新式快閃式記憶體電容耦合參數之萃取方法

Abstract

快閃式記憶體與dummy cell作用在次臨界區且操作在基體效應之情況下將面臨高估電容耦合係數的問題, 現今的次臨界電流斜率法也將面臨同樣的議題。本論文利用雙參數之次臨界電流公式: = 之電流係數 與斜率係數 可將此種無法忍受之估計錯誤歸究於因製程變動而引起之快閃式記憶體與dummy cell電流不匹配, 為了降低此種製程變動的因素, 我們提出一種利用基體效應之新式萃取方法將dummy cell操控在次臨界區之上並將快閃式記憶體操作在次臨界區之內, 並利用量測背閘偏壓與臨界電壓之關係求出斜率係數n , 由此方式所萃取出之電容耦合係數將相當貼近其本身的設計值, 此種萃取方法本身的快捷性將提供一種製程生產線上對電容耦合係數之監視工具。

Overestimation of capacitance coupling coefficients in flash memory cells is encountered in the case of body effect while operating both flash memory cells and dummy transistors in subthreshold region. Existing subthreshold slope method also faces the same problem. The origin of these intolerable error can be pointed to process variations induced current mismatch between flash memory cells and dummy transistors, as explained in terms of current factor and slope factor n in a two-parameters subthreshold current model: . To minimize the effect of process variations, a new method incorporating body effect is built, in which dummy transistors are biased above-threshold while still operating flash memory cells in subthreshold. Once the slope factor n is gotten from threshold voltage versus source-to-substrate bias measurement, strikingly the resultant capacitance coupling coefficients are found to be fairly close to design value. This method is also fast and thus is able to serve as in-line monitor of capacitance coupling coefficients.