3D TCAD Simulation Dopant Distribution Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for simulating integrated circuit processing struggle to accurately transform representations of integrated circuits, particularly in generating three-dimensional dopant distributions, which are computationally expensive and often result in incomplete or inaccurate doping profiles.
Innovation Solution
A system and method that uses a non-transitory computer-readable medium with instructions to transform a first representation of an integrated circuit into a second representation by performing process simulations, generating one-dimensional and two-dimensional dopant profiles, and combining them to create a three-dimensional dopant distribution, utilizing customized lateral diffusion functions and spreading data to conserve dopant quantities and achieve results within 10% of complete 3D simulations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complete 3D simulations are performed to generate accurate three-dimensional dopant distributions, then manufacturing precision is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the three-dimensional dopant distribution generation into two independent one-dimensional simulations (lateral profile and depth profile). These 1D simulations are performed separately and then combined through convolution to produce the final 3D dopant distribution, avoiding the need for computationally intensive complete 3D simulations while maintaining accuracy within 10% of full 3D results.
2Manufacturing precision
If complete 3D simulations are performed to generate accurate three-dimensional dopant distributions, then manufacturing precision is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the three-dimensional dopant distribution generation into two independent one-dimensional simulations (lateral profile and depth profile). These 1D simulations are performed separately and then combined through convolution to produce the final 3D dopant distribution, avoiding the need for computationally intensive complete 3D simulations while maintaining accuracy within 10% of full 3D results.
3Productivity
If simplified methods are used to generate dopant distributions, then computational efficiency is improved, but accuracy of dopant distribution deteriorates
Solution Approach 1:
The patent transforms the 3D simulation problem into a combination of 1D simulations by introducing dimensional separation. The lateral diffusion function (2D convolution) combines the 1D lateral profile with mask geometry, and then this 2D lateral distribution is combined with the 1D depth profile to create the final 3D dopant distribution. This dimensional decomposition maintains accuracy while dramatically improving computational efficiency.
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 approach efficiently generates accurate three-dimensional dopant distributions with reduced computational resources, effectively simulating integrated circuit processing and improving modeling accuracy, while being applicable to various process conditions and mask types.
Implementation Method 1
generating a two-dimensional lateral dopant profile from adding the first dopant under the first set of process conditions with a first mask corresponding to the first dopant, by convolving the first mask with a lateral diffusion function in two dimensions
Data Source
AI summary
A first representation of an integrated circuit undergoing processing is transformed into a second representation. The second representation including additional dopants relative to the first representation. The transformation generates a three-dimensional dopant distribution from adding a first dopant under a first set of process conditions with a mask, by combining the two-dimensional lateral profile of the dopant with the one-dimensional depth profile of the dopant. The one-dimensional depth profile of the dopant is retrieved from a database storing selected results from earlier process simulation of the first addition of the first dopant under the first set of process conditions. The two-dimensional lateral dopant profile from adding the first dopant under the first set of process conditions with a first mask corresponding to the first dopant, is generated by convolving the mask with a lateral diffusion function, or from at least one solution to the 2D diffusion equation without convolution.