新的電荷幫浦技術暨應用於帶對帶穿隧所引發之熱載子應力的可靠性分析

Abstract

在本論文中,我們發展了一種可以萃取界面陷阱和氧化層被捕捉電荷的新的電荷幫浦技術.藉由此新的技術,我們探討了帶對帶穿隧所引發之熱載子效應.此外,為了能正確的計測出界面陷阱和氧化層被捕捉電荷,我們也探討了逆短通道效應,提出一個新的模式來模擬此效應.再者,我們亦推導了一個考慮此逆短通道效應的新的解析臨界電壓模式. 第一章包含有關於我們研究動機的概括論述及介紹本論文的組織架構.在第二章中,根據費斯克第一與第二定律,我們對n-通道與p-通道金氧半電晶體的逆短通道效應發展了一個摻雜重新分佈模式,此模式不僅考慮橫向的摻雜重新分佈,亦考慮了縱向的摻雜重新分佈.在此,我們用一個空乏參數來考慮因基層的摻雜重新分佈所導致的摻雜空乏效應.此模式引進四個參數來描述逆短通道效應,每一個參數皆有其物理意並且經由二維元件模擬器來驗證.模擬的結果顯示與n-通道和p-通道金氧半電晶體元件的實驗數據有很好的吻合.顯式可用此模式來描述具有逆短通道效應之金氧半電晶體元件的特徵. 在第三章中, 以格林定律為基礎並結合適當之邊界條件,我們對於具有逆短通道效應的次微米n-通道金氧半電晶體元件提出了一個臨界電壓模式.此臨界電壓模式不僅考慮橫向摻雜重新分佈,同時也考慮了縱向的摻雜空乏效應.此模式中有四個參數用來描述逆短通道效應,每一個參數都有其物理意義,可幫助我們了解逆短通道效應.經由二維數值模擬器的驗證,顯示此解析臨界電壓模式可提供很好的預測.更進一部與短通道之n-通道金氧半電晶體元件的實驗數據做比較,顯示當通道長度縮小至深次微米區域仍能得到令人滿意的吻合.此結果更增加了這個解析模式的準確性. 在第四章中,我們發展了一個新的電荷幫浦技術用以計測及分析熱載子應力所導致之局部破壞.藉由保持閘極脈波之升起與下降的斜率然而改變閘極脈波之高電壓與低電壓的準位,我們可計測出界面陷阱與氧化層被捕捉電荷的橫向分佈.實驗結果顯示當萃取氧化層被捕捉電荷時,改變閘極脈波之高電壓準位與改變閘極脈波之低電壓準位的萃取方法所得到的結果有相互矛盾之處.因此,當我們在萃取經由熱載子應力所產生之界面陷阱與氧化層被捕捉電荷時,必須做一些修正以去除界面陷阱對計測氧化層被捕捉電荷時所造成的干擾. 在第五章中,我們利用新發展之電荷幫浦技術來分析經由帶對帶穿隧所引發之熱載子應力所產生之界面陷阱與氧化層被捕捉電荷的橫向分佈.實驗結果顯示由於所產生的界面陷阱與氧化層被捕捉電荷其橫向分佈的最大值落於不同的位置,因此界面陷阱與氧化層被捕捉電荷應該分別由不同型態的載子所產生.更進一步的證據說明界面陷阱在能隙中之分佈顯示往閘極邊緣移動之載子為產生界面陷阱的主要來源,然而遠離閘極邊緣移動之載子在縱向電場的幫助之下將產生氧化層被捕捉電荷.這些實驗結果對於快閃記憶體之可靠性分析將有極大的助益. 第六章將本論文的重要貢獻作一總結,並提出後續研究方向.

A new Charge-Pumping(CP) technique is developed to extract the interface-states and oxide-trapped charges in MOSFET's in this thesis. Based on this new technique, the Band-To-Band Tunneling(BTBT)-induced hot-carrier stress is investigated. Besides, to correctly profile the interface-states and oxide-trapped charges, the Reverse Short-Channel Effect(RSCE) on the threshold-voltage of short-channel devices are investigated. An accurate model is proposed to simulate this effect. Furthermore, an analytic threshold-voltage model is derived with considering the two-dimensional effects. Chapter 1 includes a general introduction concerned with the motivation and the organization of this thesis. In Chapter 2, a new doping redistribution model based on Fisk's first and second laws is proposed for the RSCE in both p- and n-MOSFET devices, which considers not only the lateral but also vertical doping redistribution. Moreover, the doping depletion effect in the bulk due to the doping redistribution is also considered by using a depletion coefficient. This model introduces four parameters to describe the RSCE, each of them has its physical meaning and is verified by a 2-dimensional device simulator. Simulation results show good agreements as compared with experimental data of both p- and n-MOSFETs. This implies that this model can be used to characterize the short-channel MOS devices with the RSCE. In Chapter 3, a new physical model for the threshold voltage of submicrometer n-MOSFET devices considering the RSCE is presented, which is based on the Green function technique combining with the appropriate boundary conditions. The derived threshold-voltage model accounts for not only lateral doping redistribrtion but also vertical doping depletion. Four parameters are introduced to describe the RSCE, each of them has its physical meaning and can help us to get more insight into the RSCE. As verified by 2-D numerical analysis, the analytical threshold-voltage model provides a good prediction and is further compared with experimental data for short-channel n-MOSFET's. It is shown that satisfactory agreement for channel length down to deep-submicrometer range has been obtained and this comfirms the accuracy of the proposed model. In Chapter 4, a new charge-pumping method has been developed to characterize the hot-carrier induced local damages. By holding the rising and falling slopes of the gate pulse constant and then varying the high-level($V_{GH}$) and base-level($V_{GL}$) voltages, the lateral distribution of interface-states($N_{it}(x)$) and oxide-trapped charges($Q_{ox}(x)$) can be profiled. The experimental results show that during extracting $Q_{ox}(x)$ after hot-carrier stress, a contradictory result occurs between the extraction methods by varing the high-level($V_{GH}$) and base-level($V_{GL}$) voltages. As a result, some modifications are made to eliminate the perturbation induced by the generated interface-states after hot-carrier stress for extracting $Q_{ox}(x)$. In Chapter 5, the lateral distributions of interface-states($N_{it}$) and oxide-trapped charges($Q_{ox}$) generated by BTBT-induced hot-carrier stress are analyzed by the new charge-pumping method. It is shown that the interface-states and oxide-trapped charges should be originated from different types of carriers due to the separation of the locations of their peak values. The further evidence of the measured distribution of the interface-states in the band-gap shows that the carriers travelled toward the gate edge would be the dominant carrier for the generation of interface-states while the carriers travelled away from the gate edge will generate oxide-trapped charges through the help of the vertical electric field. These results should be very useful for the reliability analysis of flash memories. Finally, conclusions are given in Chapter 6, where the major contributions of this thesis and some possible future researches are proposed. 2 A new doping redistribution model for the reverse short-channel effect of short-channel MOSFET's 3 A new analytical threshold voltage model for the reverse short-channel effect of short-channel n-MOSFET's 4 A new charge-pumping technique for profiling the interface-states and oxide-trapped charges in MOSFET's 5 A new observation of band-to-band tunneling induced hot-carrier stress using charge-pumping technique 6 Conclusions