Abstract: A method of detecting a polishing endpoint includes storing a plurality of library spectra, measuring a sequence of spectra from the substrate in-situ during polishing, and for each measured spectrum of the sequence of spectra, finding a best matching library spectrum from the plurality of library spectra to generate a sequence of best matching library spectra. Each library spectrum has a stored associated value representing a degree of progress through a polishing process, and the stored associated value for the best matching library spectrum is determined for each best matching library spectrum to generate a sequence of values representing a progression of polishing of the substrate. The sequence of values is compared to a target value, and a polishing endpoint is triggered when the sequence of values reaches the target value.

Abstract: A method of detecting a polishing endpoint includes storing a plurality of library spectra, measuring a sequence of spectra from the substrate in-situ during polishing, and for each measured spectrum of the sequence of spectra, finding a best matching library spectrum from the plurality of library spectra to generate a sequence of best matching library spectra. Each library spectrum has a stored associated value representing a degree of progress through a polishing process, and the stored associated value for the best matching library spectrum is determined for each best matching library spectrum to generate a sequence of values representing a progression of polishing of the substrate. The sequence of values is compared to a target value, and a polishing endpoint is triggered when the sequence of values reaches the target value.

Abstract: A method for predicting yield for a semiconductor process. A particular type of wafer is fabricated to have a first set of features disposed on the wafer, with a wafer map identifying a location for each of the first set of features on the wafer. Data from wafer acceptance tests and circuit probe tests is collected over time for wafers of that particular type as made in a semiconductor fabrication process, and at least one training dataset and a least one validation dataset are created from the collected data. A second set of “engineered” features are created and also incorporated onto the wafer and wafer map. Important features from the first and second sets of features are identified and selected, and using those important features as inputs, a number of different process models are run, with yield as the target. The results of the different models can be combined, for example, statistically.

Abstract: Robust machine learning predictions. Temporal dependencies of process targets for different machine learning models can be captured and evaluated for the impact on process performance for target. The most robust of these different models is selected for deployment based on minimizing variance for the desired performance characteristic.

Abstract: Techniques for measuring and/or compensating for process variations in a semiconductor manufacturing processes. Machine learning algorithms are used on extensive sets of input data, including upstream data, to organize and pre-process the input data, and to correlate the input data to specific features of interest. The correlations can then be used to make process adjustments. The techniques may be applied to any feature or step of the semiconductor manufacturing process, such as overlay, critical dimension, and yield prediction.

Abstract: Among other things, a computer-based method is described. The method comprises receiving, by one or more computers, a plurality of measured spectra reflected from a substrate at a plurality of different positions on the substrate. The substrate comprises at least two regions having different structural features. The method also comprises performing, by the one or more computers, a clustering algorithm on the plurality of measured spectra to separate the plurality of measured spectra into a number of groups based on the spectral characteristics of the plurality of measured spectra; selecting one of the number of groups to provide a selected group having a subset of spectra from the plurality of measured spectra; and determining, in the one or more computers, at least one characterizing value for the substrate based on the subset of spectra of the selected group.

Abstract: A plurality of spectra reflected from one or more substrates at a plurality of different positions on the one or more substrates are represented in the form of a first matrix, and the first matrix is decomposed into products of at least two component matrixes of a first set of component matrixes. The dimensions of each of the at least two component matrixes is reduced to produce a second set of component matrixes containing the at least two matrixes with reduced dimensions.

Abstract: Among other things, a computer-based method is described. The method comprises receiving, by one or more computers, a plurality of measured spectra reflected from a substrate at a plurality of different positions on the substrate. The substrate comprises at least two regions having different structural features. The method also comprises performing, by the one or more computers, a clustering algorithm on the plurality of measured spectra to separate the plurality of measured spectra into a number of groups based on the spectral characteristics of the plurality of measured spectra; selecting one of the number of groups to provide a selected group having a subset of spectra from the plurality of measured spectra; and determining, in the one or more computers, at least one characterizing value for the substrate based on the subset of spectra of the selected group.

Abstract: A method to assist in identifying a spectral feature and a characteristic of the selected spectral feature to monitor during polishing includes polishing a test substrate and measuring a sequence of spectra of light reflected from a substrate while the substrate is being polished, where at least some of the spectra of the sequence differ due to material being removed during the polishing. The sequence of spectra are visually displayed as a contour plot.

Abstract: A plurality of spectra reflected from one or more substrates at a plurality of different positions on the one or more substrates are represented in the form of a first matrix, and the first matrix is decomposed into products of at least two component matrixes of a first set of component matrixes. The dimensions of each of the at least two component matrixes is reduced to produce a second set of component matrixes containing the at least two matrixes with reduced dimensions.

Abstract: In one aspect, a method of polishing includes polishing a substrate, and receiving an identification of a selected spectral feature and a characteristic of the selected spectral feature to monitor during polishing. The method includes measuring a sequence of spectra of light reflected from the substrate while the substrate is being polished, where at least some of the spectra of the sequence differ due to material being removed during the polishing. The method of polishing includes determining a value of a characteristic of the selected spectral feature for each of the spectra in the sequence of spectra to generate a sequence of values for the characteristic, fitting a function to the sequence of values, and determining either a polishing endpoint or an adjustment for a polishing rate based on the function.

Abstract: A plurality of spectra reflected from one or more substrates at a plurality of different positions on the one or more substrates are represented in the form of a first matrix, and the first matrix is decomposed into products of at least two component matrixes of a first set of component matrixes. The dimensions of each of the at least two component matrixes is reduced to produce a second set of component matrixes containing the at least two matrixes with reduced dimensions.

Abstract: A method of controlling polishing includes storing a base spectrum, the base spectrum being a spectrum of light reflected from a substrate after deposition of a deposited dielectric layers overlying a metallic layer or semiconductor wafer and before deposition of a non-metallic layer over the plurality of deposited dielectric layer. After deposition of the non-metallic layer and during polishing of the non-metallic layer on the substrate, measurements of a sequence of raw spectra of light reflected the substrate during polishing are received from an in-situ optical monitoring system. Each raw spectrum is normalized to generate a sequence of normalized spectra using the raw spectrum and the base spectrum. At least one of a polishing endpoint or an adjustment for a polishing rate is determined based on at least one normalized predetermined spectrum from the sequence of normalized spectra.

Abstract: A method of controlling polishing includes polishing a substrate and receiving an identification of a selected spectral feature, a wavelength range having a width, and a characteristic of the selected spectral feature to monitor during polishing. A sequence of spectra of light from the substrate is measured while the substrate is being polished. A sequence of values of the characteristic of the selected spectral feature is generated from the sequence of spectra. For at least some spectra from the sequence of spectra, a modified wavelength range is generated based on a position of the spectral feature within a previous wavelength range used for a previous spectrum in the sequence of spectra, the modified wavelength range is searched for the selected spectral feature, and a value of a characteristic of the selected spectral feature is determined.

Abstract: A method of detecting a polishing endpoint includes storing a plurality of library spectra, measuring a sequence of spectra from the substrate in-situ during polishing, and for each measured spectrum of the sequence of spectra, finding a best matching library spectrum from the plurality of library spectra to generate a sequence of best matching library spectra. Each library spectrum has a stored associated value representing a degree of progress through a polishing process, and the stored associated value for the best matching library spectrum is determined for each best matching library spectrum to generate a sequence of values representing a progression of polishing of the substrate. The sequence of values is compared to a target value, and a polishing endpoint is triggered when the sequence of values reaches the target value.

Abstract: A plurality of spectra reflected from one or more substrates at a plurality of different positions on the one or more substrates are represented in the form of a first matrix, and the first matrix is decomposed into products of at least two component matrixes of a first set of component matrixes. The dimensions of each of the at least two component matrixes is reduced to produce a second set of component matrixes containing the at least two matrixes with reduced dimensions.

Abstract: Techniques for measuring and/or compensating for process variations in a semiconductor manufacturing processes. Machine learning algorithms are used on extensive sets of input data, including upstream data, to organize and pre-process the input data, and to correlate the input data to specific features of interest. The correlations can then be used to make process adjustments. The techniques may be applied to any feature or step of the semiconductor manufacturing process, such as overlay, critical dimension, and yield prediction.

Abstract: Methods and apparatus for spectrum-based endpointing. An endpointing method includes selecting a reference spectrum. The reference spectrum is a spectrum of white light reflected from a film of interest on a first substrate and has a thickness greater than a target thickness. The reference spectrum is empirically selected for particular spectrum-based endpoint determination logic so that the target thickness is achieved when endpoint is called by applying the particular spectrum-based endpoint logic. The method includes obtaining a current spectrum. The current spectrum is a spectrum of white light reflected from a film of interest on a second substrate when the film of interest is being subjected to a polishing step and has a current thickness that is greater than the target thickness. The method includes determining, for the second substrate, when an endpoint of the polishing step has been achieved. The determining is based on the reference and current spectra.

Abstract: A method of controlling polishing includes polishing a substrate of a non-metallic layer undergoing polishing and a metal layer underlying the non-metallic layer; storing a metal reference spectrum, the metal reference spectrum being a spectrum of light reflected from a same metal material as the metal layer; measuring a sequence of raw spectra of light reflected from the substrate during polishing with an in-situ optical monitoring system; normalizing each raw spectrum in the sequence of spectra to generate a sequence of normalized spectra, of which normalizing includes a division operation where the measured spectrum is in the numerator and the metal reference spectrum is in the denominator; and determining at least one of a polishing endpoint or an adjustment for a polishing rate based on at least one normalized predetermined spectrum from the sequence of normalized spectra.

Abstract: Among other things, a machine based method includes representing a plurality of spectra reflected from one or more substrates at a plurality of different positions on the one or more substrates in the form of a first matrix; decomposing, by one or more computers, the first matrix into products of at least two component matrixes of a first set of component matrixes; reducing dimensions of each of the at least two component matrixes to produce a second set of component matrixes containing the at least two matrixes with reduced dimensions; and generating, by the one or more computers, a second matrix by taking a product of the matrixes of the second set of component matrixes.