Simulation Method for Electron Device Random Dopant Fluctuations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The drift-diffusion model, used for simulating electron devices, is limited in its ability to accurately reproduce random dopant fluctuations due to its continuum approximation, which becomes invalid as semiconductor devices are miniaturized, leading to inconsistencies between the continuous charge distribution assumption and the discrete nature of impurity ions, resulting in unreliable device simulations.
Innovation Solution
A method is introduced to simulate random dopant fluctuations by defining a screening length to distinguish between long-range and short-range terms in the drift-diffusion model, allowing for a more accurate representation of discrete impurity distributions and their impact on charge distribution, using specific formulae to calculate the screening length and incorporating an image charge to model the location-dependent behavior of impurity ions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the drift-diffusion model uses continuum approximation to simulate charge distribution, then the simulation is computationally efficient and mathematically simplified, but the model fails to accurately reproduce random dopant fluctuations in miniaturized semiconductor devices
Solution Approach 1:
The patent segments the continuous charge distribution into discrete impurity ions by introducing a screening length parameter. This allows the model to distinguish between long-range (continuum) and short-range (discrete) effects, enabling accurate representation of random dopant fluctuations while maintaining computational efficiency through selective discretization.
Solution Approach 2:
The patent applies different modeling approaches to different spatial regions based on the screening length. Within the screening length, discrete impurity ion behavior is captured; beyond this length, continuum approximation is used. This local differentiation resolves the contradiction by applying high accuracy where needed while maintaining overall computational efficiency.
2Device complexity
If the drift-diffusion model assumes continuous charge distribution, then the mathematical formulation is simplified, but the model cannot accurately represent the discrete nature of impurity ions in miniaturized devices
Solution Approach 1:
The screening length acts as an intermediary parameter that bridges the continuum approximation and discrete impurity ion representation. It provides a mathematical transition mechanism that allows the model to represent discrete ion effects without fully abandoning the continuum framework, thus maintaining formulation simplicity while improving representation accuracy.
3Productivity
If semiconductor devices are miniaturized according to device-scaling rule, then device performance and integration density improve, but random dopant fluctuations become more significant and simulation reliability deteriorates
Solution Approach 1:
The patent introduces the screening length as a new parameter that changes with device size. By adjusting this parameter according to the device scaling factor, the model adapts to miniaturization effects and maintains simulation reliability. The screening length effectively captures the enhanced impact of random dopant fluctuations in smaller devices without requiring complete model reformulation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables a more reliable simulation of random dopant fluctuations, reducing computational resources required and improving the accuracy of device simulations, particularly in capturing threshold voltage variations, while maintaining consistency with the drift-diffusion model.
Implementation Method 1
the drift-diffusion model with an external potential applied to the PN junction as a boundary condition
Implementation Method 2
the drift-diffusion model with an external potential applied to the PN junction as a boundary condition
Implementation Method 3
the Poisson equation, wherein the charge density as well as the second derivatives of potential distribution inside the electron device
Data Source
AI summary
There is a significant precaution when performing random dopant fluctuation by using the drift-diffusion model that is the basis of the conventional device simulation. Because the continuation by a long wavelength approximation was done to derive said drift-diffusion model. That is how to recover the location dependence of discrete impurity ions in the long wavelength approximation. For example, in the case that there is an impurity ion near to the interface to an insulating film, the charge density of an impurity ion, which was made continuous in the conventional method, is unable to catch the charge density change due to polarization at the interface. Because this polarization is dependent of the location of a discrete impurity ion near to the interface.A method for simply implementing the effect of polarization to the device simulation is provided by appending an image charge inside the insulating film to linearize the charge of discrete impurity ion which locates near to the interface to satisfy the consistency to the drift-diffusion model while keeping the location dependency of the discrete ion.