Metrology method and apparatus, computer program and lithographic system
A method, computer program and associated apparatuses for metrology. The method includes determining a reconstruction recipe describing at least nominal values for use in a reconstruction of a parameterization describing a target. The method includes obtaining first measurement data relating to measurements of a plurality of targets on at least one substrate, the measurement data relating to one or more acquisition settings and performing an optimization by minimizing a cost function which minimizes differences between the first measurement data and simulated measurement data based on a reconstructed parameterization for each of the plurality of targets. A constraint on the cost function is imposed based on a hierarchical prior. Also disclosed is a hybrid model method comprising obtaining a coarse model operable to provide simulated coarse data; and training a data driven model to correct the simulated coarse data so as to determine simulated data for use in reconstruction.
1 . A method of constructing a hybrid model for providing simulated data for use in parameter reconstruction of a structure, the method comprising:
obtaining a coarse model operable to provide simulated coarse data; and
training, by a hardware computer, a data driven model so as to determine the simulated data, the data driven model configured to produce a same type of data as the simulated coarse data and the same type of data is a corrected version of the simulated coarse data.
2 . The method according to claim 1 , wherein the simulated coarse data comprises simulated coarse reflectivity data.
3 . The method according to claim 1 , wherein the data driven model is configured to determine corrections to combine with the simulated coarse data.
4 . The method according to claim 1 , wherein the data driven model is configured to derive the simulated data from the simulated coarse data.
5 . The method according to claim 1 , wherein at least the data driven model is a machine learned neural network model.
6 . The method according to claim 1 , wherein the hybrid model is a machine learned neural network model and comprises a bottleneck layer having fewer nodes than one or more preceding layers.
7 . The method according to claim 6 , wherein the bottleneck layer and the one or more one or more preceding layers together learn a mapping to a restricted parameter space with respect to an input parameterization.
8 . The method according to claim 7 , wherein:
the hybrid model comprises one or more of the preceding layers, the preceding layers comprising decorrelation layers, or
the restricted parameter space represents a minimum sufficient statistic for the combination of all parameters of the input parameterization, or
the hybrid model comprises one or more modeling layers subsequent to the bottleneck layer, which models the simulated data from the restricted parameter space.
9 . The method according to claim 6 , further comprising training an additional model which is operable to invert and/or re-map a mapping which the bottleneck layer induces; and
using the additional model to determine a correlation metric for describing a correlation between a parameter and one or more other parameters.
10 . The method according to claim 1 , wherein the training step comprises performing iterations of:
determining parameter value estimates for the coarse model based on measurement data and the simulated data; and
updating the data driven model using the determined parameter value estimates such that the measurement data and the simulated data are a better match.
11 . The method according to claim 10 , wherein the measurement data comprises intensity data relating to a plurality of targets.
12 . The method according to claim 1 , further comprising using the hybrid model in a parameter reconstruction of a structure in a metrology operation.
13 . A computer program product comprising a non-transitory computer readable medium having instructions therein, the instructions, when executed by a computer system, configured to cause the computer system to at least:
obtain a coarse model operable to provide simulated coarse data; and
train a data driven model so as to determine simulated data for use in parameter reconstruction of a structure by a hybrid model, the data driven model configured to produce a same type of data as the simulated coarse data and the same type of data is a corrected version of the simulated coarse data.
14 . The computer program product of claim 13 , wherein the simulated coarse data comprises simulated coarse reflectivity data.
15 . The computer program product of claim 13 , wherein the data driven model is configured to determine corrections to combine with the simulated coarse data.
16 . The computer program product of claim 13 , wherein the data driven model is configured to derive the simulated data from the simulated coarse data.
17 . The computer program product of claim 13 , wherein at least the data driven model is a machine learned neural network model.
18 . The computer program product of claim 17 , wherein the hybrid model is a machine learned neural network model and comprises a bottleneck layer having fewer nodes than one or more preceding layers.
19 . The computer program product of claim 13 , wherein the instructions configured to cause the computer system to train the data driven model are further configured to cause the computer system to perform iterations of:
determination of parameter value estimates for the coarse model based on measurement data and the simulated data; and
updates of the data driven model using the determined parameter value estimates such that the measurement data and the simulated data are a better match.
20 . The computer program product of claim 13 , further configured to cause the computer system to use the hybrid model in a parameter reconstruction of a structure in a metrology operation.