Abstract

Massive advances in microelectronic manufacturing technology with an exponential growth of their complexity and speed are needed to ensure a continuous development of novel techniques, structures, devices, circuits and systems. This paper is intended for the introduction of a new PMOS transient model for modern microelectronic devices that provides a fast transient response. Such suggested model expresses the transient terminal currents as polynomial functions of the normalized channel charge densities at the channel bounds with the assistance of a modified version of the cubic spline collocation methodology in symmetrical telescopic fashion. Additionally, the optimum number of segments, which is suitable for the new version of the cubic spline collocation algorithm, is investigated. Moreover, the normalized channel charge density at collocation points is modeled in terms of its values at the channel bounds through the quasi-static approach. Furthermore, by means of introducing an inverse function for the normalized overdrive channel voltage, the transient terminal currents are formulated as a function of the terminal voltages. In comparison with usual cubic spline collocation structure, the novel model has much better accuracy in its application to EKV structure. The developed model has been applied to the standard 0.15 μm technology and validated by MATLAB R2014a. The obtained results demonstrate that it gives a very high degree of relative accuracy, on average of 99%, for the total time and exhibits absolute error of less than 5% of the maximum value, in its worst case.

中文翻译：

---

适用于EKV PMOS模型的现代微电子器件的瞬态模型

摘要

为了确保新技术，结构，器件，电路和系统的不断发展，需要微电子制造技术的大规模进步以及其复杂性和速度的指数级增长。本文旨在介绍一种用于现代微电子器件的新型PMOS瞬态模型，该模型可提供快速的瞬态响应。这种建议的模型借助于三次样条并置方法的修改版本以对称伸缩方式将瞬态终端电流表示为通道边界处归一化通道电荷密度的多项式函数。此外，研究了适用于新版本三次样条并置算法的最佳段数。而且，通过准静态方法，根据其在通道边界处的值对并置点处的归一化通道电荷密度进行建模。此外，通过为归一化的过驱动通道电压引入反函数，将瞬态端子电流公式化为端子电压的函数。与通常的三次样条并置结构相比，该新模型在EKV结构中的应用具有更好的精度。开发的模型已应用于标准0.15μm技术，并已通过MATLAB R2014a验证。所获得的结果表明，在整个时间段内，它的相对准确度非常高，平均为99％，在最坏的情况下，绝对误差小于最大值的5％。