Back to results
US 12,717,243 B2Granted
Methods of determining process models by machine learning
AssigneeASML Netherlands B.V.
InventorsYa Luo, Yu Cao, Jen-Shiang Wang, Yen-Wen Lu
Life of this patent
Estimating the term…
Abstract
Methods of determining, and using, a process model that is a machine learning model. The process model is trained partially based on simulation or based on a non-machine learning model. The training data may include inputs obtained from a design layout, patterning process measurements, and image measurements.
Claims19
A computer program product comprising a non-transitory computer readable medium having instructions therein, the instructions, upon execution by a computer system, configured to cause the computer system to at least:
- determine first inputs to a first process model based on a processing condition or a location on a substrate, the first process model being a machine learning model;
- generate, by execution of the first process model, an output of the first process model from the first inputs;
- determine second inputs to a second process model based on the processing condition or the location, the second process model being a non-machine learning model;
- generate, by execution of the second process model, an output of the second process model from the second inputs;
- determine a deviation between the output of the first process model and the output of the second process model;
- upon determination that the deviation satisfies a condition, adjust the first process model;
- execute the adjusted first process model to generate an output comprising at least a portion of an image to be formed on the substrate;
- and
- generate a mask pattern for exposure of a resist layer of a substrate in a photolithographic process by executing an optical proximity correction and/or source mask optimization, wherein the image that was output from the adjusted first process model is provided as input to the optical proximity correction and/or source mask optimization, and the source mask optimization comprises identifying one or more values for one or more parameters that provide an improvement to a patterning process.
6 more elements
A method comprising:
- determining first inputs to a first process model based on a processing condition or a location on a substrate, the first process model being a machine learning model;
- generating, by execution of the first process model on a hardware computer, an output of the first process model from the first inputs;
- determining second inputs to a second process model based on the processing condition or the location, the second process model being a non-machine learning model;
- generating, by execution of the second process model, an output of the second process model from the second inputs;
- determining a deviation between the output of the first process model and the output of the second process model;
- upon determination that the deviation satisfies a condition, adjusting the first process model;
- executing the adjusted first process model to generate an output comprising an image to be formed on the substrate;
- and
- generating a mask pattern for exposure of a resist layer of a substrate in a photolithographic process by executing an optical proximity correction and/or source mask optimization, wherein the image that was output from the adjusted first process model is provided as input to the optical proximity correction and/or source mask optimization, and the source mask optimization comprising identifying one or more values for one or more parameters that provide an improvement to a patterning process.
6 more elements
A computer program product comprising a non-transitory computer readable medium having instructions therein, the instructions, upon execution by a computer system, configured to cause the computer system to at least:
- determine a deviation between an output of a first process model and an output of a second process model, the first process model being a machine learning model and the second process model being a non-machine learning model and the first process model configured to output one or more physical characteristics of an image or structure formed on a substrate using a patterning process or output one or more characteristics of the patterning process to achieve a particular result of the patterning process;
- upon determination that the deviation satisfies a condition, adjust the first process model;
- execute the adjusted first process model to generate an output comprising an adjusted image or structure to be formed on the substrate;
- and
- generate a mask pattern for exposure of a resist layer of a substrate in a photolithographic process by executing an optical proximity correction and/or source mask optimization, wherein the adjusted image or structure that was output from the adjusted first process model is provided as input to the optical proximity correction and/or source mask optimization, and the source mask optimization comprises identifying one or more values for one or more parameters that provide an improvement to the patterning process.
2 more elements
Dependencies are inferred from the claim text.