Scatterometry method and system

1. A method for use in model-based optical measurements in patterned structures, the method being carried out by a computer system comprising a non-transitory computer readable memory and a data processor: the method comprising:

receiving and storing input data comprising information about a patterned structure configuration;

in response to said receiving of the input data, processing and analyzing the input data, by said data processor, thereby generating a selected optical model for interpretation of optical measured data indicative of optical response of the patterned structure configuration, said processing and analyzing comprising

based on the information about the patterned structure configuration, creating a complete optical model including parameters describing the patterned structure configuration, where a majority of the parameters are floating parameters, and defining multiple model configurations describing an optical response of the patterned structure configuration, each of the multiple model configurations created by fixing values of one or more of the parameters,

based on the complete model, generating synthetic spectra of the patterned structure configuration,

applying a fitting procedure to the synthetic spectra and said multiple model configurations,

creating a reference predictor from a subset of the multiple model configurations that meet predefined best-fit criteria,

optically measuring a patterned structure on a semiconductor device to produce preliminary measured data,

applying said reference predictor to the preliminary measured data to generate corresponding virtual reference data, and

selecting at least one of the multiple model configurations corresponding to a highest correlation to the virtual reference data among the multiple model configurations; and

providing, for process control during manufacturing of semiconductor devices, the selected model for interpretation of optical measured data of one or more parameters of corresponding patterned structures on the semiconductor devices.

2. The method of claim 1 , wherein said selection of the optical model further comprises Total Measurement Uncertainty (TMU) analysis.

3. The method according to claim 1 , wherein the creation of the reference predictor comprises:

interpreting theoretical optical responses corresponding to different combinations of said model configurations;

selecting a combination of the model configurations that best interprets the theoretical optical responses according to predetermined criteria; and

utilizing the selected combination of the model configurations to create the reference predictor.

4. The method according to claim 3 , comprising applying at least one statistical tool to the selected combination of the model configurations for creation of the reference predictor.

5. The method according to claim 3 , wherein the creation of the reference predictor from the model configurations comprises spanning a model configuration space by performing at least one selection session, the selection session comprising: selecting a preceding model configuration having highest correlation to the theoretical optical response, subsequently selecting a successive model configuration having highest correlation to the theoretical optical response and lowest correlation to the preceding model configuration.

6. The method according to claim 1 , wherein said floating parameters defining the multiple model configurations include at least some of the following parameters of a pattern in the structure: critical dimensions (CD), top CD, bottom CD, sidewall angle, spacer widths, spacer pull-down, epitaxial proximity, footing/undercut, overfill/underfill parameters, rounding.

7. The method according to claim 1 , wherein at least some of said floating parameters defining the multiple model configurations are parameters of interest in the corresponding patterned structures being measured.

8. A system for carrying out the method of claim 1 , the system being configured as a computer comprising: a data input utility configured and operable to receive the input data comprising apriori knowledge about one or more parameters of the patterned structure configuration; a non-transitory computer readable memory for storing the input data; and a processor utility comprising; a reference predictor module adapted for utilizing said preliminary measured data and said apriori knowledge and creating the virtual reference data; and a verification module for verifying the preliminary measured data with respect to the virtual reference data, and determining and storing the selected optical model for interpretation of the optical measured data.