Abstract: A metrology system may include a spectroscopic metrology sub-system and a controller, the controller including one or more processors configured to execute program instructions configured to cause the one or more processors to: generate a tool-induced shift (TIS) model of a training sample by the metrology sub-system comprising: receiving training data from metrology measurements of the training sample, the training data comprising spectral data associated with at least one off-diagonal Mueller matrix element generated by one or more first measurements of the training sample at a first azimuthal angle and one or more second measurements of the training sample at a second azimuthal angle, deriving overlay spectra data and TIS spectra data from the training data, decomposing the overlay spectra data and the TIS spectra data, and inferring a TIS signature for the training sample; and to remove the TIS signature from a test sample.

Abstract: A metrology module includes an estimation model that is configured to provide an estimation of independent overlay with tool induced shift on received wafers based on only one azimuth angle spectra. The estimation model can use at least one machine learning algorithm. The estimation model can be derived by the machine learning algorithm applied to calculated training data based on a first training sample set from initial metrology measurements and an additional tool induced shift training sample.

Abstract: Methods and systems for determining information for a specimen are provided. One system includes an output acquisition subsystem configured to generate output for a specimen at one or more target locations on the specimen and one or more temperature sensors configured to measure one or more temperatures within the system. The system also includes a deep learning model configured for predicting error in at least one of the one or more target locations based on at least one of the one or more measured temperatures input to the deep learning model by the computer subsystem. The computer subsystem is configured for determining a corrected target location for the at least one of the one or more target locations by applying the predicted error to the at least one of the one or more target locations.

Abstract: A method for improving measurement performance of characterization systems is disclosed. The method may include training a plurality of machine learning models based on a set of training data, where each machine learning model is capable of generating an uncertainty estimator and a first machine learning model is different from one or more additional machine learning models based on one of a set of hyperparameters or a dataset. The method may further include receiving a plurality of sample measurement datasets from one or more test samples. For each of the plurality of sample measurement datasets, the method may further include applying each trained machine learning model to determine a measurement value and the uncertainty estimator for each trained machine learning model and generating a measurement output based on N trained machine learning models with the lowest uncertainty estimators.

Abstract: A characterization system is disclosed. The system may include one or more controllers including one or more processors configured to execute a set of program instructions stored in memory. The controller may be configured to train a machine learning-based characterization library based on a set of training data. The controller may be configured to generate one or more characterization measurements using the trained machine learning-based characterization library based on the real-time characterization data associated with the one or more samples from the characterization sub-system. The controller may be configured to determine one or more additional characterization measurements based on a non-machine learning-based technique. The controller may be configured to compare the one or more characterization measurements and the one or more additional characterization measurements to monitor a measurement uncertainty of the machine learning-based characterization library.

Abstract: Methods and systems for determining information for a specimen are provided. One system includes an output acquisition subsystem configured to generate output for a specimen at one or more target locations on the specimen and one or more temperature sensors configured to measure one or more temperatures within the system. The system also includes a deep learning model configured for predicting error in at least one of the one or more target locations based on at least one of the one or more measured temperatures input to the deep learning model by the computer subsystem. The computer subsystem is configured for determining a corrected target location for the at least one of the one or more target locations by applying the predicted error to the at least one of the one or more target locations.

Abstract: A self-calibrating overlay metrology system may receive device overlay data for a device targets on a sample from an overlay metrology tool, determine preliminary device overlay measurements for the device targets including device-scale features using an overlay recipe with the device overlay data as inputs, receive assist overlay data for one or more assist targets on the sample including device-scale features from the overlay metrology tool, where at least one of the one or more assist targets has a programmed overlay offset of a selected value along a particular measurement direction, determine self-calibrating assist overlay measurements for the one or more assist targets based on the assist overlay data, where the self-calibrating assist overlay measurements are linearly proportional to overlay on the sample, and generate corrected overlay measurements for the device targets by adjusting the preliminary device overlay measurements based on the self-calibrating assist overlay measurements.

Abstract: A spectroscopic metrology system includes a spectroscopic metrology tool and a controller. The controller generates a model of a multilayer grating including two or more layers, the model including geometric parameters indicative of a geometry of a test layer of the multilayer grating and dispersion parameters indicative of a dispersion of the test layer. The controller further receives a spectroscopic signal of a fabricated multilayer grating corresponding to the modeled multilayer grating from the spectroscopic metrology tool. The controller further determines values of the one or more parameters of the modeled multilayer grating providing a simulated spectroscopic signal corresponding to the measured spectroscopic signal within a selected tolerance. The controller further predicts a bandgap of the test layer of the fabricated multilayer grating based on the determined values of the one or more parameters of the test layer of the fabricated structure.

Abstract: A method and system for measuring overlay in a semiconductor manufacturing process comprise capturing an image of a feature in an article at a predetermined manufacturing stage, deriving a quantity of an image parameter from the image and converting the quantity into an overlay measurement. The conversion is by reference to an image parameter quantity derived from a reference image of a feature at the same predetermined manufacturing stage with known overlay (“OVL”). There is also disclosed a method of determining a device inspection recipe for use by an inspection tool comprising identifying device patterns as candidate device care areas that may be sensitive to OVL, deriving an OVL response for each identified pattern, correlating the OVL response with measured OVL, and selecting some or all of the device patterns as device care areas based on the correlation.

Abstract: A metrology module includes an estimation model that is configured to provide an estimation of independent overlay with tool induced shift on received wafers based on only one azimuth angle spectra. The estimation model can use at least one machine learning algorithm. The estimation model can be derived by the machine learning algorithm applied to calculated training data based on a first training sample set from initial metrology measurements and an additional tool induced shift training sample.

Abstract: A self-calibrating overlay metrology system may receive device overlay data for a device targets on a sample from an overlay metrology tool, determine preliminary device overlay measurements for the device targets including device-scale features using an overlay recipe with the device overlay data as inputs, receive assist overlay data for one or more assist targets on the sample including device-scale features from the overlay metrology tool, where at least one of the one or more assist targets has a programmed overlay offset of a selected value along a particular measurement direction, determine self-calibrating assist overlay measurements for the one or more assist targets based on the assist overlay data, where the self-calibrating assist overlay measurements are linearly proportional to overlay on the sample, and generate corrected overlay measurements for the device targets by adjusting the preliminary device overlay measurements based on the self-calibrating assist overlay measurements.

Abstract: An overlay metrology system may receive overlay data for in-die overlay targets within various fields on a skew training sample from one or more overlay metrology tools, wherein the in-die overlay targets within the fields have a range programmed overlay offsets, wherein the fields are fabricated with a range of programmed skew offsets. The system may further generate asymmetric target signals for the in-die overlay targets using an asymmetric function providing a value of zero when physical overlay is zero and a sign indicative of a direction of physical overlay. The system may further generate corrected overlay offsets for the in-die overlay targets on the asymmetric target signals, generate self-calibrated overlay offsets for the in-die overlay targets based on the programmed overlay offsets and the corrected overlay offsets, generate a trained overlay recipe, and generate overlay measurements for in-die overlay targets on additional samples using the trained overlay recipe.

Abstract: A self-calibrating overlay metrology system may receive device overlay data from device targets on a sample, determine preliminary device overlay measurements for the device targets including device-scale features using an overlay recipe with the device overlay data as inputs, receive assist overlay data from sets of assist targets on the sample including device-scale features, where a particular set of assist targets includes one or more target pairs formed with two overlay targets having programmed overlay offsets of a selected value with opposite signs along a particular measurement direction.

Abstract: An overlay metrology system may receive overlay data for in-die overlay targets within various fields on a skew training sample from one or more overlay metrology tools, wherein the in-die overlay targets within the fields have a range programmed overlay offsets, wherein the fields are fabricated with a range of programmed skew offsets. The system may further generate asymmetric target signals for the in-die overlay targets using an asymmetric function providing a value of zero when physical overlay is zero and a sign indicative of a direction of physical overlay. The system may further generate corrected overlay offsets for the in-die overlay targets on the asymmetric target signals, generate self-calibrated overlay offsets for the in-die overlay targets based on the programmed overlay offsets and the corrected overlay offsets, generate a trained overlay recipe, and generate overlay measurements for in-die overlay targets on additional samples using the trained overlay recipe.

Abstract: A self-calibrating overlay metrology system may receive device overlay data from device targets on a sample, determine preliminary device overlay measurements for the device targets including device-scale features using an overlay recipe with the device overlay data as inputs, receive assist overlay data from sets of assist targets on the sample including device-scale features, where a particular set of assist targets includes one or more target pairs formed with two overlay targets having programmed overlay offsets of a selected value with opposite signs along a particular measurement direction.

Abstract: A spectroscopic metrology system includes a spectroscopic metrology tool and a controller. The controller generates a model of a multilayer grating including two or more layers, the model including geometric parameters indicative of a geometry of a test layer of the multilayer grating and dispersion parameters indicative of a dispersion of the test layer. The controller further receives a spectroscopic signal of a fabricated multilayer grating corresponding to the modeled multilayer grating from the spectroscopic metrology tool. The controller further determines values of the one or more parameters of the modeled multilayer grating providing a simulated spectroscopic signal corresponding to the measured spectroscopic signal within a selected tolerance. The controller further predicts a bandgap of the test layer of the fabricated multilayer grating based on the determined values of the one or more parameters of the test layer of the fabricated structure.

Abstract: A die screening system may receive die-resolved metrology data for a population of dies on one or more samples from the one or more in-line metrology tools after one or more fabrication steps, where the die-resolved metrology data includes images generated using one or more measurement configurations of the one or more in-line metrology tools. In this way, the die-resolved metrology data provides many measurement channels per die, where a particular measurement channel includes data from a particular pixel of a particular image. The controller may then generate screening data for the population of dies from the die-resolved metrology data, where the screening data includes a subset of the plurality of measurement channels of the die-resolved metrology data, and screen the plurality of dies into two or more disposition classes including at least outlier dies based on variability in the screening data.

Abstract: First and second metrology data are used to train a machine-learning model to predict metrology data for a metrology target based on metrology data for a device area. The first metrology data are for a plurality of instances of a device area on semiconductor die fabricated using a fabrication process. The second metrology data are for a plurality of instances of a metrology target that contains structures distinct from structures in the device area. Using the trained machine-learning model, fourth metrology data are predicted for the metrology target based on third metrology data for an instance of the device area. Using a recipe for the metrology target, one or more parameters of the metrology target are determined based on the fourth metrology data. The fabrication process is monitored and controlled based at least in part on the one or more parameters.

Abstract: A spectroscopic metrology system includes a spectroscopic metrology tool and a controller. The controller generates a model of a multilayer grating including two or more layers, the model including geometric parameters indicative of a geometry of a test layer of the multilayer grating and dispersion parameters indicative of a dispersion of the test layer. The controller further receives a spectroscopic signal of a fabricated multilayer grating corresponding to the modeled multilayer grating from the spectroscopic metrology tool. The controller further determines values of the one or more parameters of the modeled multilayer grating providing a simulated spectroscopic signal corresponding to the measured spectroscopic signal within a selected tolerance. The controller further predicts a bandgap of the test layer of the fabricated multilayer grating based on the determined values of the one or more parameters of the test layer of the fabricated structure.

Abstract: A method and system for measuring overlay in a semiconductor manufacturing process comprise capturing an image of a feature in an article at a predetermined manufacturing stage, deriving a quantity of an image parameter from the image and converting the quantity into an overlay measurement. The conversion is by reference to an image parameter quantity derived from a reference image of a feature at the same predetermined manufacturing stage with known overlay (“OVL”). There is also disclosed a method of determining a device inspection recipe for use by an inspection tool comprising identifying device patterns as candidate device care areas that may be sensitive to OVL, deriving an OVL response for each identified pattern, correlating the OVL response with measured OVL, and selecting some or all of the device patterns as device care areas based on the correlation.