研究目的
To present and evaluate a noninvasive optical plasma monitoring method for detecting variations and faults in the plasma-enhanced atomic layer deposition (PEALD) process, specifically for Al2O3 nanoscale water vapor barrier films, to ensure process quality and enable advanced control.
研究成果
The OPMS technique effectively monitors PEALD processes by detecting plasma ignition variations and abnormalities in real-time. It can distinguish between normal and abnormal deposition cycles based on pulse parameters, aiding in fault detection and potentially classification. Future work should involve statistical modeling for fault diagnosis with more diverse fault conditions.
研究不足
The study is limited to Al2O3 deposition and may not generalize to other materials. The OPMS cannot identify specific chemical species like optical emission spectroscopy (OES). Further experiments with various induced faults are needed for comprehensive fault classification and diagnosis.
1：Experimental Design and Method Selection:
The study employed an optical plasma monitoring system (OPMS) based on a photomultiplier tube for high-speed monitoring of plasma emission in PEALD processes. The methodology included real-time data acquisition and pulse parameter analysis to detect abnormalities.
2：Sample Selection and Data Sources:
Al2O3 films were deposited on transparent flexible substrates using trimethylaluminum (TMA) and oxygen in a low RF frequency PEALD reactor. Two experimental runs (Run 1 and Run 2) with different recipes were conducted to compare normal and abnormal deposition conditions.
3：List of Experimental Equipment and Materials:
Equipment included a PEALD reactor, OPMS with photomultiplier tube, collimating lens, fiber-optic cable, high-speed analog-to-digital converter, and data analysis software. Materials included TMA, oxygen gas, and flexible substrates.
4：Experimental Procedures and Operational Workflow:
For Run 1, the deposition cycle consisted of TMA feed, purge, O2 feed with plasma ignition, and purge steps over 75 cycles. Run 2 had a modified recipe with reduced O2 stabilization time over 71 cycles. Plasma emission was monitored in real-time using OPMS, and data were analyzed for pulse parameters such as rise time, fall time, and standard deviation.
5：Data Analysis Methods:
Data were digitized at 250 kHz sampling rate, and a pulse detection algorithm identified individual pulses. Parameters like rise time, fall time, pulse duration, peak value, area under pulse, and on-time standard deviation were calculated to detect variations and faults. Statistical analysis and visualization (e.g., min-max charts) were used to compare runs.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
