Abstract: A method and system for predicting process information (e.g., phase data) using a given input (e.g., intensity) to a parameterized model are described. A latent space of a given input is determined based on dimensional data in a latent space of the parameterized model for a given input to the parameterized model. Further, an optimum latent space is determined by constraining the latent space with prior information (e.g., wavelength) that enables converging to a solution that causes more accurate predictions of the process information. The optimum latent space is used to predict the process information. The given input may be a measured amplitude (e.g., intensity) associated with the complex electric field image. The predicted process information can be complex electric field image having amplitude data and phase data. The parameterized model comprises variational encoder-decoder architecture.

Abstract: A method and apparatus for analyzing an input electron microscope image of a first area on a first wafer are disclosed. The method comprises obtaining a plurality of mode images from the input electron microscope image corresponding to a plurality of interpretable modes. The method further comprises evaluating the plurality of mode images, and determining, based on evaluation results, contributions from the plurality of interpretable modes to the input electron microscope image. The method also comprises predicting one or more characteristics in the first area on the first wafer based on the determined contributions. In some embodiments, a method and apparatus for performing an automatic root cause analysis based on an input electron microscope image of a wafer are also disclosed.

Abstract: A system, method, and apparatus for determining three-dimensional (3D) information of a structure of a patterned substrate. The 3D information can be determined using one or more models configured to generate 3D information (e.g., depth information) using only a single image of a patterned substrate. In a method, the model is trained by obtaining a pair of stereo images of a structure of a patterned substrate. The model generates, using a first image of the pair of stereo images as input, disparity data between the first image and a second image, the disparity data being indicative of depth information associated with the first image. The disparity data is combined with the second image to generate a reconstructed image corresponding to the first image. Further, one or more model parameters are adjusted based on the disparity data, the reconstructed image, and the first image.

Abstract: A method for training a machine learning model includes obtaining a set of unpaired after-development (AD) images and after-etch (AE) images associated with a substrate. Each AD image in the set is obtained at a location on the substrate that is different from the location at which any of the AE images is obtained. The method further includes training the machine learning model to generate a predicted AE image based on the AD images and the AE images, wherein the predicted AE image corresponds to a location from which an input AD image of the AD images is obtained.

Abstract: A method for categorizing a substrate subject to a semiconductor manufacturing process including multiple operations, the method including: obtaining values of functional indicators derived from data generated during one or more of the multiple operations on the substrate, the functional indicators characterizing at least one operation; applying a decision model including one or more threshold values to the values of the functional indicators to obtain one or more categorical indicators; and assigning a category to the substrate based on the one or more categorical indicators.

Abstract: A method and apparatus for selecting patterns from an image such as a design layout. The method includes obtaining an image (e.g., of a target layout) having a plurality of patterns; determining, based on pixel intensities within the image, a metric (e.g., entropy) indicative of an amount of information contained in one or more portions of the image; and selecting, based on the metric, a sub-set of the plurality of patterns from the one or more portions of the image having values of the metric within a specified range. The sub-set of patterns can be provided as training data for training a model associated with a patterning process.

Abstract: A method for categorizing a substrate subject to a semiconductor manufacturing process including multiple operations, the method including: obtaining values of functional indicators derived from data generated during one or more of the multiple operations on the substrate, the functional indicators characterizing at least one operation; applying a decision model including one or more threshold values to the values of the functional indicators to obtain one or more categorical indicators; and assigning a category to the substrate based on the one or more categorical indicators.

Abstract: Described herein is a method for training a machine learning model to determine a source of error contribution to multiple features of a pattern printed on a substrate. The method includes obtaining training data having multiple datasets, wherein each dataset has error contribution values representative of an error contribution from one of multiple sources to the features, and wherein each dataset is associated with an actual classification that identifies a source of the error contribution of the corresponding dataset; and training, based on the training data, a machine learning model to predict a classification of a reference dataset of the datasets such that a cost function that determines a difference between the predicted classification and the actual classification of the reference dataset is reduced.

Abstract: A method for determining an optimized weighting of an encoder and decoder network; the method comprising: for each of a plurality of test weightings, performing the following steps with the encoder and decoder operating using the test weighting: (a) encoding, using the encoder, a reference image and a distorted image into a latent space to form an encoding; (b) decoding the encoding, using the decoder, to form a distortion map indicative of a difference between the reference image and a distorted image; (c) spatially transforming the distorted image by the distortion map to obtain an aligned image; (d) comparing the aligned image to the reference image to obtain a similarity metric; and (e) determining a loss function which is at least partially defined by the similarity metric; wherein the optimized weighting is determined to be the test weighting which has an optimized loss function.

Abstract: An inspection tool comprises an imaging system configured to image a portion of a semiconductor substrate. The inspection tool may further comprise an image analysis system configured to obtain an image of a structure on the semiconductor substrate from the imaging system, encode the image of the structure into a latent space thereby forming a first encoding. the image analysis system may subtract an artifact vector, representative of an artifact in the image, from the encoding thereby forming a second encoding; and decode the second encoding to obtain a decoded image.

Abstract: A method and system for predicting complex electric field images with a parameterized model are described. A latent space representation of a complex electric field image is determined based on dimensional data in a latent space of the parameterized model for a given input to the parameterized model. The given input may be a measured amplitude (e.g., intensity) associated with the complex electric field image. The complex electric field image is predicted based on the latent space representation of the complex electric field image. The predicted complex electric field image includes an amplitude and a phase. The parameterized model comprises encoder-decoder architecture. In some embodiments, determining the latent space representation of the electric field image comprises minimizing a function constrained by a set of electric field images that could be predicted by the parameterized model based on the dimensional data in the latent space and the given input.

Abstract: An acoustic scatterometer has an acoustic source operable to project acoustic radiation onto a periodic structure and formed on a substrate. An acoustic detector is operable to detect the ?1st acoustic diffraction order diffracted by the periodic structure and while discriminating from specular reflection (0th order). Another acoustic detector is operable to detect the +1st acoustic diffraction order diffracted by the periodic structure, again while discriminating from the specular reflection (0th order). The acoustic source and acoustic detector may be piezo transducers. The angle of incidence of the projected acoustic radiation and location of the detectors and are arranged with respect to the periodic structure and such that the detection of the ?1st and +1st acoustic diffraction orders and discriminates from the 0th order specular reflection.

Abstract: A system and method for generating predictive images for wafer inspection using machine learning are provided. Some embodiments of the system and method include acquiring the wafer after a photoresist applied to the wafer has been developed; imaging a portion of a segment of the developed wafer; acquiring the wafer after the wafer has been etched; imaging the segment of the etched wafer; training a machine learning model using the imaged portion of the developed wafer and the imaged segment of the etched wafer; and applying the trained machine learning model using the imaged segment of the etched wafer to generate predictive images of a developed wafer. Some embodiments include imaging a segment of the developed wafer; imaging a portion of the segment of the etched wafer; training a machine learning model; and applying the trained machine learning model to generate predictive after-etch images of the developed wafer.

Abstract: A method for applying a deposition model in a semiconductor manufacturing process. The method includes predicting a deposition profile of a substrate using the deposition model; and using the predicted deposition profile to enhance a metrology target design. The deposition model can be calibrated using experimental cross-section profile information from a layer of a physical substrate. In some embodiments, the deposition model is a machine-learning model, and calibrating the deposition model includes training the machine-learning model. The metrology target design may include an alignment metrology target design or an overlay metrology target design, for example.

Abstract: A method, system and program for determining a fingerprint of a parameter. The method includes determining a contribution from a device out of a plurality of devices to a fingerprint of a parameter. The method includes obtaining parameter data and usage data, wherein the parameter data is based on measurements for multiple substrates having been processed by the plurality of devices, and the usage data indicates which of the devices out of the plurality of the devices were used in the processing of each substrate; and determining the contribution using the usage data and parameter data.

Abstract: Described herein is a method of training a model configured to predict whether a feature associated with an imaged substrate will be defective after etching of the imaged substrate and determining etch conditions based on the trained model. The method includes obtaining, via a metrology tool, (i) an after development image of the imaged substrate at a given location, the after development image including a plurality of features, and (ii) an after etch image of the imaged substrate at the given location; and training, using the after development image and the after etch image, the model configured to determine defectiveness of a given feature of the plurality of features in the after development image. In an embodiment, the determining of defectiveness is based on comparing the given feature in the after development image with a corresponding etch feature in the after etch image.

Abstract: A method for increasing certainty in parameterized model predictions. The method includes clustering dimensional data in a latent space associated with a parameterized model into clusters. Different clusters correspond to different portions of a given input. The method includes predicting, with the parameterized model, an output based on the dimensional data in the latent space. The method includes transforming, with the parameterized model, the dimensional data in the latent space into a recovered version of the given input that corresponds to one or more of the clusters. In some embodiments, the method includes determining which one or more clusters correspond to predicted outputs with higher variance, and making the parameterized model more descriptive by adding to the dimensionality of the latent space, and/or training the parameterized model with more diverse training data associated with one or more determined clusters or parts thereof associated with predicted outputs with the higher variance.

Abstract: Parameters of a structure (900) are measured by reconstruction from observed diffracted radiation. The method includes the steps: (a) defining a structure model to represent the structure in a two- or three-dimensional model space; (b) using the structure model to simulate interaction of radiation with the structure; and (c) repeating step (b) while varying parameters of the structure model. The structure model is divided into a series of slices (a-f) along at least a first dimension (Z) of the model space. By the division into slices, a sloping face (904, 906) of at least one sub-structure is approximated by a series of steps (904?, 906?) along at least a second dimension of the model space (X). The number of slices may vary dynamically as the parameters vary. The number of steps approximating said sloping face is maintained constant. Additional cuts (1302, 1304) are introduced, without introducing corresponding steps.

Abstract: A method, system and program for determining a fingerprint of a parameter. The method includes determining a contribution from a device out of a plurality of devices to a fingerprint of a parameter. The method includes obtaining parameter data and usage data, wherein the parameter data is based on measurements for multiple substrates having been processed by the plurality of devices, and the usage data indicates which of the devices out of the plurality of the devices were used in the processing of each substrate; and determining the contribution using the usage data and parameter data.

Abstract: A method for training a machine learning model configured to predict a substrate image corresponding to a printed pattern of a substrate as measured via a metrology tool. The method involves obtaining a training data set including (i) metrology data of the metrology tool used to measure the printed pattern of the substrate, and (ii) a representation of a mask pattern employed for imaging the printed pattern on the substrate; and training, based on the training data set, a machine learning model to predict the substrate image of the substrate as measured by the metrology tool such that a cost function is improved, wherein the cost function includes a relationship between the predicted substrate image and the metrology data.