RF Generator Arc Detection Using Multi-Frequency Impedance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current arc detection methods in plasma chambers, particularly in semiconductor fabrication processes, face challenges in accurately identifying arcs due to false detections and inefficiencies with pulsing signals, which can lead to damage and contamination.
Innovation Solution
A novel arc management algorithm utilizing a machine learning model that analyzes impedance, voltage, and current data to determine arc events, allowing for precise detection and suppression of arcs, including microarcs, and subsequent plasma re-ignition if necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If voltage and current detection methods are used to detect arcs, then arc detection capability is provided, but false arc detections occur and real arcs are inaccurately detected
Solution Approach 1:
The patent changes the detection parameter from voltage/current values to impedance values. By measuring impedance at different frequencies (including fundamental and harmonic frequencies), the system can distinguish between actual arcs and normal plasma variations, thereby improving both reliability and measurement precision of arc detection.
Solution Approach 2:
The patent introduces impedance as an intermediary parameter between voltage/current measurements and arc detection. Instead of directly detecting arcs through voltage/current, the system measures impedance changes caused by arcs, which provides a more reliable and precise indication of actual arc events while filtering out false detections.
2Reliability
If traditional arc detection methods are used, then some arc detection is achieved, but detection fails with pulsing signals and microarcs
Solution Approach 1:
The patent employs periodic impedance measurements at multiple frequency points (fundamental frequency and harmonics) to detect arcs. This periodic multi-frequency measurement approach enables reliable detection of both sustained arcs and transient microarcs, as well as arcs occurring during pulsing signals, by capturing impedance variations across different temporal and spectral domains.
Solution Approach 2:
The patent adds spectral dimension to arc detection by measuring impedance at multiple frequencies rather than a single frequency or time point. This multi-dimensional approach (time + frequency spectrum) provides enhanced adaptability to detect various arc types including microarcs and arcs during pulsing signals, which single-parameter methods cannot reliably detect.
3Measurement precision
If arc detection sensitivity is increased to detect all arc events, then more arcs are detected, but false detections increase
Solution Approach 1:
The patent uses feedback from multi-frequency impedance measurements to distinguish true arcs from false detection events. By analyzing impedance changes across multiple frequency points and comparing them against expected patterns, the system can confirm actual arc events while filtering out false detections, thereby maintaining high sensitivity without sacrificing reliability.
Solution Approach 2:
The patent performs preliminary impedance characterization during normal operation to establish baseline values before arc events occur. This preliminary action enables the system to quickly identify deviations indicating actual arcs while recognizing normal variations, thus maintaining high detection sensitivity without increasing false detection rates.
Data Source
AI summary
Methods, apparatuses and systems for detecting and managing arc events during a plasma chamber process include receiving impedance data measured during a plasma chamber process, analyzing the impedance data to determine if an arc event is occurring during the plasma chamber process, and if it is determined that an arc event is occurring, an action is taken to suppress an arc of the arc event. In some instances, a machine learning model that has been trained to recognize when an arc event is occurring from received measurement data is used to determine if an arc event is occurring.