEU敏感性。我们使用SRIM-2013详细地计算了重离子参数,确定获得每个能量条件所需的铝箔降能片和空气层厚度,通过调整被测器件和离子束流出窗口之间的铝箔降能片和空气层厚度以获得能量较低、LET较大的实验条件。原子能研究院HI-13加速器开展辐照实验时,我们选用了四种LET较低的重离子进行实验,实验中通过改变重离子类型来改变LET。
辐照前先让测试系统进入正常工作状态,辐照开始后翻转数达到400个或重离子总注量达到1×107 ions/cm2时,停止辐照并记录数据,以保证实验数据具有统计意义。
-
如表2所列,我们在HIRFL和HI-13进行了不同LET的重离子辐照实验,重离子射程范围从30到116.6 μm,所有离子射程都能到达器件的灵敏区。除了空气层厚度为70 mm的Ta离子实验进行了入射角度的改变,其余实验均使重离子沿器件表面的法线方向入射。在I/O电压1.62 V、核心电压0.72 V条件下研究LET较低时器件SEU截面和MCU比例随重离子LET增加的变化趋势,在I/O电压1.8 V、核心电压0.9 V条件下研究高LET下器件的SEU截面和MCU比例随入射角度改变的变化趋势。如图3所示,当离子入射方位角α分别为0°和90°,离子将沿芯片的两个互相垂直的方向入射,研究离子入射方位角对器件SEU敏感性的影响;随后,改变这两个方向下离子入射倾角β的大小,研究离子入射方向和芯片表面法线所成角度对SEU敏感性的影响。
表 2 重离子辐照实验参数
离子
种类空气
厚度/mm铝箔
厚度/μm能量/MeV LET/
(MeV·cm2·mg-1)Si中
射程/μmTa 40 0 2 006.4 75.4 116.6 70 0 1 721.4 78.1 100.7 Al 0 0 110.0 8.7 48.0 Cl 0 0 150.0 13.4 42.8 Ti 0 0 160.0 22.2 32.9 Ge 0 0 205.0 37.4 30.0 -
单粒子测试系统记录了出现SEU的逻辑地址和错误数据,我们根据在同一时间范围内的翻转数量和逻辑位图,对不同翻转位数的MCU的数量进行了统计分析和研究。计算并研究了比特截面(SEU bit cross section)、事件截面(SEU event cross section)以及MCU比例。计算方式如下。
比特截面:
$$ \begin{split} {\sigma }_{\mathrm{U}-\mathrm{S}\mathrm{E}\mathrm{U}}=& \sum _{i=1}^{\mathrm{\infty }}\frac{i\times {E}_{i-\mathrm{b}\mathrm{i}\mathrm{t}}}{\varPhi \times \mathrm{c}\mathrm{o}\mathrm{s}\beta } \\ =&\frac{{1\times E}_{1-\mathrm{b}\mathrm{i}\mathrm{t}}+{2\times E}_{2-\mathrm{b}\mathrm{i}\mathrm{t}}+{3\times E}_{3-\mathrm{b}\mathrm{i}\mathrm{t}}+\dots }{\varPhi \times \mathrm{c}\mathrm{o}\mathrm{s}\beta }, \end{split}$$ (1) 事件截面:
$$ \begin{split} {\sigma }_{\mathrm{E}-\mathrm{S}\mathrm{E}\mathrm{U}}=& \sum _{i=1}^{\mathrm{\infty }}\frac{{E}_{i-bit}}{\varPhi \times \mathrm{c}\mathrm{o}\mathrm{s}\beta } \\ =&\frac{{E}_{1-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{2-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{3-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{4-\mathrm{b}\mathrm{i}\mathrm{t}}+\dots }{\varPhi \times \mathrm{c}\mathrm{o}\mathrm{s}\beta }, \end{split} $$ (2) 其中:i是一个SEU事件中受到影响的单元数;
$ {E}_{i-\mathrm{b}\mathrm{i}\mathrm{t}} $ 是一个事件中的翻转单元数;Ф是重离子注量和器件存储容量的乘积,如图3所示入射倾角β是重离子入射方向与芯片表面法线的夹角。计算MCU比例的公式如下所示:$$ {P}_{n-\mathrm{b}\mathrm{i}\mathrm{t}} = \frac{{E}_{n-\mathrm{b}\mathrm{i}\mathrm{t}}}{{E}_{1-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{2-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{3-\mathrm{b}\mathrm{i}\mathrm{t}}+{E}_{4-\mathrm{b}\mathrm{i}\mathrm{t}}+\dots }, $$ (3) 单粒子事件数据分布通常为高斯(正态)分布;N个事件的标准偏差为N1/2;百分比标准偏差为N1/2/N=1/N1/2。
-
摘要: 全耗尽绝缘体上硅(FDSOI)工艺是制备高可靠宇航电子器件的理想半导体工艺,因此深入揭示FDSOI工艺器件的单粒子效应机理对抗辐射加固设计具有理论指导意义。针对22 nm FDSOI SRAM测试器件,研究了不同重离子及电学参数对器件单粒子翻转敏感性的影响规律及物理机制。实验结果表明,高LET值区域单粒子多单元翻转事件占比可达20%,且核心电压对单粒子翻转类型比例及发生概率影响较小;重离子倾角入射会显著增大器件的单粒子翻转截面,且重离子沿平行与垂直衬底阱区方向入射时的单粒子翻转截面差异可达130%。因此,FDSOI器件单粒子效应建模及抗辐射加固设计,必须考虑非直接扩散型电荷共享机制、衬底电势畸变触发寄生电流机制对单粒子瞬态电离电荷收集过程的影响。Abstract: The Fully Depleted Silicon on Insulator(FDSOI) process is considered an ideal semiconductor technology for producing highly reliable aerospace electronic devices. Therefore, a comprehensive understanding of the single event effects mechanism in FDSOI devices is of theoretical significance for radiation-hardened design. This paper focuses on 22 nm FDSOI SRAM test devices and investigates the impact patterns and physical mechanisms of different heavy ions and electrical parameters on the sensitivity of Single Event Upset(SEU) in the devices. Experimental results indicate that in regions with high Linear Energy Transfer(LET) values, the proportion of Multi-Cell Upset(MCU) can reach 20%. Additionally, the core voltage has a relatively minor impact on the type proportion and occurrence probability of SEU. The incidence angle of heavy ions significantly increases the SEU cross-section of the devices, with a 130% difference observed when heavy ions are incident along parallel and perpendicular directions to the substrate well region. Therefore, when modeling Single Event Effect in FDSOI devices and designing for radiation hardening, it is imperative to consider the influence of non-direct diffusion charge sharing mechanisms and substrate potential distortion-triggered parasitic current mechanisms on the transient ionization charge collection process.
-
Key words:
- single event effect /
- heavy ions /
- multi-cell Upset /
- FDSOI /
- SRAM /
- angular incidence
-
表 1 SRAM实验条件及参数
实验条件 实验参数 角度/(°) α = 0,90;β = 0,30,60 核心电压/V 0.54,0.72,0.80,0.90,1.08,1.26 LET/(MeV·cm2·mg−1) 8.7,13.4,22.2,37.4,75.4,78.1 表 2 重离子辐照实验参数
离子
种类空气
厚度/mm铝箔
厚度/μm能量/MeV LET/
(MeV·cm2·mg-1)Si中
射程/μmTa 40 0 2 006.4 75.4 116.6 70 0 1 721.4 78.1 100.7 Al 0 0 110.0 8.7 48.0 Cl 0 0 150.0 13.4 42.8 Ti 0 0 160.0 22.2 32.9 Ge 0 0 205.0 37.4 30.0 -
[1] BARILLOT C, CALVEL P. IEEE Transactions on Nuclear Science, 1996, 43(2): 453. doi: 10.1109/23.490914 [2] BINDER D, SMITH E C, HOLMAN A B. IEEE Transactions on Nuclear Science, 1975, 22(6): 2675. doi: 10.1109/TNS.1975.4328188. [3] ECOFFET R. IEEE Transactions on Nuclear Science, 2013, 60(3): 1791. doi: 10.1109/TNS.2013.2262002. [4] HARBOE-SORENSEN R, POIVEY C, ZADEH A, et al. IEEE Transactions on Nuclear Science, 2012, 59(4): 1086. doi: 10.1109/TNS.2012.2185062 [5] SIMOEN E, GAILLARDIN M, PAILLET P, et al. IEEE Trans Nucl Sci, 2013, 60(3 Part2): 1970. doi: 10.1109/TNS.2013.2255313 [6] VITALE S, WYATT P, CHECKA N, et al. Proceedings of the IEEE, 2010, 98(2): 333. doi: 10.1109/JPROC.2009.2034476 [7] GASIOT G, SOUSSAN D, GLORIEUX M, et al. SER/SELPerformances of SRAMs in UTBB FDSOI28 and Comparisons with PDSOI and BULK Counterparts[C]//IEEE International Reliability Physics Symposium, New York: IEEE, 2014. [8] ASPRILLA A, CORDOVA D, DEVAL Y, et al. IEEE Transactions on Circuits and Systems II Express Briefs, 2021, 68(2): 602. doi: 10.1109/TCSII.2020.3047464 [9] SCHWANK J R, FERLET-CAVROIS V, SHANEYFELT M R, et al. IEEE Transactions on Nuclear Science, 2003, 50(3): 522. doi: 10.1109/TNS.2003.812930 [10] DAVIS G E, HITE L R, BLAKE T, et al. IEEE Transactions on Nuclear Science, 1985, 32(6): 4431. doi: 10.1109/TNS.1985.4334137 [11] MUSSEAU O. IEEE Transactions on Nuclear Science, 1996, 43(2): 603. doi: 10.1109/23.490904 [12] GASIOT G, SOUSSAN D, GLORIEUX M, et al. SER/SEL Performances of SRAMs in UTBB FDSOI 28nm and Comparisons with PDSOI and BULK Counterparts[C]//IEEE International Reliability Physics Symposium. New York: Institute of Electrical and Electronics Engineer Inc, 2014 [13] CAI Chang, ZHAO Peixiong, XU Liwei, et al. Microelectronics Reliablity, 2019, 100-101: 113322. doi: 10.1016/j.microrel.2019.06.014 [14] MALHERBE V, GASIOT G, SOUSSAN D, et al. Alpha Soft Error Rate of FDSOI 28 nm SRAMs: Experimental Testing and Simulation Analysis[C]. New York: IEEE, 2015. [15] HEIDEL DAVID F, MARSHALL PAUL W, PELLISH JONATHAN A. et al. IEEE Transactions on Nuclear Science, 2009, 56(6): 3499. doi: 10.1109/TNS.2009.2033796 [16] AKKERMAN A, BARAK J. Nuclear Science IEEE Transactions on, 2002, 49(6): 3022. doi: 10.1109/TNS.2002.805366 [17] GASPARD N J, JAGANNATHAN S, DIGGINS Z J, et al. IEEE Transactions on Nuclear Science, 2013, 60(6 Part1): 4368. doi: 10.1109/TNS.2013.2289745 [18] DENNARD R H, GAENSSLEN F H, YU H N, et al. IEEE Solid-State Circuits Society Newsletter, 2007, 12(1): 38. doi: 10.1109/N-SSC.2007.4785543 [19] RAINE M, HUBERT G, GAILLARDIN M, et al. IEEE Transactions on Nuclear Science, 2011, 58(6): 2607. doi: 10.1109/TNS.2011.2168238 [20] RAINE M, GAILLARDIN M, LAGUTERE T, et al. IEEE Transactions on Nuclear Science, 2017, 65(1): 339. doi: 10.1109/TNS.2017.2779786 [21] 赵凯, 高见头, 杨波, 等. 信息与电子工程, 2010, 8(01): 91. doi: cnki:sun:xxyd.0.2010-01-023 ZHAO Kai, GAO Jiantou, YANG Bo, et al. Information and Electronic Engineering, 2010, 8(01): 91. (in Chinese) doi: cnki:sun:xxyd.0.2010-01-023 [22] RAINE M, HUBERT G, GAILLARDIN M, et al. IEEE Transactions on Nuclear Science, 2011, 58(3): 840. doi: 10.1109/TNS.2011.2109966 [23] RAINE M, VALENTIN A, GAILLARDIN M, et al. IEEE Transactions on Nuclear Science, 2012, 59(6): 2697. doi: 10.1109/TNS.2012.2220783 [24] 贺泽. 纳米存储器件和组合逻辑器件的单粒子效应研究[D]. 兰州: 中国科学院近代物理研究所, 2022. HE Ze. Study of Single Event Effects on Nanoscale Memory Device and Combinational Logic Device[D]. Lanzhou: Institute of Modern Physics, Chinese Academy of Science, 2022. (in Chinese) [25] 莫莉华. 新型FinFET器件和3D堆叠器件的单粒子效应研究[D]. 兰州: 中国科学院近代物理研究所, 2021. MO Lihua. Investigation of Single Event Effects in New FinFET Device and 3D Stacked Device[D]. Lanzhou: Institute of Modern Physics, Chinese Academy of Science, 2021.(in Chinese) [26] 蔡畅. 纳米SRAM型FPGA的单粒子效应及其加固技术研究[D]. 兰州: 中国科学院近代物理研究所, 2021. CAI Chang. Study of Single Event Effects and Hardening Strategies on Nanoscale SRAM-based FPGA[D]. Lanzhou: Institute of Modern Physics, Chinese Academy of Science, 2021.(in Chinese) [27] ROCHE P, GASIOT G, FORBES K, et al. IEEE Transactions on Nuclear Science, 2003, 50(6): 2046. doi: 10.1109/TNS.2003.821588 [28] MERELLE T, SAIGNE F, SAGNES B, et al. IEEE Transactions on Nuclear Science, 2005, 52(5): 1538. doi: 10.1109/tns.2005.855823 [29] TIPTON A D, PELLISH J A, HUTSON J M, et al. IEEE Transactions on Nuclear Science, 2008, 55(6): 2880. doi: 10.1109/TNS.2008.2006503 [30] GIOT D, ROCHE P, GASIOT G, et al. IEEE Transactions on Nuclear Science, 2007, 54(4): 904. doi: 10.1109/tns.2007.902360