工艺偏差下PMOS器件的负偏置温度不稳定效应分布特性

当器件特征尺寸进入纳米级,负偏置温度不稳定性(NBTI)效应和工艺偏差都会导致p型金属氧化层半导体(PMOS)器件性能和可靠性的下降.基于反应-扩散(R-D)模型,本文分析了工艺偏差对NBTI效应的影响;在此基础上将氧化层厚度误差和初始阈值电压误差引入到R-D模型中,提出了在工艺偏差下PMOS器件的NBTI效应统计模型.基于65 nm工艺,首先蒙特卡罗仿真表明在工艺偏差和NBTI效应共同作用下,PMOS器件阈值电压虽然会随着应力时间增大而沿着负方向增加,但是阈值电压的匹配性却随着时间推移而变好;其次验证本文提出的统计模型准确性,以R-D模型为参考,在10~4s应力时间内,PMOS器件阈值电压退化量平均值和均方差的最大相对误差分别为0.058%和0.91%;最后将此模型应用到电流舵型数模转换器中,仿真结果显示在工艺偏差和NBTI效应共同作用下,数模转换器的增益误差会随着应力时间的推移而增大,而线性误差会逐渐减小.     

具有HfN/HfO2栅结构的p型MOSFET中的负偏置-温度不稳定性研究

研究了HfN/HfO₂高K栅结构p型金属-氧化物-半导体(MOS)晶体管(MOSFET)中，负 偏置-温度应力引起的阈值电压不稳定性(NBTI)特征.HfN/HfO₂高K栅结构的等效 氧化层厚度(EOT)为1.3nm，内含原生缺陷密度较低.研究表明，由于所制备的HfN/HfO₂ 高K栅结构具有低的原生缺陷密度，因此在p-MOSFET器件中观察到的NBTI属HfN/HfO2高K栅结构的本征特征，而非工艺缺陷引起的；进一步研究表明，该HfN/HfO₂高K栅结构中观察到的NBTI与传统的SiO₂基栅介质p-MOSFET器件中观察 到的NBTI具有类似的特征，可以被所谓的反应-扩散(R-D)模型表征： HfN/HfO₂ 栅结构p-MOSFET器件的NBTI效应的起源可以归为衬底注入空穴诱导的界面反应机理，即在负 偏置和温度应力作用下，从Si衬底注入的空穴诱导了Si衬底界面Si-H键断裂这一化学反应的 发生，并由此产生了Si^＋陷阱在Si衬底界面的积累和H原子在介质层内部的扩散 ，这种Si^＋陷阱的界面积累和H原子的扩散导致了器件NBTI效应的发生. 关键词： 高K栅介质 负偏置-温度不稳定性(NBTI) 反应-扩散(R-D)模型     

pMOS器件直流应力负偏置温度不稳定性效应随器件基本参数变化的分析

应用负偏置温度不稳定性(negative bias temperature instability,NBTI),退化氢分子的漂移扩散模型,与器件二维数值模拟软件结合在一起进行计算,并利用已有的实验数据和基本器件物理和规律,分析直流应力NBTI效应随器件沟道长度、栅氧层厚度和掺杂浓度等基本参数的变化规律,是研究NBTI可靠性问题发生和发展机理变化的一种有效方法.分析结果显示,NBTI效应不受器件沟道长度变化的影响,而主要受到栅氧化层厚度变化的影响;栅氧化层厚度的减薄和栅氧化层电场增强的影响是一致的,决定了器件退化按指数规律变化;当沟道掺杂浓度提高,NBTI效应将减弱,这是因为器件沟道表面空穴浓度降低引起的;然而当掺杂浓度提高到器件的源漏泄漏电流很小时(小泄露电流器件),NBTI效应有明显的增强.这些结论对认识NBTI效应的发展规律以及对高性能器件的设计具有重要的指导意义.     

超深亚微米PMOS器件的NBTI退化机理

对超深亚微米PMOS器件的负栅压温度不稳定性(NBTI)退化机理进行了研究.主要集中在对器件施加NBT和随后的PBT应力后器件阈值电压的漂移上.实验证明反型沟道中空穴在栅氧中的俘获以及氢分子在栅氧中的扩散是引起NBTI退化的主要原因.当应力条件变为PBT时，陷落的空穴可以快速退陷，但只有部分氢分子可以扩散回栅氧与衬底界面钝化硅悬挂键，这就导致了PBT条件下阈值电压只能部分恢复. 关键词： 超深亚微米PMOS器件 负偏压温度不稳定性 界面陷阱 氢气     

具有HfN/HfO2栅结构的p型MOSFET中的负偏置-温度不稳定性研究

研究了HfN/HfO2高K栅结构p型金属-氧化物-半导体(MOS)晶体管(MOSFET)中,负偏置-温度应力引起的阈值电压不稳定性(NBTI)特征.HfN/HfO2高K栅结构的等效氧化层厚度(EOT)为1.3nm,内含原生缺陷密度较低.研究表明,由于所制备的HfN/HfO2高K栅结构具有低的原生缺陷密度,因此在p-MOSFET器件中观察到的NBTI属HfN/HfO2高K栅结构的本征特征,而非工艺缺陷引起的;进一步研究表明,该HfN/HfO2高K栅结构中观察到的NBTI与传统的SiO2基栅介质p-MOSFET器件中观察到的NBTI具有类似的特征,可以被所谓的反应-扩散(R-D)模型表征: HfN/HfO2栅结构p-MOSFET器件的NBTI效应的起源可以归为衬底注入空穴诱导的界面反应机理,即在负偏置和温度应力作用下,从Si衬底注入的空穴诱导了Si衬底界面Si-H键断裂这一化学反应的发生,并由此产生了Si+陷阱在Si衬底界面的积累和H原子在介质层内部的扩散,这种Si+陷阱的界面积累和H原子的扩散导致了器件NBTI效应的发生.     

功率MOSFET的负偏置温度不稳定性效应中的平衡现象

通过对功率金属氧化物半导体场效应晶体管在静态应力下的负偏置温度不稳定性的实验研究, 发现器件参数的退化随时间的关系遵循反应扩散模型所描述的幂函数关系, 并且在不同栅压应力下, 实验结果中均可观察到平台阶段的出现. 基于反应扩散理论的模型进行了仿真研究, 通过仿真结果分析和验证了此平台阶段对应于反应平衡阶段, 并且解释了栅压应力导致平台阶段持续时间不同的原因. 关键词： 功率金属氧化物半导体场效应晶体管 负偏置温度不稳定性 反应扩散模型     

NBT导致的深亚微米PMOS器件退化与物理机理

研究了深亚微米PMOS器件在负偏压温度(negative bias temperature, NBT) 应力前后的电流电压特性随应力时间的退化，重点分析了NBT应力对PMOS器件阈值电压漂移的影响，通过实验证明了在栅氧化层和衬底界面附近的电化学反应和栅氧化层内与氢相关的元素的扩散，是PMOS器件中NBT效应产生的主要原因.指出NBT导致的PMOS器件退化依赖于反应机理和扩散机理两种机理的平衡. 关键词： 深亚微米PMOS器件 负偏压温度不稳定性 界面态 氧化层固定正电荷     

p型金属氧化物半导体场效应晶体管界面态的积累对单粒子电荷共享收集的影响

由于负偏置温度不稳定性和热载流子注入,p型金属氧化物半导体场效应晶体管(pMOSFET)将在工作中不断退化,而其SiO₂/Si界面处界面态的积累是导致其退化的主要原因之一. 采用三维器件数值模拟方法,基于130 nm体硅工艺,研究了界面态的积累对相邻pMOSFET之间单粒子电荷共享收集的影响. 研究发现,随着pMOSFET SiO₂/Si界面处界面态的积累,相邻pMOSFET漏端的单粒子电荷共享收集量均减少. 还研究了界面态的积累对相邻反相器中单粒子电荷共享收集 关键词： 负偏置温度不稳定性 电荷共享收集 双极放大效应 单粒子多瞬态     

国产星用VDMOS器件总剂量辐射损伤效应研究

研究了两种国产星用VDMOS器件在不同偏置条件下的总剂量辐射损伤效应, 探讨了器件的阈值电压、击穿电压、导通电阻、漏电流等电参数随累积剂量、退火时间的变化关系. 实验结果表明这两种国产星用VDMOS器件辐照后电参数符合技术指标, 满足在复杂空间电离辐射环境下工作的要求.此外, 通过对器件在不同偏置条件下的总剂量辐射损伤效应进行研究, 对其他型号星用VDMOS器件工艺和设计的进一步改进, 具有参考作用.     

辐照下背栅偏置对部分耗尽型绝缘层上硅器件背栅效应影响及机理分析

基于部分耗尽型绝缘层上硅（SOI）器件的能带结构,从电荷堆积机理的电场因素入手,为改善辐照条件下背栅Si/SiO₂界面的电场分布,将半导体金属氧化物（MOS）器件和平板电容模型相结合,建立了背栅偏置模型.为验证模型,利用合金烧结法将背栅引出加负偏置,对NMOS和PMOS进行辐照试验,得出:NMOS背栅接负压,可消除背栅效应对器件性能的影响,改善器件的前栅I-V特性;而PMOS背栅接负压,则会使器件的前栅I-V性能恶化.因此,在利用背栅偏置技术改善SOI/NMOS器件性能的同时,也需要考虑背栅偏置对PMOS的影响,折中选取偏置电压.该研究结果为辐照条件下部分耗尽型SOI/MOS器件背栅效应的改善提供了设计加固方案,也为宇航级集成电路设计和制造提供了理论支持.     

Study on the recovery of NBTI of ultra-deep sub-micro MOSFETs

Taking the actual operating condition of complementary metal oxide semiconductor (CMOS) circuit into account, conventional direct current (DC) stress study on negative bias temperature instability (NBTI) neglects the detrapping of oxide positive charges and the recovery of interface states under the `low' state of p-channel metal oxide semiconductor field effect transistors (MOSFETs) inverter operation. In this paper we have studied the degradation and recovery of NBTI under alternating stress, and presented a possible recovery mechanism. The three stages of recovery mechanism under positive bias are fast recovery, slow recovery and recovery saturation.     

ICP发射光谱测定保偏光纤应力棒中的二氧化二硼

本文提出了用ＩＣＰ发射光谱仪测定保偏光纤应力棒中的二氧化二硼的方法。将应力棒粉碎后，使用ＨＦ在压力溶样器中消解，最后溶液的介质为０．５％氢氟酸，并采用耐腐蚀的样品引进系统进行ＩＣＰ光谱测定。分析程序简单，准确度高，该方法回收率为９９～１０１％，相对标准偏差为１．０４％     

Mechanism of NBTI Recovery under Negative Voltage Stress

Recovery phenomenon is observed under negative gate voltage stress which is smaller than the previous degradation stress. We focus on the drain current to study the degradation and recovery of negative bias temperature instability (NBTI) with a real-time method. By this method, different recovery phenomena among different size devices are observed. Under negative recovery stress, the drain current gradually recovers for the large size devices and gets into recovery saturation when long recovery time is involved. For small-size devices, a step-like recovery of drain current is observed. The recovery of the drain current is mainly caused by the holes detrapping and tunnelling back to the channel surface which are trapped in oxide. The model of hole detrapping explains the recovery under negative voltage stress reasonably.     

Stress-Strain Behaviors and Crosslinked Networks Studies of Natural Rubber-Zinc Dimethacrylate Composites

The stress-strain behaviors of natural rubber (NR)-zinc methacrylate (ZDMA) composite have been studied by uniaxial tension. The results indicated that there was a large reinforcement by ZDMA and the NR/ZDMA composites exhibited a high stress-softening effect. Meanwhile, the recovery stretch curve was close to the second stretch curve; thus a weak stress recovery of the composites was shown. The analysis of crosslink density indicated that the damage to the crosslink network was mainly due to the breakage of ionic crosslinks at low strain (100%). A more developed ionic crosslink network was formed at a higher content of ZDMA. When the vulcanizate is subjected to loading in tension, the ionic crosslink network will suffer the force first. Next, the slippage of ionic bonds will take place under the stress. A new ionic crosslink network might be formed rapidly after the ionic bonds were broken during the stretching. Therefore, it could not return to the initial state. The analysis of crosslink density and stress recovery indicated that the rubber chains could be adsorbed to the ZDMA aggregates due to the formation of poly-zinc methacrylate (PZDMA). A molecular analysis of NR/ZDMA composites is proposed in the last part of this article.     

Chain orientation in natural rubber,Part I: The inverse yielding effect

Inhomogeneous deformations are observed in stretched natural rubber of different crosslink density; the conditions of observation, nucleation and propagation are given in the first part of the paper. In samples of low crosslink density these inhomogeneities recall necking observed in others materials and in glassy polymers when the materials are drawn above a critical draw ratio. The difference is that in natural rubbers, NR, they nucleate and propagate at constant stress during unloading. This phenomenon, called inverse yielding appears during recovery only if the samples have been drawn previously in the hardening domain. During necking propagation the stress is constant. The mechanical and crystallinity properties of samples with and without inverse yielding are studied as a function of draw ratio, crosslink density and temperature. In the second part of the paper this transition zone (neck) of thickness 2 mm is studied by WAXS at the synchrotron source. From the orientation of NR crystallites and from the orientation of the stearic acid (2%, present in this type of rubber) we conclude that the deformation in the neck follows the flow lines. From the local crystallinity of the NR crystallites one deduces the local draw ratio across this transition zone. We suggest that in all these rubbers, which present a plateau of the recovery stress strain curve, micronecking exists. This effect is discussed in the framework of the Flory theory.-1     

Dislocation arrangements during cyclic loading of pure iron

Fatigue induced substructure has been studied in polycrystalline iron (99.95 per cent) cycled at constant low stress amplitude. The main experimental technique was transmission electron microscopy. The results showed the dominant role of friction stress during cyclic loading and the importance of cross slip during dynamic recovery.     

Hot carrier injection degradation under dynamic stress

In this paper,we have studied hot carrier injection(HCI) under alternant stress.Under different stress modes,different degradations are obtained from the experiment results.The different alternate stresses can reduce or enhance the HC effect,which mainly depends on the latter condition of the stress cycle.In the stress mode A(DC stress with electron injection),the degradation keeps increasing.In the stress modes B(DC stress and then stress with the smallest gate injection) and C(DC stress and then stress with hole injection under V g = 0 V and V d = 1.8 V),recovery appears in the second stress period.And in the stress mode D(DC stress and then stress with hole injection under V g = 1.8 V and V d = 1.8 V),as the traps filled in by holes can be smaller or greater than the generated interface states,the continued degradation or recovery in different stress periods can be obtained.     

A dislocation-based model for creep recovery in ice

Creep recovery strain is significant in polycrystalline ice, and its stress dependence is strongly dependent on ice type and deformation history. Although it is generally recognized that creep recovery strain is largely attributable to dislocations, a dislocation-based model that rectifies the observed differences in two important ice types (freshwater and sea ice) has not previously emerged; and the development of such a model forms the goal of the present effort. The model considers basal dislocation distributions, employs a dislocation density–stress relationship from previous work, and uses an empirical expression for the decrease in slip-line spacing with increasing stress. The dislocation processes are taken to operate over a subgrain-sized domain, and the strain associated with the relative motion of neighbouring domains is considered. The model accounts for dislocation multiplication and the decrease in slip-line spacing if the applied stress is sufficiently high. The model explains the observed differences in creep recovery strain for freshwater and sea ice and adequately reproduces both the stress dependence and the limiting value of the experimentally observed behaviour. It also reproduces the shift from nonlinear to linear behaviour that has been observed in prestrain experiments on freshwater and sea ice cores.     

Study on the negative bias temperature instability effect under dynamic stress

This paper studies negative bias temperature instability (NBTI) under alternant and alternating current (AC) stress.Under alternant stress,the degradation smaller than that of single negative stress is obtained.The smaller degradation is resulted from the recovery of positive stress.There are two reasons for the recovery.One is the passivation of H dangling bonds,and another is the detrapping of charges trapped in the oxide.Under different frequencies of AC stress,the parameters all show regular degradation,and also smaller than that of the direct current stress.The higher the frequency is,the smaller the degradation becomes.As the negative stress time is too small under higher frequency,the deeper defects are hard to be filled in.Therefore,the detrapping of oxide charges is easy to occur under positive bias and the degradation is smaller with higher frequency.     

On polymeric materials containing fibrils with a phase transition. II. The mechanical consequences of matrix shear

The mechanical behavior of a system of parallel fibrils, capable of a phase change and embedded in a matrix, is analyzed by considering a simplified model. The model consists of a single infinitely long isolated fibril together with its associated matrix. Furthermore, it is assumed that slippage cannot occur at the fibril-matrix interface but that the matrix deforms in both tension and in shear. The phase change occurring in the fibrils gives rise to large discontinuous local strains in the fibril and is characterized by a “critical stress” and an “equilibrium stress.” Analysis of the model shows that distinct zones of the secondary phase will form consecutively along the fibril length when the system is extended. These zones are mobile and the stress fields in the matrix associated with each zone result in the mutual repulsion of adjacent zones. Stress-strain curves for both extension and recovery are calculated using linear and nonlinear forms of the system parameters. It is suggested that the model studied is relevant to the mechanical behavior of keratin fibers.