System, method and computer program product for calibration of metrology tools

Patent metadata

Field	Value
Publication number	US-10295342-B2
Application number	US-201615236334-A
Country	US
Kind code	B2
Filing date	Aug 12, 2016
Priority date	Aug 14, 2015
Publication date	May 21, 2019
Grant date	May 21, 2019

Abstract

A system, method and computer program product are provided for calibrating metrology tools. One or more design-of-experiments wafers is received for calibrating a metrology tool. A set of signals is collected by measuring the one or more wafers utilizing the metrology tool. A first transformation is determined to convert the set of signals to components, and a second transformation is determined to convert a set of reference signals to reference components. The set of reference signals is collected by measuring the one or more wafers utilizing a well-calibrated reference tool. A model is trained based on the reference components that maps the components to converted components, and the model, first transformation, and second transformation are stored in a memory associated with the metrology tool.

First claim

What is claimed is:

1. A method, comprising:
receiving one or more DoE (design-of-experiments) wafers for calibrating a metrology tool, the one or more DoE wafers including a plurality of metrology targets and the plurality of metrology targets including at least one of a periodic grating, a FinFet structure, an SRAM device structure, a Flash memory structure, and a DRAM memory structure;
collecting a set of signals by measuring the one or more DoE wafers utilizing the metrology tool;
analyzing the set of signals, utilizing analysis logic or an autoencoder, to determine a portion of principal components of the set of signals;
determining a first transformation that converts the set of signals to the determined portion of principal components;
converting the set of signals to components, using the first transformation;
identifying a set of reference signals collected by either:
a reference metrology tool measuring the one or more DoE wafers, or a simulation performed based on measurement parameters associated with the metrology tool and associated with the one or more DoE wafers;

analyzing the set of reference signals, utilizing the analysis logic or the autoencoder, to determine a portion of principal reference components of the set of reference signals;
determining a second transformation that converts the set of reference signals to the determined portion of principal reference components;
converting the set of reference signals to reference components, using the second transformation;
training a model, using a machine learning algorithm or a neural network, that takes as input the components and the reference components to map the components to converted components; and
storing the model, first transformation, and second transformation in a memory associated with the metrology tool for use in calibrating the metrology tool.

2. The method of claim 1 , wherein the model comprises one of a linear model and a nonlinear model.

3. The method of claim 1 , wherein the model comprises one of a neural network, a random forest, a support vector machine (SVM), a deep network, and a convolution network.

4. The method of claim 1 , wherein the set of signals is analyzing using principal component analysis (PCA).

5. The method of claim 1 , wherein the metrology tool is selected from one of:
a spectroscopic ellipsometer (SE);
a SE with multiple angles of illumination;
a SE measuring Mueller matrix elements;
a single-wavelength ellipsometer;
a beam profile ellipsometer;
a beam profile reflectometer;
a broadband reflective spectrometer;
a single-wavelength reflectometer;
an angle-resolved reflectometer;
an imaging system;
a scatterometer;
a small-angle x-ray scattering (SAXS) device;
an x-ray powder diffraction (XRD) device;
an x-ray Fluorescence (XRF) device;
an x-ray photoelectron spectroscopy (XPS) device;
an x-ray reflectivity (XRR) device;
a Raman spectroscopy device;
a scanning electron microscopy (SEM) device;
a tunneling electron microscope (TEM) device; and
an atomic force microscope (AFM) device.

6. The method of claim 1 , wherein at least one of the first transformation and the second transformation incorporates noise reduction.

7. The method of claim 1 , wherein training the model comprises minimizing a difference between the converted components and the reference components.

8. The method of claim 1 , wherein training the model comprises minimizing a difference between structural parameters based on the converted components and structural parameters based on the reference components.

9. A computer program product embodied on a non-transitory computer readable medium, the computer program product including code adapted to be executed by a computer to perform a method comprising:
receiving one or more DoE (design-of-experiments) wafers for calibrating a metrology tool, the one or more DoE wafers including a plurality of metrology targets and the plurality of metrology targets including at least one of a periodic grating, a FinFet structure, an SRAM device structure, a Flash memory structure, and a DRAM memory structure;
collecting a set of signals by measuring the one or more DoE wafers utilizing the metrology tool;
analyzing the set of signals, utilizing analysis logic or an autoencoder, to determine a portion of principal components of the set of signals;
determining a first transformation that converts the set of signals to the determined portion of principal components;
converting the set of signals to components, using the first transformation;
identifying a set of reference signals collected by either:
a reference metrology tool measuring the one or more DoE wafers, or a simulation performed based on measurement parameters associated with the metrology tool and associated with the one or more DoE wafers;

analyzing the set of reference signals, utilizing the analysis logic or the autoencoder, to determine a portion of principal reference components of the set of reference signals;
determining a second transformation that converts the set of reference signals to the determined portion of principal reference components;
converting the set of reference signals to reference components, using the second transformation;
training a model, using a machine learning algorithm or a neural network, that takes as input the components and the reference components to map the components to converted components; and
storing the model, first transformation, and second transformation in a memory associated with the metrology tool for use in calibrating the metrology tool.

10. The computer program product of claim 9 , wherein training the model comprises minimizing a difference between the converted components and the reference components.

11. A system, comprising:
one or more DoE (design-of-experiments) wafers;
a metrology tool configured to collect a set of signals by measuring the one or more DoE wafers, the one or more DoE wafers including a plurality of metrology targets and the plurality of metrology targets including at least one of a periodic grating, a FinFet structure, an SRAM device structure, a Flash memory structure, and a DRAM memory structure;
a memory associated with the metrology tool; and
a processor configured to:
analyze the set of signals, utilizing analysis logic or an autoencoder, to determine a portion of principal components of the set of signals;
determine a first transformation that converts the set of signals to the determined portion of principal components,
convert the set of signals to components, using the first transformation, identify a set of reference signals collected by either:
a reference metrology tool measuring the one or more DoE wafers, or
a simulation performed based on measurement parameters associated with the metrology tool and associated with the one or more DoE wafers,

analyze the set of reference signals, utilizing the analysis logic or the autoencoder, to determine a portion of principal reference components of the set of reference signals,
determine a second transformation that converts the set of reference signals to the determined portion of principal reference components,
convert the set of reference signals to reference components, using the second transformation,
train a model, using a machine learning algorithm or a neural network, that takes as input the components and the reference components to map the components to converted components, and
store the model, first transformation, and second transformation in the memory for use in calibrating the metrology tool.

12. The system of claim 11 , wherein the processor executes:
a calibration module confi

Assignees

• Kla Tencor Corp

Inventors

• Pandev Stilian Ivanov
• Sanko Dzmitry

Classifications

• G01B2210/56

  Measuring geometric parameters of semiconductor structures, e.g. profile, critical dimensions or trench depth · CPC title

• G01B21/042 Primary

  Calibration or calibration artifacts (G01B3/30, G01B9/02072 take precedence) · CPC title