不同鰭寬度之鰭狀場效電晶體的低頻雜訊與可靠度分析
陳旼志＊ 方炎坤＊＊
國立成功大學微電子工程研究所
摘要
本論文主要研究方向為針對不同元件尺寸之鰭狀場效電晶體的電性量測以及可靠度分析。本論文所使用的元件是為使用二氧化鉿高介電常數材料作為閘極氧化層，另外使用氮化鈦當作金屬閘極的鰭狀場效電晶體。研究將進行基本電性分析以及可靠度分析。其基本電性部分包括臨界電壓(Vth)、次臨界擺幅(SS)，汲極引發能障下降(DIBL)、驅動電流(ID,sat)、串聯電阻(RSD)以及低頻雜訊(LFN)；可靠度部分包括正/負偏壓不穩定性(P/NBTI)以及熱載子量測(HCI)，其中熱載子量測我們分成原始熱載子可靠度分析以及考慮串聯電阻效應後之定通道電場熱載子可靠度分析，而我們也利用低頻雜訊量測來評估可靠度量測前後閘極氧化層內缺陷的變化。
基本電性，在臨界電壓的部分，n型元件之臨界電壓隨著通道長度微縮呈現先上升後下降的反向短通道效應(Reverse Short Channel Effect)，而p型元件則是隨通道微縮而臨界電壓跟著下降之典型的短通道效應。次臨界擺幅、汲極引發能障下降以及驅動電流則都是隨著元件通道長度的微縮而呈現變大的趨勢。而鰭寬度較寬之元件，其臨界電壓比鰭寬度較窄之元件要大。而在次臨界擺幅以及汲極引發能障下降方面，鰭寬度較窄之元件其閘極控制通道電荷的能力較佳，且鰭體內漏電流的路徑也比較少，以至於能夠得到比較好的改善。雖然鰭寬度較窄的元件能夠提供比較好的短通道效應抑制，但其元件驅動電流以及源極/汲極寄生電阻卻沒有鰭寬度較寬的元件之特性佳。元件尺寸持續微縮所帶來的短通道效應日益嚴重，雖說可以透過鰭狀場效電晶體之其寬度微縮來達到有效的改善，但鰭寬度微縮到某個限度之後，量子效應對於元件的影響也會越來越明顯，元件尺寸的微縮造成的量子侷限(Structure Confinement)之影響勢必也得納入考慮。
可靠度量測，正/負偏壓不穩定性的部分，鰭寬度較寬之元件，較不受量子效應影響，使得載子傾向於表面反轉，當載子於表面反轉，則較容易與氧化層內缺陷交互作用而產生元件特性退化；鰭寬度較窄之元件由於受到量子效應之影響，載子於鰭體內反轉(volume inversion)，載子較不易與氧化層缺陷交互作用使得元件退化較不嚴重，而在偏壓不穩定性前後皆以低頻雜訊分析發現，輸入參照電壓頻譜曲線斜率，在鰭寬度較寬的元件有增加的趨勢；而鰭寬度較窄的元件則受到量子效應的影響，曲線斜率較無明顯變化。熱載子分析方面，以往傳統的量測方式是固定閘極與汲極端電壓，而此量測方法得出的結果是鰭寬度較寬的元件退化特性較明顯。但此種量測方法無考慮到源極/汲極寄生電阻效應，使得實際跨在通道的電場不相等，而當我們萃取了不同鰭寬度之元件寄生電阻後，經過計算，固定了不同鰭寬度的通道電場，發現鰭寬度較窄的元件因為受到量子效應的影響，載子於鰭體內反轉，而汲極端靠近鰭體內的空乏區也較大，使得載子流到汲極端造成較嚴重的撞擊游離使得元件退化較嚴重。

Studies of reliability and low-frequency-noise of the FinFETs with different fin widths

Author : Min-Chih Chen
Advisors : Professor Yean-Kuen Fang
Professor Wen-Kuan Yeh
Professor Cheng-Li Lin
Electrical Engineering & Institute of Microelectrinics

SUMARY

In this thesis, we mainly measured the basic electrical characteristics and reliability items including bias temperature instability and hot carrier effect. We also used the low-frequency-noise system to analyze the devices characteristics before and after the reliability tests. For the basic electrical characteristics, the thinner fin device has a better improvement in short channel effect but poor driving ability than that of the thicker device. Besides, the thinner fin device also has a higher quantum confinement effect, which gathers the inversion carriers toward the center of the fin body instead of the interface between the dielectric layer and the fin body, and thus called volume inversion. Based on this phenomenon, the thinner fin devices would suffer less characteristic degradations; because the inversion carriers gathering at the center of the fin body will become difficultly interact with defects in the dielectric layer. The thinner fin device suffers more characteristics degradation for its larger area of the depletion region in the center of the fin body. Besides, the volume inversion in a thinner fin device make it not significant coulomb scattering as found in the thicker fin device.

Key words: bias temperature instability, hot carrier effect, quantum confinement, low-frequency-noise

INTRODUCTION

FinFET is one of the most promising candidates for the next generation for its good improvement of short channel effect. Shrinking the fin width can improve the short channel effect, but the quantum confinement effect will be arisen. Although there were many simulation results showed the quantum confinement effect previously, but there are few measured data to prove this effect. In this thesis, we used low-frequency-noise, bias temperature instability and hot carrier test to analyze the different fin width of the devices characteristics.

MATERIALS AND METHODS

The samples we used were the bulk-silicon FinFET with the HfO2 as the dielectric layer and TiN as the metal gate. The size of the fin widths are 10 nm and 25 nm. The most critical gate length is 40 nm. The fin height is 30 nm. Followings are our measurement conditions for the n-type devices. For measuring ID-VG, the gate was added the ramp voltage from -1V to 1.5 V. The drain was fixed at voltage of 0.05 V and 1 V for devices working at the linear region and saturation region, respectively. The source and body were grounded. For the ID-VD measurements, the gate was added the overdrive voltages from 0.2 V to 1 V, while the drain was added the ramp voltage from 0 V to 1.5 V. The source and body were grounded. For the IG-VG measurement, the gate was added the ramp voltage from -1.5 V to 1.5 V. The drain, source and body were grounded. For bias temperature instability, the gate was fixed at a gate overdrive voltage of 1.6 V. The drain, source and body were grounded. In the hot carrier effect test, the gate was fixed at the gate overdrive voltage 1.5 V. In order to get the same electrical field in the channel of the different fin width, the drain was added at voltage of 1.64 V and 1.73 V, respectively for the thicker and thinner fin device. The source and body were grounded. In the reliability tests, the total stress time were 6000 seconds, and separated it in 5 cycles. Each cycle was followed by a set of basic I-V measurement. The low-frequency-noise, the gate was added from 0 V to 1.2 V. The drain was fixed at a voltage of 0.05 V, and grounded the source and body.

RESULTS AND DISCUSSON

Table 1 summary the hot carrier test and bias temperature instability results. In the hot carrier test, we can see the worse degradation in the thinner fin devices due to the volume inversion and the larger depletion region inside the fin body. The thicker fin devices are dominated by the interface inversion and thus have the worse degradation, because of the interaction between the dielectric traps and the interface inversion carriers. The thinner fin devices are the volume inversion dominating, hence a few carriers to interact with the dielectric traps.
Wfin/Lg (nm)
25/100 25/40 10/100 10/40
Items
RSD HCI_high Vd
10.22% 12.41%
PBTI 5.97% 2.46% 3.06% 1.63%
NBTI 24.24% 13.41% 16.04 8.85%
Table 1 The reliability items caused ID degradations
Thicker fin devices are dominated by the interface inversion, the inversion carriers that flowing in the channel easily interacts with the dielectric traps to cause the coulomb scattering. But thinner fin devices are the volume inversion dominating, the carriers can’t easily interact with the dielectric traps. With the coulomb scattering, the device would show the better linearity in the square root of input-referred voltage noise versus gate overdrive voltage plot that was showed in the following. From figure 1, we find the better linearity in the thicker fin devices than the thinner fin devices.

Figure 1. Square root of input-referred voltage noise versus gate overdrive voltage for n-type FinFETs with different fin widths.

CONCLUSION

In our study, the thinner fin devices improve the short channel effects obviously such as subthreshold swing, drain-induced-barrier-lowering and the threshold voltage roll-off. The thicker fin devices are dominated by the interface inversion, but the volume inversion dominates the thinner fin devices for the quantum confinement effect. Thus the thicker fin devices have worse degradation in the bias temperature instability due to their larger interaction of dielectric traps and the interface inversion carriers. In contrast, the thinner fin devices become worse degradation in the fixed-channel electrical field hot carrier test. This is owing to the larger depletion region inside the fin body instead of the interface between the dielectric layer and fin body. The volume carriers flow toward the larger depletion region to induce the worse degradation. In the low-frequency-noise analysis, the thicker fin devices are the interface inversion dominating, the inversion carriers interact with the traps in the dielectric layer more easily, so it can induce more obvious coulomb scattering.

參考文獻
[1] PROBIR K. & BONDYOPADHYAY, “Moore’s Law Governs the Silicon Revolution,” Proceedings of the IEEE, vol. 86, pp.78-81, 1998.
[2] ITRS. (2013). Available: http://www.itrs.net/Links/2013ITRS/Home2013.htm
[3] Gaurav Saini and Ashwani K Rana, “Physical Scaling Limits of FinFET Structure: A Simulation Study,” International Journal of VLSI design & Communication Systems (VLSICS), vol. 2, No1, pp.26-35, 2011.
[4] G. He and Z. Sun, “High-k Gate Dielectrics for CMOS Technology,” Wiley, 2012.
[5] Jha, K.K, ; Srivastava, A. ; Pattanaik, M. “Design and estimation of drive current of scaled devices by selection of suitable metal gate with ZrO2 dielectric in nm regime,” Nanoscience, Engineering and Technology (ICONSET), International Conferenct on pp.202-204, 2011.
[6] Yew, K.S. ;Ang, D.S. ;Tang, L.J. “A New Method for Enhancing High-k/Metal-Gate Stack Performance and Reliability for High-k Last Integration,” Electron Device Letters, IEEE, vol. 34, pp.295-297, 2013.
[7] Hsing-Huang Tseng ; Tobin, Philip J., et. al, “Defect Passivation With Fluorine and Interface Engineering for Hf-Based High-k/Metal Gate Stack Device Reliability and Performance Enhancement,” Electron Devices, IEEE Transactions, vol. 54, pp3267-3275, 2007.
[8] 張文亮, “Ultrashallow Junction Formation in Deca-nanometer CMOS Technology,” 毫微米通訊, vol. 8, pp.9-17, 2001.
[9] Adan, A.O. ; Naka, T., et.al, “SOI as a mainstream IC technology,” Proceedings., 1998 IEEE International, pp.9-12, 1998.
[10] D.A. Antoniadis, A. Wei, A. Lochtefeld, "SOI Devices and Technology," Proc. 29th European Solid-State Device Research Conf., pp. 81-87, 1999.
[11] Chenming Hu, “SOI and Nanoscale MOSFETs,” Device Research Conference, pp.3-4, 2001.
[12] Scott E. Thompson, Mark Armstrong, et. al, “A Logic Nanotechnology Featuring Strained-Silicon,” Electron Devices Letters, IEEE, vol. 25, pp.191-193.
[13] Jong-Tae Park and Jean-Pierre Colinge, “Multiple-Gate SOI MOSFETs:
Device Design Guidelines,” Electron Devices, IEEE Transactions, vol. 49, pp.2222-2229.
[14] Chang, L. ; Yang-Kyu, et.al, “Moore’s law lives on [CMOS transistors],” Circuits and Devices Magazine, IEEE, vol. 19, pp. 35-42, 2003.
[15] Rasouli, S.H. ; Endo, K. et.al, “Variability analysis of FinFET-based devices and circuits considering electrical confinement and width quantization,” Computer-Aided Design – Digest of Technical Papers, pp.505-512, 2009.
[16] Alexander Kloes, Mike Schwarz et. al, “Quantum Confinement and Volume Inversion in MOS3 Model for Short-Channel Tri-Gate MOSFETs,” Electron Devices, IEEE Transactions, vol. 60, pp.2691-2694, 2013.
[17] Jeffrey Lutze and Suresh Venkatesan, “Techniques for Reducing the Reverse Short Channel Effect in Sub-0.5 prn CMOS,” Electron Device Letters, IEEE, vol. 16, pp.373-375, 1995.
[18] Bin Yu, Clement H. J. Wann, et. al, “Techniques for Reducing the Reverse Short
Channel Effect in Sub-0.5 prn CMOS,” Electron Devices, IEEE Transactions, vol. 44, pp.627-634, 1997.
[19] Byung-Kil Choi, Kyoung-Rok Han, et. al, “Threshold-Voltage Modeling of Body-Tied FinFETs (Bulk FinFETs),” Electron Devices, IEEE Transactions, vol.54, pp.537-545, 2007.
[20] Mirko Poljak, Vladimir Jovanovic, et. al, “Improving bulk FinFET DC performance in comparison to SOI FinFET,” Microelectronic Eng., vol. 86, pp.2078-2085, 2009.
[21] S. Chabukswar, D. Maji, C. R. Manoj, K. G. Anil, V. R. Rao, F. Crupi, P. Magnone, G. Giusi, C. Pace, and N. Collaert, “Implications of fin width scaling on variability and reliability of high-k metal gate FinFETs,” Microelectron. Eng., vol. 87, no. 10, pp. 1963–1967, Oct. 2010.
[22] Cheng-Li Lin, Member, Po-Hsiu Hsiao, Wen-Kuan Yeh, et. al “ Effects of Fin Width on Device Performance and Reliability of Double-Gate n-Type FinFETs” Electron Devices, IEEE Transactions, vol. 60, pp. 3639-3644, 2013
[23] Joshi, K., Mukhopadhyay, S., Goel, N., Mahapatra, S., “A consistent physical framework for N and P BTI in HKMG MOSFETs,” Reliability Physics Symposium (IRPS), 2012 IEEE, Page(s): 5A.3.1 – 5A.3.10.
[24] Pei-Yu Wang, Bing-Yue Tsui, “SixGe1-x Epitaxial Tunnel Layer Structure for P-Channel Tunnel FET Improvement,” Electron Devices, IEEE Transactions, vol. 60, pp.4098-4104, 2013