Open Access System for Information Sharing

Login Library

 

Thesis
Cited 0 time in webofscience Cited 0 time in scopus
Metadata Downloads

Fast method of Lifetime estimation for High-k nMOSFET and Study on electron trap distributions under PBTI

Title
Fast method of Lifetime estimation for High-k nMOSFET and Study on electron trap distributions under PBTI
Authors
노기윤
Date Issued
2020
Publisher
포항공과대학교
Abstract
이 논문은 PBTI 하에서 HfSiON/SiO2 dielectric nMOSFETs의 수명을 빠르고 정확하게 예측하는 방법을 제안하고, electron trap분포를 분석합니다. PBTI 하에서 HfSiON/SiO2 dielectric nMOSFETs의 수명을 빠르고 정확하게 예측하는 방법이 제안되었습니다. VRS열화는 n << 1 이고 ΔVg,str ≈ 50mV 일 때, 마지막 VRS pulse에 의존적으로 결정되었습니다. Power law constants는 n << 1 이고 ΔVg,str ≈ 50mV인 상황이 보장되지 않더라도 빠르고 정확하게 추출될 수 있어야 합니다. 제안된 방법은 단일 MOSFET의 ΔVth vs. t data를 측정하고, 각 각의 Vg,str에서 전압 의존적인 power law constants a와 n을 얻기 위해 각 VRS pulse의 fitting은 regression method를 사용합니다. 제안된 방법을 사용하여 추출된 power-law constants는 constant voltage stress (CVS)를 이용하여 얻어진 값과 매우 잘 일치합니다. 그래서, 제안된 VRS-regression방법은 PBTI하에서 nMOSFETs을 screening하는데 매우 유용합니다. HfSiON/SiO2 dielectric nMOSFETs에서 기존에 존재하고 있던 oxide traps과 stress에 의해 발생한 oxide traps의 정량적인 oxide trap profiling은 trap spectroscopy by charge injection and sensing (TSCIS)에 의해 얻어질 수 있었습니다. PBTI하에서 HfSiON/SiO2 nMOSFETs의 큰 분극성에 의해 야기된 lattice relaxation때문에 electron traps의 energy level은 아래쪽 방향으로 이동될 수 있습니다. 실험에 사용된 nMOSFETs에 대하여, interfacial layer를 형성하기 위해 두가지 다른 공정들이 사용되었습니다 (P1은 plasma oxidation이며, P2는 thermal oxidation). P1과 P2의 electron trap profiles은 TSCIS를 사용하여 stress이전 상태와 PBTI stress하에서 얻었습니다. P1은 P2에 비해 더 낮은 trap density를 가지고 있었으며, 더 trap level shift는 더 느리게 발생하였습니다. PBTI하에서 HfSiON두께가 HfSiON/ SiO2 nMOSFET의 electron trap분포에 미치는 영향은 carrier charging/discharging방법을 이용해 연구되었습니다. HfSiON의 두께가 증가할수록 결정성의 정도가 증가하기 때문에, 두꺼운 nMOS는 얇은 nMOS에 비해 trap density가 더 많았습니다. 하지만 얇은 nMOS보다 두꺼운 nMOS에서 trap energy level은 깊은 방향으로 더 빨리 shift하였습니다. 그 결과, 두꺼운 nMOS는 얇은 nMOS에 비해 더 큰 ΔVth과 더 낮은 n을 가지게 되었습니다. 신뢰성있는 HfSiON/SiO2 nMOSFETs을 달성하기 위해서는, HfSiON두께는 반드시 최적화되어야 하며, 따라서 electron traps분포는 반드시 정량화 되어야 합니다. 이 논문은 VRS-regression방법이 high-k MOSFET의 수명을 빠르게 예측하는데 매우 유용하다는 것을 보여주며, 큰 분극성을 가진 high-k nMOSFETs의 신뢰성을 확인하기 위해서 TSCIS나 carrier charging and discharging방법을 이용한 electron trap profiling은 필수적이라는 것을 보여줍니다
This thesis proposes a fast and accurate method to estimate lifetime tL, and investigates electron trap distributions for HfSiON/SiO2 dielectric nMOSFETs under positive bias temperature instability (PBTI). The proposed method to extract lifetime requires one nMOSFET only, uses a voltage ramp stress (VRS), measures ΔVth vs. t data during VRS, uses a regression method to fit the data for each VRS pulse to the power law to obtain a and n at each stress voltage Vg,str, then obtains five voltage-independent constants for the power law after fitting the curves of a and n vs. Vg,str to empirical models. The extracted power-law constants using the proposed method agreed very well with those obtained using the constant voltage stress (CVS) method. After obtaining the voltage-independent power-law constants, the tL at an operating voltage Vop was estimated using the power law. The predicted tL = 1.67ⅹ108 s was quite close to tL = 1.74ⅹ108 s predicted using CVS and to tL = 1.72ⅹ108 s measured at Vop. The time required for measurement was 900 s, compared to 30,000 s for the CVS method. These experimental results show that the proposed VRS-regression method is very useful for screening nMOSFETs under PBTI. A quantitative oxide trap profiling method for both pre-existing and stress-induced oxide traps using TSCIS and investigates PBTI mechanism of HfSiON/SiO2 nMOSFETs based on the extracted oxide traps. For extracting initial trap and stress-induced trap, the charge voltages are determined by stress-induced leakage current (SILC). The experimental results indicate that the energy level of electron traps can be shifted downward due to lattice relaxation once electrons are trapped at pre-existing traps by PBTI stress, demonstrating that the proposed method is very useful for oxide trap profiling of high-k nMOSFETs with large polarizability. The effect of HfSiON thickness on electron trap distributions under positive bias temperature instability (PBTI) was investigated for HfSiON/SiO2 nMOSFET. Trap distributions of HfSiON/SiO2 nMOSFET were observed by carrier charging/discharging method. Experimental results show that the peak values of electron trap density shifted to deeper electron trap energy level (Et) with increasing stress field Estr and stress time ts. Compared to the Thick HfSiON device, the Thin HfSiON device had lower trap density and slower Et-shift; as a result, the Thin HfSiON device had lower threshold voltage-shift ΔVth and larger power-law time exponent n of PBTI than the Thick HfSiON device. Low ΔVth is beneficial for tL of HfSiON/SiO2 nMOSFET but large n is not, so the effect of HfSiON thickness on distribution of electron trap must be quantified to enable optimization of HfSiON thickness to yield reliable HfSiON/SiO2 nMOSFETs. This thesis show that VRS-regression method is very useful to estimate tL rapidly for high-k MOSFETs regardless of the range of n and the number of samples, and electron trap profiling using TSCIS or carrier charging/discharging method is neccesary to evaluate reliability of high-k nMOSFETs which has large polarizability.
URI
http://postech.dcollection.net/common/orgView/200000288325
https://oasis.postech.ac.kr/handle/2014.oak/111270
Article Type
Thesis
Files in This Item:
There are no files associated with this item.

qr_code

  • mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Views & Downloads

Browse