FIB Endpoint Detection via Firefly Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Focused ion beam (FIB) systems face challenges in accurately detecting the endpoint of material removal or deposition due to reduced secondary electrons and ions available for imaging, especially when processing small features below 100 nanometers, leading to difficulties in discerning when the intended device feature is uncovered, which can result in overetching or underetching.
Innovation Solution
The implementation of a method that analyzes real-time images from FIB systems by superimposing graphical elements, known as fireflies, on the image based on intensity and positional weighting, allowing for the automatic detection of endpoint events by highlighting intensity differences and converging fireflies towards exposed features, facilitating precise control of the milling process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If beam flux is decreased to process small features below 100 nanometers, then manufacturing precision is improved, but measurement precision deteriorates due to reduced secondary electrons and ions available for imaging and endpoint detection
Solution Approach 1:
The patent introduces an intermediary image processing algorithm that processes the low-signal images obtained during FIB processing. The algorithm enhances edge detection and feature visibility through computational methods, acting as a mediator between the limited physical signal and the required measurement precision. This allows accurate endpoint detection even when secondary electron and ion signals are reduced due to low beam flux requirements for sub-100nm features.
2Manufacturing precision
If beam flux is decreased for processing small features, then manufacturing precision is improved, but reliability deteriorates due to difficulty in discerning when the intended device feature is uncovered
Solution Approach 1:
The patent implements a feedback mechanism where the image processing algorithm continuously analyzes images during the FIB processing sequence. By enhancing edge detection and feature visibility in real-time, the system provides feedback on the processing state, enabling reliable determination of when the intended device feature is uncovered. This feedback loop compensates for the reduced signal quality caused by decreased beam flux.
3Manufacturing precision
If conventional imaging methods are used with reduced beam flux, then manufacturing precision is improved for small features, but device complexity increases due to overetching or underetching requiring additional process steps
Solution Approach 1:
The patent replaces the conventional mechanical/optical imaging system with a computational image processing approach. Instead of relying on physical signal intensity from secondary electrons and ions, the system uses algorithms to extract feature information from low-signal images. This substitution eliminates the need for additional process steps to compensate for poor imaging, thereby reducing overall device complexity while maintaining precision for sub-100nm features.
Data Source
AI summary
A method, system and apparatus are presented for real time analysis of images in a focused beam system. In various embodiments, marker positions are displayed as graphical elements on the image of a sample being processed. Selected characteristics of all or a portion of the pixels in the image are used to determine the positions. The marker positions are used to detect the occurrence of an event such as an endpoint. In some embodiments attributes of the of the graphic elements change based on the occurrence of selected events and in further embodiments an action is initiated.