Holistic analysis of multidimensional sensor data for substrate processing equipment
Inventors
Liu, Chao • Hao, Yudong • Li, Shifang • Schulze, Andreas
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12386342-B2
Publication Date
2025-08-12
Expiration Date
Abstract
A method includes receiving, by a processing device, first data. The first data includes data from one or more sensors of a processing chamber and is associated with a processing operation. The first data is resolved in at least two dimensions, one of which is time. The method further includes providing the first data to a model. The method further includes receiving from the model second data. The second data includes an indication of an evolution of a processing parameter during the processing operation. The method further includes causing performance of a corrective action in view of the second data.
Core Innovation
The invention fits temporal evolution of multi-dimensional in-chamber sensor data to obtain an indication of how a processing parameter evolves during a processing operation. The first data is received from one or more sensors of a processing chamber and is indicative of a first property of a substrate. The first data is multi-dimensional data comprising at least two dimensions, one of which is time, and is provided as input to a model configured to fit temporal evolution over a duration associated with the processing operation.
The model outputs second data that comprises an indication of an evolution of a processing parameter comprising the first property of the substrate during the processing operation. Based on the second data, a temporal evolution of the first property is determined to be indicative of a corrective action to be performed. The corrective action is performed in view of the second data.
Performing the corrective action comprises one or more of updating a process recipe associated with the substrate, scheduling corrective maintenance of the processing chamber, or scheduling preventative maintenance of the processing chamber. The document further supports that the multi-dimensional sensor data can be spectrally resolved electromagnetic radiation data and that the modeling can be constrained to a physics-based model configured to fit a temporal evolution over the processing-associated duration.
Claims Coverage
The independent claims (method, system, and non-transitory machine-readable storage medium) share a common inventive concept: time-inclusive multi-dimensional sensor data from a processing chamber is provided to a model to fit temporal evolution and to drive corrective actions via recipe updates and/or maintenance scheduling. Across the independent claims, at least three main inventive features recur: receiving and modeling time-inclusive multi-dimensional data to produce an evolution indication, determining that the temporal evolution indicates a corrective action, and performing corrective actions selected from recipe updating and maintenance scheduling.
Temporal evolution fitting of multi-dimensional sensor data with time
Receiving first data from one or more sensors of a processing chamber associated with a processing operation, where the first data is multi-dimensional data comprising at least two dimensions including time, providing the first data to a model configured to fit a temporal evolution of the first data over a duration associated with the processing operation, and obtaining second data as output of the model.
Corrective action determination from processing-parameter evolution indication
Obtaining second data comprising an indication of an evolution of a processing parameter comprising a first property of the substrate during the processing operation, and determining, based on the second data, that a temporal evolution of the first property is indicative of a corrective action to be performed.
Corrective action execution via recipe update and maintenance scheduling
Causing performance of the corrective action in view of the second data, wherein causing performance of the corrective action comprises one or more of updating a process recipe associated with the substrate, scheduling corrective maintenance of the processing chamber, or scheduling preventative maintenance of the processing chamber.
Together, the independent claims define an end-to-end approach in which time-inclusive multi-dimensional sensor data is modeled to output an indication of processing-parameter evolution, that indication is used to determine that corrective action is needed, and the corrective action is carried out through process recipe updating and/or scheduling corrective or preventative maintenance.
Stated Advantages
Improved robustness to sparsity/noise compared to conventional approaches.
Better extraction of temporal process behavior compared to conventional time-independent or frame-by-frame analysis.
Documented Applications
Triggering corrective actions (e.g., recipe updates and/or corrective/preventative maintenance scheduling) based on an indication of evolution of processing parameters derived from holistic temporal analysis of multi-dimensional in-chamber sensor data.
Interested in licensing this patent?