积累模式p沟道SOI基围栅Fin-FET驱动电流的研究
- 格式:pdf
- 大小:747.43 KB
- 文档页数:5
SiGe高迁移率沟道钝化关键技术及其FinFET器件应用研究摘要:随着集成电路技术的不断发展,如何提高半导体器件的性能成为一个重要的研究方向。
硅锗(SiGe)材料作为一种高迁移率半导体材料,被广泛应用于FinFET器件中,以提高器件的速度和功耗。
本文介绍了SiGe高迁移率沟道钝化的关键技术,并探讨了其在FinFET器件中的应用研究。
关键词:SiGe,高迁移率沟道,钝化,FinFET器件一、引言FinFET器件是一种新型的三维晶体管结构,具有较低的漏电流和更好的控制能力。
然而,传统的FinFET器件在速度和功耗方面仍然有待提高。
SiGe高迁移率沟道钝化技术提供了一种改善器件性能的方法。
二、SiGe高迁移率沟道钝化技术SiGe材料具有比传统硅材料更高的电子迁移率,可以显著提高器件的速度。
然而,SiGe材料表面的氧化物会导致界面态和电子散射,降低材料的性能。
因此,SiGe高迁移率沟道钝化技术是非常关键的。
SiGe高迁移率沟道钝化技术主要包括两个方面的研究:表面钝化和界面钝化。
表面钝化通过在SiGe材料表面形成一层氧化物或硝化物薄膜,阻止杂质的扩散和电子散射。
界面钝化则通过在SiGe材料和硅基底之间形成一层界面层,减少界面态和电子散射。
三、SiGe高迁移率沟道钝化技术在FinFET器件中的应用研究SiGe高迁移率沟道钝化技术在FinFET器件中的应用主要包括两个方面:材料选择和制备工艺。
在材料选择方面,需要选择合适的SiGe材料,以获得最佳的迁移率提升效果。
同时,还需要考虑材料的稳定性和成本因素。
在制备工艺方面,需要优化钝化层的厚度和制备条件,以达到最佳的效果。
同时,还需要考虑钝化层与其他器件材料之间的界面匹配性。
通过上述研究,SiGe高迁移率沟道钝化技术在FinFET器件中取得了显著的性能提升。
SiGe材料的高迁移率使得FinFET器件具有更快的速度和更低的功耗,适用于高性能应用。
四、结论SiGe高迁移率沟道钝化技术是提高集成电路性能的关键技术之一。
SOI FET NMOS电路等效一、背景介绍SOI(Silicon-on-Insulator)技术是近年来在半导体器件制造领域取得重大突破的一种先进技术。
SOI技术以绝缘层为衬底,将硅薄膜直接生长在绝缘层上,形成具有较低漏电流、低电容和高速特性的SOI器件。
其中,SOI场效应晶体管(FET)是SOI技术中的重要应用之一。
二、FET NMOS电路的基本特性FET(Field Effect Transistor)是一种三极管,由栅极、漏极和源极组成。
NMOS(N-type Metal-Oxide-Semiconductor Field-Effect Transistor)是以n型半导体为基础制造的FET。
NMOS电路主要用于数字电路中,具有高的开关速度和较低的功耗。
在SOI技术中,NMOS电路的等效模型和传统的CMOS(Complementary Metal-Oxide-Semiconductor)电路有一定差异,需要进行适当的等效处理。
三、SOI FET NMOS电路的等效理论1. SOI FET NMOS电路的等效模型在SOI技术中,NMOS电路的等效模型需要考虑到SOI结构的特殊性,包括绝缘层的影响、二维电荷分布等因素。
SOI FET NMOS电路的等效模型与传统NMOS电路的等效模型有所不同,需要进行适当的简化和修正。
2. SOI FET NMOS电路的等效处理方法针对SOI FET NMOS电路的特殊结构和工作原理,可以采用不同的等效处理方法。
可以通过使用适当的SOI NMOS器件等效模型进行仿真分析,以验证其在SOI技术下的性能表现。
也可以通过实际测试和数据采集,对SOI FET NMOS电路的实际特性进行评估和分析。
四、SOI FET NMOS电路等效的应用1. SOI FET NMOS电路在RF集成电路中的应用由于SOI技术具有较低的电容和较高的工作频率特性,SOI FET NMOS电路在射频(RF)集成电路中具有广泛的应用前景。
finfet工作原理FinFET工作原理。
FinFET(Fin Field-Effect Transistor)是一种三维晶体管结构,它是一种改进型的MOSFET(金属-氧化物-半导体场效应晶体管),可以在集成电路中实现更好的性能和功耗效率。
在FinFET的结构中,栅极控制区域被放置在两个细长的“鳍”上,这种设计可以有效地控制电流流动,提高晶体管的性能。
下面我们将详细介绍FinFET的工作原理。
首先,FinFET的工作原理基于栅极对源漏区的控制。
当栅极施加电压时,它会形成电场,这个电场会影响源漏区的电荷分布,从而控制电流的流动。
由于FinFET的结构中有两个鳍,因此可以更有效地控制电流,提高晶体管的开关速度和功耗效率。
其次,FinFET的三维结构也是其工作原理的关键。
与传统的平面MOSFET相比,FinFET的三维结构可以提供更大的通道长度,从而减小了漏电流的问题。
此外,FinFET还可以实现更好的电子掺杂效果,提高了晶体管的性能。
另外,FinFET还具有优异的抑制短沟道效应的特点。
由于FinFET的结构可以有效地控制电场分布,因此可以减小短沟道效应对晶体管性能的影响。
这使得FinFET在集成电路中可以实现更小的尺寸,从而提高了集成度和性能。
最后,FinFET的工作原理还与材料和制造工艺密切相关。
FinFET需要采用先进的工艺技术来实现其复杂的三维结构,同时还需要优质的半导体材料来保证其性能。
因此,FinFET的工作原理也受到材料和工艺的限制,需要在这些方面不断进行创新和改进。
总之,FinFET作为一种新型的三维晶体管结构,其工作原理基于栅极对源漏区的控制,利用三维结构和优异的抑制短沟道效应的特点,可以实现更好的性能和功耗效率。
同时,FinFET的工作原理还与材料和制造工艺密切相关,需要不断进行创新和改进。
相信随着技术的不断发展,FinFET将在集成电路中发挥越来越重要的作用。
vdmosfet原理VDMOSFET原理引言:VDMOSFET是一种金属氧化物半导体场效应晶体管,在电子设备中被广泛应用。
它具有低开关损耗、快速开关速度和高功率密度等优点,因此在功率放大和开关电路中被广泛采用。
本文将介绍VDMOSFET的原理及其工作机制。
一、VDMOSFET结构VDMOSFET是由P型衬底上形成的N型沟道和漏极、栅极以及漏源区组成的。
具体结构如下:1. P型衬底:VDMOSFET的衬底是P型材料,它是整个器件的基础。
衬底的掺杂浓度和厚度会影响器件的电特性。
2. N型沟道:N型沟道是在P型衬底上生长的,它负责电流的导通。
沟道的深度和宽度决定了导通能力。
3. 栅极:栅极是控制VDMOSFET导通和截止的关键部分。
通过给栅极施加电压,可以改变沟道的导通能力。
4. 漏源区:漏源区是电流的输入和输出区域。
漏极是输出端,而源极是输入端。
二、VDMOSFET工作原理VDMOSFET的工作原理可以分为导通和截止两个阶段。
1. 导通阶段:当栅极施加正向电压时,栅极和沟道之间会形成一个电场。
这个电场会吸引N型沟道上的自由电子,使其移动到P型衬底上,从而形成一个导电通道。
当漏源区施加正向电压时,电子会从源极注入到沟道中,经过漏极流出。
此时,VDMOSFET导通,电流通过。
2. 截止阶段:当栅极施加负向电压时,栅极和沟道之间的电场会阻止自由电子的移动。
此时,导电通道被切断,VDMOSFET截止,电流不再通过。
三、VDMOSFET特点VDMOSFET具有以下特点:1. 低开关损耗:由于VDMOSFET导通的电阻很小,因此开关过程中的功率损耗较低。
2. 快速开关速度:VDMOSFET的导通和截止速度很快,可以实现快速开关。
3. 高功率密度:由于VDMOSFET的导通能力强,可以承受较大的功率。
4. 低漏极电阻:VDMOSFET的漏极电阻较低,可以降低功耗和热耗散。
四、VDMOSFET应用VDMOSFET广泛应用于功率放大和开关电路中,包括:1. 电源管理:VDMOSFET在电源开关和稳压器中被广泛采用,可以实现高效率的能量转换。
专利名称:基于SOI的后栅型积累模式Si-NWFET制备方法专利类型:发明专利
发明人:黄晓橹
申请号:CN201210135272.3
申请日:20120503
公开号:CN102646642A
公开日:
20120822
专利内容由知识产权出版社提供
摘要:本发明公开了一种基于SOI的后栅型积累模式Si-NWFET制备方法,通过刻蚀SOI衬底上形成的硅层和锗硅层,形成鳍形有源区;在鳍形有源区内形成硅纳米线;接着,在SOI衬底沟道区沉积无定形碳;在栅极沟槽中形成栅极;进行金半合金工艺,去除无定形碳;同时进行沟道隔离介质与层间隔离介质的沉积,形成PMOSFET;接着形成NMOSFET;最后进行合金以及金属互连工艺。
本发明基于SOI的后栅型积累模式Si-NWFET制备方法实现了NMOSFET与PMOSFET结构分离,从而能够独立工艺调试,有效减小NMOSFET的接触孔电阻以提高NMOSFET性能,提高载流子迁移率。
申请人:上海华力微电子有限公司
地址:201203 上海市浦东新区张江高科技园区高斯路497号
国籍:CN
代理机构:上海思微知识产权代理事务所(普通合伙)
代理人:陆花
更多信息请下载全文后查看。
finfet结构类型
FinFET,全称为鳍式场效应晶体管,是一种新结构的互补式金属氧化物半导体晶体管。
它的发明人是加州大学伯克利分校的胡正明教授。
FinFET结构是一种多栅极、垂直通道的晶体管结构,主要包括以下部分:
1. 通道:位于硅基底上,是电流在晶体管中流动的区域。
2. 源极和漏极:分别位于通道两侧,用于注入和收集电流。
3. 栅极:垂直覆盖整个通道,用于控制通道中的电流。
根据不同的分类标准,FinFET可以分为多种类型。
例如,根据是否有SiO2埋氧层以及其特点,可以分为Silicon-on-Insulator(SOI)FinFET、Bulk FinFET、Body-on-Insulator(BOI)FinFET等。
此外,从Gate的数量和形状出发,可以分为双栅、三栅、Ω 栅、环栅等。
以上内容仅供参考,如需更准确的信息,可以查阅科技类文献或咨询该领域的专家。
结型场效应管p沟道的工作原理
(原创实用版)
目录
1.结型场效应管的基本结构
2.场效应管的工作原理
3.P 沟道的形成条件
4.P 沟道场效应管的导电过程
5.P 沟道场效应管的应用
正文
结型场效应管是一种具有特殊电学性能的半导体器件,它利用多数载流子导电,故又称单极型半导体器件。
由于它仅有一个电极(基极),所
以称为结型场效应管(junction,fet)。
场效应管的结构包括源极、漏极
和栅极三部分。
源极是工作电流的来源;漏级为输入端;栅级为输出端,栅压的大小取决于输入电压的大小,通常由外加电压控制其通断状态。
场效应管的工作原理是基于半导体材料的电子运动特性。
在结型场效应管中,当栅极施加正向电压时,栅极附近的电子被吸引到栅极处,形成一个导电通道,使得源极和漏极之间的电流得以流通。
当栅极电压为负时,栅极处的电子减少,导电通道消失,源极和漏极之间的电流中断。
P 沟道是场效应管的一种导电模式,其形成条件是栅极施加正向电压。
在 P 沟道场效应管中,空穴是多数载流子,它们在栅极正向电压的作用下,从源极向漏极方向运动,形成一个导电通道。
与 N 沟道相反,P 沟
道的导电通道是由空穴形成的。
P 沟道场效应管的导电过程可以分为两个阶段。
第一阶段是栅极施加正向电压,空穴被吸引到栅极处,形成一个导电通道。
第二阶段是源极和漏极之间的电流开始流动,直到栅极电压变为负向,导电通道消失,电流
中断。
P 沟道场效应管广泛应用于放大、开关、调制等电路中,它的单极性导电特性使得电路结构简单,工作可靠。
nmosfet的i-v方程的具体数学推导过程
N型金属氧化物半导体场效应晶体管(NMOSFET)的I-V关
系可以通过以下步骤进行数学推导:
1. 先了解NMOSFET的基本结构和工作原理。
NMOSFET由一个N型沟道、一个P型掺杂的基座和两个控制电极(栅极和
源漏极)组成。
栅极电压可以控制沟道中电荷的密度以及导通状态。
2. 假设NMOSFET处于饱和区工作状态,即沟道中的电压小
于栅极电压,沟道导通。
我们将分析沟道中的电流和栅极电压之间的关系。
3. 分析沟道中的电流。
根据欧姆定律,沟道的电流与电压成正比,且在给定电压下,沟道电流正比于沟道截面的宽度和厚度。
4. 根据Fermi-Dirac分布函数,沟道电流正比于沟道电荷浓度、沟道截面积和迁移率(与材料的物理特性有关)。
5. 根据接近室温下的平衡条件,可以假设沟道电荷密度与沟道电压之间的关系遵循一个非线性函数。
通常,这个函数可以使用二次式来近似。
6. 对沟道电流进行数学建模。
通过对沟道电流进行近似,可以得到一个与栅极电压和源漏电压相关的I-V关系的方程。
请注意,上述仅为简要的数学推导过程,并且实际的I-V关系
方程可能还涉及到更多的修正项和对材料特性的详细考虑。
具体的推导过程可能因材料和器件结构的不同而有所变化。
FinFET基本原理
1。
1 引言
随着半导体器件特征尺寸按摩尔定律等比缩小,芯片集成度不断提高,出现的众多负面效应使传统的平面型MOSFET在半导体技术发展到22nm时遇到了瓶颈,尤其是短沟道效应显著增大,导致器件关态电流急剧增加.尽管提高掺杂浓度可以在一定程度上抑制短沟道效应,然而高掺杂沟道会增大库伦散射,使载流子迁移率下降.
目前,针对此问题已经提出了多种可能的解决措施,主要包括全耗尽绝缘体上硅技术(FDSOI)及三维立体FinFET等.
FinFET最早由胡正明教授提出,是一种新的互补式金氧半导体(CMOS)晶体管,是由一个垂直硅鳍控制的自对准双栅[1].
1。
2 FinFET原理
FinFET与平面型MOSFET结构的主要区别在于其沟道由绝缘衬底上凸起的高而薄的鳍(fin)构成,源漏两极分别在其两端,三栅极紧贴其侧壁和顶部,用于辅助电流控制,如图1。
这种鳍形结构增大了栅围绕沟道的面,加强了栅对沟道的控制,从而可以有效缓解平面器件中出现的短沟道效应,大幅改善电路控制并减少漏电流(leakage),也可以大幅缩短晶体管的闸长也正由于该特性,FinFET无需高掺杂沟道,因此能够有效降低杂质离子散射效应,提高沟道载流子迁移率[1]。
图1 FinFET typical layout and schematic cross sectional
structures。
这种结构的特点有以下5点:
1)用来抑制短沟道效应的超薄si鳍
2)两个栅极是自对准的,同时和源漏也是对准的3)在源漏区生长多晶硅,可以减少寄生电阻
4) 短的(50nm)si鳍是一个准平面结构
5) 栅极后制作工艺,和低温,高k栅介质相兼容。
Vol.32,No.9Journal of Semiconductors September2011 Drive current of accumulation-mode p-channel SOI-based wrap-gated Fin-FETs Zhang Yanbo(张严波)1;2,Du Yandong(杜彦东)1,Xiong Ying(熊莹)1,Yang Xiang(杨香)1,Han Weihua(韩伟华)1; ,and Yang Fuhua(杨富华)1;31Engineering Research Center for Semiconductor Integration Technology,Institute of Semiconductors,Chinese Academy of Sciences,Beijing100083,China2Department of Electronic Engineering,Tsinghua University,Beijing100084,China3State Key Laboratory for Superlattices and Microstructures,Beijing100083,ChinaAbstract:Comparisons are performed to study the drive current of accumulation-mode(AM)p-channel wrap-gated Fin-FETs.The drive current of the AM p-channel FET is15%–26%larger than that of the inversion-mode(IM)p-channel FET with the same wrap-gated fin channel,because of the body current component in the AM FET,which becomes less dominative as the gate overdrive becomes larger.The drive currents of the AM p-channel wrap-gated Fin-FETs are50%larger than those of the AM p-channel planar FETs,which arises from effective conductingsurface broadening and volume accumulation in the AM wrap-gated Fin-FETs.The effective conducting surfacebroadening is due to wrap-gate-induced multi-surface conduction,while the volume accumulation,namely themajority carrier concentration anywhere in the fin cross section exceeding the fin doping density,is due to thecoupling of electric fields from different parts of the wrap gate.Moreover,for AM p-channel wrap-gated Fin-FETs,the current in channel along h110i is larger than that in channel along h100i,which arises from the surface mobilitydifference due to different transport directions and surface orientations.That is more obvious as the gate overdrivebecomes larger,when the surface current component plays a more dominative role in the total current.Key words:accumulation mode;inversion mode;wrap gate;Fin-FET;volume accumulationDOI:10.1088/1674-4926/32/9/094001EEACC:2550N;2560R1.IntroductionNowadays,the gate lengths of field effect transistors (FETs)have scaled down to tens of nanometers and are pre-dicted to scale further in the future.One important issue of fur-ther scaling is the short channel effect(SCE),where the gate encounters competition with the drain in controlling the chan-nel.Therefore,structural improvements are essential for the enhancement of FET performance.The wrap-gated Fin-FET has proven to be a promising candidate thanks to its excellent gate electrostatic control and thus good SCE immunityŒ1;2 .Another important issue of nano-FET scaling is doping profile adjustment in the channel regionŒ3 .Controlling the overlap between the gate and the source/drain extension is a challenge for inversion-mode(IM)FETs.An alternative so-lution is to use an accumulation-mode(AM)FET,whose channel region is doped with the same type of dopant as the source/drain region.When the AM FET is turned off,the chan-nel is fully depleted.When the AM FET is turned on,for a pla-nar structure,carriers transport along both the surface accumu-lation layer and the neutral part of the bodyŒ4 ;for a wrap-gated fin structure,however,the whole fin may be in accumulation (i.e.volume accumulation)thanks to the coupling of electric fields from different parts of the wrap gate,causing an obvi-ous current enhancement because of the higher mobility in the fin center with reduced scatteringŒ4 ,which still needs to be demonstrated.In contrast with the planar channel,where carriers trans-port along one surface,carriers in the wrap-gated fin channel transport along the surfaces all around it.The surfaces may have different orientations,which results in mobility differ-enceŒ5 7 and thus current difference.This may also happen to AM wrap-gated Fin-FETs and needs to be verified.In this work,comparisons are performed to study the drive current of AM p-channel wrap-gated Fin-FETs.Firstly,with the same wrap-gated fin structure,AM p-channel FETs are compared with IM p-channel FETs in drive current to ver-ify the body current component in the AM FETs.Secondly, AM p-channel wrap-gated Fin-FETs are compared with AM p-channel planar FETs to study the body current difference and the possible volume accumulation.Moreover,AM p-channel wrap-gated Fin-FETs with h110i channel are compared with those with h100i channel to check the current difference aris-ing from different orientations of fin surfaces.2.Device fabricationAM p-channel wrap-gated Fin-FETs were fabricated on a boron-doped(1 1015cm 3/,(110)oriented silicon-on-insulator(SOI)wafer with88-nm-thick top silicon.First,e-beam lithography and inductively coupled plasma(ICP)etch-ing were performed to define the active region and the chan-nel fins,followed by a sacrificial oxidation(900ıC,30min) to eliminate the surface damage induced by the etching.Dur-*Project supported by the National High Technology Research and Development Program of China(No.2007AA03Z303)and the National Basic Research Program of China(No.2010CB934104).Corresponding author.Email:weihua@Received30March2011,revised manuscript received29April2011c 2011Chinese Institute of ElectronicsTable1.Structural and electrical parameters of IM/AM wrap-gated Fin-FETs and planar FETs.Section FET type Fin width/height(nm)The number offins in an FETGate length( m)ChanneldirectionV T(V)I on( A)3.1IM Fin-FET90/60110.3h110i 0:9156.2AM Fin-FET90/60110.3h110i 0:6964.63.2AM planar FET5000/6010.30.51h110i 0:700:690:7468.558.949.2AM Fin-FET90/60210.30.51h110i 0:670:750:75102.693.678.53.3AM Fin-FET90/6011h110i 0:689.8AM Fin-FET90/6011h100i 0:758.8Fig.1.Cross section SEM of a wrap-gated fin channel.ing the following sacrificial oxide removal in the buffered hy-drofluoric solution,an under-etching happened in the buriedoxide,causing the silicon fins to be suspended.Then,a30-nm-thick gate oxide was grown in dry oxygen at900ıC,and 150-nm-thick boron-doped polysilicon was deposited by low-pressure chemical vapor deposition(LPCVD),wrapping thefins with the conformal coverage.The polysilicon gate was de-fined by a round of second e-beam lithography and ICP etch-ing.Source/drain doping was carried out by self-aligned BF C2 implantation and rapid annealing at1050ıC.After deposit-ing200-nm-thick silicon dioxide as the protective layer,con-tact holes were opened using optical lithography and ICP etch-ing.Following20-nm-thicknickel film evaporation,the ohmic contact layer i.e.nickel silicide(NiSi)was formed by alloy at 500ıC.The unreacted nickel was removed by a selective etch-ing solution(H2SO4:H2O2D3:1)at80ıC without attack-ing the silicide.To form the electrode pads of source/drain and gate,optical lithography was used to define the pattern in the photoresist and400-nm-thick aluminum was evaporated and lifted off.The device fabrication process was finished with a 400ıC annealing process.Figure1shows the cross section of the wrap-gated fin channel.The width and height of the fin are 90nm and60nm,respectively.For comparison,AM p-channel planar FETs with the same width of active region(W D5 m)as the AM p-channel wrap-Fig.2.Transfer characteristics of the AM and the IM p-channel wrap-gated Fin-FETs for0.3- m-long gate.gated Fin-FETs were fabricated on the same SOI wafer.IM p-channel wrap-gated Fin-FETs were fabricated using the same process as the AM p-channel ones except for an additional n-well doping process(1 1017cm 3/before the channel fin definition.3.Results and discussionStructural and electrical parameters of the devices are shown in Table1,where V T is the threshold voltage extracted using second-derivative methodŒ8 and I on is the drain current when the drain bias(V ds/is–3V and the gate overdrive(V gs–V T/is–2.3V.The results will be discussed by comparison in three aspects.3.1.Accumulation-mode versus inversion-modeFigure2shows transfer curves of the AM and the IM p-channel FETs with the same wrap-gated fin channel along h110i.The V T difference between the AM and the IM FET re-sults from the difference in doping type and density in the chan-nel region.The AM FET is compared with the IM FET in drive current under the same gate overdrive,as shown in Fig.3.The current of the AM FET is15%–26%larger than that of the IM FET,and the percentage gets smaller as the gate overdrive in-creases,indicating that the factor causing the larger current in the AM FET becomes less dominative.The factor is the bodyFig.3.Output characteristics of the same AM and IM p-channel wrap-gated Fin-FETs in Fig.2.Fig.4.Drive currents of AM p-channel wrap-gated Fin-FETs and AM p-channel planar FETs for different gate lengths.current in the AM FET,an additional current component com-pared with the IM FET.As the gate overdrive increases,more holes are accumulated at the surfaces,which increases the sur-face current,but screens the body from the gate electric field and thus limits the body hole increasing.So the body current percentage decreases with gate overdrive increasing,making the current difference between the AM FET and the IM FET smaller.3.2.Wrap-gated fin channel versus planar channelWith the same width of active region(W D5 m)and fab-ricated through the same process,AM p-channel wrap-gated Fin-FETs and AM p-channel planar FETs are compared in drive current for different gate lengths.All the AM p-channel FETs here have channels along h110i and have threshold volt-ages(V T/around 0:7V.As shown in Fig.4,the drive cur-rents of the AM p-channel wrap-gated Fin-FETs are about50% larger than those of the AM p-channel planar FETs,which in-dicate a drive current improvement induced by the wrap-gated fin structure.The current improvement arises from two aspects.First, in contrast with the planar channel with only one conducting surface,the wrap-gated fin has an additional bottom surface Fig.5.Simulated hole distribution over the cross section in the middle point of(a)the fin channel and(b)the planar channel,with both the gate and the drain biased at 3V.and side surfaces that broaden the effective conducting surface. In our case,neglecting the hole mobility difference on differ-ent surfaces,which will not affect the final conclusion because of the largest hole mobility in h110i direction on f110g sur-faceŒ5 7 ,the effective width of the conducting surface in the AM planar FETs is5 m,while that in the AM wrap-gated Fin-FETs is6.3 m.So the effective conducting surface is broad-ened by26%.Obviously,the effective conduction surface broadening is not enough to cause the50%current difference between the wrap-gated fin channel and the planar channel.Therefore,the body current difference plays another important role,namely, the body current in the wrap-gated fin channel is larger than that in the planar channel.Whereas,owing to the spaces between adjacent fins,the physical cross section of the wrap-gated fin channel is smaller than that of the planar channel.In addition, because the holes in the fin body are closer to the side sur-faces than the holes in the body of planar structure,they are more likely to be affected by the surface charge scattering,al-though this scattering effect is very small due to the surface holes’screening.So the body mobility in the fin channel will not exceed that in the planar channel and the only factor result-ing in the body current difference is the body hole concentra-tion,which in the wrap-gated fin channel is larger than that in the planar channel.To understand that point,the hole distribution over the cross section in the middle point of the channel was simulated using an ATLAS3D device simulatorŒ9 ,on which the linear iterative method was used to solve the combination of Pois-son’s equation,drift-diffusion equations including Lombardi’s mobility modelŒ10 ,and carrier continuity equations including Shockley’s carrier generation-recombination modelŒ11 .The structural parameters were set to be60nm in channel thick-ness,90nm in fin width,5 m in planar channel width,0.5 m in gate length and30nm in gate oxide thickness.The doping density of the channel region was1 1015cm 3and the inter-face charge density was set to be consistent with the threshold voltage(V T D 0:7V).Both the gate bias and the drain bias were set to be 3V.As shown in Fig.5,for the AM planarFig.6.Schematic of surface orientations of fins along different direc-tions.channel the hole concentration gets minimum at the bottom in-terface,and the minimum concentration is less than the dopingdensity (1 1015cm 3/owing to the depletion effect of the interface charge,while for the AM wrap-gated fin channel the hole concentration reaches a minimum in the center of the fin cross section and the minimum concentration is two orders of magnitude greater than the fin doping density,which is due to the coupling of electric fields from different parts of the wrap gate.Therefore,hole accumulation occurs all over the cross section of the wrap-gated fin,in other words volume accumu-lation occurs,which we think is another reason why the drive currents of the AM p-channel wrap-gated Fin-FETs are larger than those of the AM p-channel planar FETs.3.3.Considering the different surface orientations of thefin To check the effect of surface orientation on the drive cur-rent of AM p-channel wrap-gated Fin-FETs,fin channels along different directions,h 110i and h 100i ,are designed on the same wafer,which result in f 100g and f 110g side surfaces,respec-tively,after the fins are etched out,as shown schematically in Fig.6.Considering different the mobility on surfaces of differ-ent orientations,the current in a wrap-gated fin is given byI D I body CX jW j 2L effS ;j C OX ;j V gs V T 2;where I body is the body current,W j , S ;j ,C OX ;j are the width,the mobility and the gate oxide capacitance per unit area re-spectively of the j th surface,and L eff ,V gs ,V T are the effective channel length,gate bias and threshold voltage respectively.In our case,because of the same gate oxide thickness and thus the same C OX ;j ,the current along different surfaces is proportional to W j S ;j .It has been reported that the ratio of hole mobility along different surfaces and transport directions is h 110i =.110/: h 100i =.110/: h 110i =.100/D 2.5:1.6:1Œ5 7 .Accordingly,the ratio of surface current is evaluated as shown in Table 2.Figure 7shows the output characteristics of two AM p-channel wrap-gated Fin-FETs with different channel direc-tions,h 110i and h 100i .Owing to the body current component,the total drain current difference between these two directions is less than 19%,i.e.the surface current difference in Table 2.As the gate overdrive gets larger,the surface current compo-nent plays a more dominative role in the total current and thus the surface mobility difference appears more obviously in the total current.Table 2.Evaluation of the ratio of surface current along h 110i to that along h 100i .Current direction Surface orientation Surface mobility ratio Surface width (nm)Surface currentratioh 110i f 110g 2.51801.19f 100g 1120h 100i f 110g 1.63001Fig.7.Output characteristics of AM p-channel wrap-gated Fin-FETs with channel along h 110i and h 100i .4.ConclusionComparisons are performed to study the drive current of AM p-channel wrap-gated Fin-FETs.The drive current of the AM p-channel FET is 15%–26%larger than that of the IM p-channel FET with the same wrap-gated fin channel,due to the body current component in the AM FET,which becomes less dominative as the gate overdrive increases.The drive currents of the AM p-channel wrap-gated Fin-FETs are 50%larger than those of the AM p-channel planar FETs,which arises from ef-fective conduction surface broadening and volume accumula-tion.Moreover,for the AM p-channel wrap-gated Fin-FETs,the current in the channel along h 110i is larger than that in the channel along h 100i ,which arises from the surface mobil-ity difference due to different transport directions and surface orientations.That is more obvious as the gate overdrive gets larger,when the surface current component plays a more dom-inative role in the total current.AcknowledgementsThe authors are grateful to Dr.Li Y .Q.and Dr.Luo Q.for their help with SEM and Dr.Song R.L.for his help with TCAD simulation.References[1]Colinge J P.Multiple-gate SOI MOSFETs.Solid-State Electron,2004,48(6):897[2]Colinge J P.Multi-gate SOI MOSFETs.Microelectron Eng,2007,84(9/10):2071[3]Yang F L,Lee D H,Chen H Y ,et al.5-nm-gate nanowire FinFET.Symposium on VLSI Technology,2004:196[4]Colinge J P.Silicon-on-insulator technology:materials to VLSI.2nd ed.Norwell,Massachusetts:Kluwer Academic Publishers, 1997[5]Colman D,Bate R T,Mize J P.Mobility anisotropy and piezore-sistance in silicon p-type inversion layers.J Appl Phys,1968, 39(4):1923[6]Sato T,Takeishi Y,Hara H.Effects of crystallographic orientationon mobility,surface state density,and noise in p-type inversion layers on oxidized silicon surfaces.Jpn J Appl Phys,1969,8(5): 588[7]Yang M,Ieong M,Shi L,et al.High performance CMOS fab-ricated on hybrid substrate with different crystal orientations.IEDM Tech Dig,2003:453[8]Ortiz-Conde A,Garc´ıa S´a nchez F J,Liou J J,et al.A review ofrecent MOSFET threshold voltage extraction methods.Micro-electron Reliab,2002,42(4/5):583[9]SILV ACO International.ATLAS User’s Manual:Device simula-tion software,Santa Clara,California,2007[10]Lombardi C,Manzini S,Saporito A,et al.A physically basedmobility model for numerical simulation of non-planar devices.IEEE Trans CAD,1988,7(11):1164[11]Shockley W,Read W T.Statistics of the recombination of holesand electrons.Phys Rev,1952,87(5):835。