Polishing Apparatus Predictive Scheduling for Copper Corrosion Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional schedulers for wafer processing systems maximize throughput but result in cleaning wait times that can lead to corrosion of copper interconnects in semiconductor devices, failing to fully prevent corrosion during the time gap between polishing and cleaning.
Innovation Solution
A polishing apparatus with a control section that predicts and optimizes the polishing, transport, and cleaning times to eliminate cleaning wait times, ensuring immediate or near-immediate cleaning of wafers after polishing, thereby minimizing the time between polishing and cleaning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional schedulers maximize throughput in wafer processing, then productivity is improved, but cleaning wait time increases leading to copper corrosion
Solution Approach 1:
The control section performs preliminary calculation of polishing end times and cleaning start times before actual processing. By predicting polishing end times based on start times and polishing times, and determining cleaning start times in advance, the system prepares the timing schedule proactively to eliminate wait times and prevent corrosion while maintaining throughput.
Solution Approach 2:
The control section continuously monitors and adjusts the timing based on feedback from polishing end times and cleaning start times. By comparing predicted times with actual times and dynamically adjusting subsequent wafer processing schedules, the system maintains optimal timing coordination between polishing and cleaning operations.
2Reliability
If cleaning wait time is eliminated to prevent copper corrosion, then reliability is improved, but device complexity increases due to predictive timing control
Solution Approach 1:
The control section autonomously calculates and determines optimal timing schedules without requiring external intervention. It self-manages the coordination between polishing and cleaning operations by automatically computing polishing end times, predicting cleaning start times, and adjusting schedules based on internal data, thereby reducing the need for complex external control mechanisms.
3Productivity
If multiple polishing lines feed one cleaning line, then productivity is improved through parallel processing, but cleaning wait time occurs when cleaning line is unavailable
Solution Approach 1:
The control section calculates polishing end times and predicts cleaning start times for multiple wafers from different polishing lines in advance. By determining the optimal timing schedule before processing begins, the system coordinates the feeding of wafers to the cleaning line to minimize or eliminate wait times, ensuring continuous operation and maintaining high throughput.
Data Source
Figure 1A~1D
Figure 2
Figure 3
AI summary
A polishing apparatus includes a loading section (14) for placing therein a cassette (12) in which a plurality of polishing objects are housed; a first polishing line (20) and a second polishing line (30) for polishing a polishing object; a cleaning line (40) having cleaning machines (42a, 42b, 42c, 42d) for cleaning the polishing object after polishing and a transport unit (44) for transporting the polishing object; a transport mechanism (50) for transporting the polishing object between the loading section (14), the polishing lines (20, 30) and the cleaning line (40); and a control section for controlling the polishing lines (20, 30), the cleaning line (40) and the transport mechanism (50). The control section determines a polishing start time in each of the first and second polishing lines (20, 30) based on a predicted polishing time in each of the first and second polishing lines (20, 30), a predicted transport time in the transport mechanism (50), a predicted cleaning time in the cleaning line (40) and a predicted cleaning start time to start cleaning by driving the transport unit (44) of the cleaning line (40).