Simulator for Semiconductor Damage Distribution Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current simulation techniques for evaluating damage from ion injection in device manufacturing, such as etching and PVD, face challenges in accurately and efficiently calculating damage distribution due to high computational complexity and time requirements, particularly when dealing with light particles like hydrogen and complex patterns.
Innovation Solution
A simulator utilizing the Flux method to calculate the relationship between injected and released substances during a process, reducing computational burden by using a database of reaction parameters and considering solid angle effects for patterned surfaces, allowing for faster damage distribution prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If molecular dynamics simulation technique is used to predict crystal lattice disorder in target film, then measurement precision of damage distribution is improved, but calculation time increases enormously
Solution Approach 1:
The patent pre-calculates damage distribution data for various incident ion conditions (energy, angle, film type) and stores them in a database before actual simulation runs. This preliminary action eliminates the need to perform computationally intensive molecular dynamics calculations during each simulation, reducing calculation time from weeks to minutes while maintaining prediction accuracy through database lookup based on Monte Carlo calculated parameters.
Solution Approach 2:
The patent divides the simulation process into two separate stages: (1) Monte Carlo method to calculate incident particle position and angle on the pattern surface, and (2) database lookup to retrieve pre-calculated damage distribution data. This segmentation allows each method to be optimized independently, combining the geometric accuracy of Monte Carlo with the speed of pre-computed data.
2Adaptability or versatility
If molecular dynamics simulation is applied to real pattern with light particles like hydrogen, then damage distribution of complex pattern is evaluated, but calculation time becomes excessively long due to long flying distance of incident ions
Solution Approach 1:
The patent pre-calculates and stores damage distribution data for various incident ion conditions including light particles like hydrogen in a database. This preliminary computation covers the wide range of conditions before actual simulation, so that during practical use, only fast database lookup and Monte Carlo calculation are needed, regardless of particle type or pattern complexity.
Solution Approach 2:
The patent introduces a database as an intermediary between the Monte Carlo calculation and the damage prediction. The database stores pre-computed damage distribution data that acts as a mediator, translating the geometric parameters from Monte Carlo into damage predictions without requiring real-time molecular dynamics simulation, thus enabling fast evaluation for any particle type.
3Measurement precision
If Monte Carlo method is used to calculate incident particle position and angle, then transportation route in real pattern is considered, but calculation time is still long when large number of particles are required for high accuracy
Solution Approach 1:
The patent performs Monte Carlo calculation only for the specific parameters needed (incident position and angle on the pattern surface) rather than tracking the complete transportation route of all particles through the entire film thickness. This partial action approach focuses computational effort only on the geometric parameters that affect damage distribution, eliminating unnecessary calculations while maintaining accuracy.
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
The simulator significantly reduces calculation time and improves accuracy in predicting damage distribution across larger areas, enabling quicker optimization of processing conditions and mask pattern layouts.
Implementation Method 1
calculating, using the Flux method, a relationship between an amount of a first substance externally injected onto a given evaluation point on a workpiece during a given process and an amount of a second substance released from the given evaluation point on the workpiece as a result of the injection of the first substance
Data Source
AI summary
Disclosed herein is a simulator including: an input section adapted to acquire processing conditions for a given process performed on a workpiece; and a damage calculation section adapted to acquire the damage of the workpiece, based on the processing conditions, by calculating, using a Flux method, the relationship between the amount of a first substance externally injected onto a given evaluation point on the workpiece during the given process and the amount of a second substance released from the given evaluation point on the workpiece as a result of the injection of the first substance.