本文選題：MOSFET電勢模型 + 短溝道效應(yīng)　； 參考：《華北電力大學(xué)(北京)》2017年碩士論文

【摘要】：隨著半導(dǎo)體器件的尺寸不斷減小,短溝道效應(yīng)和漏致勢壘下降效應(yīng)等次級物理效應(yīng)不斷增強,嚴(yán)重影響了器件性能,給器件和電路的仿真模擬帶來了一系列的挑戰(zhàn)。解析模型能夠給出描述MOSFET在亞閾值狀態(tài)下短溝效應(yīng)對器件性能影響的解析表達(dá)式,因此精準(zhǔn)的MOSFET解析電勢模型是迫切需要的。論文首先對伯克利短溝閾值電壓模型、電壓-摻雜轉(zhuǎn)換模型以及二維雙區(qū)模型、單區(qū)模型在模型構(gòu)建和解析方程方面做了詳細(xì)描述。針對二維雙區(qū)模型,為了提高其模型精度,改進(jìn)了以下三個方面:一、增加了特征函數(shù)的展開項數(shù);二、修改了源漏邊界條件,小于結(jié)深的邊界依然是常數(shù)電勢,大于結(jié)深部分的邊界條件采用了耗盡近似,稱為淺結(jié)條件;三、考慮到小尺寸器件的耗盡層厚度會受漏電壓等因素的影響,用電壓-摻雜轉(zhuǎn)換模型耗盡層厚度的計算方法校正了雙區(qū)模型。同時定義了一個平均誤差,該誤差可以評測模型滿足源漏邊界條件的情況。由于伯克利短溝閾值電壓模型是準(zhǔn)二維模型,只能描述表面處的電勢分布,本文將其拓展為二維模型使之能夠得到二維電勢分布,這樣可研究襯偏效應(yīng)對亞閾值擺幅的影響。和二維雙區(qū)模型不同的是二維拓展的伯克利短溝閾值電壓模型中源漏邊界條件是常數(shù)。本文在不同襯底電壓、溝道摻雜濃度和漏電壓下,分別計算了雙區(qū)和單區(qū)模型的源漏平均誤差,發(fā)現(xiàn)雙區(qū)模型的誤差要小得多,并給出了兩模型隨柵長和漏電壓減小的閾值電壓滾降趨勢,初步確定了雙區(qū)模型具有更高的精度。計算了雙區(qū)模型的亞閾值斜率,與伯克利短溝閾值電壓模型及其二維拓展模型、電壓-摻雜轉(zhuǎn)換模型做了比較,結(jié)果顯示雙區(qū)模型的亞閾值斜率無論是變化趨勢還是數(shù)值大小都與實驗數(shù)據(jù)最為接近,且淺結(jié)條件要比常數(shù)條件更適合于小尺寸器件。同時發(fā)現(xiàn)對于二維拓展的伯克利短溝閾值電壓模型,當(dāng)結(jié)深遠(yuǎn)小于耗盡區(qū)厚度時,矩形求解區(qū)域的襯底邊界電勢斜率非零,表明源漏常數(shù)邊界條件近似不成立。綜上結(jié)果表明,校正后的雙區(qū)模型在淺結(jié)邊界條件時的精度最好。因為BSIM6.1模型在電路設(shè)計中更加方便,最后一章利用基因算法實現(xiàn)了BSIM6.1模型與BSIM3v3.2模型的參數(shù)轉(zhuǎn)換。作為初步工作,我們在不考慮襯偏效應(yīng)的情況下實現(xiàn)了利用BSIM6.1模型擬合BSIM3v3.2模型的I-V曲線。
[Abstract]:With the decrease of semiconductor device size, the secondary physical effects, such as short channel effect and leakage barrier drop effect, are increasing, which seriously affect the performance of semiconductor devices and bring a series of challenges to the simulation of devices and circuits. The analytical model can give an analytical expression to describe the effect of short-channel effect on the performance of MOSFET in sub-threshold state, so precise analytical potential model of MOSFET is urgently needed. In this paper, the Berkeley short-channel threshold voltage model, the voltage-doping conversion model, the two-dimensional two-zone model and the single-zone model are described in detail in terms of model construction and analytical equations. In order to improve the accuracy of the two-dimensional two-zone model, the following three aspects are improved: first, the expansion term of the eigenfunction is increased; second, the boundary condition of source and drain is modified, and the boundary less than the junction depth is still a constant potential. The boundary condition which is larger than the junction depth adopts the depletion approximation, which is called the shallow junction condition. Thirdly, considering that the thickness of the depletion layer of the small-sized device is affected by the leakage voltage and other factors, The two-zone model is corrected by calculating the depletion layer thickness of the voltage-dopant conversion model. At the same time, an average error is defined, which can be used to evaluate the condition that the model satisfies the boundary condition of source and drain. Since the Berkeley short-channel threshold voltage model is a quasi-two-dimensional model, it can only describe the potential distribution on the surface. In this paper, it is extended to a two-dimensional model so that the two-dimensional potential distribution can be obtained, so that the influence of the bias effect on the sub-threshold swing can be studied. Different from the two-dimensional two-zone model, the source and drain boundary conditions are constant in the two-dimension extended Berkeley short-channel threshold voltage model. In this paper, under different substrate voltages, channel doping concentrations and leakage voltages, the source and drain mean errors of two-zone model and single-zone model are calculated, respectively. It is found that the error of two-zone model is much smaller. The trend of threshold voltage rolling with the decrease of gate length and leakage voltage is given, and it is preliminarily determined that the two-zone model has higher accuracy. The sub-threshold slope of the two-zone model is calculated and compared with the Berkeley short-channel threshold voltage model and its two-dimensional expansion model, and the voltage-doping conversion model. The results show that the sub-threshold slope of the two-zone model is most close to the experimental data, and the shallow junction condition is more suitable for small-sized devices than the constant condition. It is also found that for the two-dimensional extended Berkeley short-channel threshold voltage model, when the junction depth is less than the depletion region thickness, the potential slope of the substrate boundary of the rectangular solution region is non-zero, which indicates that the boundary condition of the source and drain constant is approximately untenable. The above results show that the corrected two-zone model has the best accuracy under the shallow junction boundary condition. Because BSIM 6.1 model is more convenient in circuit design, in the last chapter, the parameter conversion between BSIM 6.1 model and BSIM3 v3.2 model is realized by genetic algorithm. As a preliminary work, the I-V curve of BSIM3 v3.2 model is fitted by BSIM6.1 model without considering the bias effect.
【學(xué)位授予單位】：華北電力大學(xué)(北京)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TN386

【參考文獻(xiàn)】

相關(guān)期刊論文 前7條

1 韓名君;柯導(dǎo)明;;超淺結(jié)亞45nm MOSFET亞閾值區(qū)二維電勢模型[J];電子學(xué)報;2015年01期

2 孟堅;柯導(dǎo)明;韓名君;;亞90nm溝道MOSFET亞閾值狀態(tài)下二維電勢和閾值電壓的半解析模型[J];中國科學(xué):信息科學(xué);2013年11期

3 韓名君;柯導(dǎo)明;遲曉麗;王敏;王保童;;超短溝道MOSFET電勢的二維半解析模型[J];物理學(xué)報;2013年09期

4 黃力;黃安平;鄭曉虎;肖志松;王玫;;高k介質(zhì)在新型半導(dǎo)體器件中的應(yīng)用[J];物理學(xué)報;2012年13期

5 秦珊珊;張鶴鳴;胡輝勇;屈江濤;王冠宇;肖慶;舒鈺;;應(yīng)變Si全耗盡SOI MOSFET二維亞閾電流模型[J];物理學(xué)報;2011年05期

6 陳志堅,鄭學(xué)仁,姚若河,李斌;深亞微米CMOS器件建模與BSIM模型[J];微納電子技術(shù);2004年06期

7 夏增浪,胡貴才;亞微米MOSFET的BSIM3模型參數(shù)提取[J];微電子學(xué)與計算機;1999年04期

相關(guān)碩士學(xué)位論文 前1條

1 王保童;超短溝道MOSFET二維模型的半解析法求解[D];安徽大學(xué);2014年

，

本文編號：1994454