US 12,443,840 B2
Dynamic control of machine learning based measurement recipe optimization
Stilian Ivanov Pandev, Santa Clara, CA (US); and Arvind Jayaraman, New Hudson, MI (US)
Assigned to KLA Corporation, Milpitas, CA (US)
Filed by KLA Corporation, Milpitas, CA (US)
Filed on Dec. 2, 2020, as Appl. No. 17/110,005.
Claims priority of provisional application 63/089,550, filed on Oct. 9, 2020.
Prior Publication US 2022/0114438 A1, Apr. 14, 2022
Int. Cl. G06N 3/08 (2023.01); G01N 21/21 (2006.01); G01N 21/95 (2006.01)
CPC G06N 3/08 (2013.01) [G01N 21/211 (2013.01); G01N 21/9501 (2013.01); G01N 2201/126 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a metrology tool including an illumination source and a detector configured to collect an amount of regularization measurement data from measurements of one or more regularization structures disposed on a first wafer; and
a computing system configured to:
receive an amount of Design of Experiments (DOE) measurement data associated with measurements of one or more Design of Experiments (DOE) metrology targets;
receive known, reference values of one or more parameters of interest associated with the DOE metrology targets;
receive the regularization measurement data;
receive values of one or more measurement performance metrics associated with the regularization measurement data, each of the one or more measurement performance metrics characterizing a measurement performance associated with values of physical attributes characterizing the one or more regularization structures; and
iteratively train a measurement model based on an optimization function including the amount of Design of Experiments (DOE) measurement data, the reference values of one or more parameters of interest, the regularization measurement data, and the one or more measurement performance metrics, wherein the optimization function includes a regularization term associated with each of the one or more measurement performance metrics, and wherein a weighting value associated with each of the regularization terms is dynamically controlled during iteration of the measurement model based on a difference between an achieved value of the measurement performance metric associated with each of the regularization terms and a desired value of the measurement performance metric associated with each of the regularization terms.