Systems and methods for controlling device performance variability during manufacturing of a device on wafers are disclosed. The system includes a process platform, on-board metrology (OBM) tools, and a first server that stores a machine-learning based process control model. The first server combines virtual metrology (VM) data and OBM data to predict a spatial distribution of one or more dimensions of interest on a wafer. The system further comprises an in-line metrology tool, such as SEM, to measure the one or more dimensions of interest on a subset of wafers sampled from each lot. A second server having a machine-learning engine receives from the first server the predicted spatial distribution of the one or more dimensions of interest based on VM and OBM, and also receives SEM metrology data, and updates the process control model periodically (e.g., to account for chamber-to-chamber variability) using machine learning techniques.

There is provided a system that includes a review tool configured to review at least part of potential defects of an examined object, and assign each of the at least part of the potential defects with a multiplicity of attribute values. The system also includes a computer-based classifier configured to classify, based on the attribute values as assigned, the at least part of potential defects into a set of classes, the set comprising at least a first major class, a second major class and a first minor class, the classifier trained based on a training set comprising a multiplicity of training defects with assigned attribute values, the training defects classified into the set of classes.

In the measurement of properties of a wafer substrate, such as Critical Dimension or overlay a sampling plan is produced defined for measuring a property of a substrate, wherein the sampling plan comprises a plurality of sub-sampling plans. The sampling plan may be constrained to a predetermined fixed number of measurement points and is used to control an inspection apparatus to perform a plurality of measurements of the property of a plurality of substrates using different sub-sampling plans for respective substrates, optionally, the results are stacked to at least partially recompose the measurement results according to the sample plan.

A system for auto defect screening using adaptive machine learning includes an adaptive model controller, a defect/nuisance library and a module for executing data modeling analytics. The adaptive model controller has a feed-forward path for receiving a plurality of defect candidates in wafer inspection, and a feedback path for receiving defects of interest already screened by one or more existing defect screening models after wafer inspection. The adaptive model controller selects data samples from the received data, interfaces with scanning electron microscope (SEM) review/inspection to acquire corresponding SEM results that validate if each data sample is a real defect or nuisance, and compiles model training and validation data. The module of executing data modeling analytics is adaptively controlled by the adaptive model controller to generate and validate one or more updated defect screening models using the model training and validation data according to a target specification.

In the measurement of properties of a wafer substrate, such as Critical Dimension or overlay a sampling plan is produced 2506 defined for measuring a property of a substrate, wherein the sampling plan comprises a plurality of sub-sampling plans. The sampling plan may be constrained to a predetermined fixed number of measurement points and is used 2508 to control an inspection apparatus to perform a plurality of measurements of the property of a plurality of substrates using different sub-sampling plans for respective substrates, optionally, the results are stacked 2510 to at least partially recompose the measurement results according to the sample plan.

A method of calculating an overlay correction model in a unit for the fabrication of a wafer is disclosed. The method comprises measuring overlay deviations of a subset of first overlay marks and second overlay marks by determining the differences between the subset of first overlay marks generated in the first layer and corresponding ones of the subset of second overlay marks generated in the second layer.

An inspection system may receive inspection datasets from a defect inspection system associated with inspection of one or more samples, where an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes. An inspection system may further label each of the inspection datasets with a class label based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, where each class label corresponds to a region of the signal space. An inspection system may further segment the inspection datasets into two or more defect groups by training a classifier with the values of the context attributes and corresponding class labels for the inspection datasets, where the two or more defect groups are identified based on the trained classifier.

Systems and methods for controlling device performance variability during manufacturing of a device on wafers are disclosed. The system includes a process platform, on-board metrology (OBM) tools, and a first server that stores a machine-learning based process control model. The first server combines virtual metrology (VM) data and OBM data to predict a spatial distribution of one or more dimensions of interest on a wafer. The system further comprises an in-line metrology tool, such as SEM, to measure the one or more dimensions of interest on a subset of wafers sampled from each lot. A second server having a machine-learning engine receives from the first server the predicted spatial distribution of the one or more dimensions of interest based on VM and OBM, and also receives SEM metrology data, and updates the process control model periodically (e.g., to account for chamber-to-chamber variability) using machine learning techniques.

Methods and systems for selecting a mode for inspection of a specimen are provided. One method includes determining how separable defects of interest (DOIs) and nuisances detected on a specimen are in one or more modes of an inspection subsystem. The separability of the modes for the Dais and nuisances is used to select a subset of the modes for inspection of other specimens of the same type. Other characteristics of the performance of the modes may be used in combination with the separability to select the modes. The subset of modes selected based on the separability may also be an initial subset of modes for which additional analysis is performed to determine the final subset of the modes.

A method includes receiving, into a measurement tool, a substrate having a material feature, wherein the material feature is formed on the substrate according to a design feature. The method further includes applying a source signal on the material feature, collecting a response signal from the material feature by using the measurement tool, and with a computer connected to the measurement tool, calculating a simulated response signal from the design feature. The method further includes, with the computer, in response to determining that a difference between the collected response signal and the simulated response signal exceeds a predetermined value, causing the measurement tool to re-measure the material feature.