Virtual measurement of conditions proximate to a substrate with physics-informed compressed sensing
AssigneeApplied Materials, Inc.
InventorsPreetham Prahallada Rao, Prashanth Kothnur
Life of this patent
Estimating the term…
Abstract
A method includes measuring a subset of property values within a manufacturing chamber during a process performed on a substrate within the manufacturing chamber. The method further includes determining property values in the manufacturing chamber at locations removed from the locations the measurements are taken. The method further includes performing a corrective action based on the determined properties.
Claims20
A system, comprising memory and a processing device coupled to the memory, wherein the processing device is configured to perform operations comprising:
- obtaining a set of property values generated by a plurality of sensors associated with a manufacturing chamber during a process performed in the manufacturing chamber;
- determining a property value at a location of interest proximate a substrate in the manufacturing chamber based on inputting the set of property values into a trained machine learning model, wherein the trained machine learning model is configured to represent property values in the manufacturing chamber in accordance with a reduced order model, and wherein the property value at the location of interest is determined as a weighted additive combination of a number of members of the reduced order model that are a subset of a basis set that describes combinations of property values in the manufacturing chamber;
- and
- performing an action in response to the property value at the location of interest.
A system, comprising memory and a processing device coupled to the memory, wherein the processing device is configured to perform operations comprising:
- generating training data for a machine learning model, wherein generating the training data comprises:
- identifying a first training input comprising data indicative of first property values associated with a limited subset of property values in a manufacturing chamber;
- and
- identifying a first target output for the first training input, wherein the first target output comprises a map of property values in a plurality of locations within the manufacturing chamber, wherein a number of the plurality of locations is more than a number of locations of the subset of property values used as the first training input;
- and
- training the machine learning model to generate a trained machine learning model based on (i) a set of training inputs comprising the first training input, and (ii) a set of target outputs comprising the first target output, wherein the trained machine learning model is to obtain input comprising data indicative of second property values associated with a limited subset of property values in the manufacturing chamber, the second property values being different than the first property values, and to produce output based on the input, the output comprising a target property value at a location of interest proximate to a substrate in the manufacturing chamber determined as a weighted additive combination of a number of members of a reduced order model that are a subset of a basis set.
A non-transitory machine-readable storage medium storing instructions which, when executed by a processing device, cause the processing device to perform operations comprising:
- generating training data for a machine learning model, wherein generating the training data comprises:
- identifying a first training input comprising data indicative of first property values associated with a limited subset of property values in a manufacturing chamber;
- and
- identifying a first target output for the first training input, wherein the first target output comprises a map of property values in a plurality of locations within the manufacturing chamber, wherein a number of the plurality of locations is more than a number of locations of the subset of property values used as the first training input;
- and
- training the machine learning model to generate a trained machine learning model based on (i) a set of training inputs comprising the first training input, and (ii) a set of target outputs comprising the first target output, wherein the trained machine learning model is to obtain input comprising data indicative of second property values associated with a limited subset of property values in the manufacturing chamber, the second property values being different than the first property values, and to produce output based on the input, the output comprising a target property value at a location of interest proximate to a substrate in the manufacturing chamber determined as a weighted additive combination of a number of members of a reduced order model that are a subset of a basis set.
Dependencies are inferred from the claim text.