Epitaxial Growth Parameter Tuning Using Response Function Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for determining process parameters in epitaxial growth systems are time-consuming and costly, requiring extensive data analysis with large error margins, making them impractical for rapid adjustments in large-scale production environments.
Innovation Solution
A parameter determination device and method that computes optimum process parameter values using response functions derived from measured characteristic data, allowing for quick adjustments to maintain uniform epitaxial film thickness distribution by varying specific parameters and simulating optimal conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple regression analysis is used to compute influence coefficients from past film thickness adjustment results, then the film thickness distribution can be optimized, but the computation is time-consuming and requires a large number of data points acquired over long periods
Solution Approach 1:
The patent pre-calculates and stores sensitivity coefficients (partial derivatives of film thickness with respect to each parameter) for various positions on the wafer surface before actual deposition. These pre-computed coefficients are stored in a lookup table, allowing the system to quickly determine optimal parameters by simple interpolation rather than performing time-consuming multiple regression analysis during parameter determination
Solution Approach 2:
The patent creates a simplified computational model that copies the essential relationships between parameters and film thickness distribution. Instead of using the full complex multiple regression analysis, the system uses pre-computed sensitivity coefficients that capture the key relationships, enabling rapid parameter determination while maintaining accuracy
2Adaptability or versatility
If multiple regression analysis is performed on data from slightly different environments and system conditions, then comprehensive parameter optimization is achieved, but the analysis has large error margins due to data state differences between acquisition and analysis
Solution Approach 1:
The patent transforms the approach by changing from using raw historical data requiring regression analysis to using pre-computed sensitivity coefficients. These coefficients are calculated for specific parameter variations and stored, allowing the system to adapt to different conditions by selecting appropriate pre-computed coefficients rather than re-analyzing historical data with varying environmental conditions
Solution Approach 2:
The system pre-calculates sensitivity coefficients for various parameter changes and environmental conditions, storing these in advance. When parameter determination is needed, the system selects from these pre-computed coefficients rather than performing new analysis, eliminating errors associated with data state changes between acquisition and analysis
3Manufacturing precision
If test deposition and fine adjustment of process parameters are performed based on predicted values, then the film thickness distribution is optimized, but the process is time-consuming and costly when performed at every maintenance point and product changeover
Solution Approach 1:
The patent implements a self-service system where the determination device automatically computes optimal parameters using pre-stored sensitivity coefficients and current system state information. This eliminates the need for manual test depositions and expert adjustment, allowing the system to self-optimize parameters quickly at maintenance points and product changeovers without time-consuming trial-and-error processes
Solution Approach 2:
The patent replaces the mechanical trial-and-error adjustment process with a computational system. Instead of performing physical test depositions and manually adjusting parameters based on experience, the system uses computational algorithms with pre-stored sensitivity coefficients to directly determine optimal parameters, substituting computational analysis for physical experimentation
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
Enables the derivation of optimum process parameter values in a short period and at low costs, reducing the time and expense associated with maintaining uniform film thickness distribution in epitaxial growth systems.
Implementation Method 1
The computation unit has a following equation for a characteristic Qj, where Qj is the characteristic data, x is the position, Pi is the one or more process parameters, and M is a total number of the one or more parameters... when the parameterPk at i=k is varied by ΔPk, and of the one or more parameters P1, ..., PM, the parameter Pi at i≠k is not varied, an amount of variation ΔQj,k of the characteristic Qj is expressed by the following equation
Data Source
Figure 1~2
Figure 3~4
Figure 5~6A
AI summary
A parameter determination device, a parameter determination method, and a parameter determination program for computing optimum values of one or more parameters for an epitaxial growth system in a short time and at low costs are provided. The device includes an input unit that accepts a plurality of values for each of one or more parameters, position data of an epitaxial film, and characteristic data measured by a characteristic measurement device for a deposited epitaxial film; and a computation unit that computes an optimum value of each of the one or more parameters from the position data, the characteristic data, and the values for each of the one or more parameters. The computation unit derives the response function FQj(x, ΔPk) to compute the optimum value of the one or more parameters by simulation.