Etch End-Point Detection Pipeline Using RF Signal Pixel Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional end-point detection systems for semiconductor etching lack sensitivity and rely on timed recipes, leading to increased risk of process faults and decreased uniformity, as they fail to effectively utilize multivariate sensor output and do not incorporate real-time analysis of RF signals generated by electric and magnetic field antenna configurations.
Innovation Solution
A system utilizing E-field and B-field antennas to generate RF signals, combined with a computing device for real-time processing, including spectrum analysis and machine learning, to predict the end-point of semiconductor etching processes by preprocessing sensor data, selecting best scoring signals, and performing an end-point prediction algorithm in pixel space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional end-point detection systems are used, then the system simplicity is maintained, but the detection precision and reliability deteriorate due to lack of sensitivity and reliance on timed recipes
Solution Approach 1:
The patent segments the complex sensor output into multiple individual signal channels, each representing different physical phenomena during etching. By processing each signal separately through the pipeline and then combining results, the system achieves precise end-point detection while managing complexity through modular signal handling rather than treating the multivariate output as a single complex dataset
Solution Approach 2:
The patent introduces an intermediary processing pipeline including spectrum analyzer circuitry, preprocessing module, and machine learning inference engine. This intermediary layer transforms raw multivariate sensor data into actionable end-point predictions, bridging the gap between complex sensor output and reliable detection without requiring direct complex system architecture
2Manufacturing precision
If timed recipes are used for etching control, then the ease of operation is maintained, but the manufacturing precision deteriorates due to increased risk of over- or under-etching
Solution Approach 1:
The patent implements real-time feedback by continuously monitoring sensor data during the etching process and dynamically adjusting the end-point determination. The machine learning model analyzes incoming sensor signals and provides feedback control, allowing the system to adapt to actual process conditions rather than relying on predetermined timed recipes, thereby improving etching uniformity
Solution Approach 2:
The patent performs preliminary training of the machine learning model using historical sensor data and known end-point conditions. This preliminary action prepares the system to automatically recognize end-point patterns during actual etching processes, reducing the need for complex real-time operator intervention while maintaining high manufacturing precision
3Reliability
If multivariate sensor output is not fully utilized, then the device complexity is reduced, but the information completeness deteriorates leading to process faults
Solution Approach 1:
The patent segments the multivariate sensor output into multiple individual signal channels, processing each through the analysis pipeline separately. This segmentation allows comprehensive utilization of all sensor information while managing processing complexity through modular, independent signal handling that can be parallelized and optimized
Solution Approach 2:
The patent applies partial processing to the full sensor output by selecting and prioritizing certain signal channels or frequency components that are most indicative of end-point conditions. This partial action approach utilizes the most critical information from the multivariate output while reducing unnecessary processing complexity from less relevant signals
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 precise, real-time detection of etching end-points, reducing the risk of over- or under-etching and enhancing process uniformity, thereby improving chip quality and manufacturing yield.
Implementation Method 1
A system utilizing E-field and B-field antennas to generate RF signals
Data Source
AI summary
A semiconductor etching process end-point detection system includes a sensor configured to measure in real-time electrical properties associated with a semiconductor etching process and to generate raw sensor data, and a computing device configured to receive the raw sensor data. The computing device includes spectrum analyzer circuitry and a memory configured to execute instructions of the spectrum analyzer circuitry via at least one processor. The instructions include obtaining, via the sensor, the raw sensor data; receiving, at the computing device, the raw sensor data; preprocessing the raw sensor data to normalize the raw sensor data to a standard range or distribution; selecting at least one best scoring signal of the normalized sensor data; converting the at least one best scoring signal to a pixel space; and performing an end-point prediction algorithm on the pixel space to predict the end-point of the semiconductor etching process.